netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* V2.4 policy router operates faster/better than V2.6
@ 2005-01-03 20:55 Jeremy M. Guthrie
  2005-01-03 22:51 ` Stephen Hemminger
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-03 20:55 UTC (permalink / raw)
  To: netdev


[-- Attachment #1.1: Type: text/plain, Size: 3833 bytes --]

I have a dual processor box running Suse 9.1 Ent. that I changed over to the 
V2.6.10 kernel.  The box has two interfaces in it, both E1000s.  The box 
receives anywhere from  200mbit to 500+ mbit that it needs to route out to 
other boxes.  The policy routing table is running ~ 150-200 rules.  ie.  data 
comes in E3(e1000), is policy routed to a destination sent out E2(e1000).

Under V2.4 kernels, the system will operate just fine and drop few packets if 
any.  ie. right now under V2.4, I have dropped all of three packets.  Under 
2.6, I can watch the RX drop counter increment.  See below.

[h-pr-msn-1 guthrie 1:48pm]~-> ifconfig eth3 ; sleep 10 ; ifconfig eth3
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
          inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:132919934 errors:311285 dropped:311285 overruns:247225 
frame:0
          TX packets:6 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:2630721320 (2508.8 Mb)  TX bytes:484 (484.0 b)
          Base address:0x22a0 Memory:eff80000-effa0000

eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
          inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:133847068 errors:325697 dropped:325697 overruns:258546 
frame:0
          TX packets:6 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:3102796062 (2959.0 Mb)  TX bytes:484 (484.0 b)
          Base address:0x22a0 Memory:eff80000-effa0000

If I turn off the policy routing, I instantly stop getting RX errors or 
overruns as it appears the CPU can now pay attention to the packets coming in 
and drop them(as I turned off IP forwarding as well).

V2.4 Kernel mpstat data:
command:  mpstat -P ALL 60
Linux 2.4.21-251-smp (h-pr-msn-1) 12/15/2004

01:16:24 PM  CPU   %user   %nice %system   %idle    intr/s
01:17:19 PM  all    0.16    0.00   50.12   49.72  42114.18
01:17:19 PM    0    0.12    0.00   55.60   44.28  42114.18
01:17:19 PM    1    0.20    0.00   44.65   55.15  42114.18

01:17:19 PM  CPU   %user   %nice %system   %idle    intr/s
01:18:19 PM  all    0.13    0.00   48.49   51.38  42103.08
01:18:19 PM    0    0.13    0.00   31.88   67.98  42103.08
01:18:19 PM    1    0.13    0.00   65.10   34.77  42103.08

V2.6 kernel mpstat data:
command:  mpstat -P ALL 60
Linux 2.6.5-7.111.5-smp (h-pr-msn-1)         12/15/04

13:36:25     CPU   %user   %nice %system %iowait    %irq   %soft   %idle    
intr/s
13:37:25     all    0.13    0.00    0.15    0.09    2.03   43.14   54.45  
25506.53
13:37:25       0    0.17    0.00    0.08    0.18    0.00   16.81   82.76   
2215.63
13:37:25       1    0.08    0.00    0.20    0.00    4.08   69.49   26.14  
23291.34

13:37:25     CPU   %user   %nice %system %iowait    %irq   %soft   %idle    
intr/s
13:38:24     all    0.14    0.00    0.12    0.12    2.02   42.89   54.71  
25900.70
13:38:24       0    0.03    0.00    0.05    0.22    0.00   16.67   83.03   
2246.10
13:38:24       1    0.25    0.00    0.20    0.03    4.02   69.12   26.40  
23654.55

Any insights as to why there would be such a stark difference in performance 
between V2.6 and V2.4?

Please advise.

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #1.2: 0x719905E5.asc --]
[-- Type: application/pgp-keys, Size: 1734 bytes --]

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-03 20:55 V2.4 policy router operates faster/better than V2.6 Jeremy M. Guthrie
@ 2005-01-03 22:51 ` Stephen Hemminger
  2005-01-03 22:56   ` Jeremy M. Guthrie
  2005-01-04 15:07   ` Jeremy M. Guthrie
  0 siblings, 2 replies; 88+ messages in thread
From: Stephen Hemminger @ 2005-01-03 22:51 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev

On Mon, 3 Jan 2005 14:55:24 -0600
"Jeremy M. Guthrie" <jeremy.guthrie@berbee.com> wrote:

> I have a dual processor box running Suse 9.1 Ent. that I changed over to the 
> V2.6.10 kernel.  The box has two interfaces in it, both E1000s.  The box 
> receives anywhere from  200mbit to 500+ mbit that it needs to route out to 
> other boxes.  The policy routing table is running ~ 150-200 rules.  ie.  data 
> comes in E3(e1000), is policy routed to a destination sent out E2(e1000).
> 
> Under V2.4 kernels, the system will operate just fine and drop few packets if 
> any.  ie. right now under V2.4, I have dropped all of three packets.  Under 
> 2.6, I can watch the RX drop counter increment.  See below.
> 
> [h-pr-msn-1 guthrie 1:48pm]~-> ifconfig eth3 ; sleep 10 ; ifconfig eth3
> eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
>           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
>           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:132919934 errors:311285 dropped:311285 overruns:247225 
> frame:0
>           TX packets:6 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:1000
>           RX bytes:2630721320 (2508.8 Mb)  TX bytes:484 (484.0 b)
>           Base address:0x22a0 Memory:eff80000-effa0000
> 
> eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
>           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
>           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:133847068 errors:325697 dropped:325697 overruns:258546 
> frame:0
>           TX packets:6 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:1000
>           RX bytes:3102796062 (2959.0 Mb)  TX bytes:484 (484.0 b)
>           Base address:0x22a0 Memory:eff80000-effa0000
> 
> If I turn off the policy routing, I instantly stop getting RX errors or 
> overruns as it appears the CPU can now pay attention to the packets coming in 
> and drop them(as I turned off IP forwarding as well).
> 
> V2.4 Kernel mpstat data:
> command:  mpstat -P ALL 60
> Linux 2.4.21-251-smp (h-pr-msn-1) 12/15/2004
> 
> 01:16:24 PM  CPU   %user   %nice %system   %idle    intr/s
> 01:17:19 PM  all    0.16    0.00   50.12   49.72  42114.18
> 01:17:19 PM    0    0.12    0.00   55.60   44.28  42114.18
> 01:17:19 PM    1    0.20    0.00   44.65   55.15  42114.18
> 
> 01:17:19 PM  CPU   %user   %nice %system   %idle    intr/s
> 01:18:19 PM  all    0.13    0.00   48.49   51.38  42103.08
> 01:18:19 PM    0    0.13    0.00   31.88   67.98  42103.08
> 01:18:19 PM    1    0.13    0.00   65.10   34.77  42103.08
> 
> V2.6 kernel mpstat data:
> command:  mpstat -P ALL 60
> Linux 2.6.5-7.111.5-smp (h-pr-msn-1)         12/15/04
> 
> 13:36:25     CPU   %user   %nice %system %iowait    %irq   %soft   %idle    
> intr/s
> 13:37:25     all    0.13    0.00    0.15    0.09    2.03   43.14   54.45  
> 25506.53
> 13:37:25       0    0.17    0.00    0.08    0.18    0.00   16.81   82.76   
> 2215.63
> 13:37:25       1    0.08    0.00    0.20    0.00    4.08   69.49   26.14  
> 23291.34
> 
> 13:37:25     CPU   %user   %nice %system %iowait    %irq   %soft   %idle    
> intr/s
> 13:38:24     all    0.14    0.00    0.12    0.12    2.02   42.89   54.71  
> 25900.70
> 13:38:24       0    0.03    0.00    0.05    0.22    0.00   16.67   83.03   
> 2246.10
> 13:38:24       1    0.25    0.00    0.20    0.03    4.02   69.12   26.40  
> 23654.55
> 
> Any insights as to why there would be such a stark difference in performance 
> between V2.6 and V2.4?

How many flows are going through the router? The neighbour cache
can get to be a bottleneck. Perhaps Robert "the Router Man" Olssen can
give some hints.

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-03 22:51 ` Stephen Hemminger
@ 2005-01-03 22:56   ` Jeremy M. Guthrie
  2005-01-05 13:18     ` Robert Olsson
  2005-01-04 15:07   ` Jeremy M. Guthrie
  1 sibling, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-03 22:56 UTC (permalink / raw)
  To: netdev; +Cc: Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 789 bytes --]

How would I check?  It should be in the hundreds of thousands.

On Monday 03 January 2005 04:51 pm, Stephen Hemminger wrote:
> On Mon, 3 Jan 2005 14:55:24 -0600
>
> "Jeremy M. Guthrie" <jeremy.guthrie@berbee.com> wrote:
> > Any insights as to why there would be such a stark difference in
> > performance between V2.6 and V2.4?
>
> How many flows are going through the router? The neighbour cache
> can get to be a bottleneck. Perhaps Robert "the Router Man" Olssen can
> give some hints.

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-03 22:51 ` Stephen Hemminger
  2005-01-03 22:56   ` Jeremy M. Guthrie
@ 2005-01-04 15:07   ` Jeremy M. Guthrie
  1 sibling, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-04 15:07 UTC (permalink / raw)
  To: netdev

[-- Attachment #1: Type: text/plain, Size: 1022 bytes --]

On Monday 03 January 2005 04:51 pm, Stephen Hemminger wrote:
> How many flows are going through the router? The neighbour cache
> can get to be a bottleneck. Perhaps Robert "the Router Man" Olssen can
> give some hints.
Tue Jan  4 08:59:12 CST 2005
58406 rt_cache
Tue Jan  4 08:59:48 CST 2005
60636 rt_cache
Tue Jan  4 09:00:34 CST 2005
63891 rt_cache
Tue Jan  4 09:01:02 CST 2005
64635 rt_cache
Tue Jan  4 09:01:29 CST 2005
65689 rt_cache
Tue Jan  4 09:01:58 CST 2005
63426 rt_cache
Tue Jan  4 09:02:25 CST 2005
64139 rt_cache
Tue Jan  4 09:02:53 CST 2005
63860 rt_cache
Tue Jan  4 09:03:20 CST 2005
65719 rt_cache
Tue Jan  4 09:03:48 CST 2005
62465 rt_cache
Tue Jan  4 09:04:15 CST 2005
39339 rt_cache



-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-03 22:56   ` Jeremy M. Guthrie
@ 2005-01-05 13:18     ` Robert Olsson
  2005-01-05 15:18       ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-05 13:18 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Stephen Hemminger


Jeremy M. Guthrie writes:
 > How would I check?  It should be in the hundreds of thousands.

 Good question  Stephen,..

 Yes it seems like this pretty hefty load. Forwarding rate of 92k kpps 
 and a drop rate of 10 kpps and dst hash mostly at 50-60 kentries if 
 I read the stats correctly.
 
 And 2.4 were able handle this but not 2.6.10?  
 
 Assuming things are uses and setup identically. 2.6 uses RCU for route hash 
 locking. Any dst cache overslow messages seen?

 A couple of lines of rtstat would be very interesing from this box.

 Also check that the CPU shares the RX packet load. CPU0 affinty to eth0 
 and CPU1 to eth1 seems to be best. It gives cache bouncing at "TX" and
 slab jobs but we have accept that for now. 

 13:37:25     CPU   %user   %nice %system %iowait    %irq   %soft   %idle    
> intr/s
> 13:38:24     all    0.14    0.00    0.12    0.12    2.02   42.89   54.71  
> 25900.70
> 13:38:24       0    0.03    0.00    0.05    0.22    0.00   16.67   83.03   
> 2246.10
> 13:38:24       1    0.25    0.00    0.20    0.03    4.02   69.12   26.40  
> 23654.55

This looks weird to me... we cannot have CPU left? Due to the imbalance?

Check /proc/net/softnet_stat,

Haven't used mpstat. %soft is that *all* softirq's or only softirq's deferred 
to ksoftird only?

						--ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-05 13:18     ` Robert Olsson
@ 2005-01-05 15:18       ` Jeremy M. Guthrie
  2005-01-05 16:30         ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-05 15:18 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 2632 bytes --]

On Wednesday 05 January 2005 07:18 am, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > How would I check?  It should be in the hundreds of thousands.
>
>  Good question  Stephen,..
>
>  Yes it seems like this pretty hefty load. Forwarding rate of 92k kpps
>  and a drop rate of 10 kpps and dst hash mostly at 50-60 kentries if
>  I read the stats correctly.
Yeah, the load will be high.  I'm expecting this to be watching ~ 750 mbps by 
next December.  The app profiles all traffic going in and out of our data 
centers.

>  And 2.4 were able handle this but not 2.6.10?
Yes.  It does handle it.  It runs harder ie. 2.6 caps out at ~ 50% utilization 
where 2.4 might run 60-75% utilized.

>  Assuming things are uses and setup identically. 2.6 uses RCU for route
> hash locking. Any dst cache overslow messages seen?
No.

>  A couple of lines of rtstat would be very interesing from this box.
I'm not showing the /proc/net/rt_cache_stat file.  Was there a kernel option I 
need to recompile with for rt_cache_stat to show up in proc?

>  Also check that the CPU shares the RX packet load. CPU0 affinty to eth0
>  and CPU1 to eth1 seems to be best. It gives cache bouncing at "TX" and
>  slab jobs but we have accept that for now.
How would I go about doing this?

>  13:37:25     CPU   %user   %nice %system %iowait    %irq   %soft   %idle
>
> > intr/s
> > 13:38:24     all    0.14    0.00    0.12    0.12    2.02   42.89   54.71
> > 25900.70
> > 13:38:24       0    0.03    0.00    0.05    0.22    0.00   16.67   83.03
> > 2246.10
> > 13:38:24       1    0.25    0.00    0.20    0.03    4.02   69.12   26.40
> > 23654.55
>
> This looks weird to me... we cannot have CPU left? Due to the imbalance?

> Check /proc/net/softnet_stat,
cat /proc/net/softnet_stat
5592c972 00000000 00001fc8 00000000 00000000 00000000 00000000 00000000 
00391c3f
000f1991 00000000 00000000 00000000 00000000 00000000 00000000 00000000 
001292ba


> Haven't used mpstat. %soft is that *all* softirq's or only softirq's
> deferred to ksoftird only?
"%soft"
Show  the percentage of time spent by the CPU or CPUs to service               
softirqs.  A softirq (software interrupt) is one  of  up  to  32               
enumerated software interrupts which can run on multiple CPUs at               
once.


-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-05 15:18       ` Jeremy M. Guthrie
@ 2005-01-05 16:30         ` Robert Olsson
  2005-01-05 17:35           ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-05 16:30 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger


Jeremy M. Guthrie writes:

 > Yeah, the load will be high.  I'm expecting this to be watching ~ 750 mbps by 
 > next December.  The app profiles all traffic going in and out of our data 
 > centers.

 BW itself or pps is that not much of challange as handling of concurrent 
 flows.

 > I'm not showing the /proc/net/rt_cache_stat file.  Was there a kernel 
 > option I need to recompile with for rt_cache_stat to show up in proc?


 No it's there without any options. Would be nice to the output from rtstat


 > >  Also check that the CPU shares the RX packet load. CPU0 affinty to eth0
 > >  and CPU1 to eth1 seems to be best. It gives cache bouncing at "TX" and
 > >  slab jobs but we have accept that for now.

 > How would I go about doing this?

 Assume you route packets between eth0 <-> eth1 
 Set eth0 irq to CPU0 and eth1 to CPU1 with /proc/irq/XX/smp_affinity
 Disable irqbalancer etc.

 > cat /proc/net/softnet_stat

 total    droppped tsquz    Throttl  FR_hit   FR_succe FR_defer FR_def_o 
 cpu_coll 

 > 5592c972 00000000 00001fc8 00000000 00000000 00000000 00000000 00000000 
 > 00391c3f
 > 000f1991 00000000 00000000 00000000 00000000 00000000 00000000 00000000 
 > 001292ba

 See! One line per CPU. So CPU0 is handing almost all packets.

 > "%soft"
 > Show  the percentage of time spent by the CPU or CPUs to service
 > softirqs.  A softirq (software interrupt) is one  of  up  to  32
 > enumerated software interrupts which can run on multiple CPUs

 Well yes. I had a more specific question. I'll look into mpstat where do 
 find it? Kernel pacthes?

 Be also aware that packet forwarding with SMP/NUMA is very much research
 today it is not that easy or not even possible to get aggregated performance 
 from several CPU's. in any setup. Well anyway we are beginning to see some 
 benefits now as we better understand the problems.

      
						--ro

  
 

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-05 16:30         ` Robert Olsson
@ 2005-01-05 17:35           ` Jeremy M. Guthrie
  2005-01-05 19:25             ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-05 17:35 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 6120 bytes --]

On Wednesday 05 January 2005 10:30 am, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > Yeah, the load will be high.  I'm expecting this to be watching ~ 750
>  > mbps by next December.  The app profiles all traffic going in and out of
>  > our data centers.
>  BW itself or pps is that not much of challange as handling of concurrent
>  flows.
Roger that.

>  > I'm not showing the /proc/net/rt_cache_stat file.  Was there a kernel
>  > option I need to recompile with for rt_cache_stat to show up in proc?
>
>  No it's there without any options. Would be nice to the output from rtstat
Output from rtstat:
rtstat
fopen: No such file or directory

cd /proc/net/
ls -la
total 0
dr-xr-xr-x   5 root root 0 2005-01-04 08:53 .
dr-xr-xr-x  72 root root 0 2005-01-04 08:53 ..
-r--r--r--   1 root root 0 2005-01-05 10:32 anycast6
-r--r--r--   1 root root 0 2005-01-05 10:32 arp
dr-xr-xr-x   2 root root 0 2005-01-05 10:32 atm
-r--r--r--   1 root root 0 2005-01-05 10:32 dev
-r--r--r--   1 root root 0 2005-01-05 10:32 dev_mcast
dr-xr-xr-x   2 root root 0 2005-01-05 10:32 dev_snmp6
-r--r--r--   1 root root 0 2005-01-04 08:53 if_inet6
-r--r--r--   1 root root 0 2005-01-05 10:32 igmp
-r--r--r--   1 root root 0 2005-01-05 10:32 igmp6
-r--r--r--   1 root root 0 2005-01-05 10:32 ip6_flowlabel
-r--r--r--   1 root root 0 2005-01-05 10:32 ip_mr_cache
-r--r--r--   1 root root 0 2005-01-05 10:32 ip_mr_vif
-r--r--r--   1 root root 0 2005-01-05 10:32 ip_tables_matches
-r--r--r--   1 root root 0 2005-01-05 10:32 ip_tables_names
-r--r--r--   1 root root 0 2005-01-05 10:32 ip_tables_targets
-r--r--r--   1 root root 0 2005-01-05 10:32 ipv6_route
-r--r--r--   1 root root 0 2005-01-05 10:32 mcfilter
-r--r--r--   1 root root 0 2005-01-05 10:32 mcfilter6
-r--r--r--   1 root root 0 2005-01-05 10:32 netlink
-r--r--r--   1 root root 0 2005-01-05 10:32 netstat
-r--r--r--   1 root root 0 2005-01-05 10:32 psched
-r--r--r--   1 root root 0 2005-01-05 10:32 raw
-r--r--r--   1 root root 0 2005-01-05 10:32 raw6
-r--r--r--   1 root root 0 2005-01-05 10:32 route
dr-xr-xr-x   6 root root 0 2005-01-05 10:32 rpc
-r--r--r--   1 root root 0 2005-01-05 10:32 rt6_stats
-r--r--r--   1 root root 0 2005-01-05 10:32 rt_acct
-r--r--r--   1 root root 0 2005-01-05 10:32 rt_cache
-r--r--r--   1 root root 0 2005-01-05 10:32 snmp
-r--r--r--   1 root root 0 2005-01-05 10:32 snmp6
-r--r--r--   1 root root 0 2005-01-05 10:32 sockstat
-r--r--r--   1 root root 0 2005-01-05 10:32 sockstat6
-r--r--r--   1 root root 0 2005-01-05 10:32 softnet_stat
dr-xr-xr-x   2 root root 0 2005-01-05 10:32 stat
-r--r--r--   1 root root 0 2005-01-04 08:53 tcp
-r--r--r--   1 root root 0 2005-01-05 10:32 tcp6
-r--r--r--   1 root root 0 2005-01-05 10:32 tr_rif
-r--r--r--   1 root root 0 2005-01-04 08:53 udp
-r--r--r--   1 root root 0 2005-01-05 10:32 udp6
-r--r--r--   1 root root 0 2005-01-05 10:32 unix
-r--r--r--   1 root root 0 2005-01-05 10:32 wireless

>  > >  Also check that the CPU shares the RX packet load. CPU0 affinty to
>  > > eth0 and CPU1 to eth1 seems to be best. It gives cache bouncing at
>  > > "TX" and slab jobs but we have accept that for now.
>  >
>  > How would I go about doing this?
>
>  Assume you route packets between eth0 <-> eth1
Yup

>  Set eth0 irq to CPU0 and eth1 to CPU1 with /proc/irq/XX/smp_affinity
Done

>  Disable irqbalancer etc.
Done

I'll let you know what I see for stats once I get some collected.

>  > cat /proc/net/softnet_stat
>
>  total    droppped tsquz    Throttl  FR_hit   FR_succe FR_defer FR_def_o
>  cpu_coll
>
>  > 5592c972 00000000 00001fc8 00000000 00000000 00000000 00000000 00000000
>  > 00391c3f
>  > 000f1991 00000000 00000000 00000000 00000000 00000000 00000000 00000000
>  > 001292ba
>
>  See! One line per CPU. So CPU0 is handing almost all packets.
cat /proc/interrupts
           CPU0       CPU1
  0:     674862   93484967    IO-APIC-edge  timer
  1:        564          9    IO-APIC-edge  i8042
  7:          0          0   IO-APIC-level  ohci_hcd
  8:          1          1    IO-APIC-edge  rtc
 12:        268         62    IO-APIC-edge  i8042
 14:          2          0    IO-APIC-edge  ide0
 18: 2105131410    9140835   IO-APIC-level  eth3
 20:       1077  248075156   IO-APIC-level  eth2
 27:     118224          1   IO-APIC-level  eth0
 28:      36298         49   IO-APIC-level  aic7xxx
 30:          0          0   IO-APIC-level  acpi
NMI:          0          0
LOC:   94168097   94168094
ERR:          0
MIS:          0

>  > "%soft"
>  > Show  the percentage of time spent by the CPU or CPUs to service
>  > softirqs.  A softirq (software interrupt) is one  of  up  to  32
>  > enumerated software interrupts which can run on multiple CPUs
>
>  Well yes. I had a more specific question. I'll look into mpstat where do
>  find it? Kernel pacthes?
Sorry about that.  New to the list.  I'm not suggesting anything.  I 
appreciate the help!

Suse listed mpstat as part of sysstat 5.1.2.  I'm running stock 2.6.10.

>  Be also aware that packet forwarding with SMP/NUMA is very much research
>  today it is not that easy or not even possible to get aggregated
> performance from several CPU's. in any setup. Well anyway we are beginning
> to see some benefits now as we better understand the problems.
Understood.  As long as I know this, I can articulate this to my uppers for 
bigger hardware.  My current system is a dual P-III 700mhz.  May be time for 
an upgrade.  However, I figure this may also offer a good environment to help 
provide you guys with a taxed system running an load of flows.  Nothing like 
finding fun stuff while a system is ready to fall over.

Would a single hyper threaded CPU help this or should I default to a normal 
dual-cpu system?

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-05 17:35           ` Jeremy M. Guthrie
@ 2005-01-05 19:25             ` Jeremy M. Guthrie
  2005-01-05 20:22               ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-05 19:25 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 927 bytes --]

After smp_affinity adjustments and turning off IRQ balancing.

eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
          inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:537271635 errors:2377048 dropped:2377048 overruns:1849169 
frame:0
          TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:1804727106 (1721.1 Mb)  TX bytes:398 (398.0 b)
          Base address:0x22a0 Memory:eff80000-effa0000


-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-05 19:25             ` Jeremy M. Guthrie
@ 2005-01-05 20:22               ` Robert Olsson
  2005-01-05 20:52                 ` Jeremy M. Guthrie
  2005-01-06 15:26                 ` Jeremy M. Guthrie
  0 siblings, 2 replies; 88+ messages in thread
From: Robert Olsson @ 2005-01-05 20:22 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger


Jeremy M. Guthrie writes:
 > After smp_affinity adjustments and turning off IRQ balancing.
 > 
 > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30

 >           RX packets:537271635 errors:2377048 dropped:2377048 overruns:1849169 

 Worse?

 Yes. I'll remember Hararld moved rt_cache_stat /proc/net/stats/ and wrote a 
 new utility for it. Should be in iproute2 package.  Stephen knows the
 details. Otherwise change path in rtstat. You need this to relate and 
 verify the load.

 Check throughtput and how packet load is used with 2.4. /proc/net/softnet_stat
 and rtstat. Also meditate how the comparison can be fair wrt taffic patterns.
 Do smame with 2.6.

 Prepare for a oprofile. I'm lazy and compile the stuff I use into kernel if 
 you can get the same result w. nodules it's ok.


						  --ro

 

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-05 20:22               ` Robert Olsson
@ 2005-01-05 20:52                 ` Jeremy M. Guthrie
  2005-01-06 15:26                 ` Jeremy M. Guthrie
  1 sibling, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-05 20:52 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 1881 bytes --]

On Wednesday 05 January 2005 02:22 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > After smp_affinity adjustments and turning off IRQ balancing.
>  >
>  > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
>  >
>  >           RX packets:537271635 errors:2377048 dropped:2377048
>  > overruns:1849169
>
>  Worse?
>
>  Yes. I'll remember Hararld moved rt_cache_stat /proc/net/stats/ and wrote
> a new utility for it. Should be in iproute2 package.  Stephen knows the
> details. Otherwise change path in rtstat. You need this to relate and
> verify the load.
Hm...
ls -la /proc/net/stat/
total 0
dr-xr-xr-x  2 root root 0 Jan  5 14:47 .
dr-xr-xr-x  5 root root 0 Jan  5 11:50 ..
-r--r--r--  1 root root 0 Jan  5 14:47 arp_cache
-r--r--r--  1 root root 0 Jan  5 14:47 clip_arp_cache
-r--r--r--  1 root root 0 Jan  5 14:47 ndisc_cache
-r--r--r--  1 root root 0 Jan  5 14:47 rt_cache


>  Check throughtput and how packet load is used with 2.4.
> /proc/net/softnet_stat and rtstat. 
Will do.

> Also meditate how the comparison can be 
> fair wrt taffic patterns. Do smame with 2.6.
Not sure I follow.  I can see higher pps/byte through puts on switch port 
counters when I run with 2.4 vs 2.6.  I'll double check though.  I am the 
'dev' state so I can swap between V2.4 and V2.6 as necessary to compare 
during roughly equivalent times.  The only delay is the time between reboots.

>  Prepare for a oprofile. I'm lazy and compile the stuff I use into kernel
> if you can get the same result w. nodules it's ok.
Okay.

>
>
>         --ro

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-05 20:22               ` Robert Olsson
  2005-01-05 20:52                 ` Jeremy M. Guthrie
@ 2005-01-06 15:26                 ` Jeremy M. Guthrie
  2005-01-06 18:15                   ` Robert Olsson
  1 sibling, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-06 15:26 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 3954 bytes --]

On Wednesday 05 January 2005 02:22 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > After smp_affinity adjustments and turning off IRQ balancing.
>  >
>  > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
>  >
>  >           RX packets:537271635 errors:2377048 dropped:2377048
>  > overruns:1849169
>
>  Worse?
>
>  Yes. I'll remember Hararld moved rt_cache_stat /proc/net/stats/ and wrote
> a new utility for it. Should be in iproute2 package.  Stephen knows the
> details. Otherwise change path in rtstat. You need this to relate and
> verify the load.
I still don't see the rt_cache_stat file even under 2.4.28 stock.  See below.

>  Check throughtput and how packet load is used with 2.4.
> /proc/net/softnet_stat and rtstat. Also meditate how the comparison can be
> fair wrt taffic patterns. Do smame with 2.6.
wc -l /proc/net/rt_cache
Segmentation fault
cat /proc/net/rt_cache > /tmp/rt_cache ; wc -l /tmp/rt_cache
cat: write error: Bad address
57664 /tmp/rt_cache

ls -la /proc/net/
total 0
dr-xr-xr-x   6 root root 0 Jan  6 09:22 .
dr-xr-xr-x  55 root root 0 Jan  5 15:23 ..
-r--r--r--   1 root root 0 Jan  6 09:22 arp
dr-xr-xr-x   2 root root 0 Jan  6 09:22 atm
-r--r--r--   1 root root 0 Jan  6 09:22 dev
-r--r--r--   1 root root 0 Jan  6 09:22 dev_mcast
dr-xr-xr-x   2 root root 0 Jan  6 09:22 drivers
-r--r--r--   1 root root 0 Jan  6 09:22 igmp
-r--r--r--   1 root root 0 Jan  6 09:22 ip_mr_cache
-r--r--r--   1 root root 0 Jan  6 09:22 ip_mr_vif
-r--r--r--   1 root root 0 Jan  6 09:22 ip_queue
-r--r--r--   1 root root 0 Jan  6 09:22 ip_tables_matches
-r--r--r--   1 root root 0 Jan  6 09:22 ip_tables_names
-r--r--r--   1 root root 0 Jan  6 09:22 ip_tables_targets
-r--r--r--   1 root root 0 Jan  6 09:22 mcfilter
-r--r--r--   1 root root 0 Jan  6 09:22 netlink
-r--r--r--   1 root root 0 Jan  6 09:22 netstat
-r--r--r--   1 root root 0 Jan  6 09:22 pnp
-r--r--r--   1 root root 0 Jan  6 09:22 psched
-r--r--r--   1 root root 0 Jan  6 09:22 raw
-r--r--r--   1 root root 0 Jan  6 09:22 route
dr-xr-xr-x   2 root root 0 Jan  6 09:22 rpc
-r--r--r--   1 root root 0 Jan  6 09:22 rt_acct
-r--r--r--   1 root root 0 Jan  6 09:22 rt_cache
-r--r--r--   1 root root 0 Jan  6 09:22 snmp
-r--r--r--   1 root root 0 Jan  6 09:22 sockstat
-r--r--r--   1 root root 0 Jan  6 09:22 softnet_stat
dr-xr-xr-x   2 root root 0 Jan  6 09:22 stat
-r--r--r--   1 root root 0 Jan  6 09:22 tcp
-r--r--r--   1 root root 0 Jan  6 09:22 udp
-r--r--r--   1 root root 0 Jan  6 09:22 unix
-r--r--r--   1 root root 0 Jan  6 09:22 wireless
ls -la /proc/net/stat/
total 0
dr-xr-xr-x  2 root root 0 Jan  6 09:22 .
dr-xr-xr-x  6 root root 0 Jan  6 09:22 ..
-r--r--r--  1 root root 0 Jan  6 09:22 arp_cache
-r--r--r--  1 root root 0 Jan  6 09:22 clip_arp_cache
-r--r--r--  1 root root 0 Jan  6 09:22 rt_cache
cat /proc/net/softnet_stat
96deb140 0032844e 00012bbb 00000fd8 00000000 00000000 00000000 00000000 
00000dd4
00013dda 00000000 00000000 00000000 00000000 00000000 00000000 00000000 
0000041f
cat /proc/interrupts
           CPU0       CPU1
  0:     821852    5664906    IO-APIC-edge  timer
  1:        484       1247    IO-APIC-edge  keyboard
  8:          0          2    IO-APIC-edge  rtc
 14:          5         16    IO-APIC-edge  ide0
 18: 1928608529       1374   IO-APIC-level  eth3
 20:          1  226113098   IO-APIC-level  eth2
 27:       7950      34312   IO-APIC-level  eth0
 28:       2718       8097   IO-APIC-level  aic7xxx
 30:          0          0   IO-APIC-level  acpi
NMI:          0          0
LOC:    6487240    6487239
ERR:          0
MIS:          0


-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-06 15:26                 ` Jeremy M. Guthrie
@ 2005-01-06 18:15                   ` Robert Olsson
  2005-01-06 19:35                     ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-06 18:15 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger


Jeremy M. Guthrie writes:

 > I still don't see the rt_cache_stat file even under 2.4.28 stock.  See below.

 > ls -la /proc/net/stat/
 > total 0
 > -r--r--r--  1 root root 0 Jan  6 09:22 rt_cache


 rtstat is replaced by lnstat which is in iproute2 package:
 http://developer.osdl.org/dev/iproute2/download/

 (Old version  ftp://robur.slu.se/pub/Linux/net-development/rt_cache_stat/rtstat.c)
 

 > cat /proc/net/softnet_stat
 > 96deb140 0032844e 00012bbb 00000fd8 00000000 00000000 00000000 00000000 
 > 00000dd4
 > 00013dda 00000000 00000000 00000000 00000000 00000000 00000000 00000000 
 > 0000041f


 You only use CPU0 for packet processing. Also it seems you use  the non-NAPI 
 version of e1000.

 > cat /proc/interrupts
 >            CPU0       CPU1
 >   0:     821852    5664906    IO-APIC-edge  timer
 >   1:        484       1247    IO-APIC-edge  keyboard
 >   8:          0          2    IO-APIC-edge  rtc
 >  14:          5         16    IO-APIC-edge  ide0
 >  18: 1928608529       1374   IO-APIC-level  eth3
 >  20:          1  226113098   IO-APIC-level  eth2
 >  27:       7950      34312   IO-APIC-level  eth0


 Traffic is flowing from eth3->eth2 Why all the interrupts on eth2/CPU1? 
 Is traffic most unidirectional?


						--ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-06 18:15                   ` Robert Olsson
@ 2005-01-06 19:35                     ` Jeremy M. Guthrie
  2005-01-06 20:29                       ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-06 19:35 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger


[-- Attachment #1.1: Type: text/plain, Size: 1938 bytes --]

On Thursday 06 January 2005 12:15 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > I still don't see the rt_cache_stat file even under 2.4.28 stock.  See
>  > below.
>  >
>  > ls -la /proc/net/stat/
>  > total 0
>  > -r--r--r--  1 root root 0 Jan  6 09:22 rt_cache
>
>  rtstat is replaced by lnstat which is in iproute2 package:
>  http://developer.osdl.org/dev/iproute2/download/
>
>  (Old version 
> ftp://robur.slu.se/pub/Linux/net-development/rt_cache_stat/rtstat.c)
Please see the attachment.

>  > cat /proc/net/softnet_stat
>  > 96deb140 0032844e 00012bbb 00000fd8 00000000 00000000 00000000 00000000
>  > 00000dd4
>  > 00013dda 00000000 00000000 00000000 00000000 00000000 00000000 00000000
>  > 0000041f
>
>  You only use CPU0 for packet processing. Also it seems you use  the
> non-NAPI version of e1000.
The E1000 driver is the stock driver in 2.4.28.

>
>  > cat /proc/interrupts
>  >            CPU0       CPU1
>  >   0:     821852    5664906    IO-APIC-edge  timer
>  >   1:        484       1247    IO-APIC-edge  keyboard
>  >   8:          0          2    IO-APIC-edge  rtc
>  >  14:          5         16    IO-APIC-edge  ide0
>  >  18: 1928608529       1374   IO-APIC-level  eth3
>  >  20:          1  226113098   IO-APIC-level  eth2
>  >  27:       7950      34312   IO-APIC-level  eth0
>
>  Traffic is flowing from eth3->eth2 Why all the interrupts on eth2/CPU1?
>  Is traffic most unidirectional?
eth2 is TX only.  We don't receive anything on it.  This system should only 
ever RX on eth3 and TX on eth2 as part of its function.  Eth0 is the 
management interface on the host.

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #1.2: results-2.4.28.txt --]
[-- Type: text/plain, Size: 26640 bytes --]

arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|
 entries|  allocs|destroys|hash_gro| lookups|    hits|res_fail|rcv_prob|rcv_prob|periodic|forced_g|forced_g| entries|  in_hit|in_slow_|in_slow_|in_no_ro|  in_brd|in_marti|in_marti| out_hit|out_slow|out_slow|gc_total|gc_ignor|gc_goal_|gc_dst_o|in_hlist|out_hlis| entries|  allocs|destroys|hash_gro| lookups|    hits|res_fail|rcv_prob|rcv_prob|periodic|forced_g|forced_g|
        |        |        |      ws|        |        |      ed|es_mcast|es_ucast|_gc_runs|  c_runs|c_goal_m|        |        |     tot|      mc|     ute|        |  an_dst|  an_src|        |    _tot|     _mc|        |      ed|    miss| verflow| _search|t_search|        |        |        |      ws|        |        |      ed|es_mcast|es_ucast|_gc_runs|  c_runs|c_goal_m|
        |        |        |        |        |        |        |        |        |        |        |     iss|        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |     iss|
      22|       9|     425|       1|   17202|   14599|       0|       0|       0|   75743|       0|       0|   61129|    6329|   20428|       0|       0|    1539|       0|       0|       8|      28|       0|    1677|    1675|       0|       0|  143038|     145|       0|       0|       0|       0|       0|       0|       0|       0|       0|    4649|       0|       0|
      22|       0|       0|       0|       1|       1|       0|       0|       0|       2|       0|       0|   62833|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      10|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       1|       1|       0|       0|       0|       1|       0|       0|   64306|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   65602|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   59707|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   60684|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   62249|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   62249|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|       0|   65000|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   58800|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   58800|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|       0|   62135|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   63677|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   63677|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|       0|   59013|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|       0|   59013|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      24|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|   62462|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      24|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|   62462|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      24|       0|       0|       0|       1|       0|       0|       0|       0|       4|       0|       0|   65225|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      24|       0|       0|       0|       1|       0|       0|       0|       0|       4|       0|       0|   65225|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|
 entries|  allocs|destroys|hash_gro| lookups|    hits|res_fail|rcv_prob|rcv_prob|periodic|forced_g|forced_g| entries|  in_hit|in_slow_|in_slow_|in_no_ro|  in_brd|in_marti|in_marti| out_hit|out_slow|out_slow|gc_total|gc_ignor|gc_goal_|gc_dst_o|in_hlist|out_hlis| entries|  allocs|destroys|hash_gro| lookups|    hits|res_fail|rcv_prob|rcv_prob|periodic|forced_g|forced_g|
        |        |        |      ws|        |        |      ed|es_mcast|es_ucast|_gc_runs|  c_runs|c_goal_m|        |        |     tot|      mc|     ute|        |  an_dst|  an_src|        |    _tot|     _mc|        |      ed|    miss| verflow| _search|t_search|        |        |        |      ws|        |        |      ed|es_mcast|es_ucast|_gc_runs|  c_runs|c_goal_m|
        |        |        |        |        |        |        |        |        |        |        |     iss|        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |     iss|
      24|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|   60904|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      24|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|   60904|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|       0|   63890|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|       0|   63890|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|   59738|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|   59738|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       1|       0|       0|       0|       4|       0|       0|   63346|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      13|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       1|       0|       0|       0|       4|       0|       0|   63346|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      13|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       0|       0|       0|       0|       4|       0|       0|   58798|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      13|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       0|       0|       0|       0|       4|       0|       0|   58798|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      13|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      24|       0|       0|       0|       1|       1|       0|       0|       0|       4|       0|       0|   62518|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      13|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      24|       0|       0|       0|       1|       1|       0|       0|       0|       4|       0|       0|   62518|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      13|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       2|       2|       0|       0|       0|       3|       0|       0|   65364|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      12|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       2|       2|       0|       0|       0|       3|       0|       0|   65364|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      12|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       2|       2|       0|       0|       0|       4|       0|       0|   61111|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      11|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       2|       2|       0|       0|       0|       4|       0|       0|   61111|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      11|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       1|       0|       0|       0|       4|       0|       0|   63152|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       6|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       1|       0|       0|       0|       4|       0|       0|   63152|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       6|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       1|       0|       0|       0|       4|       0|       0|   58223|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       1|       0|       0|       0|       4|       0|       0|   58223|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|arp_cach|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|clip_arp|
 entries|  allocs|destroys|hash_gro| lookups|    hits|res_fail|rcv_prob|rcv_prob|periodic|forced_g|forced_g| entries|  in_hit|in_slow_|in_slow_|in_no_ro|  in_brd|in_marti|in_marti| out_hit|out_slow|out_slow|gc_total|gc_ignor|gc_goal_|gc_dst_o|in_hlist|out_hlis| entries|  allocs|destroys|hash_gro| lookups|    hits|res_fail|rcv_prob|rcv_prob|periodic|forced_g|forced_g|
        |        |        |      ws|        |        |      ed|es_mcast|es_ucast|_gc_runs|  c_runs|c_goal_m|        |        |     tot|      mc|     ute|        |  an_dst|  an_src|        |    _tot|     _mc|        |      ed|    miss| verflow| _search|t_search|        |        |        |      ws|        |        |      ed|es_mcast|es_ucast|_gc_runs|  c_runs|c_goal_m|
        |        |        |        |        |        |        |        |        |        |        |     iss|        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |        |     iss|
      24|       0|       0|       0|       2|       2|       0|       0|       0|       4|       0|       0|   62044|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      15|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      24|       0|       0|       0|       2|       2|       0|       0|       0|       4|       0|       0|   62044|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      15|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      24|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|       0|   64932|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      24|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|       0|   64932|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      25|       0|       0|       0|       2|       2|       0|       0|       0|       4|       0|       0|   60712|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      14|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       1|       0|       1|       1|       0|       0|       0|       2|       0|       0|   62304|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      10|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   63758|       1|       1|       0|       0|       1|       0|       0|       0|       0|       0|       1|       1|       0|       0|       8|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       1|       0|       0|   64967|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       1|       0|       0|       0|       2|       0|       0|   58931|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   60836|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       6|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       1|       0|       0|       0|       2|       0|       0|   62574|       1|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      11|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   64121|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       1|       0|       0|       0|       1|       0|       0|   65615|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      10|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       1|       0|       0|       0|       2|       0|       0|   60057|       1|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      11|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      23|       0|       0|       0|       1|       1|       0|       0|       0|       1|       0|       0|   62008|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       1|       1|       0|       0|       0|       1|       0|       0|   63651|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       1|       1|       0|       0|       0|       1|       0|       0|   63651|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      22|       0|       0|       0|       1|       1|       0|       0|       0|       1|       0|       0|   63651|       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      25|       0|       0|       0|       2|       2|       0|       0|       0|       6|       0|       0|   61300|       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      12|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
      24|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|       0|   63087|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-06 19:35                     ` Jeremy M. Guthrie
@ 2005-01-06 20:29                       ` Robert Olsson
  2005-01-06 20:54                         ` Jeremy M. Guthrie
                                           ` (2 more replies)
  0 siblings, 3 replies; 88+ messages in thread
From: Robert Olsson @ 2005-01-06 20:29 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger


Jeremy M. Guthrie writes:

 > >  You only use CPU0 for packet processing. Also it seems you use  the
 > > non-NAPI version of e1000.
 > The E1000 driver is the stock driver in 2.4.28.


 There is a kernel config option.  Use Rx Polling (NAPI)
 
 > eth2 is TX only.  We don't receive anything on it.  This system should only 
 > ever RX on eth3 and TX on eth2 as part of its function.  

 Ok! So unidirectional traffic and CPU0 processes all skb's and passes them 
 over to CPU1 which touches and frees them. You could try some other 
 affinity setting later on but there are so many options depending what 
 we want do. If we just want to compare 2.4/2.6 we can start out simple 
 with setting affinity so only one CPU. Which is almost what you got.
 We reach all the complicated problems immediately... Now we pass skb's
 between the CPU's which release cache bouncing and makes slab rebalance it's
 pools. So using just one CPU is probably better... If we had incoming pkts
 on eth2 eventually we could have had any use for CPU1.

 Install NAPI-driver, set affinity so eth2/eth3 uses same CPU to start with.

 The stats info you sent was almost impossible to read. See if you can get 
 only the rtstats.

						--ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-06 20:29                       ` Robert Olsson
@ 2005-01-06 20:54                         ` Jeremy M. Guthrie
  2005-01-06 20:55                         ` Jeremy M. Guthrie
  2005-01-06 21:19                         ` Jeremy M. Guthrie
  2 siblings, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-06 20:54 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 1969 bytes --]

On Thursday 06 January 2005 02:29 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > >  You only use CPU0 for packet processing. Also it seems you use  the
>  > > non-NAPI version of e1000.
>  >
>  > The E1000 driver is the stock driver in 2.4.28.
>  There is a kernel config option.  Use Rx Polling (NAPI)
I'm recompiling the module and will reload the module.

>  > eth2 is TX only.  We don't receive anything on it.  This system should
>  > only ever RX on eth3 and TX on eth2 as part of its function.
>
>  Ok! So unidirectional traffic and CPU0 processes all skb's and passes them
>  over to CPU1 which touches and frees them. You could try some other
>  affinity setting later on but there are so many options depending what
>  we want do. If we just want to compare 2.4/2.6 we can start out simple
>  with setting affinity so only one CPU. Which is almost what you got.
>  We reach all the complicated problems immediately... Now we pass skb's
>  between the CPU's which release cache bouncing and makes slab rebalance
> it's pools. So using just one CPU is probably better... If we had incoming
> pkts on eth2 eventually we could have had any use for CPU1.
>
>  Install NAPI-driver, set affinity so eth2/eth3 uses same CPU to start
> with.
I installed the NAPI driver and now I drop packets under 2.4.28.  I tried 
setting eth2/3 onto the same CPU.  If both eth2/3 are on CPU0, I drop a lot 
of packets.  If CPU0-eth2 & CPU1-eth3, I still drop packets, just not as 
fast.

I am going to try V2.6 w/o NAPI to see how much that helps.

>  The stats info you sent was almost impossible to read. See if you can get
>  only the rtstats.


-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-06 20:29                       ` Robert Olsson
  2005-01-06 20:54                         ` Jeremy M. Guthrie
@ 2005-01-06 20:55                         ` Jeremy M. Guthrie
  2005-01-06 21:19                         ` Jeremy M. Guthrie
  2 siblings, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-06 20:55 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger


[-- Attachment #1.1: Type: text/plain, Size: 1664 bytes --]

Forgot the attachment.  


On Thursday 06 January 2005 02:29 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > >  You only use CPU0 for packet processing. Also it seems you use  the
>  > > non-NAPI version of e1000.
>  >
>  > The E1000 driver is the stock driver in 2.4.28.
>
>  There is a kernel config option.  Use Rx Polling (NAPI)
>
>  > eth2 is TX only.  We don't receive anything on it.  This system should
>  > only ever RX on eth3 and TX on eth2 as part of its function.
>
>  Ok! So unidirectional traffic and CPU0 processes all skb's and passes them
>  over to CPU1 which touches and frees them. You could try some other
>  affinity setting later on but there are so many options depending what
>  we want do. If we just want to compare 2.4/2.6 we can start out simple
>  with setting affinity so only one CPU. Which is almost what you got.
>  We reach all the complicated problems immediately... Now we pass skb's
>  between the CPU's which release cache bouncing and makes slab rebalance
> it's pools. So using just one CPU is probably better... If we had incoming
> pkts on eth2 eventually we could have had any use for CPU1.
>
>  Install NAPI-driver, set affinity so eth2/eth3 uses same CPU to start
> with.
>
>  The stats info you sent was almost impossible to read. See if you can get
>  only the rtstats.
>
>       --ro

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #1.2: results-2.4.28.txt --]
[-- Type: text/plain, Size: 10005 bytes --]

rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|
  in_hit|in_slow_|in_slow_|in_no_ro|  in_brd|in_marti|in_marti| out_hit|out_slow|out_slow|gc_total|gc_ignor|gc_goal_|gc_dst_o|in_hlist|out_hlis|
        |     tot|      mc|     ute|        |  an_dst|  an_src|        |    _tot|     _mc|        |      ed|    miss| verflow| _search|t_search|
    9092|   21888|       0|       0|    1656|       0|       0|       9|      42|       0|    1836|    1833|       0|       0|  158622|     206|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       1|       1|       0|       0|       9|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       1|       1|       0|       0|       9|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      14|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      14|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|
  in_hit|in_slow_|in_slow_|in_no_ro|  in_brd|in_marti|in_marti| out_hit|out_slow|out_slow|gc_total|gc_ignor|gc_goal_|gc_dst_o|in_hlist|out_hlis|
        |     tot|      mc|     ute|        |  an_dst|  an_src|        |    _tot|     _mc|        |      ed|    miss| verflow| _search|t_search|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       9|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       9|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       9|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       9|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       9|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       9|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       9|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       9|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       9|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|
       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       1|       1|       0|       0|      14|       0|
       0|       2|       0|       0|       0|       0|       0|       0|       0|       0|       1|       1|       0|       0|      14|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       7|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       3|       0|
rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|rt_cache|
  in_hit|in_slow_|in_slow_|in_no_ro|  in_brd|in_marti|in_marti| out_hit|out_slow|out_slow|gc_total|gc_ignor|gc_goal_|gc_dst_o|in_hlist|out_hlis|
        |     tot|      mc|     ute|        |  an_dst|  an_src|        |    _tot|     _mc|        |      ed|    miss| verflow| _search|t_search|
       1|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      16|       0|
       1|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      16|       0|
       1|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      16|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       8|       0|
       0|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       8|       0|
       2|       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|      16|       0|
       1|       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       8|       0|
       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       6|       0|
       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       6|       0|
       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       6|       0|
       2|       2|       0|       0|       1|       0|       0|       0|       0|       0|       1|       1|       0|       0|      24|       0|
       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       2|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
       1|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|
       0|       2|       0|       0|       1|       0|       0|       0|       0|       0|       1|       1|       0|       0|      16|       0|
       0|       2|       0|       0|       1|       0|       0|       0|       0|       0|       1|       1|       0|       0|      16|       0|
       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|
       2|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       4|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|
       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|       0|

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-06 20:29                       ` Robert Olsson
  2005-01-06 20:54                         ` Jeremy M. Guthrie
  2005-01-06 20:55                         ` Jeremy M. Guthrie
@ 2005-01-06 21:19                         ` Jeremy M. Guthrie
  2005-01-06 21:36                           ` Robert Olsson
  2 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-06 21:19 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 2005 bytes --]

Pulling Rx Polling has kept V2.6 from dropping packets.  argh?!?!?!  I should 
have caught that sooner.   8(  Well, I have a window of opportunity here.  If 
you want I can still provide a broken environment but that seems pointless 
while NAPI is not acting right... what is evident is that this will run 
better till we start to bump up against the top of CPU0.

On Thursday 06 January 2005 02:29 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > >  You only use CPU0 for packet processing. Also it seems you use  the
>  > > non-NAPI version of e1000.
>  >
>  > The E1000 driver is the stock driver in 2.4.28.
>
>  There is a kernel config option.  Use Rx Polling (NAPI)
>
>  > eth2 is TX only.  We don't receive anything on it.  This system should
>  > only ever RX on eth3 and TX on eth2 as part of its function.
>
>  Ok! So unidirectional traffic and CPU0 processes all skb's and passes them
>  over to CPU1 which touches and frees them. You could try some other
>  affinity setting later on but there are so many options depending what
>  we want do. If we just want to compare 2.4/2.6 we can start out simple
>  with setting affinity so only one CPU. Which is almost what you got.
>  We reach all the complicated problems immediately... Now we pass skb's
>  between the CPU's which release cache bouncing and makes slab rebalance
> it's pools. So using just one CPU is probably better... If we had incoming
> pkts on eth2 eventually we could have had any use for CPU1.
>
>  Install NAPI-driver, set affinity so eth2/eth3 uses same CPU to start
> with.
>
>  The stats info you sent was almost impossible to read. See if you can get
>  only the rtstats.
>
>       --ro

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-06 21:19                         ` Jeremy M. Guthrie
@ 2005-01-06 21:36                           ` Robert Olsson
  2005-01-06 21:46                             ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-06 21:36 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger


Jeremy M. Guthrie writes:

 > Pulling Rx Polling has kept V2.6 from dropping packets.  argh?!?!?!  

 Thats fine...

 > I should have caught that sooner.   8(  Well, I have a window of opportunity
 > here.  If you want I can still provide a broken environment but that seems 
 > pointless while NAPI is not acting right... what is evident is that this 
 > will run better till we start to bump up against the top of CPU0.
 
 I don't follow... 

 The stats didn't show any numbers so we don't know your load.


					       --ro
 

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-06 21:36                           ` Robert Olsson
@ 2005-01-06 21:46                             ` Jeremy M. Guthrie
  2005-01-06 22:11                               ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-06 21:46 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 1016 bytes --]

On Thursday 06 January 2005 03:36 pm, Robert Olsson wrote:
>  > I should have caught that sooner.   8(  Well, I have a window of
>  > opportunity here.  If you want I can still provide a broken environment
>  > but that seems pointless while NAPI is not acting right... what is
>  > evident is that this will run better till we start to bump up against
>  > the top of CPU0.
>
>  I don't follow...
Technically by turning off NAPI, I have 'solved' my short term packet loss 
problem.

However I can still provide you a window to go through further troubleshooting 
if it is beneficial.

>  The stats didn't show any numbers so we don't know your load.
Are you referring to the # of entries in the rt_cache table?

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-06 21:46                             ` Jeremy M. Guthrie
@ 2005-01-06 22:11                               ` Robert Olsson
  2005-01-06 22:18                                 ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-06 22:11 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger


Jeremy M. Guthrie writes:

 > >  I don't follow...
 > Technically by turning off NAPI, I have 'solved' my short term packet loss 
 > problem.

 You need NAPI of course just wait until your packet load increases just
 a little bit... I saw drops in your 2.4 /proc/net/softnet_stat which 
 indicates you are close to your system performance. With NAPI you keep
 up your system system performance regardless of incoming load. 

 The e1000 driver has some bugs in "your setup" as it enables irq's when
 there is only RX and no TX and should have irq's disabled. I sent patch 
 to intel.

					      --ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-06 22:11                               ` Robert Olsson
@ 2005-01-06 22:18                                 ` Jeremy M. Guthrie
  2005-01-06 22:35                                   ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-06 22:18 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 1292 bytes --]

On Thursday 06 January 2005 04:11 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > >  I don't follow...
>  >
>  > Technically by turning off NAPI, I have 'solved' my short term packet
>  > loss problem.
>
>  You need NAPI of course just wait until your packet load increases just
>  a little bit... 
I agree.  Running mpstat now shows that even w/o packet drops, I'm running 
2-5% free CPU.

>  I saw drops in your 2.4 /proc/net/softnet_stat which 
>  indicates you are close to your system performance. 
Should those drops show up in RX drops in ifconfig?

>  With NAPI you keep 
>  up your system system performance regardless of incoming load.
You mentioned before that " The stats didn't show any numbers so we don't know 
your load."  Was there a command you wanted me to re-run?

>  The e1000 driver has some bugs in "your setup" as it enables irq's when
>  there is only RX and no TX and should have irq's disabled. I sent patch
>  to intel.
>
>            --ro

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-06 22:18                                 ` Jeremy M. Guthrie
@ 2005-01-06 22:35                                   ` Robert Olsson
  2005-01-07 16:17                                     ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-06 22:35 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger


Jeremy M. Guthrie writes:

 > You mentioned before that " The stats didn't show any numbers so we don't 
 > know your load."  Was there a command you wanted me to re-run?

 rtstat should show the routing/packet load.

 From a systems like yours 2*933 MHz PIII in production for tens of thoundans
 of users many filter and full BGP routing. Current (late here) use.


 ifstat2 eth*
             RX --------------------------   TX -------------------------
eth0          272.8 M bit/s       51 k pps    350.7 M bit/s       51 k pps 
eth1          371.9 M bit/s       55 k pps    293.6 M bit/s       55 k pps 
eth2            6.7 M bit/s     1348   pps      3.0 M bit/s      991   pps 
eth3            472   bit/s        0   pps      600   bit/s        0   pps 

rtstat
 size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc GC: tot ignored goal_miss ovrf HASH: in_search out_search
21007    114060     748     0     5     0     0     0        11       5      0       0       0         0    0           58280          4
22683    112556     827     0     6     0     0     0         5       5      0       0       0         0    0           60841          7
24230    111083     765     0     4     0     0     0        13       4      0       0       0         0    0           66628          7


Around 110 kpps hitting warm cache entries and ~800 new lookups/sec. I was
expecting so see something similar from your system. 


FYI.
 cat /proc/net/softnet_stat 
9ba490f3 00000000 01281572 00000000 00000000 00000000 00000000 00000000 002562c2
9939268d 00000000 010e42e9 00000000 00000000 00000000 00000000 00000000 0028fe72


Good Night.


						    --ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-06 22:35                                   ` Robert Olsson
@ 2005-01-07 16:17                                     ` Jeremy M. Guthrie
  2005-01-07 19:18                                       ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-07 16:17 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 3111 bytes --]

On Thursday 06 January 2005 04:35 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > You mentioned before that " The stats didn't show any numbers so we
>  > don't know your load."  Was there a command you wanted me to re-run?
>
>  rtstat should show the routing/packet load.
>
>  From a systems like yours 2*933 MHz PIII in production for tens of
> thoundans of users many filter and full BGP routing. Current (late here)
> use.
>
>
>  ifstat2 eth*
>              RX --------------------------   TX -------------------------
> eth0          272.8 M bit/s       51 k pps    350.7 M bit/s       51 k pps
> eth1          371.9 M bit/s       55 k pps    293.6 M bit/s       55 k pps
> eth2            6.7 M bit/s     1348   pps      3.0 M bit/s      991   pps
> eth3            472   bit/s        0   pps      600   bit/s        0   pps
Ifstat output:
#kernel
Interface        RX Pkts/Rate    TX Pkts/Rate    RX Data/Rate    TX Data/Rate
                 RX Errs/Drop    TX Errs/Drop    RX Over/Rate    TX Coll/Rate
lo                     0 0             0 0             0 0             0 0
                       0 0             0 0             0 0             0 0
eth0                   1 0             1 0            66 0           178 0
                       0 0             0 0             0 0             0 0
eth2                   0 0         30770 0             0 0        13361K 0
                       0 0             0 0             0 0             0 0
eth3               81019 0             0 0        41740K 0             0 0
                       0 0             0 0             0 0             0 0

> rtstat
>  size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot    
> mc GC: tot ignored goal_miss ovrf HASH: in_search out_search 21007   
> 114060     748     0     5     0     0     0        11       5      0      
> 0       0         0    0           58280          4 22683    112556     827
>     0     6     0     0     0         5       5      0       0       0     
>    0    0           60841          7 24230    111083     765     0     4   
>  0     0     0        13       4      0       0       0         0    0     
>      66628          7
> Around 110 kpps hitting warm cache entries and ~800 new lookups/sec. I was
> expecting so see something similar from your system.
Did my second email w/ the lnstat data not make it?

>
> FYI.
>  cat /proc/net/softnet_stat


total    droppped tsquz    Throttl  FR_hit   FR_succe FR_defer FR_def_o 
 cpu_coll 
> 9ba490f3 00000000 01281572 00000000 00000000 00000000 00000000 00000000
> 002562c2 9939268d 00000000 010e42e9 00000000 00000000 00000000 00000000 
> 00000000 0028fe72
Why do these drops not show up in the interface drop?

> Good Night.
>
>
>           --ro

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 16:17                                     ` Jeremy M. Guthrie
@ 2005-01-07 19:18                                       ` Robert Olsson
  2005-01-07 19:38                                         ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-07 19:18 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger


Jeremy M. Guthrie writes:

 > Did my second email w/ the lnstat data not make it?

 I see most zeroes

 > >  cat /proc/net/softnet_stat
 > total    droppped tsquz    Throttl  FR_hit   FR_succe FR_defer FR_def_o 
 >  cpu_coll 
 > > 9ba490f3 00000000 01281572 00000000 00000000 00000000 00000000 00000000
 > > 002562c2 9939268d 00000000 010e42e9 00000000 00000000 00000000 00000000 
 > > 00000000 0028fe72

 > Why do these drops not show up in the interface drop?

 These are drops in the network stack not in the devices.


					     --ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 19:18                                       ` Robert Olsson
@ 2005-01-07 19:38                                         ` Jeremy M. Guthrie
  2005-01-07 20:07                                           ` Robert Olsson
  2005-01-07 20:14                                           ` Jeremy M. Guthrie
  0 siblings, 2 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-07 19:38 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 1043 bytes --]

On Friday 07 January 2005 01:18 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > Did my second email w/ the lnstat data not make it?
>  I see most zeroes
Is there a particular duration and count you want me to run for?  The snapshot 
should have been 60 one-second snapshots.  

>  > >  cat /proc/net/softnet_stat
>  >
>  > total    droppped tsquz    Throttl  FR_hit   FR_succe FR_defer FR_def_o
>  >  cpu_coll
>  >
>  > > 9ba490f3 00000000 01281572 00000000 00000000 00000000 00000000
>  > > 00000000 002562c2 9939268d 00000000 010e42e9 00000000 00000000
>  > > 00000000 00000000 00000000 0028fe72
>  >
>  > Why do these drops not show up in the interface drop?
>
>  These are drops in the network stack not in the devices.


-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 19:38                                         ` Jeremy M. Guthrie
@ 2005-01-07 20:07                                           ` Robert Olsson
  2005-01-07 20:14                                           ` Jeremy M. Guthrie
  1 sibling, 0 replies; 88+ messages in thread
From: Robert Olsson @ 2005-01-07 20:07 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger


Jeremy M. Guthrie writes:

 > Is there a particular duration and count you want me to run for?  
 > The snapshot should have been 60 one-second snapshots.  

 That should be OK as we would see one GC going and some flows recreated 
 after this. 

						--ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 19:38                                         ` Jeremy M. Guthrie
  2005-01-07 20:07                                           ` Robert Olsson
@ 2005-01-07 20:14                                           ` Jeremy M. Guthrie
  2005-01-07 20:40                                             ` Robert Olsson
  2005-01-07 22:28                                             ` Jesse Brandeburg
  1 sibling, 2 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-07 20:14 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 2638 bytes --]

I just updated the Intel Drivers to the latest on source-forge(5.6.10.1).  I 
now see lower CPU usage but I am still dropping.

During a 60 second window the machine received 5,110,164 packets and dropped 
20461 or roughly 0.4% packet loss.

cat /proc/net/softnet_stat
2f9c9bd0 150dc67c 00c4701b 000d2659 00000000 00000000 00000000 00000000 
00097049
00010f9b 00000000 0000003a 00000000 00000000 00000000 00000000 00000000 
000002fc
It has been at 150dc67c for a while now.  So while I am dropping at the card, 
I am not dropping in the stack.

mpstat -P ALL 2
Linux 2.6.10 (h-idspr-msn-1)    01/07/05

14:06:45     CPU   %user   %nice %system %iowait    %irq   %soft   %idle    
intr/s
14:06:47     all    0.00    0.00    0.00    0.75    1.74   43.03   54.48  
22598.51
14:06:47       0    0.00    0.00    0.00    0.00    2.49   78.61   18.91  
17570.65
14:06:47       1    0.00    0.00    0.00    1.49    1.49    7.46   90.05   
5023.88

14:06:47     CPU   %user   %nice %system %iowait    %irq   %soft   %idle    
intr/s
14:06:49     all    0.00    0.00    0.00    0.00    1.50   45.75   52.75  
19352.00
14:06:49       0    0.00    0.00    0.00    0.00    2.50   83.00   14.50  
14467.00
14:06:49       1    0.00    0.00    0.00    0.00    0.50    9.00   90.50   
4886.50

14:06:49     CPU   %user   %nice %system %iowait    %irq   %soft   %idle    
intr/s
14:06:51     all    0.00    0.00    0.00    0.00    1.75   42.50   55.75  
22482.59
14:06:51       0    0.00    0.00    0.00    0.00    2.49   77.61   19.90  
17572.64
14:06:51       1    0.00    0.00    0.50    0.00    1.00    6.97   91.54   
4919.90

14:06:51     CPU   %user   %nice %system %iowait    %irq   %soft   %idle    
intr/s
14:06:53     all    0.00    0.00    0.25    0.00    1.99   43.03   54.73  
22458.00
14:06:53       0    0.00    0.00    0.00    0.00    3.00   78.50   18.50  
17456.00
14:06:53       1    0.00    0.00    0.00    0.00    0.50    8.00   91.00   
4992.50

14:06:53     CPU   %user   %nice %system %iowait    %irq   %soft   %idle    
intr/s
14:06:55     all    0.00    0.00    0.00    0.00    1.75   42.75   55.50  
22854.00
14:06:55       0    0.00    0.00    0.00    0.00    3.00   77.00   20.00  
17838.00
14:06:55       1    0.00    0.00    0.00    0.00    1.00    8.50   91.00   
5012.50


-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 20:14                                           ` Jeremy M. Guthrie
@ 2005-01-07 20:40                                             ` Robert Olsson
  2005-01-07 21:06                                               ` Jeremy M. Guthrie
  2005-01-07 22:28                                             ` Jesse Brandeburg
  1 sibling, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-07 20:40 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger


Jeremy M. Guthrie writes:

 > During a 60 second window the machine received 5,110,164 packets and 
 > dropped 20461 or roughly 0.4% packet loss.

 Around 85 kpps. If you run rtstat we could a feeling how may slow-pathx
 that are taken. Or save the /proc/net/stat/rt_cache before you 60 sec
 run.

 mpstat I don't trust in this context.

 > It has been at 150dc67c for a while now.  So while I am dropping at the 
 > card, I am not dropping in the stack.

 You use NAPI driver then... 
 Check if the patch below is in your e1000 driver. 

				       --ro


 --- drivers/net/e1000/e1000_main.c.orig	2004-02-16 14:46:16.000000000 +0100
+++ drivers/net/e1000/e1000_main.c	2004-02-16 15:45:05.000000000 +0100
@@ -2161,19 +2161,21 @@
 	struct e1000_adapter *adapter = netdev->priv;
 	int work_to_do = min(*budget, netdev->quota);
 	int work_done = 0;
-	
-	e1000_clean_tx_irq(adapter);
+	static boolean_t tx_cleaned;	
+
+	tx_cleaned = e1000_clean_tx_irq(adapter);
 	e1000_clean_rx_irq(adapter, &work_done, work_to_do);
 
 	*budget -= work_done;
 	netdev->quota -= work_done;
 	
-	if(work_done < work_to_do || !netif_running(netdev)) {
+	if( (!tx_cleaned && (work_done == 0)) || !netif_running(netdev)) {
 		netif_rx_complete(netdev);
 		e1000_irq_enable(adapter);
+		return 0;
 	}
 
-	return (work_done >= work_to_do);
+	return 1;
 }
 #endif
 

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 20:40                                             ` Robert Olsson
@ 2005-01-07 21:06                                               ` Jeremy M. Guthrie
  2005-01-07 21:30                                                 ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-07 21:06 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 3962 bytes --]

On Friday 07 January 2005 02:40 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > During a 60 second window the machine received 5,110,164 packets and
>  > dropped 20461 or roughly 0.4% packet loss.
>
>  Around 85 kpps. If you run rtstat we could a feeling how may slow-pathx
>  that are taken. 
I can't run rstat.  The rt_cache_stat file doesn't exist in /proc/net 
or /proc/net/stat in either V2.4 or V2.6.

>  Or save the /proc/net/stat/rt_cache before you 60 sec 
>  run.
lnstat isn't giving me anything other than zeros.  

entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
gc_dst_overflow in_hlist_search out_hlist_search
0000f91e  15b1ddab 0b69a32c 00000000 00000000 00001b0a 00001fbd 00000000  
00066f65 0000b943 00000000 087b80a1 0878e81a 00000256 00000000 ca08c77d 
0020e549
0000f91e  00005c59 0000b3ef 00000000 00000000 00000cf7 00000000 00000000  
0000000f 0000004c 00000002 00000fb9 00000fb5 00000000 00000000 0004e8d9 
000000b2

60 seconds......

entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
gc_dst_overflow in_hlist_search out_hlist_search
0000fb3e  15ff4696 0b6bbb65 00000000 00000000 00001b10 00001fc1 00000000  
00067111 0000b974 00000000 087d9903 087b0003 00000256 00000000 cb58b399 
0020ee79
0000fb3e  00005c5b 0000b40d 00000000 00000000 00000cf9 00000000 00000000  
0000000f 0000004c 00000002 00000fbb 00000fb7 00000000 00000000 0004e9ba 
000000b2



>  mpstat I don't trust in this context.
>
>  > It has been at 150dc67c for a while now.  So while I am dropping at the
>  > card, I am not dropping in the stack.
>
>  You use NAPI driver then...
>  Check if the patch below is in your e1000 driver.
>            --ro
The drivers should be built as NAPI.

Here is the snippet:
static int
e1000_clean(struct net_device *netdev, int *budget)
{
        struct e1000_adapter *adapter = netdev->priv;
        int work_to_do = min(*budget, netdev->quota);
        int tx_cleaned;
        int work_done = 0;

        if (!netif_carrier_ok(netdev))
                goto quit_polling;

        tx_cleaned = e1000_clean_tx_irq(adapter);
        e1000_clean_rx_irq(adapter, &work_done, work_to_do);

        *budget -= work_done;
        netdev->quota -= work_done;

        /* if no Rx and Tx cleanup work was done, exit the polling mode */
        if(!tx_cleaned || (work_done < work_to_do) ||
                                !netif_running(netdev)) {
quit_polling:   netif_rx_complete(netdev);
                e1000_irq_enable(adapter);
                return 0;
        }

        return (work_done >= work_to_do);
}



>
>  --- drivers/net/e1000/e1000_main.c.orig 2004-02-16 14:46:16.000000000
> +0100 +++ drivers/net/e1000/e1000_main.c 2004-02-16 15:45:05.000000000
> +0100 @@ -2161,19 +2161,21 @@
>   struct e1000_adapter *adapter = netdev->priv;
>   int work_to_do = min(*budget, netdev->quota);
>   int work_done = 0;
> -
> - e1000_clean_tx_irq(adapter);
> + static boolean_t tx_cleaned;
> +
> + tx_cleaned = e1000_clean_tx_irq(adapter);
>   e1000_clean_rx_irq(adapter, &work_done, work_to_do);
>
>   *budget -= work_done;
>   netdev->quota -= work_done;
>
> - if(work_done < work_to_do || !netif_running(netdev)) {
> + if( (!tx_cleaned && (work_done == 0)) || !netif_running(netdev)) {
>    netif_rx_complete(netdev);
>    e1000_irq_enable(adapter);
> +  return 0;
>   }
>
> - return (work_done >= work_to_do);
> +	return 1;
>  }
>  #endif

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 21:06                                               ` Jeremy M. Guthrie
@ 2005-01-07 21:30                                                 ` Robert Olsson
  2005-01-11 15:11                                                   ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-07 21:30 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger


Jeremy M. Guthrie writes:
 > lnstat isn't giving me anything other than zeros.  

 Crap.

 > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
 > out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
 > gc_dst_overflow in_hlist_search out_hlist_search
 
 > 0000f91e  15b1ddab 0b69a32c 00000000 00000000 00001b0a 00001fbd 00000000  
 > 00066f65 0000b943 00000000 087b80a1 0878e81a 00000256 00000000 ca08c77d 
 > 0020e549
 
 
 > 0000fb3e  15ff4696 0b6bbb65 00000000 00000000 00001b10 00001fc1 00000000  
 > 00067111 0000b974 00000000 087d9903 087b0003 00000256 00000000 cb58b399 
 > 0020ee79
 
 85 kpps and 2287 lookups/sec as a 60 sec average. Pretty nice load yes. 
 Where did you get the load? 
 Try see if you fix lnstat :-) it would be nice to see the route dynamics. 
 Or use the try the rtstat I pointed to. 

 And apply the e1000 patch I sent and make a test run. 

 I'll have a beer and give up for tonight...

		  				--ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 20:14                                           ` Jeremy M. Guthrie
  2005-01-07 20:40                                             ` Robert Olsson
@ 2005-01-07 22:28                                             ` Jesse Brandeburg
  2005-01-07 22:50                                               ` Jeremy M. Guthrie
  1 sibling, 1 reply; 88+ messages in thread
From: Jesse Brandeburg @ 2005-01-07 22:28 UTC (permalink / raw)
  To: Jeremy M. Guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger

On Fri, 7 Jan 2005, Jeremy M. Guthrie wrote:
> I just updated the Intel Drivers to the latest on source-forge(5.6.10.1).  I 
> now see lower CPU usage but I am still dropping.
> 
> During a 60 second window the machine received 5,110,164 packets and dropped 
> 20461 or roughly 0.4% packet loss.
> 
NAPI will cause a very busy stack to force the network card to drop the 
data instead of the stack.  This is supposed to be a good thing because 
the hardware will be forced to drop the packet (hopefully) instead of 
interrupt rate thrashing your machine right when it needs the cpu to do 
other stuff.  This is supposed to be the saving grace of NAPI.

So, not to distract from the conversation, but in the interest (my
interest :-) ) of tuning your E1000, can you please send the output of
ethtool -S ethX for each interface?  This will help us figure out if there
is anything to tune in the driver (like number of rx buffers, etc)

Thanks,
  Jesse

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 22:28                                             ` Jesse Brandeburg
@ 2005-01-07 22:50                                               ` Jeremy M. Guthrie
  2005-01-07 22:57                                                 ` Stephen Hemminger
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-07 22:50 UTC (permalink / raw)
  To: netdev; +Cc: Jesse Brandeburg, Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 3369 bytes --]

On Friday 07 January 2005 04:28 pm, Jesse Brandeburg wrote:
> On Fri, 7 Jan 2005, Jeremy M. Guthrie wrote:
> > I just updated the Intel Drivers to the latest on source-forge(5.6.10.1).
> >  I now see lower CPU usage but I am still dropping.
> >
> > During a 60 second window the machine received 5,110,164 packets and
> > dropped 20461 or roughly 0.4% packet loss.
>
> NAPI will cause a very busy stack to force the network card to drop the
> data instead of the stack.  This is supposed to be a good thing because
> the hardware will be forced to drop the packet (hopefully) instead of
> interrupt rate thrashing your machine right when it needs the cpu to do
> other stuff.  This is supposed to be the saving grace of NAPI.
Makes sense.

> So, not to distract from the conversation, but in the interest (my
> interest :-) ) of tuning your E1000, can you please send the output of
> ethtool -S ethX for each interface?  This will help us figure out if there
> is anything to tune in the driver (like number of rx buffers, etc)
ethtool -S eth2
NIC statistics:
     rx_packets: 0
     tx_packets: 314550698
     rx_bytes: 0
     tx_bytes: 4290523139
     rx_errors: 0
     tx_errors: 0
     rx_dropped: 0
     tx_dropped: 0
     multicast: 0
     collisions: 0
     rx_length_errors: 0
     rx_over_errors: 0
     rx_crc_errors: 0
     rx_frame_errors: 0
     rx_fifo_errors: 0
     rx_missed_errors: 0
     tx_aborted_errors: 0
     tx_carrier_errors: 0
     tx_fifo_errors: 0
     tx_heartbeat_errors: 0
     tx_window_errors: 0
     tx_abort_late_coll: 0
     tx_deferred_ok: 0
     tx_single_coll_ok: 0
     tx_multi_coll_ok: 0
     rx_long_length_errors: 0
     rx_short_length_errors: 0
     rx_align_errors: 0
     tx_tcp_seg_good: 0
     tx_tcp_seg_failed: 0
     rx_flow_control_xon: 0
     rx_flow_control_xoff: 0
     tx_flow_control_xon: 0
     tx_flow_control_xoff: 0
     rx_long_byte_count: 0
     rx_csum_offload_good: 0
     rx_csum_offload_errors: 0
ethtool -S eth3
NIC statistics:
     rx_packets: 719847127
     tx_packets: 5
     rx_bytes: 1880301945
     tx_bytes: 398
     rx_errors: 3368295
     tx_errors: 0
     rx_dropped: 1478044
     tx_dropped: 0
     multicast: 0
     collisions: 0
     rx_length_errors: 0
     rx_over_errors: 0
     rx_crc_errors: 0
     rx_frame_errors: 0
     rx_fifo_errors: 1890251
     rx_missed_errors: 1890251
     tx_aborted_errors: 0
     tx_carrier_errors: 0
     tx_fifo_errors: 0
     tx_heartbeat_errors: 0
     tx_window_errors: 0
     tx_abort_late_coll: 0
     tx_deferred_ok: 0
     tx_single_coll_ok: 0
     tx_multi_coll_ok: 0
     rx_long_length_errors: 0
     rx_short_length_errors: 0
     rx_align_errors: 0
     tx_tcp_seg_good: 0
     tx_tcp_seg_failed: 0
     rx_flow_control_xon: 0
     rx_flow_control_xoff: 0
     tx_flow_control_xon: 0
     tx_flow_control_xoff: 0
     rx_long_byte_count: 379837423993
     rx_csum_offload_good: 672891990
     rx_csum_offload_errors: 178666



> Thanks,
>   Jesse

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 22:50                                               ` Jeremy M. Guthrie
@ 2005-01-07 22:57                                                 ` Stephen Hemminger
  2005-01-11 15:17                                                   ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Stephen Hemminger @ 2005-01-07 22:57 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Jesse Brandeburg, Robert Olsson

On Fri, 7 Jan 2005 16:50:51 -0600
"Jeremy M. Guthrie" <jeremy.guthrie@berbee.com> wrote:

> On Friday 07 January 2005 04:28 pm, Jesse Brandeburg wrote:
> > On Fri, 7 Jan 2005, Jeremy M. Guthrie wrote:
> > > I just updated the Intel Drivers to the latest on source-forge(5.6.10.1).
> > >  I now see lower CPU usage but I am still dropping.
> > >
> > > During a 60 second window the machine received 5,110,164 packets and
> > > dropped 20461 or roughly 0.4% packet loss.
> >
> > NAPI will cause a very busy stack to force the network card to drop the
> > data instead of the stack.  This is supposed to be a good thing because
> > the hardware will be forced to drop the packet (hopefully) instead of
> > interrupt rate thrashing your machine right when it needs the cpu to do
> > other stuff.  This is supposed to be the saving grace of NAPI.
> Makes sense.

Not sure if NAPI makes sense on transmit because it causes the transmit
ring to grow and freeing the transmit skb should be quick. Perhaps
other interrupt moderation schemes work better.

On receive the processing could take longer and NAPI is a real win.

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
       [not found] <200501071619.54566.jeremy.guthrie@berbee.com>
@ 2005-01-07 23:23 ` Jesse Brandeburg
  2005-01-10 21:11   ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Jesse Brandeburg @ 2005-01-07 23:23 UTC (permalink / raw)
  To: Jeremy M. Guthrie; +Cc: Brandeburg, Jesse, netdev, Robert.Olsson, shemminger

On Fri, 7 Jan 2005, Jeremy M. Guthrie wrote:
> > Very interesting that with your workload the napi driver doesn't keep up
> > (from looking at your posts in netdev) and yet the interrupt mode driver
> > can.
> Well, I need to do more digging.   One scenario the interrupt mode can hand 
> stuff off to the CPU but the network stack still drops.  The other scenario, 
> the card drops.  
> 
> ethtool -S eth2
> NIC statistics:
<snip!>
>      tx_packets: 314550698
>      tx_bytes: 4290523139
> ethtool -S eth3
> NIC statistics:
<snip!>
>      rx_packets: 719847127
>      tx_packets: 5
>      rx_bytes: 1880301945
>      tx_bytes: 398
>      rx_errors: 3368295
>      rx_dropped: 1478044
>      rx_fifo_errors: 1890251
>      rx_missed_errors: 1890251
>      rx_long_byte_count: 379837423993
>      rx_csum_offload_good: 672891990
>      rx_csum_offload_errors: 178666

Okay, well, rx_dropped means "receive no buffers count" in our driver, but
doesn't necessarily mean that the packet was completely dropped it just
means that the fifo may have had to queue up the packet on the adapter as
best it could and wait for the OS to provide more skb's, this may or may 
not lead to further errors...

Now, the rx_fifo errors is from hardware counter "missed packet count"
which means that a packet was dropped because the fifo was full (probably
indicated at least some of the time because the card was in "receive no
buffers" state) btw fifo errors and rx_missed are both being fed by the 
same hardware counter.

rx_csum_offload_errors is somewhat worrisome because it means that you're
receiving packets that appear to be corrupted.  This could be from any
number of sources, but is most likely a misconfigured station on your lan
or something is corrupting checksums (a tcpdump would help debug here, but
would really slow things down)  The packets are NOT dropped, but handed to
the stack to decide what to do.

So, to mitigate the rnbc "receive no buffers count" a little you can
provide some more buffering on the receive side by loading the module with
RxDescriptors=2048 or something larger than the default of 256.  this will 
help (if you haven't already) but will also probably increase the amount 
of work your system has to do, as more packets will be able to be stored 
up.

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 23:23 ` Jesse Brandeburg
@ 2005-01-10 21:11   ` Jeremy M. Guthrie
  0 siblings, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-10 21:11 UTC (permalink / raw)
  To: netdev; +Cc: Jesse Brandeburg, Robert.Olsson, shemminger

[-- Attachment #1: Type: text/plain, Size: 6206 bytes --]

I rebound the card with 2048 RxDescriptors.  There appears to be a burst 
errors right when the port comes online.  Down below is an ifconfig ; sleep 
60 ; ifconfig.

I'll be tesitng Robert's patch shortly.

ethtool -S eth2
NIC statistics:
     rx_packets: 0
     tx_packets: 7295976
     rx_bytes: 0
     tx_bytes: 3413882143
     rx_errors: 0
     tx_errors: 0
     rx_dropped: 0
     tx_dropped: 0
     multicast: 0
     collisions: 0
     rx_length_errors: 0
     rx_over_errors: 0
     rx_crc_errors: 0
     rx_frame_errors: 0
     rx_fifo_errors: 0
     rx_missed_errors: 0
     tx_aborted_errors: 0
     tx_carrier_errors: 0
     tx_fifo_errors: 0
     tx_heartbeat_errors: 0
     tx_window_errors: 0
     tx_abort_late_coll: 0
     tx_deferred_ok: 0
     tx_single_coll_ok: 0
     tx_multi_coll_ok: 0
     rx_long_length_errors: 0
     rx_short_length_errors: 0
     rx_align_errors: 0
     tx_tcp_seg_good: 0
     tx_tcp_seg_failed: 0
     rx_flow_control_xon: 0
     rx_flow_control_xoff: 0
     tx_flow_control_xon: 0
     tx_flow_control_xoff: 0
     rx_long_byte_count: 0
     rx_csum_offload_good: 0
     rx_csum_offload_errors: 0

ethtool -S eth3
NIC statistics:
     rx_packets: 19231078
     tx_packets: 5
     rx_bytes: 1350479823
     tx_bytes: 398
     rx_errors: 522159
     tx_errors: 0
     rx_dropped: 320198
     tx_dropped: 0
     multicast: 0
     collisions: 0
     rx_length_errors: 0
     rx_over_errors: 0
     rx_crc_errors: 0
     rx_frame_errors: 0
     rx_fifo_errors: 201961
     rx_missed_errors: 201961
     tx_aborted_errors: 0
     tx_carrier_errors: 0
     tx_fifo_errors: 0
     tx_heartbeat_errors: 0
     tx_window_errors: 0
     tx_abort_late_coll: 0
     tx_deferred_ok: 0
     tx_single_coll_ok: 0
     tx_multi_coll_ok: 0
     rx_long_length_errors: 0
     rx_short_length_errors: 0
     rx_align_errors: 0
     tx_tcp_seg_good: 0
     tx_tcp_seg_failed: 0
     rx_flow_control_xon: 0
     rx_flow_control_xoff: 0
     tx_flow_control_xon: 0
     tx_flow_control_xoff: 0
     rx_long_byte_count: 9940414415
     rx_csum_offload_good: 17879204
     rx_csum_offload_errors: 8537

Mon Jan 10 15:06:26 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
          inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:67183383 errors:583215 dropped:583215 overruns:247258 
frame:0
          TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:1120957705 (1069.0 Mb)  TX bytes:398 (398.0 b)
          Base address:0x22a0 Memory:eff80000-effa0000

Mon Jan 10 15:07:26 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
          inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:73275567 errors:616520 dropped:616520 overruns:262517 
frame:0
          TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:337437991 (321.8 Mb)  TX bytes:398 (398.0 b)
          Base address:0x22a0 Memory:eff80000-effa0000


On Friday 07 January 2005 05:23 pm, Jesse Brandeburg wrote:
> On Fri, 7 Jan 2005, Jeremy M. Guthrie wrote:
> > > Very interesting that with your workload the napi driver doesn't keep
> > > up (from looking at your posts in netdev) and yet the interrupt mode
> > > driver can.
> >
> > Well, I need to do more digging.   One scenario the interrupt mode can
> > hand stuff off to the CPU but the network stack still drops.  The other
> > scenario, the card drops.
> >
> > ethtool -S eth2
> > NIC statistics:
>
> <snip!>
>
> >      tx_packets: 314550698
> >      tx_bytes: 4290523139
> > ethtool -S eth3
> > NIC statistics:
>
> <snip!>
>
> >      rx_packets: 719847127
> >      tx_packets: 5
> >      rx_bytes: 1880301945
> >      tx_bytes: 398
> >      rx_errors: 3368295
> >      rx_dropped: 1478044
> >      rx_fifo_errors: 1890251
> >      rx_missed_errors: 1890251
> >      rx_long_byte_count: 379837423993
> >      rx_csum_offload_good: 672891990
> >      rx_csum_offload_errors: 178666
>
> Okay, well, rx_dropped means "receive no buffers count" in our driver, but
> doesn't necessarily mean that the packet was completely dropped it just
> means that the fifo may have had to queue up the packet on the adapter as
> best it could and wait for the OS to provide more skb's, this may or may
> not lead to further errors...
>
> Now, the rx_fifo errors is from hardware counter "missed packet count"
> which means that a packet was dropped because the fifo was full (probably
> indicated at least some of the time because the card was in "receive no
> buffers" state) btw fifo errors and rx_missed are both being fed by the
> same hardware counter.
>
> rx_csum_offload_errors is somewhat worrisome because it means that you're
> receiving packets that appear to be corrupted.  This could be from any
> number of sources, but is most likely a misconfigured station on your lan
> or something is corrupting checksums (a tcpdump would help debug here, but
> would really slow things down)  The packets are NOT dropped, but handed to
> the stack to decide what to do.
>
> So, to mitigate the rnbc "receive no buffers count" a little you can
> provide some more buffering on the receive side by loading the module with
> RxDescriptors=2048 or something larger than the default of 256.  this will
> help (if you haven't already) but will also probably increase the amount
> of work your system has to do, as more packets will be able to be stored
> up.

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 21:30                                                 ` Robert Olsson
@ 2005-01-11 15:11                                                   ` Jeremy M. Guthrie
  0 siblings, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-11 15:11 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger

[-- Attachment #1: Type: text/plain, Size: 799 bytes --]

On Friday 07 January 2005 03:30 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  85 kpps and 2287 lookups/sec as a 60 sec average. Pretty nice load yes.
>  Where did you get the load?
>  Try see if you fix lnstat :-) it would be nice to see the route dynamics.
>  Or use the try the rtstat I pointed to.
>
>  And apply the e1000 patch I sent and make a test run.
The code Jesse Brandeburg had me download has the patch in it.

>  I'll have a beer and give up for tonight...
>
>         --ro

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-07 22:57                                                 ` Stephen Hemminger
@ 2005-01-11 15:17                                                   ` Jeremy M. Guthrie
  2005-01-11 16:40                                                     ` Robert Olsson
  2005-01-11 17:17                                                     ` Jeremy M. Guthrie
  0 siblings, 2 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-11 15:17 UTC (permalink / raw)
  To: netdev; +Cc: Stephen Hemminger, Jesse Brandeburg, Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 3052 bytes --]

date ; ifconfig eth3 ; cat /proc/net/softnet_stat ; 
cat /proc/net/stat/rt_cache ; sleep 60 ; date ; ifconfig eth3 ; 
cat /proc/net/softnet_stat ; cat /proc/net/stat/rt_cache

Tue Jan 11 09:12:21 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
          inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:3519452697 errors:5558592 dropped:5558592 
overruns:4011523 frame:0
          TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:497775695 (474.7 Mb)  TX bytes:398 (398.0 b)
          Base address:0x22a0 Memory:eff80000-effa0000

f59427dc 150dc67c 00c562ab 000d2659 00000000 00000000 00000000 00000000 
00547b5b
00038622 00000000 00000065 00000000 00000000 00000000 00000000 00000000 
0006804f
entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
gc_dst_overflow in_hlist_search out_hlist_search
0000f3c2  aeca1565 272c383d 00000000 00000000 000089a8 00005298 00000002  
001db403 0003054f 00000000 230aa73c 22fe33eb 00000cf0 00000000 b37011dc 
00957c14
0000f3c2  0000b975 00029703 00000000 00000000 000035a5 00000000 00000000  
00000015 00000083 00000002 000038c8 000038a3 00000000 00000000 0012a566 
0000014d

Tue Jan 11 09:13:21 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
          inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:3524302561 errors:5571396 dropped:5571396 
overruns:4022383 frame:0
          TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:3065327250 (2923.3 Mb)  TX bytes:398 (398.0 b)
          Base address:0x22a0 Memory:eff80000-effa0000

f5de26f5 150dc67c 00c562b4 000d2659 00000000 00000000 00000000 00000000 
00547ffd
00038632 00000000 00000065 00000000 00000000 00000000 00000000 00000000 
0006807c
entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
gc_dst_overflow in_hlist_search out_hlist_search
0000fc1e  af11f07b 272e58d3 00000000 00000000 000089a8 00005298 00000002  
001db571 00030572 00000000 230cc7f1 23005428 00000cf0 00000000 b4b5b529 
009583b9
0000fc1e  0000b977 00029710 00000000 00000000 000035a7 00000000 00000000  
00000015 00000083 00000002 000038ca 000038a5 00000000 00000000 0012a5c8 
0000014d

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-11 15:17                                                   ` Jeremy M. Guthrie
@ 2005-01-11 16:40                                                     ` Robert Olsson
  2005-01-12  1:27                                                       ` Jeremy M. Guthrie
  2005-01-11 17:17                                                     ` Jeremy M. Guthrie
  1 sibling, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-11 16:40 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Stephen Hemminger, Jesse Brandeburg, Robert Olsson



About the same load as last time 80 kpps forwarded and droprate of ~200 pps 
and 2323 fib lookups/sec. Not so bad for PIII but it maybe tuned better.

Did the e1000 patch cure the problem that interrupts got enabled for 
unidirectional traffic? Ring size? I've never seen any win in system 
performance w. RX rings larger than 256 at least not in lab.

					   --ro

Jeremy M. Guthrie writes:
 > date ; ifconfig eth3 ; cat /proc/net/softnet_stat ; 
 > cat /proc/net/stat/rt_cache ; sleep 60 ; date ; ifconfig eth3 ; 
 > cat /proc/net/softnet_stat ; cat /proc/net/stat/rt_cache
 > 
 > Tue Jan 11 09:12:21 CST 2005
 > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
 >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
 >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
 >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
 >           RX packets:3519452697 errors:5558592 dropped:5558592 
 > overruns:4011523 frame:0
 >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
 >           collisions:0 txqueuelen:1000
 >           RX bytes:497775695 (474.7 Mb)  TX bytes:398 (398.0 b)
 >           Base address:0x22a0 Memory:eff80000-effa0000
 > 
 > f59427dc 150dc67c 00c562ab 000d2659 00000000 00000000 00000000 00000000 
 > 00547b5b
 > 00038622 00000000 00000065 00000000 00000000 00000000 00000000 00000000 
 > 0006804f
 > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
 > out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
 > gc_dst_overflow in_hlist_search out_hlist_search
 > 0000f3c2  aeca1565 272c383d 00000000 00000000 000089a8 00005298 00000002  
 > 001db403 0003054f 00000000 230aa73c 22fe33eb 00000cf0 00000000 b37011dc 
 > 00957c14
 > 0000f3c2  0000b975 00029703 00000000 00000000 000035a5 00000000 00000000  
 > 00000015 00000083 00000002 000038c8 000038a3 00000000 00000000 0012a566 
 > 0000014d
 > 
 > Tue Jan 11 09:13:21 CST 2005
 > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
 >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
 >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
 >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
 >           RX packets:3524302561 errors:5571396 dropped:5571396 
 > overruns:4022383 frame:0
 >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
 >           collisions:0 txqueuelen:1000
 >           RX bytes:3065327250 (2923.3 Mb)  TX bytes:398 (398.0 b)
 >           Base address:0x22a0 Memory:eff80000-effa0000
 > 
 > f5de26f5 150dc67c 00c562b4 000d2659 00000000 00000000 00000000 00000000 
 > 00547ffd
 > 00038632 00000000 00000065 00000000 00000000 00000000 00000000 00000000 
 > 0006807c
 > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
 > out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
 > gc_dst_overflow in_hlist_search out_hlist_search
 > 0000fc1e  af11f07b 272e58d3 00000000 00000000 000089a8 00005298 00000002  
 > 001db571 00030572 00000000 230cc7f1 23005428 00000cf0 00000000 b4b5b529 
 > 009583b9
 > 0000fc1e  0000b977 00029710 00000000 00000000 000035a7 00000000 00000000  
 > 00000015 00000083 00000002 000038ca 000038a5 00000000 00000000 0012a5c8 
 > 0000014d
 > 
 > -- 
 > 
 > --------------------------------------------------
 > Jeremy M. Guthrie        jeremy.guthrie@berbee.com
 > Senior Network Engineer        Phone: 608-298-1061
 > Berbee                           Fax: 608-288-3007
 > 5520 Research Park Drive         NOC: 608-298-1102
 > Madison, WI 53711

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-11 15:17                                                   ` Jeremy M. Guthrie
  2005-01-11 16:40                                                     ` Robert Olsson
@ 2005-01-11 17:17                                                     ` Jeremy M. Guthrie
  2005-01-11 18:46                                                       ` Robert Olsson
  1 sibling, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-11 17:17 UTC (permalink / raw)
  To: netdev; +Cc: Stephen Hemminger, Jesse Brandeburg, Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 8723 bytes --]

 ./rtstat
 size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
GC: tot ignored goal_miss ovrf HASH: in_search out_search
63375     86092    2844     0     0     0     0     0         5       1 
1610369017    2846    2844         0    0          386039         17
64870     84431 537031471 537492192     0     0     0     0         1       0      
0    2848    2846         0    0          386130         10
58707     87889    2715     0     0     0     0     0         7       1      0    
2716    2714         0    0          384782         13
60617     89993    2796     0     0     0     0     0         5       0      0    
2795    2793         0    0          384432         17
62458     93221    2772     0     0     0     0     0         3       0      0    
2772    2770         0    0          399579          9
63990     93792    2712     0     0     0     0     0         6       0      0    
2713    2711         0    0          422585         12
65499     94235    2846     0     0     0     0     0         2       0      0    
2846    2844         0    0          424341          9
59755     90992    2864     0     0     0     0     0         1       1      0    
2865    2863         0    0          385888         13
61702     87681    2798     0     0     0     0     0         6       2      0    
2800    2798         0    0          383276         34
63334     90557    2734     0     0     0     0     0         5       1      0    
2735    2733         0    0          422121         28
64798     88178    2732     0     0     0     0     0         1       0      0    
2732    2730         0    0          410772         11
58574     85910    2669     0     0     0     0     0         3       0      0    
2670    2668         0    0          393331         10
60538     87238    2705     0     0     0     0     0         4       0      0    
2705    2703         0    0          382161         20
62164     85738    2628     0     0     0     0     0         4       0      0    
2628    2626         0    0          373049         25
63759     78752    2638     0     0     0     0     0         2       0      0    
2638    2636         0    0          348411          5
65052     79832    2482     0     0     0     0     0         3       0      0    
2481    2479         0    0          365871         20
58986     82611    2549     0     0     0     0     0         4       1      0    
2550    2548         0    0          366429         24
60845     85340    2620     0     0     1     0     0         2       0      0    
2620    2618         0    0          369963         16
62465     83122    2591     0     0     0     0     0         5       0      0    
2591    2589         0    0          371261         18
 size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
GC: tot ignored goal_miss ovrf HASH: in_search out_search
63998     86141    2640     0     0     0     0     0         4       0 
537114631    2640    2638         0    0          386452         15
65294     85989    2527     0     0     0     0     0         5       1 
1610369017    2528    2526         0    0          379801         26
59241     84305    2662     0     0     0     0     0         7       0      0    
2662    2660         0    0          371925         35
61194     85397    2837     0     0     0     0     0         3       0      0    
2837    2835         0    0          374788         17
62908     88133    2688     0     0     1     0     0         4       1      0    
2689    2687         0    0          382411         19
64329     88495    2648     0     0     0     0     0         5       1      0    
2649    2647         0    0          398763         30
65560     84994    2447     0     0     1     0     0         6       2      0    
2449    2447         0    0          394745         33
59677     81478    2568     0     0     0     0     0         5       0      0    
2569    2567         0    0          341439         22
61459     81225    2623     0     0     0     0     0         5       0      0    
2623    2621         0    0          345996         30
63038     83660    2544     0     0     0     0     0         3       0      0    
2544    2542         0    0          381567         21
64494     86321    2620     0     0     0     0     0         4       0      0    
2620    2618         0    0          395278         19
65528     88754    2766     0     0     0     0     0         6       0      0    
2765    2763         0    0          421075         34
59948     86443    2990     0     0     0     0     0         5       0      0    
2990    2988         0    0          373605         29
61860     81912    2786     0     0     0     0     0         5       1      0    
2787    2785         0    0          354678         34
63630     80399    2981     0     0     0     0     0         4       0      0    
2979    2977         0    0          354983         32
65215     83600    2934     0     0     0     0     0         6       3      0    
2937    2935         0    0          371497         41
59322     85583    2821     0     0     0     0     0         7       0      0    
2822    2820         0    0          382896         40
61268     82602    2817     0     0     0     0     0         9       2      0    
2819    2817         0    0          368078         54
62989     85336    2760     0     0     1     0     0         9       0      0    
2760    2758         0    0          394212         39
64387     84861    2603     0     0     0     0     0         7       0      0    
2602    2600         0    0          390852         36
 size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
GC: tot ignored goal_miss ovrf HASH: in_search out_search
65681     82848    2625     0     0     0     0     0         8       0 
537114631    2625    2623         0    0          370769         49
59923     82826    2685     0     0     0     0     0         6       0 
1610369017    2685    2683         0    0          344094         41
61700     82705    2706     0     0     0     0     0         7       0      0    
2706    2704         0    0          368474         39
63247     91042    2592     0     0     0     0     0         9       1      0    
2593    2591         0    0          399393         57
64636     93050    2530     0     0     0     0     0         6       0      0    
2530    2528         0    0          405903         24
58301     90662    2554     0     0     0     0     0         5       0      0    
2555    2553         0    0          404296         39
60146     88405    2567     0     0     0     0     0         8       1      0    
2567    2565         0    0          370227         49
61822     86417    2579     0     0     0     0     0         7       1      0    
2580    2578         0    0          366323         43
63351     86517    2594     0     0     1     0     0         9       0      0    
2594    2592         0    0          363340         51
64739     87879    2669     0     0     0     0     0         3       0      0    
2670    2668         0    0          387580         26
58397     88297    2566     0     0     0     0     0         4       0      0    
2566    2564         0    0          396980         34
60352     85039    2707     0     0     0     0     0         6       1      0    
2709    2707         0    0          367790         47
62169     87083    2779     0     0     0     0     0         6       0      0    
2775    2773         0    0          377163         31
63808     88700    2752     0     0     0     0     0         4       0      0    
2751    2749         0    0          392830         28
65189     91242    2725     0     0     0     0     0         5       1      0    
2727    2725         0    0          416871         31
59326     84893    2834     0     0     0     0     0         5       1      0    
2828    2826         0    0          380811         28
61064     87351    2553     0     0     0     0     0         9       0      0    
2552    2550         0    0          382053         37
62637     88761    2551     0     0     0     0     0         5       1      0    
2553    2551         0    0          384974         29

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-11 17:17                                                     ` Jeremy M. Guthrie
@ 2005-01-11 18:46                                                       ` Robert Olsson
  2005-01-12  1:30                                                         ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-11 18:46 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Stephen Hemminger, Jesse Brandeburg, Robert Olsson



Yes rtstats is what we estimated. How about RX ring size and the e1000 fix?

We give up here? Or you want to try another setting of route hash? It's beyond
the scope of this subject. You have to reboot if so.

						--ro



Jeremy M. Guthrie writes:
 >  ./rtstat
 >  size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
 > GC: tot ignored goal_miss ovrf HASH: in_search out_search
 > 63375     86092    2844     0     0     0     0     0         5       1 
 > 1610369017    2846    2844         0    0          386039         17
 > 64870     84431 537031471 537492192     0     0     0     0         1       0      
 > 0    2848    2846         0    0          386130         10
 > 58707     87889    2715     0     0     0     0     0         7       1      0    
 > 2716    2714         0    0          384782         13
 > 60617     89993    2796     0     0     0     0     0         5       0      0    
 > 2795    2793         0    0          384432         17
 > 62458     93221    2772     0     0     0     0     0         3       0      0    
 > 2772    2770         0    0          399579          9
 > 63990     93792    2712     0     0     0     0     0         6       0      0    
 > 2713    2711         0    0          422585         12
 > 65499     94235    2846     0     0     0     0     0         2       0      0    
 > 2846    2844         0    0          424341          9
 > 59755     90992    2864     0     0     0     0     0         1       1      0    
 > 2865    2863         0    0          385888         13
 > 61702     87681    2798     0     0     0     0     0         6       2      0    
 > 2800    2798         0    0          383276         34
 > 63334     90557    2734     0     0     0     0     0         5       1      0    
 > 2735    2733         0    0          422121         28
 > 64798     88178    2732     0     0     0     0     0         1       0      0    
 > 2732    2730         0    0          410772         11
 > 58574     85910    2669     0     0     0     0     0         3       0      0    
 > 2670    2668         0    0          393331         10
 > 60538     87238    2705     0     0     0     0     0         4       0      0    
 > 2705    2703         0    0          382161         20
 > 62164     85738    2628     0     0     0     0     0         4       0      0    
 > 2628    2626         0    0          373049         25
 > 63759     78752    2638     0     0     0     0     0         2       0      0    
 > 2638    2636         0    0          348411          5
 > 65052     79832    2482     0     0     0     0     0         3       0      0    
 > 2481    2479         0    0          365871         20
 > 58986     82611    2549     0     0     0     0     0         4       1      0    
 > 2550    2548         0    0          366429         24
 > 60845     85340    2620     0     0     1     0     0         2       0      0    
 > 2620    2618         0    0          369963         16
 > 62465     83122    2591     0     0     0     0     0         5       0      0    
 > 2591    2589         0    0          371261         18
 >  size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
 > GC: tot ignored goal_miss ovrf HASH: in_search out_search
 > 63998     86141    2640     0     0     0     0     0         4       0 
 > 537114631    2640    2638         0    0          386452         15
 > 65294     85989    2527     0     0     0     0     0         5       1 
 > 1610369017    2528    2526         0    0          379801         26
 > 59241     84305    2662     0     0     0     0     0         7       0      0    
 > 2662    2660         0    0          371925         35
 > 61194     85397    2837     0     0     0     0     0         3       0      0    
 > 2837    2835         0    0          374788         17
 > 62908     88133    2688     0     0     1     0     0         4       1      0    
 > 2689    2687         0    0          382411         19
 > 64329     88495    2648     0     0     0     0     0         5       1      0    
 > 2649    2647         0    0          398763         30
 > 65560     84994    2447     0     0     1     0     0         6       2      0    
 > 2449    2447         0    0          394745         33
 > 59677     81478    2568     0     0     0     0     0         5       0      0    
 > 2569    2567         0    0          341439         22
 > 61459     81225    2623     0     0     0     0     0         5       0      0    
 > 2623    2621         0    0          345996         30
 > 63038     83660    2544     0     0     0     0     0         3       0      0    
 > 2544    2542         0    0          381567         21
 > 64494     86321    2620     0     0     0     0     0         4       0      0    
 > 2620    2618         0    0          395278         19
 > 65528     88754    2766     0     0     0     0     0         6       0      0    
 > 2765    2763         0    0          421075         34
 > 59948     86443    2990     0     0     0     0     0         5       0      0    
 > 2990    2988         0    0          373605         29
 > 61860     81912    2786     0     0     0     0     0         5       1      0    
 > 2787    2785         0    0          354678         34
 > 63630     80399    2981     0     0     0     0     0         4       0      0    
 > 2979    2977         0    0          354983         32
 > 65215     83600    2934     0     0     0     0     0         6       3      0    
 > 2937    2935         0    0          371497         41
 > 59322     85583    2821     0     0     0     0     0         7       0      0    
 > 2822    2820         0    0          382896         40
 > 61268     82602    2817     0     0     0     0     0         9       2      0    
 > 2819    2817         0    0          368078         54
 > 62989     85336    2760     0     0     1     0     0         9       0      0    
 > 2760    2758         0    0          394212         39
 > 64387     84861    2603     0     0     0     0     0         7       0      0    
 > 2602    2600         0    0          390852         36
 >  size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
 > GC: tot ignored goal_miss ovrf HASH: in_search out_search
 > 65681     82848    2625     0     0     0     0     0         8       0 
 > 537114631    2625    2623         0    0          370769         49
 > 59923     82826    2685     0     0     0     0     0         6       0 
 > 1610369017    2685    2683         0    0          344094         41
 > 61700     82705    2706     0     0     0     0     0         7       0      0    
 > 2706    2704         0    0          368474         39
 > 63247     91042    2592     0     0     0     0     0         9       1      0    
 > 2593    2591         0    0          399393         57
 > 64636     93050    2530     0     0     0     0     0         6       0      0    
 > 2530    2528         0    0          405903         24
 > 58301     90662    2554     0     0     0     0     0         5       0      0    
 > 2555    2553         0    0          404296         39
 > 60146     88405    2567     0     0     0     0     0         8       1      0    
 > 2567    2565         0    0          370227         49
 > 61822     86417    2579     0     0     0     0     0         7       1      0    
 > 2580    2578         0    0          366323         43
 > 63351     86517    2594     0     0     1     0     0         9       0      0    
 > 2594    2592         0    0          363340         51
 > 64739     87879    2669     0     0     0     0     0         3       0      0    
 > 2670    2668         0    0          387580         26
 > 58397     88297    2566     0     0     0     0     0         4       0      0    
 > 2566    2564         0    0          396980         34
 > 60352     85039    2707     0     0     0     0     0         6       1      0    
 > 2709    2707         0    0          367790         47
 > 62169     87083    2779     0     0     0     0     0         6       0      0    
 > 2775    2773         0    0          377163         31
 > 63808     88700    2752     0     0     0     0     0         4       0      0    
 > 2751    2749         0    0          392830         28
 > 65189     91242    2725     0     0     0     0     0         5       1      0    
 > 2727    2725         0    0          416871         31
 > 59326     84893    2834     0     0     0     0     0         5       1      0    
 > 2828    2826         0    0          380811         28
 > 61064     87351    2553     0     0     0     0     0         9       0      0    
 > 2552    2550         0    0          382053         37
 > 62637     88761    2551     0     0     0     0     0         5       1      0    
 > 2553    2551         0    0          384974         29
 > 
 > -- 
 > 
 > --------------------------------------------------
 > Jeremy M. Guthrie        jeremy.guthrie@berbee.com
 > Senior Network Engineer        Phone: 608-298-1061
 > Berbee                           Fax: 608-288-3007
 > 5520 Research Park Drive         NOC: 608-298-1102
 > Madison, WI 53711

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-11 16:40                                                     ` Robert Olsson
@ 2005-01-12  1:27                                                       ` Jeremy M. Guthrie
  2005-01-12 15:11                                                         ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-12  1:27 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger, Jesse Brandeburg

On Tuesday 11 January 2005 10:40 am, Robert Olsson wrote:
> About the same load as last time 80 kpps forwarded and droprate of ~200 pps
> and 2323 fib lookups/sec. Not so bad for PIII but it maybe tuned better.
>
> Did the e1000 patch cure the problem that interrupts got enabled for
> unidirectional traffic? Ring size? I've never seen any win in system
> performance w. RX rings larger than 256 at least not in lab.

ETH3 Interrupts(calc'd from below):  1479968
ETH2 Interrupts:  261543
Packets RX'd on ETH3:  3892720
Packets dropped on RX on ETH3:  10305

This equates to about a 0.26% drop rate. W/ 256 packet RX ring size I see 
about a 0.42% drop rate.

This is using both the newest Intel driver w/ your patch and an increased ring 
size of 2048.

Tue Jan 11 19:15:04 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
          inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:2337377824 errors:14992144 dropped:14992144 
overruns:9643826 frame:0
          TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:3984064976 (3799.5 Mb)  TX bytes:398 (398.0 b)
          Base address:0x22a0 Memory:eff80000-effa0000

           CPU0       CPU1
  0:   93627934  353254270    IO-APIC-edge  timer
  1:         35        507    IO-APIC-edge  i8042
  7:          0          0   IO-APIC-level  ohci_hcd
  8:          0          2    IO-APIC-edge  rtc
 12:         73        145    IO-APIC-edge  i8042
 14:        120        313    IO-APIC-edge  ide0
 18: 2158179576       1815   IO-APIC-level  eth3
 20:          2 2136514988   IO-APIC-level  eth2
 27:     204201     371301   IO-APIC-level  eth0
 28:      14585      75320   IO-APIC-level  aic7xxx
 30:          0          0   IO-APIC-level  acpi
NMI:          0          0
LOC:  446922783  446921227
ERR:          0
MIS:          0

Tue Jan 11 19:16:05 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
          inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:2341270544 errors:15002449 dropped:15002449 
overruns:9652393 frame:0
          TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:1664751812 (1587.6 Mb)  TX bytes:398 (398.0 b)
          Base address:0x22a0 Memory:eff80000-effa0000

           CPU0       CPU1
  0:   93639955  353302319    IO-APIC-edge  timer
  1:         35        507    IO-APIC-edge  i8042
  7:          0          0   IO-APIC-level  ohci_hcd
  8:          0          2    IO-APIC-edge  rtc
 12:         73        145    IO-APIC-edge  i8042
 14:        120        313    IO-APIC-edge  ide0
 18: 2159659544       1815   IO-APIC-level  eth3
 20:          2 2136776531   IO-APIC-level  eth2
 27:     204245     371369   IO-APIC-level  eth0
 28:      14593      75343   IO-APIC-level  aic7xxx
 30:          0          0   IO-APIC-level  acpi
NMI:          0          0
LOC:  446982858  446981302
ERR:          0
MIS:          0

       --ro
>
> Jeremy M. Guthrie writes:
>  > date ; ifconfig eth3 ; cat /proc/net/softnet_stat ;
>  > cat /proc/net/stat/rt_cache ; sleep 60 ; date ; ifconfig eth3 ;
>  > cat /proc/net/softnet_stat ; cat /proc/net/stat/rt_cache
>  >
>  > Tue Jan 11 09:12:21 CST 2005
>  > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
>  >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
>  >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
>  >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>  >           RX packets:3519452697 errors:5558592 dropped:5558592
>  > overruns:4011523 frame:0
>  >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
>  >           collisions:0 txqueuelen:1000
>  >           RX bytes:497775695 (474.7 Mb)  TX bytes:398 (398.0 b)
>  >           Base address:0x22a0 Memory:eff80000-effa0000
>  >
>  > f59427dc 150dc67c 00c562ab 000d2659 00000000 00000000 00000000 00000000
>  > 00547b5b
>  > 00038622 00000000 00000065 00000000 00000000 00000000 00000000 00000000
>  > 0006804f
>  > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst
>  > in_martian_src out_hit out_slow_tot out_slow_mc  gc_total gc_ignored
>  > gc_goal_miss gc_dst_overflow in_hlist_search out_hlist_search
>  > 0000f3c2  aeca1565 272c383d 00000000 00000000 000089a8 00005298 00000002
>  > 001db403 0003054f 00000000 230aa73c 22fe33eb 00000cf0 00000000 b37011dc
>  > 00957c14
>  > 0000f3c2  0000b975 00029703 00000000 00000000 000035a5 00000000 00000000
>  > 00000015 00000083 00000002 000038c8 000038a3 00000000 00000000 0012a566
>  > 0000014d
>  >
>  > Tue Jan 11 09:13:21 CST 2005
>  > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
>  >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
>  >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
>  >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>  >           RX packets:3524302561 errors:5571396 dropped:5571396
>  > overruns:4022383 frame:0
>  >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
>  >           collisions:0 txqueuelen:1000
>  >           RX bytes:3065327250 (2923.3 Mb)  TX bytes:398 (398.0 b)
>  >           Base address:0x22a0 Memory:eff80000-effa0000
>  >
>  > f5de26f5 150dc67c 00c562b4 000d2659 00000000 00000000 00000000 00000000
>  > 00547ffd
>  > 00038632 00000000 00000065 00000000 00000000 00000000 00000000 00000000
>  > 0006807c
>  > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst
>  > in_martian_src out_hit out_slow_tot out_slow_mc  gc_total gc_ignored
>  > gc_goal_miss gc_dst_overflow in_hlist_search out_hlist_search
>  > 0000fc1e  af11f07b 272e58d3 00000000 00000000 000089a8 00005298 00000002
>  > 001db571 00030572 00000000 230cc7f1 23005428 00000cf0 00000000 b4b5b529
>  > 009583b9
>  > 0000fc1e  0000b977 00029710 00000000 00000000 000035a7 00000000 00000000
>  > 00000015 00000083 00000002 000038ca 000038a5 00000000 00000000 0012a5c8
>  > 0000014d
>  >
>  > --
>  >
>  > --------------------------------------------------
>  > Jeremy M. Guthrie        jeremy.guthrie@berbee.com
>  > Senior Network Engineer        Phone: 608-298-1061
>  > Berbee                           Fax: 608-288-3007
>  > 5520 Research Park Drive         NOC: 608-298-1102
>  > Madison, WI 53711

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-11 18:46                                                       ` Robert Olsson
@ 2005-01-12  1:30                                                         ` Jeremy M. Guthrie
  2005-01-12 16:02                                                           ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-12  1:30 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger, Jesse Brandeburg

On Tuesday 11 January 2005 12:46 pm, Robert Olsson wrote:
> We give up here? Or you want to try another setting of route hash? It's
> beyond the scope of this subject. You have to reboot if so.
I am up for pushing it if you do not think it is a waste of time.  Based on 
what I am seeing it looks like I just need a faster CPU to do the work.  My 
goal would be to hit zero dropped packets w/ 10-15% CPU to spare but I fail 
to see how that will happen on this box.  Do you concur that it would be 
highly unlikely I would be able to get that kind of performance increase?

> Jeremy M. Guthrie writes:
>  >  ./rtstat
>  >  size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot 
>  >    mc GC: tot ignored goal_miss ovrf HASH: in_search out_search
>  > 63375     86092    2844     0     0     0     0     0         5       1
>  > 1610369017    2846    2844         0    0          386039         17
>  > 64870     84431 537031471 537492192     0     0     0     0         1   
>  >    0 0    2848    2846         0    0          386130         10
>  > 58707     87889    2715     0     0     0     0     0         7       1 
>  >     0 2716    2714         0    0          384782         13
>  > 60617     89993    2796     0     0     0     0     0         5       0 
>  >     0 2795    2793         0    0          384432         17
>  > 62458     93221    2772     0     0     0     0     0         3       0 
>  >     0 2772    2770         0    0          399579          9
>  > 63990     93792    2712     0     0     0     0     0         6       0 
>  >     0 2713    2711         0    0          422585         12
>  > 65499     94235    2846     0     0     0     0     0         2       0 
>  >     0 2846    2844         0    0          424341          9
>  > 59755     90992    2864     0     0     0     0     0         1       1 
>  >     0 2865    2863         0    0          385888         13
>  > 61702     87681    2798     0     0     0     0     0         6       2 
>  >     0 2800    2798         0    0          383276         34
>  > 63334     90557    2734     0     0     0     0     0         5       1 
>  >     0 2735    2733         0    0          422121         28
>  > 64798     88178    2732     0     0     0     0     0         1       0 
>  >     0 2732    2730         0    0          410772         11
>  > 58574     85910    2669     0     0     0     0     0         3       0 
>  >     0 2670    2668         0    0          393331         10
>  > 60538     87238    2705     0     0     0     0     0         4       0 
>  >     0 2705    2703         0    0          382161         20
>  > 62164     85738    2628     0     0     0     0     0         4       0 
>  >     0 2628    2626         0    0          373049         25
>  > 63759     78752    2638     0     0     0     0     0         2       0 
>  >     0 2638    2636         0    0          348411          5
>  > 65052     79832    2482     0     0     0     0     0         3       0 
>  >     0 2481    2479         0    0          365871         20
>  > 58986     82611    2549     0     0     0     0     0         4       1 
>  >     0 2550    2548         0    0          366429         24
>  > 60845     85340    2620     0     0     1     0     0         2       0 
>  >     0 2620    2618         0    0          369963         16
>  > 62465     83122    2591     0     0     0     0     0         5       0 
>  >     0 2591    2589         0    0          371261         18
>  >  size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot 
>  >    mc GC: tot ignored goal_miss ovrf HASH: in_search out_search
>  > 63998     86141    2640     0     0     0     0     0         4       0
>  > 537114631    2640    2638         0    0          386452         15
>  > 65294     85989    2527     0     0     0     0     0         5       1
>  > 1610369017    2528    2526         0    0          379801         26
>  > 59241     84305    2662     0     0     0     0     0         7       0 
>  >     0 2662    2660         0    0          371925         35
>  > 61194     85397    2837     0     0     0     0     0         3       0 
>  >     0 2837    2835         0    0          374788         17
>  > 62908     88133    2688     0     0     1     0     0         4       1 
>  >     0 2689    2687         0    0          382411         19
>  > 64329     88495    2648     0     0     0     0     0         5       1 
>  >     0 2649    2647         0    0          398763         30
>  > 65560     84994    2447     0     0     1     0     0         6       2 
>  >     0 2449    2447         0    0          394745         33
>  > 59677     81478    2568     0     0     0     0     0         5       0 
>  >     0 2569    2567         0    0          341439         22
>  > 61459     81225    2623     0     0     0     0     0         5       0 
>  >     0 2623    2621         0    0          345996         30
>  > 63038     83660    2544     0     0     0     0     0         3       0 
>  >     0 2544    2542         0    0          381567         21
>  > 64494     86321    2620     0     0     0     0     0         4       0 
>  >     0 2620    2618         0    0          395278         19
>  > 65528     88754    2766     0     0     0     0     0         6       0 
>  >     0 2765    2763         0    0          421075         34
>  > 59948     86443    2990     0     0     0     0     0         5       0 
>  >     0 2990    2988         0    0          373605         29
>  > 61860     81912    2786     0     0     0     0     0         5       1 
>  >     0 2787    2785         0    0          354678         34
>  > 63630     80399    2981     0     0     0     0     0         4       0 
>  >     0 2979    2977         0    0          354983         32
>  > 65215     83600    2934     0     0     0     0     0         6       3 
>  >     0 2937    2935         0    0          371497         41
>  > 59322     85583    2821     0     0     0     0     0         7       0 
>  >     0 2822    2820         0    0          382896         40
>  > 61268     82602    2817     0     0     0     0     0         9       2 
>  >     0 2819    2817         0    0          368078         54
>  > 62989     85336    2760     0     0     1     0     0         9       0 
>  >     0 2760    2758         0    0          394212         39
>  > 64387     84861    2603     0     0     0     0     0         7       0 
>  >     0 2602    2600         0    0          390852         36
>  >  size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot 
>  >    mc GC: tot ignored goal_miss ovrf HASH: in_search out_search
>  > 65681     82848    2625     0     0     0     0     0         8       0
>  > 537114631    2625    2623         0    0          370769         49
>  > 59923     82826    2685     0     0     0     0     0         6       0
>  > 1610369017    2685    2683         0    0          344094         41
>  > 61700     82705    2706     0     0     0     0     0         7       0 
>  >     0 2706    2704         0    0          368474         39
>  > 63247     91042    2592     0     0     0     0     0         9       1 
>  >     0 2593    2591         0    0          399393         57
>  > 64636     93050    2530     0     0     0     0     0         6       0 
>  >     0 2530    2528         0    0          405903         24
>  > 58301     90662    2554     0     0     0     0     0         5       0 
>  >     0 2555    2553         0    0          404296         39
>  > 60146     88405    2567     0     0     0     0     0         8       1 
>  >     0 2567    2565         0    0          370227         49
>  > 61822     86417    2579     0     0     0     0     0         7       1 
>  >     0 2580    2578         0    0          366323         43
>  > 63351     86517    2594     0     0     1     0     0         9       0 
>  >     0 2594    2592         0    0          363340         51
>  > 64739     87879    2669     0     0     0     0     0         3       0 
>  >     0 2670    2668         0    0          387580         26
>  > 58397     88297    2566     0     0     0     0     0         4       0 
>  >     0 2566    2564         0    0          396980         34
>  > 60352     85039    2707     0     0     0     0     0         6       1 
>  >     0 2709    2707         0    0          367790         47
>  > 62169     87083    2779     0     0     0     0     0         6       0 
>  >     0 2775    2773         0    0          377163         31
>  > 63808     88700    2752     0     0     0     0     0         4       0 
>  >     0 2751    2749         0    0          392830         28
>  > 65189     91242    2725     0     0     0     0     0         5       1 
>  >     0 2727    2725         0    0          416871         31
>  > 59326     84893    2834     0     0     0     0     0         5       1 
>  >     0 2828    2826         0    0          380811         28
>  > 61064     87351    2553     0     0     0     0     0         9       0 
>  >     0 2552    2550         0    0          382053         37
>  > 62637     88761    2551     0     0     0     0     0         5       1 
>  >     0 2553    2551         0    0          384974         29
>  >
>  > --
>  >
>  > --------------------------------------------------
>  > Jeremy M. Guthrie        jeremy.guthrie@berbee.com
>  > Senior Network Engineer        Phone: 608-298-1061
>  > Berbee                           Fax: 608-288-3007
>  > 5520 Research Park Drive         NOC: 608-298-1102
>  > Madison, WI 53711

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12  1:27                                                       ` Jeremy M. Guthrie
@ 2005-01-12 15:11                                                         ` Robert Olsson
  2005-01-12 16:24                                                           ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-12 15:11 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger, Jesse Brandeburg


Jeremy M. Guthrie writes:

 > > Did the e1000 patch cure the problem that interrupts got enabled for
 > > unidirectional traffic? Ring size? I've never seen any win in system
 > > performance w. RX rings larger than 256 at least not in lab.

 > ETH3 Interrupts(calc'd from below):  1479968
 > ETH2 Interrupts:  261543
 > Packets RX'd on ETH3:  3892720
 > Packets dropped on RX on ETH3:  10305

 Very strange...
 
 eth3 is bound to CPU0 which in turn has all packet load... If we were
 to believe your CPU0 was saturated (due to the drops). We should see no 
 (RX) interrupts on eth3. But there is a lot... one irq per every three
 packet. Why?

 Can you investigate? e1000 has problem like this w. unidirectional traffic
 w/o the patch I sent.

 Or is your traffic so extremely bursty. No?

						--ro


 
 > This equates to about a 0.26% drop rate. W/ 256 packet RX ring size I see 
 > about a 0.42% drop rate.
 > 
 > This is using both the newest Intel driver w/ your patch and an increased ring 
 > size of 2048.
 > 
 > Tue Jan 11 19:15:04 CST 2005
 > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
 >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
 >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
 >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
 >           RX packets:2337377824 errors:14992144 dropped:14992144 
 > overruns:9643826 frame:0
 >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
 >           collisions:0 txqueuelen:1000
 >           RX bytes:3984064976 (3799.5 Mb)  TX bytes:398 (398.0 b)
 >           Base address:0x22a0 Memory:eff80000-effa0000
 > 
 >            CPU0       CPU1
 >   0:   93627934  353254270    IO-APIC-edge  timer
 >   1:         35        507    IO-APIC-edge  i8042
 >   7:          0          0   IO-APIC-level  ohci_hcd
 >   8:          0          2    IO-APIC-edge  rtc
 >  12:         73        145    IO-APIC-edge  i8042
 >  14:        120        313    IO-APIC-edge  ide0
 >  18: 2158179576       1815   IO-APIC-level  eth3
 >  20:          2 2136514988   IO-APIC-level  eth2
 >  27:     204201     371301   IO-APIC-level  eth0
 >  28:      14585      75320   IO-APIC-level  aic7xxx
 >  30:          0          0   IO-APIC-level  acpi
 > NMI:          0          0
 > LOC:  446922783  446921227
 > ERR:          0
 > MIS:          0
 > 
 > Tue Jan 11 19:16:05 CST 2005
 > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
 >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
 >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
 >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
 >           RX packets:2341270544 errors:15002449 dropped:15002449 
 > overruns:9652393 frame:0
 >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
 >           collisions:0 txqueuelen:1000
 >           RX bytes:1664751812 (1587.6 Mb)  TX bytes:398 (398.0 b)
 >           Base address:0x22a0 Memory:eff80000-effa0000
 > 
 >            CPU0       CPU1
 >   0:   93639955  353302319    IO-APIC-edge  timer
 >   1:         35        507    IO-APIC-edge  i8042
 >   7:          0          0   IO-APIC-level  ohci_hcd
 >   8:          0          2    IO-APIC-edge  rtc
 >  12:         73        145    IO-APIC-edge  i8042
 >  14:        120        313    IO-APIC-edge  ide0
 >  18: 2159659544       1815   IO-APIC-level  eth3
 >  20:          2 2136776531   IO-APIC-level  eth2
 >  27:     204245     371369   IO-APIC-level  eth0
 >  28:      14593      75343   IO-APIC-level  aic7xxx
 >  30:          0          0   IO-APIC-level  acpi
 > NMI:          0          0
 > LOC:  446982858  446981302
 > ERR:          0
 > MIS:          0
 > 
 >        --ro
 > >
 > > Jeremy M. Guthrie writes:
 > >  > date ; ifconfig eth3 ; cat /proc/net/softnet_stat ;
 > >  > cat /proc/net/stat/rt_cache ; sleep 60 ; date ; ifconfig eth3 ;
 > >  > cat /proc/net/softnet_stat ; cat /proc/net/stat/rt_cache
 > >  >
 > >  > Tue Jan 11 09:12:21 CST 2005
 > >  > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
 > >  >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
 > >  >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
 > >  >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
 > >  >           RX packets:3519452697 errors:5558592 dropped:5558592
 > >  > overruns:4011523 frame:0
 > >  >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
 > >  >           collisions:0 txqueuelen:1000
 > >  >           RX bytes:497775695 (474.7 Mb)  TX bytes:398 (398.0 b)
 > >  >           Base address:0x22a0 Memory:eff80000-effa0000
 > >  >
 > >  > f59427dc 150dc67c 00c562ab 000d2659 00000000 00000000 00000000 00000000
 > >  > 00547b5b
 > >  > 00038622 00000000 00000065 00000000 00000000 00000000 00000000 00000000
 > >  > 0006804f
 > >  > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst
 > >  > in_martian_src out_hit out_slow_tot out_slow_mc  gc_total gc_ignored
 > >  > gc_goal_miss gc_dst_overflow in_hlist_search out_hlist_search
 > >  > 0000f3c2  aeca1565 272c383d 00000000 00000000 000089a8 00005298 00000002
 > >  > 001db403 0003054f 00000000 230aa73c 22fe33eb 00000cf0 00000000 b37011dc
 > >  > 00957c14
 > >  > 0000f3c2  0000b975 00029703 00000000 00000000 000035a5 00000000 00000000
 > >  > 00000015 00000083 00000002 000038c8 000038a3 00000000 00000000 0012a566
 > >  > 0000014d
 > >  >
 > >  > Tue Jan 11 09:13:21 CST 2005
 > >  > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
 > >  >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
 > >  >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
 > >  >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
 > >  >           RX packets:3524302561 errors:5571396 dropped:5571396
 > >  > overruns:4022383 frame:0
 > >  >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
 > >  >           collisions:0 txqueuelen:1000
 > >  >           RX bytes:3065327250 (2923.3 Mb)  TX bytes:398 (398.0 b)
 > >  >           Base address:0x22a0 Memory:eff80000-effa0000
 > >  >
 > >  > f5de26f5 150dc67c 00c562b4 000d2659 00000000 00000000 00000000 00000000
 > >  > 00547ffd
 > >  > 00038632 00000000 00000065 00000000 00000000 00000000 00000000 00000000
 > >  > 0006807c
 > >  > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst
 > >  > in_martian_src out_hit out_slow_tot out_slow_mc  gc_total gc_ignored
 > >  > gc_goal_miss gc_dst_overflow in_hlist_search out_hlist_search
 > >  > 0000fc1e  af11f07b 272e58d3 00000000 00000000 000089a8 00005298 00000002
 > >  > 001db571 00030572 00000000 230cc7f1 23005428 00000cf0 00000000 b4b5b529
 > >  > 009583b9
 > >  > 0000fc1e  0000b977 00029710 00000000 00000000 000035a7 00000000 00000000
 > >  > 00000015 00000083 00000002 000038ca 000038a5 00000000 00000000 0012a5c8
 > >  > 0000014d
 > >  >
 > >  > --
 > >  >
 > >  > --------------------------------------------------
 > >  > Jeremy M. Guthrie        jeremy.guthrie@berbee.com
 > >  > Senior Network Engineer        Phone: 608-298-1061
 > >  > Berbee                           Fax: 608-288-3007
 > >  > 5520 Research Park Drive         NOC: 608-298-1102
 > >  > Madison, WI 53711
 > 
 > -- 
 > 
 > --------------------------------------------------
 > Jeremy M. Guthrie        jeremy.guthrie@berbee.com
 > Senior Network Engineer        Phone: 608-298-1061
 > Berbee                           Fax: 608-288-3007
 > 5520 Research Park Drive         NOC: 608-298-1102
 > Madison, WI 53711

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12  1:30                                                         ` Jeremy M. Guthrie
@ 2005-01-12 16:02                                                           ` Robert Olsson
  0 siblings, 0 replies; 88+ messages in thread
From: Robert Olsson @ 2005-01-12 16:02 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger, Jesse Brandeburg


Jeremy M. Guthrie writes:

 > I am up for pushing it if you do not think it is a waste of time.  Based on 
 > what I am seeing it looks like I just need a faster CPU to do the work.  My 
 > goal would be to hit zero dropped packets w/ 10-15% CPU to spare but I fail 
 > to see how that will happen on this box.  Do you concur that it would be 
 > highly unlikely I would be able to get that kind of performance increase?

 Well we can try to decrease some of the linear search in the route hash
 once we understand why we see all the RX interrupts. 

						   --ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 15:11                                                         ` Robert Olsson
@ 2005-01-12 16:24                                                           ` Jeremy M. Guthrie
  2005-01-12 19:27                                                             ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-12 16:24 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger, Jesse Brandeburg

[-- Attachment #1: Type: text/plain, Size: 4775 bytes --]

On Wednesday 12 January 2005 09:11 am, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > > Did the e1000 patch cure the problem that interrupts got enabled for
>  > > unidirectional traffic? Ring size? I've never seen any win in system
>  > > performance w. RX rings larger than 256 at least not in lab.
>  >
>  > ETH3 Interrupts(calc'd from below):  1479968
>  > ETH2 Interrupts:  261543
>  > Packets RX'd on ETH3:  3892720
>  > Packets dropped on RX on ETH3:  10305
>
>  Very strange...
>
>  eth3 is bound to CPU0 which in turn has all packet load... If we were
>  to believe your CPU0 was saturated (due to the drops). We should see no
>  (RX) interrupts on eth3. But there is a lot... one irq per every three
>  packet. Why?
I have no idea why it would be doing this.

>  Can you investigate? e1000 has problem like this w. unidirectional traffic
>  w/o the patch I sent.
This appears to be where this problem is getting beyond my expertise.  I 
verified NAPI is turned on.  I also verified the patch is in place.  I am 
open to suggestions but otherwise I am not the worlds best coder. 

>  Or is your traffic so extremely bursty. No?
By the nature of our business it can be very bursty.  We just have so many 
sources/destinations for traffic that our traffic is generally pretty bursty.  
We might be running 300-400 mbps with spikes of 40-100 mbps.

>       --ro
>
>  > This equates to about a 0.26% drop rate. W/ 256 packet RX ring size I
>  > see about a 0.42% drop rate.
>  >
>  > This is using both the newest Intel driver w/ your patch and an
>  > increased ring size of 2048.
>  >
>  > Tue Jan 11 19:15:04 CST 2005
>  > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
>  >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
>  >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
>  >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>  >           RX packets:2337377824 errors:14992144 dropped:14992144
>  > overruns:9643826 frame:0
>  >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
>  >           collisions:0 txqueuelen:1000
>  >           RX bytes:3984064976 (3799.5 Mb)  TX bytes:398 (398.0 b)
>  >           Base address:0x22a0 Memory:eff80000-effa0000
>  >
>  >            CPU0       CPU1
>  >   0:   93627934  353254270    IO-APIC-edge  timer
>  >   1:         35        507    IO-APIC-edge  i8042
>  >   7:          0          0   IO-APIC-level  ohci_hcd
>  >   8:          0          2    IO-APIC-edge  rtc
>  >  12:         73        145    IO-APIC-edge  i8042
>  >  14:        120        313    IO-APIC-edge  ide0
>  >  18: 2158179576       1815   IO-APIC-level  eth3
>  >  20:          2 2136514988   IO-APIC-level  eth2
>  >  27:     204201     371301   IO-APIC-level  eth0
>  >  28:      14585      75320   IO-APIC-level  aic7xxx
>  >  30:          0          0   IO-APIC-level  acpi
>  > NMI:          0          0
>  > LOC:  446922783  446921227
>  > ERR:          0
>  > MIS:          0
>  >
>  > Tue Jan 11 19:16:05 CST 2005
>  > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
>  >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
>  >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
>  >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>  >           RX packets:2341270544 errors:15002449 dropped:15002449
>  > overruns:9652393 frame:0
>  >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
>  >           collisions:0 txqueuelen:1000
>  >           RX bytes:1664751812 (1587.6 Mb)  TX bytes:398 (398.0 b)
>  >           Base address:0x22a0 Memory:eff80000-effa0000
>  >
>  >            CPU0       CPU1
>  >   0:   93639955  353302319    IO-APIC-edge  timer
>  >   1:         35        507    IO-APIC-edge  i8042
>  >   7:          0          0   IO-APIC-level  ohci_hcd
>  >   8:          0          2    IO-APIC-edge  rtc
>  >  12:         73        145    IO-APIC-edge  i8042
>  >  14:        120        313    IO-APIC-edge  ide0
>  >  18: 2159659544       1815   IO-APIC-level  eth3
>  >  20:          2 2136776531   IO-APIC-level  eth2
>  >  27:     204245     371369   IO-APIC-level  eth0
>  >  28:      14593      75343   IO-APIC-level  aic7xxx
>  >  30:          0          0   IO-APIC-level  acpi
>  > NMI:          0          0
>  > LOC:  446982858  446981302
>  > ERR:          0
>  > MIS:          0

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 16:24                                                           ` Jeremy M. Guthrie
@ 2005-01-12 19:27                                                             ` Robert Olsson
  2005-01-12 20:11                                                               ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-12 19:27 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger, Jesse Brandeburg


Jeremy M. Guthrie writes:

 > >  > ETH3 Interrupts(calc'd from below):  1479968

 > >  Very strange...
 > >
 > >  eth3 is bound to CPU0 which in turn has all packet load... If we were
 > >  to believe your CPU0 was saturated (due to the drops). We should see no
 > >  (RX) interrupts on eth3. But there is a lot... one irq per every three
 > >  packet. Why?

 > I have no idea why it would be doing this.

 Huh seems you didn't add the patch I sent. Below is diff from my editor to your
 e1000_main.c

							--ro


--- e1000_main.c.jmg	2005-01-12 20:14:08.324168072 +0100
+++ e1000_main.c	2005-01-12 20:17:24.777302656 +0100
@@ -2264,14 +2264,13 @@
 	netdev->quota -= work_done;
 	
 	/* if no Rx and Tx cleanup work was done, exit the polling mode */
-	if(!tx_cleaned || (work_done < work_to_do) || 
-				!netif_running(netdev)) {
+	 if( (!tx_cleaned && (work_done == 0)) || !netif_running(netdev)) {
 quit_polling:	netif_rx_complete(netdev);
 		e1000_irq_enable(adapter);
 		return 0;
 	}
 
-	return (work_done >= work_to_do);
+	return 1;
 }
 
 #endif

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 19:27                                                             ` Robert Olsson
@ 2005-01-12 20:11                                                               ` Jeremy M. Guthrie
  2005-01-12 20:21                                                                 ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-12 20:11 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger, Jesse Brandeburg

[-- Attachment #1: Type: text/plain, Size: 4141 bytes --]


Latest numbers after your patch Robert.

Wed Jan 12 14:05:36 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
          inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:21951496 errors:2412189 dropped:2412189 overruns:377090 
frame:0
          TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:3455362966 (3295.2 Mb)  TX bytes:398 (398.0 b)
          Base address:0x22a0 Memory:eff80000-effa0000

95b83157 150dc67c 00c734a1 000d2659 00000000 00000000 00000000 00000000 
0073cee4
00044494 00000000 00000075 00000000 00000000 00000000 00000000 00000000 
00097c77
entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
gc_dst_overflow in_hlist_search out_hlist_search
0000f8cd  41403122 34c30e49 00000000 00000000 0000b0dd 00006a38 00000002  
0027279d 0004147b 00000000 2f528a5f 2f42efcf 0000104f 00000000 81d262df 
00c5d75c
0000f8cd  0000e332 0003238f 00000000 00000000 00004263 00000000 00000000  
0000001e 000000c0 00000002 00004650 00004626 00000000 00000000 0016b879 
0000024c





Wed Jan 12 14:06:36 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
          inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:28053281 errors:2899869 dropped:2899869 overruns:427243 
frame:0
          TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:2607094069 (2486.3 Mb)  TX bytes:398 (398.0 b)
          Base address:0x22a0 Memory:eff80000-effa0000

96154d20 150dc67c 00c78bc2 000d2659 00000000 00000000 00000000 00000000 
0073d31d
00044499 00000000 00000075 00000000 00000000 00000000 00000000 00000000 
00097d2d
entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
gc_dst_overflow in_hlist_search out_hlist_search
0000fc6a  419ad196 34c586ba 00000000 00000000 0000b0e0 00006a40 00000002  
002729c5 000414cb 00000000 2f5502ff 2f4567f7 0000104f 00000000 83790467 
00c5e33c
0000fc6a  0000e333 00032393 00000000 00000000 00004263 00000000 00000000  
0000001e 000000c0 00000002 00004650 00004626 00000000 00000000 0016b898 
0000024c


On Wednesday 12 January 2005 01:27 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > >  > ETH3 Interrupts(calc'd from below):  1479968
>  > >
>  > >  Very strange...
>  > >
>  > >  eth3 is bound to CPU0 which in turn has all packet load... If we were
>  > >  to believe your CPU0 was saturated (due to the drops). We should see
>  > > no (RX) interrupts on eth3. But there is a lot... one irq per every
>  > > three packet. Why?
>  >
>  > I have no idea why it would be doing this.
>
>  Huh seems you didn't add the patch I sent. Below is diff from my editor to
> your e1000_main.c
>
>        --ro
>
>
> --- e1000_main.c.jmg 2005-01-12 20:14:08.324168072 +0100
> +++ e1000_main.c 2005-01-12 20:17:24.777302656 +0100
> @@ -2264,14 +2264,13 @@
>   netdev->quota -= work_done;
>
>   /* if no Rx and Tx cleanup work was done, exit the polling mode */
> - if(!tx_cleaned || (work_done < work_to_do) ||
> -    !netif_running(netdev)) {
> +  if( (!tx_cleaned && (work_done == 0)) || !netif_running(netdev)) {
>  quit_polling: netif_rx_complete(netdev);
>    e1000_irq_enable(adapter);
>    return 0;
>   }
>
> - return (work_done >= work_to_do);
> + return 1;
>  }
>
>  #endif

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 20:11                                                               ` Jeremy M. Guthrie
@ 2005-01-12 20:21                                                                 ` Robert Olsson
  2005-01-12 20:30                                                                   ` Jeremy M. Guthrie
  2005-01-12 20:45                                                                   ` Jeremy M. Guthrie
  0 siblings, 2 replies; 88+ messages in thread
From: Robert Olsson @ 2005-01-12 20:21 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger, Jesse Brandeburg


Jeremy M. Guthrie writes:
 > 
 > Latest numbers after your patch Robert.


 Did the RX interrupts go down?

						--ro


 > Wed Jan 12 14:05:36 CST 2005
 > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
 >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
 >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
 >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
 >           RX packets:21951496 errors:2412189 dropped:2412189 overruns:377090 
 > frame:0
 >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
 >           collisions:0 txqueuelen:1000
 >           RX bytes:3455362966 (3295.2 Mb)  TX bytes:398 (398.0 b)
 >           Base address:0x22a0 Memory:eff80000-effa0000
 > 
 > 95b83157 150dc67c 00c734a1 000d2659 00000000 00000000 00000000 00000000 
 > 0073cee4
 > 00044494 00000000 00000075 00000000 00000000 00000000 00000000 00000000 
 > 00097c77
 > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
 > out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
 > gc_dst_overflow in_hlist_search out_hlist_search
 > 0000f8cd  41403122 34c30e49 00000000 00000000 0000b0dd 00006a38 00000002  
 > 0027279d 0004147b 00000000 2f528a5f 2f42efcf 0000104f 00000000 81d262df 
 > 00c5d75c
 > 0000f8cd  0000e332 0003238f 00000000 00000000 00004263 00000000 00000000  
 > 0000001e 000000c0 00000002 00004650 00004626 00000000 00000000 0016b879 
 > 0000024c
 > 
 > 
 > 
 > 
 > 
 > Wed Jan 12 14:06:36 CST 2005
 > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
 >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
 >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
 >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
 >           RX packets:28053281 errors:2899869 dropped:2899869 overruns:427243 
 > frame:0
 >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
 >           collisions:0 txqueuelen:1000
 >           RX bytes:2607094069 (2486.3 Mb)  TX bytes:398 (398.0 b)
 >           Base address:0x22a0 Memory:eff80000-effa0000
 > 
 > 96154d20 150dc67c 00c78bc2 000d2659 00000000 00000000 00000000 00000000 
 > 0073d31d
 > 00044499 00000000 00000075 00000000 00000000 00000000 00000000 00000000 
 > 00097d2d
 > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
 > out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
 > gc_dst_overflow in_hlist_search out_hlist_search
 > 0000fc6a  419ad196 34c586ba 00000000 00000000 0000b0e0 00006a40 00000002  
 > 002729c5 000414cb 00000000 2f5502ff 2f4567f7 0000104f 00000000 83790467 
 > 00c5e33c
 > 0000fc6a  0000e333 00032393 00000000 00000000 00004263 00000000 00000000  
 > 0000001e 000000c0 00000002 00004650 00004626 00000000 00000000 0016b898 
 > 0000024c
 > 
 > 
 > On Wednesday 12 January 2005 01:27 pm, Robert Olsson wrote:
 > > Jeremy M. Guthrie writes:
 > >  > >  > ETH3 Interrupts(calc'd from below):  1479968
 > >  > >
 > >  > >  Very strange...
 > >  > >
 > >  > >  eth3 is bound to CPU0 which in turn has all packet load... If we were
 > >  > >  to believe your CPU0 was saturated (due to the drops). We should see
 > >  > > no (RX) interrupts on eth3. But there is a lot... one irq per every
 > >  > > three packet. Why?
 > >  >
 > >  > I have no idea why it would be doing this.
 > >
 > >  Huh seems you didn't add the patch I sent. Below is diff from my editor to
 > > your e1000_main.c
 > >
 > >        --ro
 > >
 > >
 > > --- e1000_main.c.jmg 2005-01-12 20:14:08.324168072 +0100
 > > +++ e1000_main.c 2005-01-12 20:17:24.777302656 +0100
 > > @@ -2264,14 +2264,13 @@
 > >   netdev->quota -= work_done;
 > >
 > >   /* if no Rx and Tx cleanup work was done, exit the polling mode */
 > > - if(!tx_cleaned || (work_done < work_to_do) ||
 > > -    !netif_running(netdev)) {
 > > +  if( (!tx_cleaned && (work_done == 0)) || !netif_running(netdev)) {
 > >  quit_polling: netif_rx_complete(netdev);
 > >    e1000_irq_enable(adapter);
 > >    return 0;
 > >   }
 > >
 > > - return (work_done >= work_to_do);
 > > + return 1;
 > >  }
 > >
 > >  #endif
 > 
 > -- 
 > 
 > --------------------------------------------------
 > Jeremy M. Guthrie        jeremy.guthrie@berbee.com
 > Senior Network Engineer        Phone: 608-298-1061
 > Berbee                           Fax: 608-288-3007
 > 5520 Research Park Drive         NOC: 608-298-1102
 > Madison, WI 53711

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 20:21                                                                 ` Robert Olsson
@ 2005-01-12 20:30                                                                   ` Jeremy M. Guthrie
  2005-01-12 20:45                                                                   ` Jeremy M. Guthrie
  1 sibling, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-12 20:30 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger, Jesse Brandeburg

[-- Attachment #1: Type: text/plain, Size: 7120 bytes --]

Sorry, here is the latest.  BTW, it is VERY sluggish now.

Wed Jan 12 14:25:03 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          RX packets:32030222 errors:105026820 dropped:105026820 
overruns:100748696 frame:0
9651ffeb 150dc67c 00c814c1 000d2659 00000000 00000000 00000000 00000000 
0073d527
00044681 00000000 000000f8 00000000 00000000 00000000 00000000 00000000 
00097d42
entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
gc_dst_overflow in_hlist_search out_hlist_search
00020000  41c25a6f 34daaca8 00000000 00000000 0000b11d 00006a40 00000002  
00272af5 00041525 00000000 2f6a27e5 2f4b712d 000f2bd8 000f1b0e 84036c96 
00c5e8a8
00020000  0000e336 0003256f 00000000 00000000 00004292 00000000 00000000  
0000001e 000000c0 00000002 00004762 0000462a 0000010e 0000010b 0016b898 
0000024c
           CPU0       CPU1
 18: 3586173518       1815   IO-APIC-level  eth3
 20:          2 2464382507   IO-APIC-level  eth2

Wed Jan 12 14:26:03 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          RX packets:32097966 errors:110579564 dropped:110579564 
overruns:106248073 frame:0
965308aa 150dc67c 00c818e4 000d2659 00000000 00000000 00000000 00000000 
0073d527
0004468b 00000000 000000f8 00000000 00000000 00000000 00000000 00000000 
00097d42
entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
gc_dst_overflow in_hlist_search out_hlist_search
00020000  41c2602e 34dbaf9f 00000000 00000000 0000b11e 00006a40 00000002  
00272af5 00041527 00000000 2f6b2ad8 2f4b9a88 00100571 000ff4a4 84036c9e 
00c5e8a8
00020000  0000e336 00032578 00000000 00000000 00004292 00000000 00000000  
0000001e 000000c0 00000002 00004762 0000462a 0000010e 0000010b 0016b898 
0000024c
           CPU0       CPU1
 18: 3586173518       1815   IO-APIC-level  eth3
 20:          2 2464387985   IO-APIC-level  eth2




On Wednesday 12 January 2005 02:21 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > Latest numbers after your patch Robert.
>
>  Did the RX interrupts go down?
>
>       --ro
>
>  > Wed Jan 12 14:05:36 CST 2005
>  > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
>  >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
>  >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
>  >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>  >           RX packets:21951496 errors:2412189 dropped:2412189
>  > overruns:377090 frame:0
>  >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
>  >           collisions:0 txqueuelen:1000
>  >           RX bytes:3455362966 (3295.2 Mb)  TX bytes:398 (398.0 b)
>  >           Base address:0x22a0 Memory:eff80000-effa0000
>  >
>  > 95b83157 150dc67c 00c734a1 000d2659 00000000 00000000 00000000 00000000
>  > 0073cee4
>  > 00044494 00000000 00000075 00000000 00000000 00000000 00000000 00000000
>  > 00097c77
>  > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst
>  > in_martian_src out_hit out_slow_tot out_slow_mc  gc_total gc_ignored
>  > gc_goal_miss gc_dst_overflow in_hlist_search out_hlist_search
>  > 0000f8cd  41403122 34c30e49 00000000 00000000 0000b0dd 00006a38 00000002
>  > 0027279d 0004147b 00000000 2f528a5f 2f42efcf 0000104f 00000000 81d262df
>  > 00c5d75c
>  > 0000f8cd  0000e332 0003238f 00000000 00000000 00004263 00000000 00000000
>  > 0000001e 000000c0 00000002 00004650 00004626 00000000 00000000 0016b879
>  > 0000024c
>  >
>  >
>  >
>  >
>  >
>  > Wed Jan 12 14:06:36 CST 2005
>  > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
>  >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
>  >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
>  >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>  >           RX packets:28053281 errors:2899869 dropped:2899869
>  > overruns:427243 frame:0
>  >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
>  >           collisions:0 txqueuelen:1000
>  >           RX bytes:2607094069 (2486.3 Mb)  TX bytes:398 (398.0 b)
>  >           Base address:0x22a0 Memory:eff80000-effa0000
>  >
>  > 96154d20 150dc67c 00c78bc2 000d2659 00000000 00000000 00000000 00000000
>  > 0073d31d
>  > 00044499 00000000 00000075 00000000 00000000 00000000 00000000 00000000
>  > 00097d2d
>  > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst
>  > in_martian_src out_hit out_slow_tot out_slow_mc  gc_total gc_ignored
>  > gc_goal_miss gc_dst_overflow in_hlist_search out_hlist_search
>  > 0000fc6a  419ad196 34c586ba 00000000 00000000 0000b0e0 00006a40 00000002
>  > 002729c5 000414cb 00000000 2f5502ff 2f4567f7 0000104f 00000000 83790467
>  > 00c5e33c
>  > 0000fc6a  0000e333 00032393 00000000 00000000 00004263 00000000 00000000
>  > 0000001e 000000c0 00000002 00004650 00004626 00000000 00000000 0016b898
>  > 0000024c
>  >
>  > On Wednesday 12 January 2005 01:27 pm, Robert Olsson wrote:
>  > > Jeremy M. Guthrie writes:
>  > >  > >  > ETH3 Interrupts(calc'd from below):  1479968
>  > >  > >
>  > >  > >  Very strange...
>  > >  > >
>  > >  > >  eth3 is bound to CPU0 which in turn has all packet load... If we
>  > >  > > were to believe your CPU0 was saturated (due to the drops). We
>  > >  > > should see no (RX) interrupts on eth3. But there is a lot... one
>  > >  > > irq per every three packet. Why?
>  > >  >
>  > >  > I have no idea why it would be doing this.
>  > >
>  > >  Huh seems you didn't add the patch I sent. Below is diff from my
>  > > editor to your e1000_main.c
>  > >
>  > >        --ro
>  > >
>  > >
>  > > --- e1000_main.c.jmg 2005-01-12 20:14:08.324168072 +0100
>  > > +++ e1000_main.c 2005-01-12 20:17:24.777302656 +0100
>  > > @@ -2264,14 +2264,13 @@
>  > >   netdev->quota -= work_done;
>  > >
>  > >   /* if no Rx and Tx cleanup work was done, exit the polling mode */
>  > > - if(!tx_cleaned || (work_done < work_to_do) ||
>  > > -    !netif_running(netdev)) {
>  > > +  if( (!tx_cleaned && (work_done == 0)) || !netif_running(netdev)) {
>  > >  quit_polling: netif_rx_complete(netdev);
>  > >    e1000_irq_enable(adapter);
>  > >    return 0;
>  > >   }
>  > >
>  > > - return (work_done >= work_to_do);
>  > > + return 1;
>  > >  }
>  > >
>  > >  #endif
>  >
>  > --
>  >
>  > --------------------------------------------------
>  > Jeremy M. Guthrie        jeremy.guthrie@berbee.com
>  > Senior Network Engineer        Phone: 608-298-1061
>  > Berbee                           Fax: 608-288-3007
>  > 5520 Research Park Drive         NOC: 608-298-1102
>  > Madison, WI 53711

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 20:21                                                                 ` Robert Olsson
  2005-01-12 20:30                                                                   ` Jeremy M. Guthrie
@ 2005-01-12 20:45                                                                   ` Jeremy M. Guthrie
  2005-01-12 22:02                                                                     ` Robert Olsson
  2005-01-12 22:05                                                                     ` Jeremy M. Guthrie
  1 sibling, 2 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-12 20:45 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger, Jesse Brandeburg

[-- Attachment #1: Type: text/plain, Size: 5110 bytes --]

My throughput dropped from 500 mbps to 8mbps.  8(


On Wednesday 12 January 2005 02:21 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > Latest numbers after your patch Robert.
>
>  Did the RX interrupts go down?
>
>       --ro
>
>  > Wed Jan 12 14:05:36 CST 2005
>  > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
>  >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
>  >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
>  >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>  >           RX packets:21951496 errors:2412189 dropped:2412189
>  > overruns:377090 frame:0
>  >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
>  >           collisions:0 txqueuelen:1000
>  >           RX bytes:3455362966 (3295.2 Mb)  TX bytes:398 (398.0 b)
>  >           Base address:0x22a0 Memory:eff80000-effa0000
>  >
>  > 95b83157 150dc67c 00c734a1 000d2659 00000000 00000000 00000000 00000000
>  > 0073cee4
>  > 00044494 00000000 00000075 00000000 00000000 00000000 00000000 00000000
>  > 00097c77
>  > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst
>  > in_martian_src out_hit out_slow_tot out_slow_mc  gc_total gc_ignored
>  > gc_goal_miss gc_dst_overflow in_hlist_search out_hlist_search
>  > 0000f8cd  41403122 34c30e49 00000000 00000000 0000b0dd 00006a38 00000002
>  > 0027279d 0004147b 00000000 2f528a5f 2f42efcf 0000104f 00000000 81d262df
>  > 00c5d75c
>  > 0000f8cd  0000e332 0003238f 00000000 00000000 00004263 00000000 00000000
>  > 0000001e 000000c0 00000002 00004650 00004626 00000000 00000000 0016b879
>  > 0000024c
>  >
>  >
>  >
>  >
>  >
>  > Wed Jan 12 14:06:36 CST 2005
>  > eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
>  >           inet addr:10.253.0.1  Bcast:10.255.255.255  Mask:255.255.255.0
>  >           inet6 addr: fe80::202:b3ff:fed5:7e30/64 Scope:Link
>  >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>  >           RX packets:28053281 errors:2899869 dropped:2899869
>  > overruns:427243 frame:0
>  >           TX packets:5 errors:0 dropped:0 overruns:0 carrier:0
>  >           collisions:0 txqueuelen:1000
>  >           RX bytes:2607094069 (2486.3 Mb)  TX bytes:398 (398.0 b)
>  >           Base address:0x22a0 Memory:eff80000-effa0000
>  >
>  > 96154d20 150dc67c 00c78bc2 000d2659 00000000 00000000 00000000 00000000
>  > 0073d31d
>  > 00044499 00000000 00000075 00000000 00000000 00000000 00000000 00000000
>  > 00097d2d
>  > entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst
>  > in_martian_src out_hit out_slow_tot out_slow_mc  gc_total gc_ignored
>  > gc_goal_miss gc_dst_overflow in_hlist_search out_hlist_search
>  > 0000fc6a  419ad196 34c586ba 00000000 00000000 0000b0e0 00006a40 00000002
>  > 002729c5 000414cb 00000000 2f5502ff 2f4567f7 0000104f 00000000 83790467
>  > 00c5e33c
>  > 0000fc6a  0000e333 00032393 00000000 00000000 00004263 00000000 00000000
>  > 0000001e 000000c0 00000002 00004650 00004626 00000000 00000000 0016b898
>  > 0000024c
>  >
>  > On Wednesday 12 January 2005 01:27 pm, Robert Olsson wrote:
>  > > Jeremy M. Guthrie writes:
>  > >  > >  > ETH3 Interrupts(calc'd from below):  1479968
>  > >  > >
>  > >  > >  Very strange...
>  > >  > >
>  > >  > >  eth3 is bound to CPU0 which in turn has all packet load... If we
>  > >  > > were to believe your CPU0 was saturated (due to the drops). We
>  > >  > > should see no (RX) interrupts on eth3. But there is a lot... one
>  > >  > > irq per every three packet. Why?
>  > >  >
>  > >  > I have no idea why it would be doing this.
>  > >
>  > >  Huh seems you didn't add the patch I sent. Below is diff from my
>  > > editor to your e1000_main.c
>  > >
>  > >        --ro
>  > >
>  > >
>  > > --- e1000_main.c.jmg 2005-01-12 20:14:08.324168072 +0100
>  > > +++ e1000_main.c 2005-01-12 20:17:24.777302656 +0100
>  > > @@ -2264,14 +2264,13 @@
>  > >   netdev->quota -= work_done;
>  > >
>  > >   /* if no Rx and Tx cleanup work was done, exit the polling mode */
>  > > - if(!tx_cleaned || (work_done < work_to_do) ||
>  > > -    !netif_running(netdev)) {
>  > > +  if( (!tx_cleaned && (work_done == 0)) || !netif_running(netdev)) {
>  > >  quit_polling: netif_rx_complete(netdev);
>  > >    e1000_irq_enable(adapter);
>  > >    return 0;
>  > >   }
>  > >
>  > > - return (work_done >= work_to_do);
>  > > + return 1;
>  > >  }
>  > >
>  > >  #endif
>  >
>  > --
>  >
>  > --------------------------------------------------
>  > Jeremy M. Guthrie        jeremy.guthrie@berbee.com
>  > Senior Network Engineer        Phone: 608-298-1061
>  > Berbee                           Fax: 608-288-3007
>  > 5520 Research Park Drive         NOC: 608-298-1102
>  > Madison, WI 53711

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 20:45                                                                   ` Jeremy M. Guthrie
@ 2005-01-12 22:02                                                                     ` Robert Olsson
  2005-01-12 22:21                                                                       ` Jeremy M. Guthrie
  2005-01-12 22:05                                                                     ` Jeremy M. Guthrie
  1 sibling, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-12 22:02 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger, Jesse Brandeburg



Jeremy M. Guthrie writes:

 > My throughput dropped from 500 mbps to 8mbps.  8(

 Weird!

 >            CPU0       CPU1
 >  18: 3586173518       1815   IO-APIC-level  eth3
 >  20:          2 2464382507   IO-APIC-level  eth2

 >            CPU0       CPU1
 >  18: 3586173518       1815   IO-APIC-level  eth3
 >  20:          2 2464387985   IO-APIC-level  eth2

 There are no irq's on eth3 at all so RX softirq is constantly running.
 Which means it's deferred to ksoftirqd now and running under scheduler 
 context. Do you have anything that competes with ksoftirqd for CPU0 on 
 your system?
 It used be recommended to increase the priority of ksoftirqd but I wonder 
 what's going on your system. We see interrupts on eth2...

 And tsquz (3:rd col) in /proc/net/sofnet_stat indicates there are very 
 little activity from the RX softirq. It's soon time time up here.
 

						--ro
 

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 20:45                                                                   ` Jeremy M. Guthrie
  2005-01-12 22:02                                                                     ` Robert Olsson
@ 2005-01-12 22:05                                                                     ` Jeremy M. Guthrie
  2005-01-12 22:22                                                                       ` Robert Olsson
  1 sibling, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-12 22:05 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger, Jesse Brandeburg

[-- Attachment #1: Type: text/plain, Size: 628 bytes --]

I am now getting some push back from the project manager on this performance 
problem.  I am wondering if you think faster CPUs will 
a)  help relieve the symptoms of this problem
b)  not help because now we will hit a '# of routes in the route-cache' 
problem
c)  or will help to a point till the # interrupts come back and bite us.

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 22:02                                                                     ` Robert Olsson
@ 2005-01-12 22:21                                                                       ` Jeremy M. Guthrie
       [not found]                                                                         ` <16869.42247.126428.508479@robur.slu.se>
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-12 22:21 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger, Jesse Brandeburg

[-- Attachment #1: Type: text/plain, Size: 1368 bytes --]

On Wednesday 12 January 2005 04:02 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > My throughput dropped from 500 mbps to 8mbps.  8(
>
>  Weird!
>
>  >            CPU0       CPU1
>  >  18: 3586173518       1815   IO-APIC-level  eth3
>  >  20:          2 2464382507   IO-APIC-level  eth2
>  >
>  >            CPU0       CPU1
>  >  18: 3586173518       1815   IO-APIC-level  eth3
>  >  20:          2 2464387985   IO-APIC-level  eth2
>
>  There are no irq's on eth3 at all so RX softirq is constantly running.
>  Which means it's deferred to ksoftirqd now and running under scheduler
>  context. Do you have anything that competes with ksoftirqd for CPU0 on
>  your system?
This box is primarily routing.  Nothing should be competing for CPU0

>  It used be recommended to increase the priority of ksoftirqd but I wonder
>  what's going on your system. We see interrupts on eth2...
>
>  And tsquz (3:rd col) in /proc/net/sofnet_stat indicates there are very
>  little activity from the RX softirq. It's soon time time up here.
>
>
>       --ro

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 22:05                                                                     ` Jeremy M. Guthrie
@ 2005-01-12 22:22                                                                       ` Robert Olsson
  2005-01-12 22:30                                                                         ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-12 22:22 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson, Stephen Hemminger, Jesse Brandeburg


Jeremy M. Guthrie writes:
 > I am now getting some push back from the project manager on this performance 
 > problem.  I am wondering if you think faster CPUs will 
 > a)  help relieve the symptoms of this problem
 > b)  not help because now we will hit a '# of routes in the route-cache' 
 > problem
 > c)  or will help to a point till the # interrupts come back and bite us.

 Back out the patch I sent.and have hardirq's to run RX-softirq as you 
 did before but something is very wrong. You didn't answer if there were
 other load on the machine...

 route-cache can probably be tuned you as have four times the linear seach
 I see in one PIII system at 110 kpps w. production traffic.

 Of course the non-engineering solution is to buy more CPU... :-)

					       --ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 22:22                                                                       ` Robert Olsson
@ 2005-01-12 22:30                                                                         ` Jeremy M. Guthrie
  0 siblings, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-12 22:30 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson, Stephen Hemminger, Jesse Brandeburg

[-- Attachment #1: Type: text/plain, Size: 1445 bytes --]

On Wednesday 12 January 2005 04:22 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > I am now getting some push back from the project manager on this
>  > performance problem.  I am wondering if you think faster CPUs will
>  > a)  help relieve the symptoms of this problem
>  > b)  not help because now we will hit a '# of routes in the route-cache'
>  > problem
>  > c)  or will help to a point till the # interrupts come back and bite us.
>
>  Back out the patch I sent.and have hardirq's to run RX-softirq as you
>  did before but something is very wrong. You didn't answer if there were
>  other load on the machine...
I have backed out.  As for the load, this box only does policy routing.  Any 
other functions it performs are part of its automated system to download the 
next days policy-routing config.

>  route-cache can probably be tuned you as have four times the linear seach
>  I see in one PIII system at 110 kpps w. production traffic.
How would I go about tuning that?

>  Of course the non-engineering solution is to buy more CPU... :-)
That is good to know.  This will help me calm the situation a bit.  8)

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
       [not found]                                                                         ` <16869.42247.126428.508479@robur.slu.se>
@ 2005-01-12 22:42                                                                           ` Jeremy M. Guthrie
  2005-01-12 22:47                                                                           ` Jeremy M. Guthrie
  1 sibling, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-12 22:42 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 2983 bytes --]

Wed Jan 12 16:36:29 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          RX packets:1868701 errors:894743 dropped:894743 overruns:410647 
frame:0
ba659628 150dc67c 00c8ad95 000d2659 00000000 00000000 00000000 00000000 
00764a9d
00045bd7 00000000 00000120 00000000 00000000 00000000 00000000 00000000 
0009b5c4
entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
gc_dst_overflow in_hlist_search out_hlist_search
00008e84  62d61b7d 37d9063a 00000000 00000000 0000b57e 00006c14 00000002  
0027d72c 00042dfa 00000000 307d209e 30505aea 001d0a7c 001cf933 108c7bfb 
00c976a6
00008e84  0000f0af 00032cd5 00000000 00000000 00004360 00000000 00000000  
00000020 000000df 00000002 00004874 000046dd 0000016c 00000165 0016e55d 
0000028f
           CPU0       CPU1
 18: 3680079907       1815   IO-APIC-level  eth3
 20:          2 2490755322   IO-APIC-level  eth2

Wed Jan 12 16:37:30 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          RX packets:6313404 errors:1216174 dropped:1216174 overruns:509016 
frame:0
baa97098 150dc67c 00c8e8ac 000d2659 00000000 00000000 00000000 00000000 
00764df1
00045be8 00000000 00000120 00000000 00000000 00000000 00000000 00000000 
0009b650
entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
gc_dst_overflow in_hlist_search out_hlist_search
0000e10e  63177e31 37db7aec 00000000 00000000 0000b587 00006c22 00000002  
0027d819 00042e29 00000000 307f9564 3052cf38 001d0a7f 001cf933 114eef21 
00c97a1e
0000e10e  0000f0b5 00032ce0 00000000 00000000 00004360 00000000 00000000  
00000020 000000e0 00000002 00004875 000046de 0000016c 00000165 0016e5a1 
00000295
           CPU0       CPU1
 18: 3680513406       1815   IO-APIC-level  eth3
 20:          2 2490979016   IO-APIC-level  eth2


On Wednesday 12 January 2005 04:30 pm, Robert Olsson wrote:
> Can you give this a last try. It's from an older driver
>
>
> e1000_clean(struct net_device *netdev, int *budget)
> {
>         struct e1000_adapter *adapter = netdev->priv;
>         int work_to_do = min(*budget, netdev->quota);
>         int work_done = 0;
>
>         e1000_clean_tx_irq(adapter);
>         e1000_clean_rx_irq(adapter, &work_done, work_to_do);
>
>         *budget -= work_done;
>         netdev->quota -= work_done;
>
>         if(work_done < work_to_do) {
>                 netif_rx_complete(netdev);
>                 e1000_irq_enable(adapter);
>         }
>
>         return (work_done >= work_to_do);
> }

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
       [not found]                                                                         ` <16869.42247.126428.508479@robur.slu.se>
  2005-01-12 22:42                                                                           ` Jeremy M. Guthrie
@ 2005-01-12 22:47                                                                           ` Jeremy M. Guthrie
  2005-01-12 23:19                                                                             ` Robert Olsson
  1 sibling, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-12 22:47 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 1385 bytes --]

Here is what it looked like I was getting before from the old driver... that 
took me down to 8mbps.

"printk: 4624 messages suppressed.
dst cache overflow
printk: 4630 messages suppressed.
dst cache overflow
printk: 4631 messages suppressed.
dst cache overflow
printk: 4646 messages suppressed.
dst cache overflow
printk: 4645 messages suppressed.
dst cache overflow"

On Wednesday 12 January 2005 04:30 pm, Robert Olsson wrote:
> Can you give this a last try. It's from an older driver
>
>
> e1000_clean(struct net_device *netdev, int *budget)
> {
>         struct e1000_adapter *adapter = netdev->priv;
>         int work_to_do = min(*budget, netdev->quota);
>         int work_done = 0;
>
>         e1000_clean_tx_irq(adapter);
>         e1000_clean_rx_irq(adapter, &work_done, work_to_do);
>
>         *budget -= work_done;
>         netdev->quota -= work_done;
>
>         if(work_done < work_to_do) {
>                 netif_rx_complete(netdev);
>                 e1000_irq_enable(adapter);
>         }
>
>         return (work_done >= work_to_do);
> }

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 22:47                                                                           ` Jeremy M. Guthrie
@ 2005-01-12 23:19                                                                             ` Robert Olsson
  2005-01-12 23:23                                                                               ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-12 23:19 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson


Jeremy M. Guthrie writes:
 > Here is what it looked like I was getting before from the old driver... that 
 > took me down to 8mbps.
 > 
 > "printk: 4624 messages suppressed.
 > dst cache overflow
 > printk: 4630 messages suppressed.
 > dst cache overflow
 > printk: 4631 messages suppressed.
 > dst cache overflow
 > printk: 4646 messages suppressed.
 > dst cache overflow
 > printk: 4645 messages suppressed.
 > dst cache overflow"
 
 Thanks! 

 This a known problem. Remember I asked you about this. We're now in the 
 RCU route-hash problem again. This is not necessary solved by more CPU
 as higher t-put forces more dst-entries to be freed aand we get 
 closer to max_size.

 As your traffic looks sane double your bucket size of the route hash to 
 start with. Use the boot option w. rhash_entries. look at rtstat  


						  --ro.

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 23:19                                                                             ` Robert Olsson
@ 2005-01-12 23:23                                                                               ` Jeremy M. Guthrie
  2005-01-13  8:56                                                                                 ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-12 23:23 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 1340 bytes --]

On Wednesday 12 January 2005 05:19 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > Here is what it looked like I was getting before from the old driver...
>  > that took me down to 8mbps.
>  >
>  > "printk: 4624 messages suppressed.
>  > dst cache overflow
>  > printk: 4630 messages suppressed.
>  > dst cache overflow
>  > printk: 4631 messages suppressed.
>  > dst cache overflow
>  > printk: 4646 messages suppressed.
>  > dst cache overflow
>  > printk: 4645 messages suppressed.
>  > dst cache overflow"
>
>  Thanks!
>
>  This a known problem. Remember I asked you about this. We're now in the
>  RCU route-hash problem again. This is not necessary solved by more CPU
>  as higher t-put forces more dst-entries to be freed aand we get
>  closer to max_size.
>
>  As your traffic looks sane double your bucket size of the route hash to
>  start with. Use the boot option w. rhash_entries. look at rtstat
Does it make sense that the driver would kill throughput and force us to 
output these types of messages?

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-12 23:23                                                                               ` Jeremy M. Guthrie
@ 2005-01-13  8:56                                                                                 ` Robert Olsson
  2005-01-13 19:28                                                                                   ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-13  8:56 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson


Jeremy M. Guthrie writes:

 > >  As your traffic looks sane double your bucket size of the route hash to
 > >  start with. Use the boot option w. rhash_entries. look at rtstat

 > Does it make sense that the driver would kill throughput and force us to 
 > output these types of messages?

 No the other way around... Higher (driver/cpu) throughput/load causes more 
 dst-entries to freed and you reach the max_size*ip_rt_gc_min_interval which
 is a constant and get "dst cache overflow".

 Increasing  rhash_entries is the easieast way to attack this.. Give this
 a try. Monitor with rtstat. You might even have to quadruple your size.

						--ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-13  8:56                                                                                 ` Robert Olsson
@ 2005-01-13 19:28                                                                                   ` Jeremy M. Guthrie
  2005-01-13 20:00                                                                                     ` David S. Miller
  2005-01-13 21:12                                                                                     ` Robert Olsson
  0 siblings, 2 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-13 19:28 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 5591 bytes --]

I after a few revs I just bumped rhash_entries to 2.4mil in an attempt to get 
well above my actual usage. 

You can see below I am over 600K entries before it blows them away and 
restarts.  How do I bump up the time from 10 minutes to something longer?  
With the way our system works, entries should be good for a day as we won't 
reprogram the policy route table but once a day.

I still have some instrumented network card drivers to work with but I now 
show some 30-40% idle CPU on CPU0 but still with 0.3% packet loss.  I'll post 
stats once I get the instrumented drivers in.

 size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
GC: tot ignored goal_miss ovrf HASH: in_search out_search
611167     98712     645     0     0     0     0     0        30       0 
1074229262     645     643         0    0           61922         17
611461     96203     591     0     0     0     1     0        42       1 
3220738034     592     590         0    0           63160         14
611759     93852     592     0     0     0     0     0        18       0      
0     591     589         0    0           63078          6
612094     95276     632     0     0     0     0     0        28       0      
0     632     630         0    0           63336          8
612368     94945     580     0     0     0     0     0        22       0      
0     578     576         0    0           61224         18
612670     99258     622     0     0     0     0     0        28       0      
0     621     619         0    0           63922          8
613025     93573     666     0     0     0     0     0        16       0      
0     665     663         0    0           61781          6
613394     83917     722     0     0     0     0     0         8       0      
0     721     719         0    0           55533         10
613697     85851     634     0     0     0     0     0        10       0      
0     633     631         0    0           56394         12
613986     81854     611     0     0     0     0     0         8       0      
0     610     608         0    0           54273          8
614349     81419     704     0     0     0     0     0         4       0      
0     702     700         0    0           52641          8
614651     83312     616     0     0     0     0     0        14       0      
0     617     615         0    0           54160         12
614962     83119     651     0     0     0     0     0         6       0      
0     651     649         0    0           56612          4
615264     84871     583     0     0     0     0     0        10       0      
0     583     581         0    0           56130         12
615521     83932     557     0     0     0     0     0         8       0      
0     557     555         0    0           56229         10
615852     86368     626     0     0     0     0     0         8       0      
0     624     622         0    0           56504         10
493558     47553    4603     0     0     0     0     0         2       0      
0    4166    4164         0    0           28346          0
10091     46526    7096     0     0     0     0     0         2       3      0       
0       0         0    0             554          0
16238     80565    6145     0     0     0     0     0         6       3      0       
0       0         0    0            1334          0
21754     81224    5515     0     0     0     0     0         6       2      0       
0       0         0    0            1793          0
 size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
GC: tot ignored goal_miss ovrf HASH: in_search out_search
26737     80792    4982     0     0     0     0     0         5       2 
1074229262       0       0         0    0            2669          0
31085     82895    4347     0     0     0     0     0         5       1 
3220738034       0       0         0    0            2397          0
35333     83220    4248     0     0     0     0     0         1       1      0       
0       0         0    0            2754          0
39053     83910    3720     0     0     0     0     0         7       1      0       
0       0         0    0            3328          0
42692     82373    3634     0     0     0     0     0         8       6      0       
0       0         0    0            3485          3
46404     84900    3707     0     0     0     0     0        15       7      0       
0       0         0    0            3889          1


On Thursday 13 January 2005 02:56 am, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > >  As your traffic looks sane double your bucket size of the route hash
>  > > to start with. Use the boot option w. rhash_entries. look at rtstat
>  >
>  > Does it make sense that the driver would kill throughput and force us to
>  > output these types of messages?
>
>  No the other way around... Higher (driver/cpu) throughput/load causes more
>  dst-entries to freed and you reach the max_size*ip_rt_gc_min_interval
> which is a constant and get "dst cache overflow".
>
>  Increasing  rhash_entries is the easieast way to attack this.. Give this
>  a try. Monitor with rtstat. You might even have to quadruple your size.
>
>       --ro

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-13 19:28                                                                                   ` Jeremy M. Guthrie
@ 2005-01-13 20:00                                                                                     ` David S. Miller
  2005-01-13 20:43                                                                                       ` Jeremy M. Guthrie
  2005-01-13 21:12                                                                                     ` Robert Olsson
  1 sibling, 1 reply; 88+ messages in thread
From: David S. Miller @ 2005-01-13 20:00 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert.Olsson

On Thu, 13 Jan 2005 13:28:29 -0600
"Jeremy M. Guthrie" <jeremy.guthrie@berbee.com> wrote:

> You can see below I am over 600K entries before it blows them away and 
> restarts.  How do I bump up the time from 10 minutes to something longer?  
> With the way our system works, entries should be good for a day as we won't 
> reprogram the policy route table but once a day.

Increase /proc/sys/net/ipv4/ip_rt_secret_interval

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-13 20:00                                                                                     ` David S. Miller
@ 2005-01-13 20:43                                                                                       ` Jeremy M. Guthrie
  2005-01-13 23:13                                                                                         ` David S. Miller
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-13 20:43 UTC (permalink / raw)
  To: netdev; +Cc: David S. Miller, Robert.Olsson

[-- Attachment #1: Type: text/plain, Size: 883 bytes --]

I don't show that proc file being there.  I see it in 'iproute.c' but not in 
proc.

On Thursday 13 January 2005 02:00 pm, David S. Miller wrote:
> On Thu, 13 Jan 2005 13:28:29 -0600
>
> "Jeremy M. Guthrie" <jeremy.guthrie@berbee.com> wrote:
> > You can see below I am over 600K entries before it blows them away and
> > restarts.  How do I bump up the time from 10 minutes to something longer?
> > With the way our system works, entries should be good for a day as we
> > won't reprogram the policy route table but once a day.
>
> Increase /proc/sys/net/ipv4/ip_rt_secret_interval

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-13 19:28                                                                                   ` Jeremy M. Guthrie
  2005-01-13 20:00                                                                                     ` David S. Miller
@ 2005-01-13 21:12                                                                                     ` Robert Olsson
  2005-01-13 22:27                                                                                       ` Jeremy M. Guthrie
  2005-01-14 14:59                                                                                       ` Jeremy M. Guthrie
  1 sibling, 2 replies; 88+ messages in thread
From: Robert Olsson @ 2005-01-13 21:12 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson


Jeremy M. Guthrie writes:
 > I after a few revs I just bumped rhash_entries to 2.4mil in an attempt to get 
 > well above my actual usage. 

 A bit hefty size :-) But the stats are looking much better as we do much
 less linear search (in_search) in hash and less fib lookups (tot) 

 And you have now "dst cache overflows"?
 Is the e1000 patch I sent in use?

 > You can see below I am over 600K entries before it blows them away and 
 > restarts. 

 This is a part of GC process to reclaim memory and reclaim unused dst entries.
 This 

 >  size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
 > GC: tot ignored goal_miss ovrf HASH: in_search out_search


 > 615852     86368     626     0     0     0     0     0         8       0      
 > 0     624     622         0    0           56504         10
 > 493558     47553    4603     0     0     0     0     0         2       0      
 > 0    4166    4164         0    0           28346          0
 > 10091     46526    7096     0     0     0     0     0         2       3      0       
 > 0       0         0    0             554          0
 > 16238     80565    6145     0     0     0     0     0         6       3      0       
 > 0       0         0    0            1334          0

 In short we reduce the hash size to remove unused flows and let it grow again. 
 You see from (tot) that we have recreate may of the flows at this point. Most 
 likely this is where we drop the packets. We have monitored small drops in our 
 system when GC happens. The GC can be smoothen out but we leave that for now.

 Is the e1000 patch I sent in use?

 > How do I bump up the time from 10 minutes to something longer?  

 Davem pointed out another periodic task thats flushes the cache totally it's

   /proc/sys/net/ipv4/route/secret_interval

 It flushes the cache totally we so all current flows has be recreated. You
 probably drop packets here in your setup. Yes it can be idea to increase it
 or run the flush manually. But most routers drop packets now and then.


						--ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-13 21:12                                                                                     ` Robert Olsson
@ 2005-01-13 22:27                                                                                       ` Jeremy M. Guthrie
  2005-01-14 15:44                                                                                         ` Robert Olsson
  2005-01-14 14:59                                                                                       ` Jeremy M. Guthrie
  1 sibling, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-13 22:27 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 8418 bytes --]

On Thursday 13 January 2005 03:12 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > I after a few revs I just bumped rhash_entries to 2.4mil in an attempt
>  > to get well above my actual usage.
>
>  A bit hefty size :-) But the stats are looking much better as we do much
>  less linear search (in_search) in hash and less fib lookups (tot)
Okay.

>  And you have now "dst cache overflows"?
No, I haven't gotten any of these yet.

>  Is the e1000 patch I sent in use?
yes.  I also have another E1000 driver I haven't had a chance to try yet.  It 
is a bit more instrumented.  

>  > You can see below I am over 600K entries before it blows them away and
>  > restarts.
>
>  This is a part of GC process to reclaim memory and reclaim unused dst
> entries. This
>
>  >  size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot 
>  >    mc GC: tot ignored goal_miss ovrf HASH: in_search out_search
>  >
>  >
>  > 615852     86368     626     0     0     0     0     0         8       0
>  > 0     624     622         0    0           56504         10
>  > 493558     47553    4603     0     0     0     0     0         2       0
>  > 0    4166    4164         0    0           28346          0
>  > 10091     46526    7096     0     0     0     0     0         2       3 
>  >     0 0       0         0    0             554          0
>  > 16238     80565    6145     0     0     0     0     0         6       3 
>  >     0 0       0         0    0            1334          0
>
>  In short we reduce the hash size to remove unused flows and let it grow
> again. You see from (tot) that we have recreate may of the flows at this
> point. Most likely this is where we drop the packets. We have monitored
> small drops in our system when GC happens. The GC can be smoothen out but
> we leave that for now.
Sorry, not quite following.

IN:Hits are cache hits yes?  Tot, are the total number of flows created since 
we last looked at the total flow count, correct?  What would cause a packet 
to drop in the network stack and thus showup in /proc/net/softnet_stat?

>  > How do I bump up the time from 10 minutes to something longer?
>
>  Davem pointed out another periodic task thats flushes the cache totally
> it's
>
>    /proc/sys/net/ipv4/route/secret_interval
>
>  It flushes the cache totally we so all current flows has be recreated. You
>  probably drop packets here in your setup. Yes it can be idea to increase
> it or run the flush manually. But most routers drop packets now and then.
If I set the secret_interval to 60 seconds then I drop over 1% of all packets 
coming through.  So GC isn't exactly my friend.

Performance has picked up.  I am not dropping packets anymore except during 
GC.  I upped my interval from 600 seconds to 1800 seconds.   

Here are 15 second snapshots.  Line 3 appears to be when GC take effect.  
Afterwards, everything stabilizes.  These numbers are much better.

Thu Jan 13 16:10:30 CST 2005 entries:  000de44a  Packets:  1255162  Errors:  0  
PPS:  83677  Percentage:  0.0%
Thu Jan 13 16:10:45 CST 2005 entries:  000df2ad  Packets:  1303050  Errors:  
3875  PPS:  86870  Percentage:  0.29%
Thu Jan 13 16:11:00 CST 2005 entries:  0000b053  Packets:  1265398  Errors:  
38586  PPS:  84359  Percentage:  3.04%
Thu Jan 13 16:11:15 CST 2005 entries:  00013df8  Packets:  1310618  Errors:  0  
PPS:  87374  Percentage:  0.0%
Thu Jan 13 16:11:30 CST 2005 entries:  0001b527  Packets:  1282435  Errors:  0  
PPS:  85495  Percentage:  0.0%
Thu Jan 13 16:11:45 CST 2005 entries:  000222bb  Packets:  1213217  Errors:  0  
PPS:  80881  Percentage:  0.0%
Thu Jan 13 16:12:01 CST 2005 entries:  00027c7e  Packets:  1279811  Errors:  0  
PPS:  85320  Percentage:  0.0%
Thu Jan 13 16:12:16 CST 2005 entries:  0002c5d5  Packets:  1224232  Errors:  0  
PPS:  81615  Percentage:  0.0%
Thu Jan 13 16:12:31 CST 2005 entries:  0003090c  Packets:  1243539  Errors:  0  
PPS:  82902  Percentage:  0.0%
Thu Jan 13 16:12:46 CST 2005 entries:  00034d41  Packets:  1267200  Errors:  0  
PPS:  84480  Percentage:  0.0%
Thu Jan 13 16:13:01 CST 2005 entries:  00038f82  Packets:  1238821  Errors:  0  
PPS:  82588  Percentage:  0.0%
Thu Jan 13 16:13:16 CST 2005 entries:  0003cf6a  Packets:  1245474  Errors:  0  
PPS:  83031  Percentage:  0.0%
Thu Jan 13 16:13:31 CST 2005 entries:  00040d23  Packets:  1266478  Errors:  0  
PPS:  84431  Percentage:  0.0%
Thu Jan 13 16:13:46 CST 2005 entries:  00044918  Packets:  1247576  Errors:  0  
PPS:  83171  Percentage:  0.0%
Thu Jan 13 16:14:01 CST 2005 entries:  00048520  Packets:  1223002  Errors:  0  
PPS:  81533  Percentage:  0.0%
Thu Jan 13 16:14:16 CST 2005 entries:  0004c0b6  Packets:  1303942  Errors:  
333  PPS:  86929  Percentage:  0.2%
Thu Jan 13 16:14:32 CST 2005 entries:  0004f83e  Packets:  1203334  Errors:  0  
PPS:  80222  Percentage:  0.0%
Thu Jan 13 16:14:47 CST 2005 entries:  00053241  Packets:  1216611  Errors:  0  
PPS:  81107  Percentage:  0.0%
Thu Jan 13 16:15:02 CST 2005 entries:  00056f97  Packets:  1281206  Errors:  0  
PPS:  85413  Percentage:  0.0%
Thu Jan 13 16:15:17 CST 2005 entries:  0005b020  Packets:  1270007  Errors:  0  
PPS:  84667  Percentage:  0.0%
Thu Jan 13 16:15:32 CST 2005 entries:  0005eb63  Packets:  1250099  Errors:  0  
PPS:  83339  Percentage:  0.0%
Thu Jan 13 16:15:47 CST 2005 entries:  00061e08  Packets:  1183444  Errors:  0  
PPS:  78896  Percentage:  0.0%
Thu Jan 13 16:16:02 CST 2005 entries:  0006489b  Packets:  1246170  Errors:  
3791  PPS:  83078  Percentage:  0.30%
Thu Jan 13 16:16:17 CST 2005 entries:  00066f1f  Packets:  1233601  Errors:  
4141  PPS:  82240  Percentage:  0.33%
Thu Jan 13 16:16:32 CST 2005 entries:  000695aa  Packets:  1273744  Errors:  
3798  PPS:  84916  Percentage:  0.29%
Thu Jan 13 16:16:47 CST 2005 entries:  0006ba5d  Packets:  1263619  Errors:  
4219  PPS:  84241  Percentage:  0.33%
Thu Jan 13 16:17:03 CST 2005 entries:  0006df19  Packets:  1240743  Errors:  
3616  PPS:  82716  Percentage:  0.29%

----------one other snapshot------------
Thu Jan 13 16:09:03 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          RX packets:459165122 errors:3427143 dropped:3427143 overruns:2045357 
frame:0
1b5e031d 00000000 0000a829 00000000 00000000 00000000 00000000 00000000 
0002cbd7
000072c1 00000000 00000001 00000000 00000000 00000000 00000000 00000000 
00001e00
entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
gc_dst_overflow in_hlist_search out_hlist_search
000d92e0  1a0ecfdc 014e19f9 00000000 00000000 000000a6 000000df 00000000  
00009558 00000c5e 00000000 000b7605 000b6c68 00000000 00000000 07c9547f 
0000398d
000d92e0  00001340 00005e40 00000000 00000000 0000005e 00000000 00000000  
00000007 00000036 00000002 00000002 00000002 00000000 00000000 00001542 
00000004
           CPU0       CPU1
 18:  123586344       8007   IO-APIC-level  eth3
 20:          1   18109191   IO-APIC-level  eth2

Thu Jan 13 16:10:03 CST 2005
eth3      Link encap:Ethernet  HWaddr 00:02:B3:D5:7E:30
          RX packets:464242944 errors:3427143 dropped:3427143 overruns:2045357 
frame:0
1bab839b 00000000 0000a82d 00000000 00000000 00000000 00000000 00000000 
0002d2bc
000072e3 00000000 00000001 00000000 00000000 00000000 00000000 00000000 
00001ed8
entries  in_hit in_slow_tot in_no_route in_brd in_martian_dst in_martian_src  
out_hit out_slow_tot out_slow_mc  gc_total gc_ignored gc_goal_miss 
gc_dst_overflow in_hlist_search out_hlist_search
000dcaba  1a5bd4fd 014e9288 00000000 00000000 000000a6 000000df 00000000  
00009678 00000c6a 00000000 000bee9e 000be489 00000000 00000000 08109f0f 
00003a97
000dcaba  00001349 00005e58 00000000 00000000 0000005e 00000000 00000000  
00000007 00000036 00000002 00000002 00000002 00000000 00000000 00001597 
00000004
           CPU0       CPU1
 18:  125388992       8007   IO-APIC-level  eth3
 20:          1   18340497   IO-APIC-level  eth2


-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
       [not found] <C925F8B43D79CC49ACD0601FB68FF50C02D39006@orsmsx408>
@ 2005-01-13 22:55 ` Jeremy M. Guthrie
  0 siblings, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-13 22:55 UTC (permalink / raw)
  To: netdev; +Cc: Brandeburg, Jesse, Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 5387 bytes --]

Thu Jan 13 16:50:06 CST 2005 entries:  0000cc39  Packets:  1127110  Errors:  
264136  PPS:  75140  Percentage:  23.43%
Thu Jan 13 16:50:22 CST 2005 entries:  000142c0  Packets:  1148930  Errors:  
743  PPS:  76595  Percentage:  0.6%
Thu Jan 13 16:50:37 CST 2005 entries:  0001aa91  Packets:  1158591  Errors:  
116  PPS:  77239  Percentage:  0.1%
Thu Jan 13 16:50:52 CST 2005 entries:  00021146  Packets:  1192241  Errors:  
11648  PPS:  79482  Percentage:  0.97%
Thu Jan 13 16:51:07 CST 2005 entries:  00025c42  Packets:  1227489  Errors:  
1056  PPS:  81832  Percentage:  0.8%
Thu Jan 13 16:51:22 CST 2005 entries:  00029ca0  Packets:  1217954  Errors:  
365  PPS:  81196  Percentage:  0.2%


ethtool -S eth3
NIC statistics:
     rx_packets: 8778549
     tx_packets: 5
     rx_bytes: 327267728
     tx_bytes: 398
     rx_errors: 575319
     tx_errors: 0
     rx_dropped: 360028
     tx_dropped: 0
     multicast: 0
     collisions: 0
     rx_length_errors: 0
     rx_over_errors: 0
     rx_crc_errors: 0
     rx_frame_errors: 0
     rx_fifo_errors: 215291
     rx_missed_errors: 215291
     tx_aborted_errors: 0
     tx_carrier_errors: 0
     tx_fifo_errors: 0
     tx_heartbeat_errors: 0
     tx_window_errors: 0
     tx_abort_late_coll: 0
     tx_deferred_ok: 0
     tx_single_coll_ok: 0
     tx_multi_coll_ok: 0
     rx_long_length_errors: 0
     rx_short_length_errors: 0
     rx_align_errors: 0
     tx_tcp_seg_good: 0
     tx_tcp_seg_failed: 0
     rx_flow_control_xon: 0
     rx_flow_control_xoff: 0
     tx_flow_control_xon: 0
     tx_flow_control_xoff: 0
     rx_long_byte_count: 4622235024
     rx_csum_offload_good: 8285347
     int_tx_desc: 5
     int_tx_queueempty: 6
     int_link_state: 1
     int_rx_frame_err: 0
     int_rx_desc_min_thresh: 1330
     int_rx_fifo_ovr: 47
     int_rx_timer: 1768171
     int_mdio: 0
     int_rxcfg: 1
     int_gpio_pins: 0
     rx_csum_offload_errors: 698


On Wednesday 12 January 2005 06:49 pm, Brandeburg, Jesse wrote:
> I didn't send this to netdev.... if the interrupt counting code does
> something good then we can publish it.
>
> Jeremy, I would agree a faster CPU is going to help you handle more
> traffic.  I can't speak to the routing thing.  Your test would be very
> interesting if we could set up something similar here, unfortunately
> we're mostly interested in network device performance and not so much on
> kernel policy routing.  I personally would be interested in having
> something set up to "play" with the driver on, but it may be doubtful
> how much time I would get to spend on it.
>
> Anyway, here is a driver that counts interrupt sources, you can get the
> counts from
> ethtool -S eth3
>
> you'll need to compile it like so:
> make CFLAGS_EXTRA=-DE1000_COUNT_ICR
>
> any messages in /var/log/messages from the network stack?  (I just saw
> your netdev email about dst cache overflow) This driver has what we
> think should be the correct napi code in e1000_clean.  If robert's fix
> works better for you then stick with it, and let me know cause what I'm
> sending you now is what we're going forward with unless we hear about
> problems.
>
> If you want to chat over an instant messenger of some kind here is my
> info:
> Aim: jbrandeb
> msn: go_jesse@hotmail.com
> yahoo: go_jesse
>
> I appreciate your patience as we try different stuff.  I know I'm poking
> at the driver a lot, but the high interrupt counts seem a little weird
> given the load of your system.
>
> jesse
>
> -----Original Message-----
> From: Jeremy M. Guthrie [mailto:jeremy.guthrie@berbee.com]
> Sent: Wednesday, January 12, 2005 2:31 PM
> To: netdev@oss.sgi.com
> Cc: Robert Olsson; Stephen Hemminger; Brandeburg, Jesse
> Subject: Re: V2.4 policy router operates faster/better than V2.6
>
> On Wednesday 12 January 2005 04:22 pm, Robert Olsson wrote:
> > Jeremy M. Guthrie writes:
> >  > I am now getting some push back from the project manager on this
> >  > performance problem.  I am wondering if you think faster CPUs will
> >  > a)  help relieve the symptoms of this problem
> >  > b)  not help because now we will hit a '# of routes in the
>
> route-cache'
>
> >  > problem
> >  > c)  or will help to a point till the # interrupts come back and
>
> bite us.
>
> >  Back out the patch I sent.and have hardirq's to run RX-softirq as you
> >  did before but something is very wrong. You didn't answer if there
>
> were
>
> >  other load on the machine...
>
> I have backed out.  As for the load, this box only does policy routing.
> Any
> other functions it performs are part of its automated system to download
> the
> next days policy-routing config.
>
> >  route-cache can probably be tuned you as have four times the linear
>
> seach
>
> >  I see in one PIII system at 110 kpps w. production traffic.
>
> How would I go about tuning that?
>
> >  Of course the non-engineering solution is to buy more CPU... :-)
>
> That is good to know.  This will help me calm the situation a bit.  8)

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-13 20:43                                                                                       ` Jeremy M. Guthrie
@ 2005-01-13 23:13                                                                                         ` David S. Miller
  0 siblings, 0 replies; 88+ messages in thread
From: David S. Miller @ 2005-01-13 23:13 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert.Olsson

On Thu, 13 Jan 2005 14:43:30 -0600
"Jeremy M. Guthrie" <jeremy.guthrie@berbee.com> wrote:

> I don't show that proc file being there.  I see it in 'iproute.c' but not in 
> proc.

Sorry, my bad, it's

/proc/sys/net/ipv4/route/secret_interval

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-13 21:12                                                                                     ` Robert Olsson
  2005-01-13 22:27                                                                                       ` Jeremy M. Guthrie
@ 2005-01-14 14:59                                                                                       ` Jeremy M. Guthrie
  2005-01-14 16:05                                                                                         ` Robert Olsson
  1 sibling, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-14 14:59 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 44980 bytes --]

I ran a script overnight using the modified driver you had given me Robert.  
it is interesting that there is almost always errors in the interface even 
though we aren't getting dst-cache errors and running ~ 40% free CPU now.  I 
am going to switch over to Jesse's driver to see if his instrumentation helps 
nail down where the problem is.

ethtool -S eth3
NIC statistics:
     rx_packets: 2722676103
     tx_packets: 5
     rx_bytes: 1171335471
     tx_bytes: 398
     rx_errors: 8558366
     tx_errors: 0
     rx_dropped: 1951692
     tx_dropped: 0
     multicast: 0
     collisions: 0
     rx_length_errors: 0
     rx_over_errors: 0
     rx_crc_errors: 0
     rx_frame_errors: 0
     rx_fifo_errors: 6606674
     rx_missed_errors: 6606674
     tx_aborted_errors: 0
     tx_carrier_errors: 0
     tx_fifo_errors: 0
     tx_heartbeat_errors: 0
     tx_window_errors: 0
     tx_abort_late_coll: 0
     tx_deferred_ok: 0
     tx_single_coll_ok: 0
     tx_multi_coll_ok: 0
     rx_long_length_errors: 0
     rx_short_length_errors: 0
     rx_align_errors: 0
     tx_tcp_seg_good: 0
     tx_tcp_seg_failed: 0
     rx_flow_control_xon: 0
     rx_flow_control_xoff: 0
     tx_flow_control_xon: 0
     tx_flow_control_xoff: 0
     rx_long_byte_count: 1384150804783
     rx_csum_offload_good: 2542900690
     rx_csum_offload_errors: 243513

01/13/05 19:20:13 entries:  000690f5  Pkts:  948466  Err:  2788  PPS:  63231  
Drop %:  0.29%  Eth3RXInt:  411656  Eth2TXInt:  52008
01/13/05 19:20:28 entries:  0006ac8e  Pkts:  967236  Err:  2582  PPS:  64482  
Drop %:  0.26%  Eth3RXInt:  412282  Eth2TXInt:  51066
01/13/05 19:20:43 entries:  0006c6f3  Pkts:  974209  Err:  2645  PPS:  64947  
Drop %:  0.27%  Eth3RXInt:  412606  Eth2TXInt:  51439
01/13/05 19:20:58 entries:  0006de75  Pkts:  926103  Err:  2654  PPS:  61740  
Drop %:  0.28%  Eth3RXInt:  411864  Eth2TXInt:  51888
01/13/05 19:30:03 entries:  00091311  Pkts:  991885  Err:  2507  PPS:  66125  
Drop %:  0.25%  Eth3RXInt:  420214  Eth2TXInt:  44873
01/13/05 19:30:18 entries:  00091ecc  Pkts:  935298  Err:  2656  PPS:  62353  
Drop %:  0.28%  Eth3RXInt:  411803  Eth2TXInt:  49733
01/13/05 19:30:33 entries:  00092a91  Pkts:  937829  Err:  2797  PPS:  62521  
Drop %:  0.29%  Eth3RXInt:  408503  Eth2TXInt:  49473
01/13/05 19:30:48 entries:  00093612  Pkts:  958149  Err:  2698  PPS:  63876  
Drop %:  0.28%  Eth3RXInt:  415769  Eth2TXInt:  49552
01/13/05 19:40:08 entries:  000ae2ac  Pkts:  917843  Err:  2654  PPS:  61189  
Drop %:  0.28%  Eth3RXInt:  394963  Eth2TXInt:  51912
01/13/05 19:40:23 entries:  000af369  Pkts:  917445  Err:  2117  PPS:  61163  
Drop %:  0.23%  Eth3RXInt:  394155  Eth2TXInt:  53112
01/13/05 19:40:38 entries:  000b0272  Pkts:  931697  Err:  2763  PPS:  62113  
Drop %:  0.29%  Eth3RXInt:  403602  Eth2TXInt:  51899
01/13/05 19:40:53 entries:  00005de3  Pkts:  854339  Err:  83088  PPS:  56955  
Drop %:  9.72%  Eth3RXInt:  296451  Eth2TXInt:  62172
01/13/05 19:50:13 entries:  000669e0  Pkts:  910599  Err:  2646  PPS:  60706  
Drop %:  0.29%  Eth3RXInt:  402982  Eth2TXInt:  53797
01/13/05 19:50:28 entries:  00068591  Pkts:  921526  Err:  2782  PPS:  61435  
Drop %:  0.30%  Eth3RXInt:  408122  Eth2TXInt:  51792
01/13/05 19:50:44 entries:  0006a046  Pkts:  909343  Err:  2996  PPS:  60622  
Drop %:  0.32%  Eth3RXInt:  410888  Eth2TXInt:  53315
01/13/05 19:50:59 entries:  0006b968  Pkts:  904949  Err:  2656  PPS:  60329  
Drop %:  0.29%  Eth3RXInt:  410376  Eth2TXInt:  53444
01/13/05 20:00:03 entries:  0009188b  Pkts:  907073  Err:  2539  PPS:  60471  
Drop %:  0.27%  Eth3RXInt:  408658  Eth2TXInt:  52399
01/13/05 20:00:18 entries:  000924f5  Pkts:  930178  Err:  2386  PPS:  62011  
Drop %:  0.25%  Eth3RXInt:  410595  Eth2TXInt:  51822
01/13/05 20:00:33 entries:  000931a3  Pkts:  895381  Err:  2447  PPS:  59692  
Drop %:  0.27%  Eth3RXInt:  406246  Eth2TXInt:  52993
01/13/05 20:00:49 entries:  00093c95  Pkts:  908271  Err:  2338  PPS:  60551  
Drop %:  0.25%  Eth3RXInt:  416026  Eth2TXInt:  53732
01/13/05 20:10:08 entries:  000ad87d  Pkts:  911705  Err:  2432  PPS:  60780  
Drop %:  0.26%  Eth3RXInt:  416020  Eth2TXInt:  51364
01/13/05 20:10:23 entries:  000ae3fc  Pkts:  961446  Err:  2385  PPS:  64096  
Drop %:  0.24%  Eth3RXInt:  418902  Eth2TXInt:  48358
01/13/05 20:10:39 entries:  000aee6a  Pkts:  913636  Err:  2347  PPS:  60909  
Drop %:  0.25%  Eth3RXInt:  414232  Eth2TXInt:  51236
01/13/05 20:10:54 entries:  00005e36  Pkts:  888657  Err:  88852  PPS:  59243  
Drop %:  9.99%  Eth3RXInt:  307888  Eth2TXInt:  58891
01/13/05 20:20:14 entries:  00064ea9  Pkts:  837630  Err:  2339  PPS:  55842  
Drop %:  0.27%  Eth3RXInt:  398772  Eth2TXInt:  57373
01/13/05 20:20:29 entries:  0006687f  Pkts:  863629  Err:  2205  PPS:  57575  
Drop %:  0.25%  Eth3RXInt:  403868  Eth2TXInt:  54794
01/13/05 20:20:44 entries:  00068299  Pkts:  872662  Err:  2487  PPS:  58177  
Drop %:  0.28%  Eth3RXInt:  404412  Eth2TXInt:  55279
01/13/05 20:20:59 entries:  0006995c  Pkts:  848795  Err:  2537  PPS:  56586  
Drop %:  0.29%  Eth3RXInt:  402853  Eth2TXInt:  56477
01/13/05 20:30:04 entries:  0009b042  Pkts:  875158  Err:  2970  PPS:  58343  
Drop %:  0.33%  Eth3RXInt:  382571  Eth2TXInt:  55533
01/13/05 20:30:19 entries:  0009c17f  Pkts:  880014  Err:  2311  PPS:  58667  
Drop %:  0.26%  Eth3RXInt:  385079  Eth2TXInt:  53440
01/13/05 20:30:34 entries:  0009d1d6  Pkts:  900716  Err:  2453  PPS:  60047  
Drop %:  0.27%  Eth3RXInt:  396762  Eth2TXInt:  54786
01/13/05 20:30:49 entries:  0009e22d  Pkts:  897585  Err:  2453  PPS:  59839  
Drop %:  0.27%  Eth3RXInt:  396506  Eth2TXInt:  54240
01/13/05 20:40:08 entries:  000b915f  Pkts:  827962  Err:  2565  PPS:  55197  
Drop %:  0.30%  Eth3RXInt:  404875  Eth2TXInt:  58387
01/13/05 20:40:24 entries:  000b9cf2  Pkts:  821150  Err:  2460  PPS:  54743  
Drop %:  0.29%  Eth3RXInt:  399397  Eth2TXInt:  58173
01/13/05 20:40:39 entries:  000baaef  Pkts:  878751  Err:  2733  PPS:  58583  
Drop %:  0.31%  Eth3RXInt:  402052  Eth2TXInt:  55343
01/13/05 20:40:54 entries:  000062d3  Pkts:  815754  Err:  77364  PPS:  54383  
Drop %:  9.48%  Eth3RXInt:  298308  Eth2TXInt:  63697
01/13/05 20:50:13 entries:  000655f5  Pkts:  874449  Err:  2589  PPS:  58296  
Drop %:  0.29%  Eth3RXInt:  397470  Eth2TXInt:  55040
01/13/05 20:50:29 entries:  000670ef  Pkts:  862434  Err:  2335  PPS:  57495  
Drop %:  0.27%  Eth3RXInt:  399384  Eth2TXInt:  55292
01/13/05 20:50:44 entries:  00068e59  Pkts:  904716  Err:  2520  PPS:  60314  
Drop %:  0.27%  Eth3RXInt:  401899  Eth2TXInt:  52789
01/13/05 20:50:59 entries:  0006a6af  Pkts:  877349  Err:  2660  PPS:  58489  
Drop %:  0.30%  Eth3RXInt:  401982  Eth2TXInt:  54755
01/13/05 21:00:03 entries:  0008ffba  Pkts:  835650  Err:  2127  PPS:  55710  
Drop %:  0.25%  Eth3RXInt:  406517  Eth2TXInt:  57597
01/13/05 21:00:18 entries:  00090a7f  Pkts:  842742  Err:  2102  PPS:  56182  
Drop %:  0.24%  Eth3RXInt:  407024  Eth2TXInt:  57559
01/13/05 21:00:33 entries:  000915dd  Pkts:  844459  Err:  2683  PPS:  56297  
Drop %:  0.31%  Eth3RXInt:  406624  Eth2TXInt:  58553
01/13/05 21:00:48 entries:  000920b6  Pkts:  810715  Err:  2626  PPS:  54047  
Drop %:  0.32%  Eth3RXInt:  401342  Eth2TXInt:  59538
01/13/05 21:10:08 entries:  000aa6a7  Pkts:  829066  Err:  2159  PPS:  55271  
Drop %:  0.26%  Eth3RXInt:  401210  Eth2TXInt:  56607
01/13/05 21:10:23 entries:  000ab10b  Pkts:  833478  Err:  2350  PPS:  55565  
Drop %:  0.28%  Eth3RXInt:  404397  Eth2TXInt:  56495
01/13/05 21:10:38 entries:  000abb70  Pkts:  813251  Err:  2425  PPS:  54216  
Drop %:  0.29%  Eth3RXInt:  402135  Eth2TXInt:  57976
01/13/05 21:10:53 entries:  000ac12b  Pkts:  836855  Err:  2315  PPS:  55790  
Drop %:  0.27%  Eth3RXInt:  404963  Eth2TXInt:  56857
01/13/05 21:20:13 entries:  000c8587  Pkts:  780311  Err:  2011  PPS:  52020  
Drop %:  0.25%  Eth3RXInt:  387437  Eth2TXInt:  60951
01/13/05 21:20:28 entries:  000c927c  Pkts:  781288  Err:  1957  PPS:  52085  
Drop %:  0.25%  Eth3RXInt:  386866  Eth2TXInt:  60672
01/13/05 21:20:43 entries:  000c9f80  Pkts:  771294  Err:  2185  PPS:  51419  
Drop %:  0.28%  Eth3RXInt:  386213  Eth2TXInt:  61530
01/13/05 21:20:58 entries:  000cabe0  Pkts:  761870  Err:  1847  PPS:  50791  
Drop %:  0.24%  Eth3RXInt:  383979  Eth2TXInt:  60613
01/13/05 21:30:02 entries:  000e1471  Pkts:  752945  Err:  2363  PPS:  50196  
Drop %:  0.31%  Eth3RXInt:  385445  Eth2TXInt:  62462
01/13/05 21:30:17 entries:  000e20e0  Pkts:  746890  Err:  2179  PPS:  49792  
Drop %:  0.29%  Eth3RXInt:  385645  Eth2TXInt:  62886
01/13/05 21:30:32 entries:  000e2bf5  Pkts:  759932  Err:  2353  PPS:  50662  
Drop %:  0.30%  Eth3RXInt:  390137  Eth2TXInt:  61938
01/13/05 21:30:47 entries:  000e368c  Pkts:  770359  Err:  2226  PPS:  51357  
Drop %:  0.28%  Eth3RXInt:  393035  Eth2TXInt:  62280
01/13/05 21:40:07 entries:  000f66ab  Pkts:  710993  Err:  1425  PPS:  47399  
Drop %:  0.20%  Eth3RXInt:  385138  Eth2TXInt:  66570
01/13/05 21:40:22 entries:  000f6eea  Pkts:  722199  Err:  1727  PPS:  48146  
Drop %:  0.23%  Eth3RXInt:  386471  Eth2TXInt:  65154
01/13/05 21:40:37 entries:  000f7731  Pkts:  721459  Err:  1858  PPS:  48097  
Drop %:  0.25%  Eth3RXInt:  388601  Eth2TXInt:  64437
01/13/05 21:40:52 entries:  00005160  Pkts:  717256  Err:  50031  PPS:  47817  
Drop %:  6.97%  Eth3RXInt:  301824  Eth2TXInt:  67610
01/13/05 21:50:11 entries:  00060c77  Pkts:  744283  Err:  2209  PPS:  49618  
Drop %:  0.29%  Eth3RXInt:  385202  Eth2TXInt:  61977
01/13/05 21:50:26 entries:  0006251c  Pkts:  712155  Err:  1731  PPS:  47477  
Drop %:  0.24%  Eth3RXInt:  375784  Eth2TXInt:  63593
01/13/05 21:50:42 entries:  00063f3d  Pkts:  732397  Err:  2502  PPS:  48826  
Drop %:  0.34%  Eth3RXInt:  382823  Eth2TXInt:  64629
01/13/05 21:50:57 entries:  00065ea4  Pkts:  732251  Err:  1954  PPS:  48816  
Drop %:  0.26%  Eth3RXInt:  372229  Eth2TXInt:  64307
01/13/05 22:00:01 entries:  0008ed30  Pkts:  753502  Err:  1919  PPS:  50233  
Drop %:  0.25%  Eth3RXInt:  392290  Eth2TXInt:  63585
01/13/05 22:00:16 entries:  0008f743  Pkts:  765954  Err:  1783  PPS:  51063  
Drop %:  0.23%  Eth3RXInt:  398195  Eth2TXInt:  64681
01/13/05 22:00:31 entries:  00090132  Pkts:  745575  Err:  1748  PPS:  49705  
Drop %:  0.23%  Eth3RXInt:  393391  Eth2TXInt:  63904
01/13/05 22:00:46 entries:  00090bf1  Pkts:  750483  Err:  1810  PPS:  50032  
Drop %:  0.24%  Eth3RXInt:  395678  Eth2TXInt:  65592
01/13/05 22:10:05 entries:  000a69b4  Pkts:  740410  Err:  4101  PPS:  49360  
Drop %:  0.55%  Eth3RXInt:  393872  Eth2TXInt:  66508
01/13/05 22:10:20 entries:  000a732d  Pkts:  737503  Err:  3481  PPS:  49166  
Drop %:  0.47%  Eth3RXInt:  392793  Eth2TXInt:  67370
01/13/05 22:10:35 entries:  000a7d9a  Pkts:  735129  Err:  3227  PPS:  49008  
Drop %:  0.43%  Eth3RXInt:  394918  Eth2TXInt:  67208
01/13/05 22:10:51 entries:  000a87d0  Pkts:  731617  Err:  3099  PPS:  48774  
Drop %:  0.42%  Eth3RXInt:  391164  Eth2TXInt:  67757
01/13/05 22:20:09 entries:  000bcb09  Pkts:  674113  Err:  3321  PPS:  44940  
Drop %:  0.49%  Eth3RXInt:  379562  Eth2TXInt:  69421
01/13/05 22:20:25 entries:  000bd3b8  Pkts:  666534  Err:  2801  PPS:  44435  
Drop %:  0.42%  Eth3RXInt:  377072  Eth2TXInt:  70649
01/13/05 22:20:40 entries:  000bdd19  Pkts:  678034  Err:  2940  PPS:  45202  
Drop %:  0.43%  Eth3RXInt:  378198  Eth2TXInt:  69143
01/13/05 22:20:55 entries:  000be345  Pkts:  661543  Err:  2829  PPS:  44102  
Drop %:  0.42%  Eth3RXInt:  373469  Eth2TXInt:  69994
01/13/05 22:30:14 entries:  000d1135  Pkts:  678235  Err:  1982  PPS:  45215  
Drop %:  0.29%  Eth3RXInt:  373441  Eth2TXInt:  71223
01/13/05 22:30:29 entries:  000d1bf0  Pkts:  685493  Err:  1983  PPS:  45699  
Drop %:  0.28%  Eth3RXInt:  373410  Eth2TXInt:  68586
01/13/05 22:30:44 entries:  000d26eb  Pkts:  684653  Err:  2007  PPS:  45643  
Drop %:  0.29%  Eth3RXInt:  373354  Eth2TXInt:  68304
01/13/05 22:30:59 entries:  000d2e78  Pkts:  664648  Err:  1527  PPS:  44309  
Drop %:  0.22%  Eth3RXInt:  365428  Eth2TXInt:  67413
01/13/05 22:40:03 entries:  000e44b4  Pkts:  645300  Err:  1954  PPS:  43020  
Drop %:  0.30%  Eth3RXInt:  369820  Eth2TXInt:  72192
01/13/05 22:40:18 entries:  000e4c66  Pkts:  643369  Err:  1702  PPS:  42891  
Drop %:  0.26%  Eth3RXInt:  366982  Eth2TXInt:  72002
01/13/05 22:40:33 entries:  000e5423  Pkts:  634654  Err:  1873  PPS:  42310  
Drop %:  0.29%  Eth3RXInt:  367668  Eth2TXInt:  72011
01/13/05 22:40:48 entries:  00002be0  Pkts:  626759  Err:  21655  PPS:  41783  
Drop %:  3.45%  Eth3RXInt:  317916  Eth2TXInt:  72957
01/13/05 22:50:07 entries:  00057b41  Pkts:  636115  Err:  1548  PPS:  42407  
Drop %:  0.24%  Eth3RXInt:  356743  Eth2TXInt:  70117
01/13/05 22:50:22 entries:  00059254  Pkts:  690168  Err:  2081  PPS:  46011  
Drop %:  0.30%  Eth3RXInt:  375141  Eth2TXInt:  69588
01/13/05 22:50:37 entries:  0005a923  Pkts:  667265  Err:  1647  PPS:  44484  
Drop %:  0.24%  Eth3RXInt:  367504  Eth2TXInt:  69765
01/13/05 22:50:52 entries:  0005be38  Pkts:  646829  Err:  1806  PPS:  43121  
Drop %:  0.27%  Eth3RXInt:  361265  Eth2TXInt:  72190
01/13/05 23:00:11 entries:  0008b137  Pkts:  616932  Err:  1566  PPS:  41128  
Drop %:  0.25%  Eth3RXInt:  361653  Eth2TXInt:  74776
01/13/05 23:00:26 entries:  0008bd2f  Pkts:  655216  Err:  2433  PPS:  43681  
Drop %:  0.37%  Eth3RXInt:  366736  Eth2TXInt:  72637
01/13/05 23:00:42 entries:  0008c6ce  Pkts:  650050  Err:  1792  PPS:  43336  
Drop %:  0.27%  Eth3RXInt:  369516  Eth2TXInt:  74231
01/13/05 23:00:57 entries:  0008cf8a  Pkts:  622743  Err:  1628  PPS:  41516  
Drop %:  0.26%  Eth3RXInt:  362609  Eth2TXInt:  73576
01/13/05 23:10:00 entries:  0009ec88  Pkts:  558466  Err:  1254  PPS:  37231  
Drop %:  0.22%  Eth3RXInt:  341801  Eth2TXInt:  77394
01/13/05 23:10:15 entries:  0009f446  Pkts:  583270  Err:  1727  PPS:  38884  
Drop %:  0.29%  Eth3RXInt:  351413  Eth2TXInt:  77099
01/13/05 23:10:31 entries:  0009fd65  Pkts:  585702  Err:  1378  PPS:  39046  
Drop %:  0.23%  Eth3RXInt:  352610  Eth2TXInt:  76484
01/13/05 23:10:46 entries:  000a061f  Pkts:  591411  Err:  1389  PPS:  39427  
Drop %:  0.23%  Eth3RXInt:  351609  Eth2TXInt:  76353
01/13/05 23:20:04 entries:  000b177a  Pkts:  630155  Err:  1788  PPS:  42010  
Drop %:  0.28%  Eth3RXInt:  368741  Eth2TXInt:  76933
01/13/05 23:20:19 entries:  000b2070  Pkts:  608282  Err:  1147  PPS:  40552  
Drop %:  0.18%  Eth3RXInt:  356552  Eth2TXInt:  76472
01/13/05 23:20:34 entries:  000b28ae  Pkts:  560094  Err:  1474  PPS:  37339  
Drop %:  0.26%  Eth3RXInt:  341610  Eth2TXInt:  77689
01/13/05 23:20:50 entries:  000b314f  Pkts:  559777  Err:  1268  PPS:  37318  
Drop %:  0.22%  Eth3RXInt:  342036  Eth2TXInt:  77905
01/13/05 23:30:08 entries:  000c3e43  Pkts:  602182  Err:  1643  PPS:  40145  
Drop %:  0.27%  Eth3RXInt:  360109  Eth2TXInt:  76867
01/13/05 23:30:23 entries:  000c456a  Pkts:  544079  Err:  1293  PPS:  36271  
Drop %:  0.23%  Eth3RXInt:  337942  Eth2TXInt:  78707
01/13/05 23:30:38 entries:  000c4d58  Pkts:  571859  Err:  1606  PPS:  38123  
Drop %:  0.28%  Eth3RXInt:  348591  Eth2TXInt:  78355
01/13/05 23:30:53 entries:  000c4f8d  Pkts:  553967  Err:  1299  PPS:  36931  
Drop %:  0.23%  Eth3RXInt:  340507  Eth2TXInt:  77327
01/13/05 23:40:12 entries:  000d5cd4  Pkts:  571850  Err:  1459  PPS:  38123  
Drop %:  0.25%  Eth3RXInt:  342271  Eth2TXInt:  77382
01/13/05 23:40:27 entries:  000d66eb  Pkts:  537062  Err:  1233  PPS:  35804  
Drop %:  0.22%  Eth3RXInt:  326360  Eth2TXInt:  79641
01/13/05 23:40:42 entries:  000d700d  Pkts:  538091  Err:  1640  PPS:  35872  
Drop %:  0.30%  Eth3RXInt:  331841  Eth2TXInt:  79667
01/13/05 23:40:57 entries:  00006e93  Pkts:  534710  Err:  14693  PPS:  35647  
Drop %:  2.74%  Eth3RXInt:  255177  Eth2TXInt:  78981
01/13/05 23:50:01 entries:  00056e16  Pkts:  574215  Err:  1233  PPS:  38281  
Drop %:  0.21%  Eth3RXInt:  337409  Eth2TXInt:  78770
01/13/05 23:50:16 entries:  00058557  Pkts:  537083  Err:  1186  PPS:  35805  
Drop %:  0.22%  Eth3RXInt:  324471  Eth2TXInt:  78352
01/13/05 23:50:31 entries:  00059b52  Pkts:  565715  Err:  1586  PPS:  37714  
Drop %:  0.28%  Eth3RXInt:  340689  Eth2TXInt:  79699
01/13/05 23:50:46 entries:  0005af99  Pkts:  558546  Err:  1355  PPS:  37236  
Drop %:  0.24%  Eth3RXInt:  337133  Eth2TXInt:  78779
01/14/05 00:00:05 entries:  0008504d  Pkts:  588805  Err:  1507  PPS:  39253  
Drop %:  0.25%  Eth3RXInt:  346266  Eth2TXInt:  77616
01/14/05 00:00:20 entries:  00085c27  Pkts:  570425  Err:  1672  PPS:  38028  
Drop %:  0.29%  Eth3RXInt:  338249  Eth2TXInt:  78085
01/14/05 00:00:35 entries:  00086908  Pkts:  578050  Err:  1327  PPS:  38536  
Drop %:  0.22%  Eth3RXInt:  340259  Eth2TXInt:  77927
01/14/05 00:00:50 entries:  00087438  Pkts:  566396  Err:  1135  PPS:  37759  
Drop %:  0.20%  Eth3RXInt:  338626  Eth2TXInt:  77983
01/14/05 00:10:09 entries:  0009a1f9  Pkts:  527344  Err:  1364  PPS:  35156  
Drop %:  0.25%  Eth3RXInt:  330864  Eth2TXInt:  80726
01/14/05 00:10:24 entries:  0009a950  Pkts:  519107  Err:  1251  PPS:  34607  
Drop %:  0.24%  Eth3RXInt:  327638  Eth2TXInt:  81313
01/14/05 00:10:39 entries:  0009b0d4  Pkts:  527377  Err:  1085  PPS:  35158  
Drop %:  0.20%  Eth3RXInt:  330084  Eth2TXInt:  80885
01/14/05 00:10:54 entries:  0009b7e8  Pkts:  527706  Err:  1603  PPS:  35180  
Drop %:  0.30%  Eth3RXInt:  331284  Eth2TXInt:  80577
01/14/05 00:20:13 entries:  000aa8fa  Pkts:  477646  Err:  1147  PPS:  31843  
Drop %:  0.24%  Eth3RXInt:  313514  Eth2TXInt:  81244
01/14/05 00:20:28 entries:  000aaf4c  Pkts:  506930  Err:  1148  PPS:  33795  
Drop %:  0.22%  Eth3RXInt:  324715  Eth2TXInt:  80366
01/14/05 00:20:43 entries:  000ab666  Pkts:  538447  Err:  1783  PPS:  35896  
Drop %:  0.33%  Eth3RXInt:  340587  Eth2TXInt:  80129
01/14/05 00:20:58 entries:  000abc90  Pkts:  503351  Err:  1151  PPS:  33556  
Drop %:  0.22%  Eth3RXInt:  324926  Eth2TXInt:  81205
01/14/05 00:30:01 entries:  000b95f8  Pkts:  503581  Err:  1588  PPS:  33572  
Drop %:  0.31%  Eth3RXInt:  323090  Eth2TXInt:  81310
01/14/05 00:30:16 entries:  000b9c10  Pkts:  509557  Err:  1373  PPS:  33970  
Drop %:  0.26%  Eth3RXInt:  327946  Eth2TXInt:  80876
01/14/05 00:30:31 entries:  000ba263  Pkts:  498061  Err:  1624  PPS:  33204  
Drop %:  0.32%  Eth3RXInt:  320360  Eth2TXInt:  80399
01/14/05 00:30:47 entries:  000ba8af  Pkts:  508690  Err:  1030  PPS:  33912  
Drop %:  0.20%  Eth3RXInt:  325848  Eth2TXInt:  80562
01/14/05 00:40:05 entries:  000c8eb0  Pkts:  423203  Err:  925  PPS:  28213  
Drop %:  0.21%  Eth3RXInt:  289128  Eth2TXInt:  81558
01/14/05 00:40:20 entries:  000c9460  Pkts:  435865  Err:  905  PPS:  29057  
Drop %:  0.20%  Eth3RXInt:  290040  Eth2TXInt:  81407
01/14/05 00:40:35 entries:  000c9b1d  Pkts:  470792  Err:  987  PPS:  31386  
Drop %:  0.20%  Eth3RXInt:  306570  Eth2TXInt:  80345
01/14/05 00:40:50 entries:  00003ace  Pkts:  455627  Err:  7859  PPS:  30375  
Drop %:  1.72%  Eth3RXInt:  261150  Eth2TXInt:  80456
01/14/05 00:50:09 entries:  0004b083  Pkts:  491755  Err:  1253  PPS:  32783  
Drop %:  0.25%  Eth3RXInt:  316362  Eth2TXInt:  81009
01/14/05 00:50:24 entries:  0004c2e2  Pkts:  465117  Err:  975  PPS:  31007  
Drop %:  0.20%  Eth3RXInt:  301050  Eth2TXInt:  81862
01/14/05 00:50:39 entries:  0004d6dc  Pkts:  459493  Err:  1327  PPS:  30632  
Drop %:  0.28%  Eth3RXInt:  298610  Eth2TXInt:  81555
01/14/05 00:50:54 entries:  0004eb2d  Pkts:  487190  Err:  887  PPS:  32479  
Drop %:  0.18%  Eth3RXInt:  313779  Eth2TXInt:  81653
01/14/05 01:00:12 entries:  0007333a  Pkts:  511807  Err:  1015  PPS:  34120  
Drop %:  0.19%  Eth3RXInt:  326166  Eth2TXInt:  80991
01/14/05 01:00:28 entries:  000742bb  Pkts:  490129  Err:  1441  PPS:  32675  
Drop %:  0.29%  Eth3RXInt:  315417  Eth2TXInt:  81048
01/14/05 01:00:43 entries:  00075002  Pkts:  480424  Err:  1108  PPS:  32028  
Drop %:  0.23%  Eth3RXInt:  312870  Eth2TXInt:  81955
01/14/05 01:00:58 entries:  00075ab5  Pkts:  479570  Err:  1305  PPS:  31971  
Drop %:  0.27%  Eth3RXInt:  314682  Eth2TXInt:  81203
01/14/05 01:10:01 entries:  0008979c  Pkts:  450310  Err:  1045  PPS:  30020  
Drop %:  0.23%  Eth3RXInt:  290075  Eth2TXInt:  81569
01/14/05 01:10:16 entries:  00089faf  Pkts:  433228  Err:  1053  PPS:  28881  
Drop %:  0.24%  Eth3RXInt:  285539  Eth2TXInt:  81461
01/14/05 01:10:31 entries:  0008a763  Pkts:  440048  Err:  1367  PPS:  29336  
Drop %:  0.31%  Eth3RXInt:  291495  Eth2TXInt:  81925
01/14/05 01:10:46 entries:  0008af1e  Pkts:  501093  Err:  1510  PPS:  33406  
Drop %:  0.30%  Eth3RXInt:  319267  Eth2TXInt:  82134
01/14/05 01:20:05 entries:  0009a978  Pkts:  484885  Err:  1591  PPS:  32325  
Drop %:  0.32%  Eth3RXInt:  304961  Eth2TXInt:  80773
01/14/05 01:20:20 entries:  0009b485  Pkts:  497080  Err:  1484  PPS:  33138  
Drop %:  0.29%  Eth3RXInt:  310173  Eth2TXInt:  81031
01/14/05 01:20:35 entries:  0009bfd7  Pkts:  503000  Err:  1555  PPS:  33533  
Drop %:  0.30%  Eth3RXInt:  315228  Eth2TXInt:  81955
01/14/05 01:20:50 entries:  0009cb17  Pkts:  473078  Err:  1004  PPS:  31538  
Drop %:  0.21%  Eth3RXInt:  302143  Eth2TXInt:  82036
01/14/05 01:30:09 entries:  000acb27  Pkts:  485756  Err:  1048  PPS:  32383  
Drop %:  0.21%  Eth3RXInt:  315112  Eth2TXInt:  81381
01/14/05 01:30:24 entries:  000ade3e  Pkts:  487954  Err:  1116  PPS:  32530  
Drop %:  0.22%  Eth3RXInt:  296615  Eth2TXInt:  80622
01/14/05 01:30:39 entries:  000b1d66  Pkts:  496891  Err:  1365  PPS:  33126  
Drop %:  0.27%  Eth3RXInt:  245583  Eth2TXInt:  79328
01/14/05 01:30:54 entries:  000b55c3  Pkts:  525362  Err:  1661  PPS:  35024  
Drop %:  0.31%  Eth3RXInt:  249821  Eth2TXInt:  76762
01/14/05 01:40:12 entries:  000c5d9b  Pkts:  441275  Err:  1262  PPS:  29418  
Drop %:  0.28%  Eth3RXInt:  295788  Eth2TXInt:  81607
01/14/05 01:40:27 entries:  000c638d  Pkts:  432402  Err:  818  PPS:  28826  
Drop %:  0.18%  Eth3RXInt:  292080  Eth2TXInt:  82215
01/14/05 01:40:42 entries:  000c6a20  Pkts:  475611  Err:  1137  PPS:  31707  
Drop %:  0.23%  Eth3RXInt:  312107  Eth2TXInt:  82512
01/14/05 01:40:58 entries:  000062a2  Pkts:  422599  Err:  3778  PPS:  28173  
Drop %:  0.89%  Eth3RXInt:  236232  Eth2TXInt:  81681
01/14/05 01:50:01 entries:  00047c87  Pkts:  434595  Err:  1120  PPS:  28973  
Drop %:  0.25%  Eth3RXInt:  291004  Eth2TXInt:  81547
01/14/05 01:50:16 entries:  0004924f  Pkts:  473118  Err:  1018  PPS:  31541  
Drop %:  0.21%  Eth3RXInt:  306494  Eth2TXInt:  81818
01/14/05 01:50:31 entries:  0004a615  Pkts:  454845  Err:  807  PPS:  30323  
Drop %:  0.17%  Eth3RXInt:  300121  Eth2TXInt:  81781
01/14/05 01:50:46 entries:  0004b989  Pkts:  446690  Err:  814  PPS:  29779  
Drop %:  0.18%  Eth3RXInt:  295153  Eth2TXInt:  82147
01/14/05 02:00:05 entries:  00071828  Pkts:  485939  Err:  1181  PPS:  32395  
Drop %:  0.24%  Eth3RXInt:  311765  Eth2TXInt:  81409
01/14/05 02:00:20 entries:  000727ca  Pkts:  473909  Err:  1384  PPS:  31593  
Drop %:  0.29%  Eth3RXInt:  307111  Eth2TXInt:  81430
01/14/05 02:00:35 entries:  00073806  Pkts:  477835  Err:  1173  PPS:  31855  
Drop %:  0.24%  Eth3RXInt:  308946  Eth2TXInt:  81813
01/14/05 02:00:50 entries:  00074824  Pkts:  481520  Err:  918  PPS:  32101  
Drop %:  0.19%  Eth3RXInt:  310852  Eth2TXInt:  81876
01/14/05 02:10:08 entries:  0008a6fc  Pkts:  441730  Err:  788  PPS:  29448  
Drop %:  0.17%  Eth3RXInt:  296119  Eth2TXInt:  82040
01/14/05 02:10:23 entries:  0008ae30  Pkts:  462573  Err:  964  PPS:  30838  
Drop %:  0.20%  Eth3RXInt:  305317  Eth2TXInt:  82031
01/14/05 02:10:39 entries:  0008b5ce  Pkts:  462473  Err:  1037  PPS:  30831  
Drop %:  0.22%  Eth3RXInt:  305753  Eth2TXInt:  81449
01/14/05 02:10:54 entries:  0008baee  Pkts:  433606  Err:  885  PPS:  28907  
Drop %:  0.20%  Eth3RXInt:  291876  Eth2TXInt:  81471
01/14/05 02:20:12 entries:  0009a92d  Pkts:  458960  Err:  851  PPS:  30597  
Drop %:  0.18%  Eth3RXInt:  303308  Eth2TXInt:  81503
01/14/05 02:20:27 entries:  0009b002  Pkts:  498472  Err:  1221  PPS:  33231  
Drop %:  0.24%  Eth3RXInt:  325077  Eth2TXInt:  81679
01/14/05 02:20:42 entries:  0009b68c  Pkts:  484378  Err:  1176  PPS:  32291  
Drop %:  0.24%  Eth3RXInt:  317011  Eth2TXInt:  80552
01/14/05 02:20:57 entries:  0009b9cf  Pkts:  448711  Err:  932  PPS:  29914  
Drop %:  0.20%  Eth3RXInt:  300504  Eth2TXInt:  81309
01/14/05 02:30:01 entries:  000abe83  Pkts:  485720  Err:  1211  PPS:  32381  
Drop %:  0.24%  Eth3RXInt:  317102  Eth2TXInt:  81805
01/14/05 02:30:16 entries:  000ac418  Pkts:  456824  Err:  770  PPS:  30454  
Drop %:  0.16%  Eth3RXInt:  303707  Eth2TXInt:  82063
01/14/05 02:30:31 entries:  000aca2e  Pkts:  464268  Err:  926  PPS:  30951  
Drop %:  0.19%  Eth3RXInt:  307084  Eth2TXInt:  82922
01/14/05 02:30:46 entries:  000acfd9  Pkts:  488002  Err:  1199  PPS:  32533  
Drop %:  0.24%  Eth3RXInt:  317994  Eth2TXInt:  82220
01/14/05 02:40:04 entries:  000b939f  Pkts:  426137  Err:  1071  PPS:  28409  
Drop %:  0.25%  Eth3RXInt:  291613  Eth2TXInt:  80942
01/14/05 02:40:19 entries:  000b99b8  Pkts:  435682  Err:  641  PPS:  29045  
Drop %:  0.14%  Eth3RXInt:  295451  Eth2TXInt:  81203
01/14/05 02:40:34 entries:  000ba164  Pkts:  469842  Err:  1099  PPS:  31322  
Drop %:  0.23%  Eth3RXInt:  310350  Eth2TXInt:  81309
01/14/05 02:40:50 entries:  000039d8  Pkts:  432433  Err:  4158  PPS:  28828  
Drop %:  0.96%  Eth3RXInt:  258916  Eth2TXInt:  79816
01/14/05 02:50:08 entries:  00045e02  Pkts:  411631  Err:  1073  PPS:  27442  
Drop %:  0.26%  Eth3RXInt:  278161  Eth2TXInt:  81417
01/14/05 02:50:23 entries:  00046f7b  Pkts:  410141  Err:  795  PPS:  27342  
Drop %:  0.19%  Eth3RXInt:  275966  Eth2TXInt:  81285
01/14/05 02:50:38 entries:  000481d2  Pkts:  433426  Err:  1245  PPS:  28895  
Drop %:  0.28%  Eth3RXInt:  289269  Eth2TXInt:  81567
01/14/05 02:50:53 entries:  000493db  Pkts:  412782  Err:  969  PPS:  27518  
Drop %:  0.23%  Eth3RXInt:  277938  Eth2TXInt:  81170
01/14/05 03:00:12 entries:  00075b30  Pkts:  525329  Err:  1144  PPS:  35021  
Drop %:  0.21%  Eth3RXInt:  335221  Eth2TXInt:  82922
01/14/05 03:00:27 entries:  00076ad2  Pkts:  517248  Err:  1513  PPS:  34483  
Drop %:  0.29%  Eth3RXInt:  325825  Eth2TXInt:  82644
01/14/05 03:00:42 entries:  000778c5  Pkts:  455097  Err:  1211  PPS:  30339  
Drop %:  0.26%  Eth3RXInt:  301609  Eth2TXInt:  81940
01/14/05 03:00:57 entries:  000785d8  Pkts:  472256  Err:  1428  PPS:  31483  
Drop %:  0.30%  Eth3RXInt:  308709  Eth2TXInt:  82655
01/14/05 03:10:00 entries:  0008c8fd  Pkts:  474980  Err:  1191  PPS:  31665  
Drop %:  0.25%  Eth3RXInt:  305622  Eth2TXInt:  80539
01/14/05 03:10:15 entries:  0008d062  Pkts:  457870  Err:  1130  PPS:  30524  
Drop %:  0.24%  Eth3RXInt:  296680  Eth2TXInt:  82396
01/14/05 03:10:30 entries:  0008d73b  Pkts:  480551  Err:  1381  PPS:  32036  
Drop %:  0.28%  Eth3RXInt:  309356  Eth2TXInt:  81644
01/14/05 03:10:46 entries:  0008ddd3  Pkts:  465617  Err:  950  PPS:  31041  
Drop %:  0.20%  Eth3RXInt:  304579  Eth2TXInt:  82222
01/14/05 03:20:04 entries:  0009e568  Pkts:  457937  Err:  1151  PPS:  30529  
Drop %:  0.25%  Eth3RXInt:  305244  Eth2TXInt:  81326
01/14/05 03:20:19 entries:  0009eb65  Pkts:  436354  Err:  938  PPS:  29090  
Drop %:  0.21%  Eth3RXInt:  292012  Eth2TXInt:  81560
01/14/05 03:20:34 entries:  0009f1d4  Pkts:  420677  Err:  863  PPS:  28045  
Drop %:  0.20%  Eth3RXInt:  285478  Eth2TXInt:  81628
01/14/05 03:20:49 entries:  0009f71d  Pkts:  451901  Err:  838  PPS:  30126  
Drop %:  0.18%  Eth3RXInt:  302797  Eth2TXInt:  81472
01/14/05 03:30:08 entries:  000ab0e6  Pkts:  496468  Err:  935  PPS:  33097  
Drop %:  0.18%  Eth3RXInt:  313564  Eth2TXInt:  81825
01/14/05 03:30:23 entries:  000ab6d9  Pkts:  487412  Err:  1234  PPS:  32494  
Drop %:  0.25%  Eth3RXInt:  311207  Eth2TXInt:  81957
01/14/05 03:30:38 entries:  000abd71  Pkts:  486905  Err:  1197  PPS:  32460  
Drop %:  0.24%  Eth3RXInt:  309132  Eth2TXInt:  81431
01/14/05 03:30:53 entries:  000ac18a  Pkts:  501542  Err:  1095  PPS:  33436  
Drop %:  0.21%  Eth3RXInt:  317566  Eth2TXInt:  82225
01/14/05 03:40:11 entries:  000b7c09  Pkts:  453065  Err:  1297  PPS:  30204  
Drop %:  0.28%  Eth3RXInt:  301844  Eth2TXInt:  81878
01/14/05 03:40:26 entries:  000b80c8  Pkts:  419895  Err:  794  PPS:  27993  
Drop %:  0.18%  Eth3RXInt:  286286  Eth2TXInt:  81854
01/14/05 03:40:41 entries:  000b85bf  Pkts:  437850  Err:  871  PPS:  29190  
Drop %:  0.19%  Eth3RXInt:  296486  Eth2TXInt:  81568
01/14/05 03:40:56 entries:  00005a7b  Pkts:  441752  Err:  4133  PPS:  29450  
Drop %:  0.93%  Eth3RXInt:  246594  Eth2TXInt:  80314
01/14/05 03:50:00 entries:  00046892  Pkts:  403034  Err:  1065  PPS:  26868  
Drop %:  0.26%  Eth3RXInt:  274784  Eth2TXInt:  81157
01/14/05 03:50:15 entries:  000476c7  Pkts:  435578  Err:  729  PPS:  29038  
Drop %:  0.16%  Eth3RXInt:  292309  Eth2TXInt:  81034
01/14/05 03:50:30 entries:  000486aa  Pkts:  412410  Err:  828  PPS:  27494  
Drop %:  0.20%  Eth3RXInt:  278210  Eth2TXInt:  81402
01/14/05 03:50:45 entries:  000495b6  Pkts:  416286  Err:  1004  PPS:  27752  
Drop %:  0.24%  Eth3RXInt:  281510  Eth2TXInt:  81008
01/14/05 04:00:03 entries:  0006a815  Pkts:  475366  Err:  1129  PPS:  31691  
Drop %:  0.23%  Eth3RXInt:  307887  Eth2TXInt:  81480
01/14/05 04:00:18 entries:  0006b54b  Pkts:  474234  Err:  1045  PPS:  31615  
Drop %:  0.22%  Eth3RXInt:  307872  Eth2TXInt:  81418
01/14/05 04:00:34 entries:  0006c1b3  Pkts:  468519  Err:  920  PPS:  31234  
Drop %:  0.19%  Eth3RXInt:  309532  Eth2TXInt:  81400
01/14/05 04:00:49 entries:  0006cec6  Pkts:  427386  Err:  701  PPS:  28492  
Drop %:  0.16%  Eth3RXInt:  288788  Eth2TXInt:  80951
01/14/05 04:10:07 entries:  0008b49a  Pkts:  450361  Err:  0  PPS:  30024  
Drop %:  0.0%  Eth3RXInt:  298760  Eth2TXInt:  82346
01/14/05 04:10:22 entries:  0008ba59  Pkts:  459431  Err:  0  PPS:  30628  
Drop %:  0.0%  Eth3RXInt:  301596  Eth2TXInt:  81793
01/14/05 04:10:37 entries:  0008c017  Pkts:  450040  Err:  0  PPS:  30002  
Drop %:  0.0%  Eth3RXInt:  298585  Eth2TXInt:  81996
01/14/05 04:10:52 entries:  0008c2e9  Pkts:  454740  Err:  963  PPS:  30316  
Drop %:  0.21%  Eth3RXInt:  299204  Eth2TXInt:  82428
01/14/05 04:20:11 entries:  0009746e  Pkts:  417841  Err:  1077  PPS:  27856  
Drop %:  0.25%  Eth3RXInt:  285292  Eth2TXInt:  81971
01/14/05 04:20:26 entries:  00097988  Pkts:  444129  Err:  834  PPS:  29608  
Drop %:  0.18%  Eth3RXInt:  294925  Eth2TXInt:  82172
01/14/05 04:20:41 entries:  00097e6c  Pkts:  440004  Err:  1152  PPS:  29333  
Drop %:  0.26%  Eth3RXInt:  298454  Eth2TXInt:  82420
01/14/05 04:20:56 entries:  000982c8  Pkts:  483248  Err:  1265  PPS:  32216  
Drop %:  0.26%  Eth3RXInt:  312325  Eth2TXInt:  80981
01/14/05 04:30:15 entries:  000a66f5  Pkts:  455322  Err:  1273  PPS:  30354  
Drop %:  0.27%  Eth3RXInt:  302449  Eth2TXInt:  82011
01/14/05 04:30:30 entries:  000a6c4f  Pkts:  443848  Err:  938  PPS:  29589  
Drop %:  0.21%  Eth3RXInt:  296296  Eth2TXInt:  82826
01/14/05 04:30:45 entries:  000a714d  Pkts:  457830  Err:  906  PPS:  30522  
Drop %:  0.19%  Eth3RXInt:  303239  Eth2TXInt:  82669
01/14/05 04:40:03 entries:  000b3c19  Pkts:  443891  Err:  1363  PPS:  29592  
Drop %:  0.30%  Eth3RXInt:  298216  Eth2TXInt:  81958
01/14/05 04:40:18 entries:  000b4131  Pkts:  436785  Err:  1055  PPS:  29119  
Drop %:  0.24%  Eth3RXInt:  293404  Eth2TXInt:  81866
01/14/05 04:40:33 entries:  000b4670  Pkts:  418272  Err:  886  PPS:  27884  
Drop %:  0.21%  Eth3RXInt:  284868  Eth2TXInt:  82277
01/14/05 04:40:48 entries:  00003a13  Pkts:  433659  Err:  5106  PPS:  28910  
Drop %:  1.17%  Eth3RXInt:  254964  Eth2TXInt:  80964
01/14/05 04:50:07 entries:  000487b1  Pkts:  426580  Err:  825  PPS:  28438  
Drop %:  0.19%  Eth3RXInt:  284115  Eth2TXInt:  81550
01/14/05 04:50:22 entries:  000499d9  Pkts:  429040  Err:  782  PPS:  28602  
Drop %:  0.18%  Eth3RXInt:  286171  Eth2TXInt:  81600
01/14/05 04:50:37 entries:  0004ab30  Pkts:  425725  Err:  965  PPS:  28381  
Drop %:  0.22%  Eth3RXInt:  287211  Eth2TXInt:  81419
01/14/05 04:50:52 entries:  0004bc1b  Pkts:  453751  Err:  1116  PPS:  30250  
Drop %:  0.24%  Eth3RXInt:  296099  Eth2TXInt:  81084
01/14/05 05:00:11 entries:  000744da  Pkts:  515817  Err:  1322  PPS:  34387  
Drop %:  0.25%  Eth3RXInt:  319462  Eth2TXInt:  81837
01/14/05 05:00:26 entries:  00075866  Pkts:  524724  Err:  1613  PPS:  34981  
Drop %:  0.30%  Eth3RXInt:  320024  Eth2TXInt:  81871
01/14/05 05:00:41 entries:  000768aa  Pkts:  513511  Err:  1273  PPS:  34234  
Drop %:  0.24%  Eth3RXInt:  319514  Eth2TXInt:  81763
01/14/05 05:00:56 entries:  00077652  Pkts:  521796  Err:  1088  PPS:  34786  
Drop %:  0.20%  Eth3RXInt:  322121  Eth2TXInt:  81846
01/14/05 05:10:14 entries:  0008bcd7  Pkts:  441778  Err:  1271  PPS:  29451  
Drop %:  0.28%  Eth3RXInt:  293564  Eth2TXInt:  81867
01/14/05 05:10:29 entries:  0008c48f  Pkts:  457062  Err:  1177  PPS:  30470  
Drop %:  0.25%  Eth3RXInt:  301042  Eth2TXInt:  82350
01/14/05 05:10:45 entries:  0008cb53  Pkts:  459844  Err:  940  PPS:  30656  
Drop %:  0.20%  Eth3RXInt:  303190  Eth2TXInt:  82519
01/14/05 05:20:03 entries:  0009b445  Pkts:  478691  Err:  889  PPS:  31912  
Drop %:  0.18%  Eth3RXInt:  313378  Eth2TXInt:  82914
01/14/05 05:20:18 entries:  0009ba72  Pkts:  485207  Err:  960  PPS:  32347  
Drop %:  0.19%  Eth3RXInt:  316409  Eth2TXInt:  82264
01/14/05 05:20:33 entries:  0009c150  Pkts:  486224  Err:  879  PPS:  32414  
Drop %:  0.18%  Eth3RXInt:  314240  Eth2TXInt:  81731
01/14/05 05:20:48 entries:  0009c35b  Pkts:  449497  Err:  1155  PPS:  29966  
Drop %:  0.25%  Eth3RXInt:  299010  Eth2TXInt:  82788
01/14/05 05:30:07 entries:  000ab4ea  Pkts:  485299  Err:  1095  PPS:  32353  
Drop %:  0.22%  Eth3RXInt:  314318  Eth2TXInt:  82271
01/14/05 05:30:22 entries:  000abae8  Pkts:  490269  Err:  797  PPS:  32684  
Drop %:  0.16%  Eth3RXInt:  317205  Eth2TXInt:  82521
01/14/05 05:30:37 entries:  000ac134  Pkts:  496198  Err:  1132  PPS:  33079  
Drop %:  0.22%  Eth3RXInt:  318933  Eth2TXInt:  81960
01/14/05 05:30:52 entries:  000ac2a1  Pkts:  501254  Err:  954  PPS:  33416  
Drop %:  0.19%  Eth3RXInt:  324082  Eth2TXInt:  83014
01/14/05 05:40:11 entries:  000baeeb  Pkts:  593611  Err:  1357  PPS:  39574  
Drop %:  0.22%  Eth3RXInt:  360758  Eth2TXInt:  81969
01/14/05 05:40:26 entries:  000bb507  Pkts:  567372  Err:  1385  PPS:  37824  
Drop %:  0.24%  Eth3RXInt:  350017  Eth2TXInt:  82295
01/14/05 05:40:41 entries:  000bbb4c  Pkts:  495626  Err:  1517  PPS:  33041  
Drop %:  0.30%  Eth3RXInt:  320385  Eth2TXInt:  81689
01/14/05 05:40:56 entries:  00006b5a  Pkts:  488706  Err:  6440  PPS:  32580  
Drop %:  1.31%  Eth3RXInt:  247798  Eth2TXInt:  80410
01/14/05 05:50:14 entries:  0004e410  Pkts:  500546  Err:  1278  PPS:  33369  
Drop %:  0.25%  Eth3RXInt:  310059  Eth2TXInt:  80386
01/14/05 05:50:29 entries:  0004f8e5  Pkts:  506876  Err:  1275  PPS:  33791  
Drop %:  0.25%  Eth3RXInt:  316871  Eth2TXInt:  81128
01/14/05 05:50:45 entries:  00050dc1  Pkts:  510450  Err:  1152  PPS:  34030  
Drop %:  0.22%  Eth3RXInt:  320390  Eth2TXInt:  80666
01/14/05 06:00:03 entries:  0008089f  Pkts:  519631  Err:  1531  PPS:  34642  
Drop %:  0.29%  Eth3RXInt:  323648  Eth2TXInt:  79945
01/14/05 06:00:18 entries:  00081247  Pkts:  535934  Err:  1163  PPS:  35728  
Drop %:  0.21%  Eth3RXInt:  327138  Eth2TXInt:  79790
01/14/05 06:00:33 entries:  00081bb3  Pkts:  540473  Err:  1305  PPS:  36031  
Drop %:  0.24%  Eth3RXInt:  332823  Eth2TXInt:  79750
01/14/05 06:00:48 entries:  00082320  Pkts:  541275  Err:  1622  PPS:  36085  
Drop %:  0.29%  Eth3RXInt:  334946  Eth2TXInt:  79207
01/14/05 06:10:07 entries:  00095795  Pkts:  559209  Err:  1351  PPS:  37280  
Drop %:  0.24%  Eth3RXInt:  336161  Eth2TXInt:  79479
01/14/05 06:10:22 entries:  000963c0  Pkts:  540560  Err:  1095  PPS:  36037  
Drop %:  0.20%  Eth3RXInt:  323433  Eth2TXInt:  79487
01/14/05 06:10:37 entries:  00096fa6  Pkts:  559203  Err:  1492  PPS:  37280  
Drop %:  0.26%  Eth3RXInt:  330691  Eth2TXInt:  80035
01/14/05 06:10:52 entries:  000979fb  Pkts:  542626  Err:  1438  PPS:  36175  
Drop %:  0.26%  Eth3RXInt:  325674  Eth2TXInt:  79483
01/14/05 06:20:11 entries:  000ad185  Pkts:  537524  Err:  1585  PPS:  35834  
Drop %:  0.29%  Eth3RXInt:  329987  Eth2TXInt:  79144
01/14/05 06:20:26 entries:  000adaba  Pkts:  513026  Err:  1035  PPS:  34201  
Drop %:  0.20%  Eth3RXInt:  317912  Eth2TXInt:  79128
01/14/05 06:20:41 entries:  000ae400  Pkts:  532172  Err:  1700  PPS:  35478  
Drop %:  0.31%  Eth3RXInt:  326356  Eth2TXInt:  79715
01/14/05 06:20:56 entries:  000aeb41  Pkts:  580261  Err:  1658  PPS:  38684  
Drop %:  0.28%  Eth3RXInt:  346806  Eth2TXInt:  79013
01/14/05 06:30:00 entries:  000c1594  Pkts:  572401  Err:  1903  PPS:  38160  
Drop %:  0.33%  Eth3RXInt:  341422  Eth2TXInt:  78273
01/14/05 06:30:15 entries:  000c1ea7  Pkts:  585850  Err:  1822  PPS:  39056  
Drop %:  0.31%  Eth3RXInt:  345728  Eth2TXInt:  77805
01/14/05 06:30:30 entries:  000c27b2  Pkts:  594077  Err:  1693  PPS:  39605  
Drop %:  0.28%  Eth3RXInt:  348358  Eth2TXInt:  76571
01/14/05 06:30:45 entries:  000c3123  Pkts:  582267  Err:  1706  PPS:  38817  
Drop %:  0.29%  Eth3RXInt:  342533  Eth2TXInt:  77763
01/14/05 06:40:04 entries:  000dfb48  Pkts:  710678  Err:  2075  PPS:  47378  
Drop %:  0.29%  Eth3RXInt:  377850  Eth2TXInt:  70656
01/14/05 06:40:19 entries:  000e083c  Pkts:  660634  Err:  1777  PPS:  44042  
Drop %:  0.26%  Eth3RXInt:  353285  Eth2TXInt:  71353
01/14/05 06:40:34 entries:  000e131e  Pkts:  683136  Err:  1696  PPS:  45542  
Drop %:  0.24%  Eth3RXInt:  369381  Eth2TXInt:  72622
01/14/05 06:40:49 entries:  00005c1d  Pkts:  713262  Err:  36148  PPS:  47550  
Drop %:  50.6%  Eth3RXInt:  286910  Eth2TXInt:  71729
01/14/05 06:50:08 entries:  0006066c  Pkts:  695785  Err:  1970  PPS:  46385  
Drop %:  0.28%  Eth3RXInt:  368369  Eth2TXInt:  73034
01/14/05 06:50:23 entries:  0006209f  Pkts:  705028  Err:  1715  PPS:  47001  
Drop %:  0.24%  Eth3RXInt:  368574  Eth2TXInt:  72892
01/14/05 06:50:38 entries:  000639ca  Pkts:  702841  Err:  1706  PPS:  46856  
Drop %:  0.24%  Eth3RXInt:  370024  Eth2TXInt:  72876
01/14/05 06:50:54 entries:  000651d2  Pkts:  726747  Err:  2020  PPS:  48449  
Drop %:  0.27%  Eth3RXInt:  376832  Eth2TXInt:  71366
01/14/05 07:00:13 entries:  0008f249  Pkts:  677362  Err:  1595  PPS:  45157  
Drop %:  0.23%  Eth3RXInt:  370724  Eth2TXInt:  69981
01/14/05 07:00:28 entries:  0008fe96  Pkts:  662406  Err:  1423  PPS:  44160  
Drop %:  0.21%  Eth3RXInt:  362967  Eth2TXInt:  72611
01/14/05 07:00:43 entries:  00090c13  Pkts:  662222  Err:  1844  PPS:  44148  
Drop %:  0.27%  Eth3RXInt:  359869  Eth2TXInt:  72181
01/14/05 07:00:58 entries:  00091789  Pkts:  656859  Err:  1729  PPS:  43790  
Drop %:  0.26%  Eth3RXInt:  358838  Eth2TXInt:  70074
01/14/05 07:10:02 entries:  000abdb6  Pkts:  697539  Err:  2042  PPS:  46502  
Drop %:  0.29%  Eth3RXInt:  379747  Eth2TXInt:  70269
01/14/05 07:10:17 entries:  000ac9d9  Pkts:  704847  Err:  1847  PPS:  46989  
Drop %:  0.26%  Eth3RXInt:  374670  Eth2TXInt:  69898
01/14/05 07:10:32 entries:  000ad57b  Pkts:  703755  Err:  2090  PPS:  46917  
Drop %:  0.29%  Eth3RXInt:  377955  Eth2TXInt:  70850
01/14/05 07:10:47 entries:  000adebf  Pkts:  709598  Err:  1790  PPS:  47306  
Drop %:  0.25%  Eth3RXInt:  379562  Eth2TXInt:  69334
01/14/05 07:20:06 entries:  000c9ceb  Pkts:  790839  Err:  2372  PPS:  52722  
Drop %:  0.29%  Eth3RXInt:  392123  Eth2TXInt:  63101
01/14/05 07:20:21 entries:  000caa53  Pkts:  815293  Err:  2200  PPS:  54352  
Drop %:  0.26%  Eth3RXInt:  391243  Eth2TXInt:  63859
01/14/05 07:20:37 entries:  000cb70a  Pkts:  801060  Err:  2421  PPS:  53404  
Drop %:  0.30%  Eth3RXInt:  392279  Eth2TXInt:  63283
01/14/05 07:20:52 entries:  000cc107  Pkts:  816395  Err:  2279  PPS:  54426  
Drop %:  0.27%  Eth3RXInt:  390125  Eth2TXInt:  63646
01/14/05 07:30:11 entries:  000e682c  Pkts:  817455  Err:  2501  PPS:  54497  
Drop %:  0.30%  Eth3RXInt:  396197  Eth2TXInt:  63137
01/14/05 07:30:26 entries:  000e756b  Pkts:  820908  Err:  2438  PPS:  54727  
Drop %:  0.29%  Eth3RXInt:  391849  Eth2TXInt:  63660
01/14/05 07:30:41 entries:  000e8389  Pkts:  816010  Err:  2613  PPS:  54400  
Drop %:  0.32%  Eth3RXInt:  392971  Eth2TXInt:  64416
01/14/05 07:30:56 entries:  000e8d18  Pkts:  827390  Err:  2536  PPS:  55159  
Drop %:  0.30%  Eth3RXInt:  394834  Eth2TXInt:  61765
01/14/05 07:40:01 entries:  0010436e  Pkts:  863423  Err:  2368  PPS:  57561  
Drop %:  0.27%  Eth3RXInt:  405190  Eth2TXInt:  60998
01/14/05 07:40:16 entries:  00104fda  Pkts:  858511  Err:  2494  PPS:  57234  
Drop %:  0.29%  Eth3RXInt:  403185  Eth2TXInt:  61815
01/14/05 07:40:31 entries:  00105c15  Pkts:  854565  Err:  2470  PPS:  56971  
Drop %:  0.28%  Eth3RXInt:  404894  Eth2TXInt:  61287
01/14/05 07:40:46 entries:  000057a0  Pkts:  826768  Err:  113636  PPS:  55117  
Drop %:  13.74%  Eth3RXInt:  302767  Eth2TXInt:  65567
01/14/05 07:50:06 entries:  00079684  Pkts:  913524  Err:  2245  PPS:  60901  
Drop %:  0.24%  Eth3RXInt:  400066  Eth2TXInt:  58282
01/14/05 07:50:21 entries:  0007b783  Pkts:  864505  Err:  2505  PPS:  57633  
Drop %:  0.28%  Eth3RXInt:  393922  Eth2TXInt:  61211
01/14/05 07:50:36 entries:  0007d8a8  Pkts:  883376  Err:  2693  PPS:  58891  
Drop %:  0.30%  Eth3RXInt:  397785  Eth2TXInt:  61160
01/14/05 07:50:51 entries:  0007f40c  Pkts:  858829  Err:  2394  PPS:  57255  
Drop %:  0.27%  Eth3RXInt:  398098  Eth2TXInt:  60954
01/14/05 08:00:11 entries:  000a460b  Pkts:  1000645  Err:  2906  PPS:  66709  
Drop %:  0.29%  Eth3RXInt:  406845  Eth2TXInt:  53497
01/14/05 08:00:26 entries:  000a55a7  Pkts:  965824  Err:  3062  PPS:  64388  
Drop %:  0.31%  Eth3RXInt:  402766  Eth2TXInt:  54223
01/14/05 08:00:41 entries:  000a6548  Pkts:  975763  Err:  3184  PPS:  65050  
Drop %:  0.32%  Eth3RXInt:  408337  Eth2TXInt:  55173
01/14/05 08:00:56 entries:  000a726e  Pkts:  963100  Err:  3105  PPS:  64206  
Drop %:  0.32%  Eth3RXInt:  409883  Eth2TXInt:  54808
01/14/05 08:10:01 entries:  000c7fc9  Pkts:  1045671  Err:  2629  PPS:  69711  
Drop %:  0.25%  Eth3RXInt:  412912  Eth2TXInt:  52786
01/14/05 08:10:16 entries:  000c8dd3  Pkts:  1059809  Err:  3206  PPS:  70653  
Drop %:  0.30%  Eth3RXInt:  414254  Eth2TXInt:  51833
01/14/05 08:10:32 entries:  000c9c1a  Pkts:  1026487  Err:  3277  PPS:  68432  
Drop %:  0.31%  Eth3RXInt:  414313  Eth2TXInt:  51822
01/14/05 08:10:47 entries:  000ca57f  Pkts:  1026565  Err:  3149  PPS:  68437  
Drop %:  0.30%  Eth3RXInt:  412748  Eth2TXInt:  51581
01/14/05 08:20:07 entries:  000eb744  Pkts:  1086031  Err:  3287  PPS:  72402  
Drop %:  0.30%  Eth3RXInt:  406328  Eth2TXInt:  48924
01/14/05 08:20:22 entries:  000ec65d  Pkts:  1066622  Err:  2733  PPS:  71108  
Drop %:  0.25%  Eth3RXInt:  408729  Eth2TXInt:  49285
01/14/05 08:20:37 entries:  000ed550  Pkts:  1069185  Err:  2919  PPS:  71279  
Drop %:  0.27%  Eth3RXInt:  409380  Eth2TXInt:  49829
01/14/05 08:20:52 entries:  000ede68  Pkts:  1098310  Err:  3181  PPS:  73220  
Drop %:  0.28%  Eth3RXInt:  414812  Eth2TXInt:  50289
01/14/05 08:30:12 entries:  0010ba04  Pkts:  1132309  Err:  3248  PPS:  75487  
Drop %:  0.28%  Eth3RXInt:  414595  Eth2TXInt:  45562
01/14/05 08:30:27 entries:  0010c986  Pkts:  1154479  Err:  3665  PPS:  76965  
Drop %:  0.31%  Eth3RXInt:  404918  Eth2TXInt:  45873
01/14/05 08:30:43 entries:  0010d8e5  Pkts:  1192416  Err:  4128  PPS:  79494  
Drop %:  0.34%  Eth3RXInt:  405126  Eth2TXInt:  43819
01/14/05 08:30:58 entries:  0010e633  Pkts:  1166668  Err:  3361  PPS:  77777  
Drop %:  0.28%  Eth3RXInt:  415004  Eth2TXInt:  44711
01/14/05 08:40:03 entries:  0012aad0  Pkts:  1138575  Err:  3211  PPS:  75905  
Drop %:  0.28%  Eth3RXInt:  373013  Eth2TXInt:  47718
01/14/05 08:40:18 entries:  0012c2e3  Pkts:  1158909  Err:  3481  PPS:  77260  
Drop %:  0.30%  Eth3RXInt:  366664  Eth2TXInt:  45401
01/14/05 08:40:33 entries:  0012d5fb  Pkts:  1155017  Err:  3674  PPS:  77001  
Drop %:  0.31%  Eth3RXInt:  389075  Eth2TXInt:  46785
01/14/05 08:40:48 entries:  00008aee  Pkts:  1033983  Err:  174485  PPS:  
68932  Drop %:  16.87%  Eth3RXInt:  246160  Eth2TXInt:  60997

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-13 22:27                                                                                       ` Jeremy M. Guthrie
@ 2005-01-14 15:44                                                                                         ` Robert Olsson
  0 siblings, 0 replies; 88+ messages in thread
From: Robert Olsson @ 2005-01-14 15:44 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson


Jeremy M. Guthrie writes:
 > 
 > IN:Hits are cache hits yes?  Tot, are the total number of flows created since 
 > we last looked at the total flow count, correct?  

 Thats correct.

 > What would cause a packet to drop in the network stack and thus showup in 
 > /proc/net/softnet_stat?
 
 Only netif_rx (non-NAPI) driver drops in the backlog. NAPI drivers drops early 
 in device.

 > Performance has picked up.  I am not dropping packets anymore except during 
 > GC.  I upped my interval from 600 seconds to 1800 seconds.   

 Yes and as far as I see you have CPU left the as RX interrupts on eth3 indicates
 (if the patch is applied) but also the low numbers of time squeeze in sofnet_stat

						   --ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-14 14:59                                                                                       ` Jeremy M. Guthrie
@ 2005-01-14 16:05                                                                                         ` Robert Olsson
  2005-01-14 19:00                                                                                           ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-14 16:05 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson


Jeremy M. Guthrie writes:
 > I ran a script overnight using the modified driver you had given me Robert.  
 > it is interesting that there is almost always errors in the interface even 
 > though we aren't getting dst-cache errors and running ~ 40% free CPU now.  I 
 > am going to switch over to Jesse's driver to see if his instrumentation helps 
 > nail down where the problem is.


 >      rx_packets: 2722676103
 >      tx_packets: 5
 >      rx_bytes: 1171335471
 >      tx_bytes: 398
 >      rx_errors: 8558366
 >      tx_errors: 0
 >      rx_dropped: 1951692

 It might come from be periodic work from GC process correlate drops w. rtatat.

 I think the GC process can be made more smooth but studies and experimentation
 is probably needed as GC process is quite complex. Maybe some looked into this 
 already?

 Also you reported less drops when you increased the size of the RX ring and
 I see higher system performance with smaller RX rings. Both statements may
 actually be true. Big rings size may buffer during periodic work as GC.

 Also I have an experimental patch so you route without the route hash as a 
 comparison. You have to be brave...

 BTW we had this thread going for week,
 
							--ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-14 16:05                                                                                         ` Robert Olsson
@ 2005-01-14 19:00                                                                                           ` Jeremy M. Guthrie
  2005-01-14 19:26                                                                                             ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-14 19:00 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 2686 bytes --]

These stats are from the driver Jesse instrumented.
ethtool -S eth3 | egrep -v ": 0"
NIC statistics:
     rx_packets: 1130958533
     tx_packets: 5
     rx_bytes: 2373643298
     tx_bytes: 398
     rx_errors: 4388190
     rx_dropped: 1486253
     rx_fifo_errors: 2901937
     rx_missed_errors: 2901937
     rx_long_byte_count: 582194228258
     rx_csum_offload_good: 1040376597
     int_tx_desc: 4
     int_tx_queueempty: 5
     int_link_state: 1
     int_rx_desc_min_thresh: 20704
     int_rx_fifo_ovr: 1208
     int_rx_timer: 331925913
     int_rxcfg: 1
     rx_csum_offload_errors: 325045

I am seeing more times w/ his driver where I run at zero errors.  Of the last 
877 samples since I switched drivers, I see 231 samples w/ zero errors.

On Friday 14 January 2005 10:05 am, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > I ran a script overnight using the modified driver you had given me
>  > Robert. it is interesting that there is almost always errors in the
>  > interface even though we aren't getting dst-cache errors and running ~
>  > 40% free CPU now.  I am going to switch over to Jesse's driver to see if
>  > his instrumentation helps nail down where the problem is.
>  >
>  >
>  >      rx_packets: 2722676103
>  >      tx_packets: 5
>  >      rx_bytes: 1171335471
>  >      tx_bytes: 398
>  >      rx_errors: 8558366
>  >      tx_errors: 0
>  >      rx_dropped: 1951692
>
>  It might come from be periodic work from GC process correlate drops w.
> rtatat.
>
>  I think the GC process can be made more smooth but studies and
> experimentation is probably needed as GC process is quite complex. Maybe
> some looked into this already?
>  Also you reported less drops when you increased the size of the RX ring
> and I see higher system performance with smaller RX rings. Both statements
> may actually be true. Big rings size may buffer during periodic work as GC.
I am running w/ 2048 input buffers.  I am going to increase to 10K and try 
again.

>  Also I have an experimental patch so you route without the route hash as a
>  comparison. You have to be brave...
I am about 300 miles from the machine this week and next week though I might 
be able to try it this weekend while I am home.

>  BTW we had this thread going for week,
Pardon me for being a bit naive but I am not understanding this last comment.

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-14 19:00                                                                                           ` Jeremy M. Guthrie
@ 2005-01-14 19:26                                                                                             ` Jeremy M. Guthrie
  2005-01-16 12:32                                                                                               ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-14 19:26 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 3208 bytes --]

I actually upped the buffer count to 8192 buffers instead of 10k.  Of the 74 
samples I have thus far, 57 have been clean of errors.  Most of the sample 
errors appear to be shortly after the cache flush.

I am going to let the buffers run for a bit longer then I have to go.  I have 
a five hour drive ahead of me.  

On Friday 14 January 2005 01:00 pm, Jeremy M. Guthrie wrote:
> These stats are from the driver Jesse instrumented.
> ethtool -S eth3 | egrep -v ": 0"
> NIC statistics:
>      rx_packets: 1130958533
>      tx_packets: 5
>      rx_bytes: 2373643298
>      tx_bytes: 398
>      rx_errors: 4388190
>      rx_dropped: 1486253
>      rx_fifo_errors: 2901937
>      rx_missed_errors: 2901937
>      rx_long_byte_count: 582194228258
>      rx_csum_offload_good: 1040376597
>      int_tx_desc: 4
>      int_tx_queueempty: 5
>      int_link_state: 1
>      int_rx_desc_min_thresh: 20704
>      int_rx_fifo_ovr: 1208
>      int_rx_timer: 331925913
>      int_rxcfg: 1
>      rx_csum_offload_errors: 325045
>
> I am seeing more times w/ his driver where I run at zero errors.  Of the
> last 877 samples since I switched drivers, I see 231 samples w/ zero
> errors.
>
> On Friday 14 January 2005 10:05 am, Robert Olsson wrote:
> > Jeremy M. Guthrie writes:
> >  > I ran a script overnight using the modified driver you had given me
> >  > Robert. it is interesting that there is almost always errors in the
> >  > interface even though we aren't getting dst-cache errors and running ~
> >  > 40% free CPU now.  I am going to switch over to Jesse's driver to see
> >  > if his instrumentation helps nail down where the problem is.
> >  >
> >  >
> >  >      rx_packets: 2722676103
> >  >      tx_packets: 5
> >  >      rx_bytes: 1171335471
> >  >      tx_bytes: 398
> >  >      rx_errors: 8558366
> >  >      tx_errors: 0
> >  >      rx_dropped: 1951692
> >
> >  It might come from be periodic work from GC process correlate drops w.
> > rtatat.
> >
> >  I think the GC process can be made more smooth but studies and
> > experimentation is probably needed as GC process is quite complex. Maybe
> > some looked into this already?
> >  Also you reported less drops when you increased the size of the RX ring
> > and I see higher system performance with smaller RX rings. Both
> > statements may actually be true. Big rings size may buffer during
> > periodic work as GC.
>
> I am running w/ 2048 input buffers.  I am going to increase to 10K and try
> again.
>
> >  Also I have an experimental patch so you route without the route hash as
> > a comparison. You have to be brave...
>
> I am about 300 miles from the machine this week and next week though I
> might be able to try it this weekend while I am home.
>
> >  BTW we had this thread going for week,
>
> Pardon me for being a bit naive but I am not understanding this last
> comment.

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-14 19:26                                                                                             ` Jeremy M. Guthrie
@ 2005-01-16 12:32                                                                                               ` Robert Olsson
  2005-01-16 16:22                                                                                                 ` Jeremy M. Guthrie
  2005-01-19 15:03                                                                                                 ` Jeremy M. Guthrie
  0 siblings, 2 replies; 88+ messages in thread
From: Robert Olsson @ 2005-01-16 12:32 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson


Jeremy M. Guthrie writes:
 > I actually upped the buffer count to 8192 buffers instead of 10k.  
 > Of the 74 samples I have thus far, 57 have been clean of errors.  
 > Most of the sample errors appear to be shortly after the cache flush.

 I don't really believe in increasing RX buffers to this extent. We verified
 that you have CPU available and the drops occur when the timer based GC
 happens. Increasing buffers decreases overall performance and adds jitter.

 We saw also the timed based GC were taking the dst-entries from about
 600k to 40k in one shot. I think this what we should look into. Just 
 GC is "work" also after GC a lot flows has to be recreated doing fib 
 lookup and creating new entries. We want to smoothen the GC process so 
 happen more frequent and does less work. 

 Some time ago an "in-flow" GC (as opposed to timer based) was added to
 the routing code look for cand in route.c. In setup like yours (and ours) 
 it would be better to relay on this process to a higher extent. Anyway
 in /proc/sys/net/ipv4/route/ you have the files.

 gc_elasticity, gc_interval, gc_thresh etc I would avoid gc_min_interval.

 And you can play with your running system and for drops without causing
 your users to much pain. 
 
 We save the patch for routing without route hash and GC until later,


					       --ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-16 12:32                                                                                               ` Robert Olsson
@ 2005-01-16 16:22                                                                                                 ` Jeremy M. Guthrie
  2005-01-19 15:03                                                                                                 ` Jeremy M. Guthrie
  1 sibling, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-16 16:22 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 2373 bytes --]

On Sunday 16 January 2005 06:32 am, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > I actually upped the buffer count to 8192 buffers instead of 10k.
>  > Of the 74 samples I have thus far, 57 have been clean of errors.
>  > Most of the sample errors appear to be shortly after the cache flush.
>
>  I don't really believe in increasing RX buffers to this extent. We
> verified that you have CPU available and the drops occur when the timer
> based GC happens. Increasing buffers decreases overall performance and adds
> jitter.
I just took a look at my logs, the increase to the max on the card of 4K RX 
buffers has stopped packet drops except during GC.  I agree, it isn't pretty 
and it is not the solution I would like.  In the mean time, it has at least 
stopped the 0.3% round-the-clock packet loss which I was seeing even at rates 
as low as 25K pps.

>  We saw also the timed based GC were taking the dst-entries from about
>  600k to 40k in one shot. I think this what we should look into. Just
>  GC is "work" also after GC a lot flows has to be recreated doing fib
>  lookup and creating new entries. We want to smoothen the GC process so
>  happen more frequent and does less work.
Agreed.  I went to the extreme because I can really see the % idle CPU.  If I 
am constantly setting up new flows then the % of free CPU shoots way down.  
Minus the effects of GC, a larger flow table equates into free CPU.  

>  Some time ago an "in-flow" GC (as opposed to timer based) was added to
>  the routing code look for cand in route.c. In setup like yours (and ours)
>  it would be better to relay on this process to a higher extent. Anyway
>  in /proc/sys/net/ipv4/route/ you have the files.
>  gc_elasticity, gc_interval, gc_thresh etc I would avoid gc_min_interval.
>  And you can play with your running system and for drops without causing
>  your users to much pain.
>  We save the patch for routing without route hash and GC until later,
Okay.  I will bring my interval down from an hour down to ten minutes and do 
further tuning.

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-16 12:32                                                                                               ` Robert Olsson
  2005-01-16 16:22                                                                                                 ` Jeremy M. Guthrie
@ 2005-01-19 15:03                                                                                                 ` Jeremy M. Guthrie
  2005-01-19 22:18                                                                                                   ` Robert Olsson
  1 sibling, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-19 15:03 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 1026 bytes --]

On Sunday 16 January 2005 06:32 am, Robert Olsson wrote:
>  Some time ago an "in-flow" GC (as opposed to timer based) was added to
>  the routing code look for cand in route.c. In setup like yours (and ours)
>  it would be better to relay on this process to a higher extent. Anyway
>  in /proc/sys/net/ipv4/route/ you have the files.
>
>  gc_elasticity, gc_interval, gc_thresh etc I would avoid gc_min_interval.
>
>  And you can play with your running system and for drops without causing
>  your users to much pain.
I have done a little tweaking.  I now hold at around 520K routes in the hash.  
I still drop packets every secret_interval but I've upped that counter so I 
don't whack all of my hash entries all that often.

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-19 15:03                                                                                                 ` Jeremy M. Guthrie
@ 2005-01-19 22:18                                                                                                   ` Robert Olsson
  2005-01-20  1:50                                                                                                     ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-19 22:18 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson


Jeremy M. Guthrie writes:

 > >  gc_elasticity, gc_interval, gc_thresh etc I would avoid gc_min_interval.

 > I have done a little tweaking. I now hold at around 520K routes in the hash.  
 > I still drop packets every secret_interval but I've upped that counter so I 
 > don't whack all of my hash entries all that often.


 Sounds you done progress. 
 This with relative conservative setting if RX-buffers?

 secret_interval needs special care as it flushes the cache totally.

 What did you tweak? 
 Your output from rtstat will be very interesting.


					 --ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-19 22:18                                                                                                   ` Robert Olsson
@ 2005-01-20  1:50                                                                                                     ` Jeremy M. Guthrie
  2005-01-20 11:30                                                                                                       ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-20  1:50 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

@ 14:51:44 we can see the secret_interval kick in.  Otherwise we run very 
solid.  

101Kpps  8)

1024 buffers on the RX of eth3.

echo 86400 > /proc/sys/net/ipv4/route/secret_interval
echo 524288 > /proc/sys/net/ipv4/route/gc_thresh

It appears to be cruising right along now.

-----------

01/19/05 14:49:57 entries:  0007fffa  Pkts:  1413598  Err:  0  PPS:  94239  
Drop %:  0.0%  Eth3RXInt:  309669  Eth2TXInt:  36693
01/19/05 14:50:13 entries:  0007fff7  Pkts:  1448829  Err:  0  PPS:  96588  
Drop %:  0.0%  Eth3RXInt:  295522  Eth2TXInt:  36475
01/19/05 14:50:28 entries:  0007ffff  Pkts:  1449774  Err:  0  PPS:  96651  
Drop %:  0.0%  Eth3RXInt:  292556  Eth2TXInt:  36780
01/19/05 14:50:43 entries:  00080000  Pkts:  1454807  Err:  0  PPS:  96987  
Drop %:  0.0%  Eth3RXInt:  300945  Eth2TXInt:  35358
01/19/05 14:50:58 entries:  00080000  Pkts:  1426412  Err:  0  PPS:  95094  
Drop %:  0.0%  Eth3RXInt:  303005  Eth2TXInt:  35118
01/19/05 14:51:14 entries:  00080002  Pkts:  1471279  Err:  0  PPS:  98085  
Drop %:  0.0%  Eth3RXInt:  298152  Eth2TXInt:  35573
01/19/05 14:51:29 entries:  00080002  Pkts:  1402715  Err:  0  PPS:  93514  
Drop %:  0.0%  Eth3RXInt:  314880  Eth2TXInt:  37688
01/19/05 14:51:44 entries:  00007ef6  Pkts:  1335367  Err:  204660  PPS:  
89024  Drop %:  15.32%  Eth3RXInt:  203518  Eth2TXInt:  54907
01/19/05 14:52:00 entries:  00014a3a  Pkts:  1484009  Err:  0  PPS:  98933  
Drop %:  0.0%  Eth3RXInt:  138942  Eth2TXInt:  60283
01/19/05 14:52:15 entries:  0001f092  Pkts:  1441979  Err:  0  PPS:  96131  
Drop %:  0.0%  Eth3RXInt:  175142  Eth2TXInt:  53552
01/19/05 14:52:30 entries:  00028f95  Pkts:  1460829  Err:  0  PPS:  97388  
Drop %:  0.0%  Eth3RXInt:  181189  Eth2TXInt:  53574
01/19/05 14:52:45 entries:  000303a3  Pkts:  1423530  Err:  0  PPS:  94902  
Drop %:  0.0%  Eth3RXInt:  234029  Eth2TXInt:  48046
01/19/05 14:53:01 entries:  00035167  Pkts:  1443491  Err:  0  PPS:  96232  
Drop %:  0.0%  Eth3RXInt:  270393  Eth2TXInt:  40791
01/19/05 14:53:16 entries:  0003994a  Pkts:  1455118  Err:  0  PPS:  97007  
Drop %:  0.0%  Eth3RXInt:  276102  Eth2TXInt:  39695
01/19/05 14:53:31 entries:  0003df5a  Pkts:  1405048  Err:  0  PPS:  93669  
Drop %:  0.0%  Eth3RXInt:  287641  Eth2TXInt:  39903
01/19/05 14:53:46 entries:  00041f5c  Pkts:  1392438  Err:  0  PPS:  92829  
Drop %:  0.0%  Eth3RXInt:  307840  Eth2TXInt:  40216
01/19/05 14:54:01 entries:  00045da3  Pkts:  1486506  Err:  0  PPS:  99100  
Drop %:  0.0%  Eth3RXInt:  280067  Eth2TXInt:  38030
01/19/05 14:54:17 entries:  00049a78  Pkts:  1504340  Err:  0  PPS:  100289  
Drop %:  0.0%  Eth3RXInt:  276769  Eth2TXInt:  36982
01/19/05 14:54:32 entries:  0004d72b  Pkts:  1489358  Err:  0  PPS:  99290  
Drop %:  0.0%  Eth3RXInt:  280576  Eth2TXInt:  37378
01/19/05 14:54:47 entries:  00051214  Pkts:  1475776  Err:  0  PPS:  98385  
Drop %:  0.0%  Eth3RXInt:  289098  Eth2TXInt:  36632
01/19/05 14:55:02 entries:  00054ceb  Pkts:  1527761  Err:  0  PPS:  101850  
Drop %:  0.0%  Eth3RXInt:  273478  Eth2TXInt:  35447
01/19/05 14:55:18 entries:  000584f7  Pkts:  1482312  Err:  0  PPS:  98820  
Drop %:  0.0%  Eth3RXInt:  288181  Eth2TXInt:  35792
01/19/05 14:55:33 entries:  0005be56  Pkts:  1477823  Err:  0  PPS:  98521  
Drop %:  0.0%  Eth3RXInt:  284099  Eth2TXInt:  36698
01/19/05 14:55:48 entries:  0005f55b  Pkts:  1489107  Err:  0  PPS:  99273  
Drop %:  0.0%  Eth3RXInt:  289669  Eth2TXInt:  36162
01/19/05 14:56:03 entries:  00062ae8  Pkts:  1451832  Err:  0  PPS:  96788  
Drop %:  0.0%  Eth3RXInt:  299549  Eth2TXInt:  36259
01/19/05 14:56:19 entries:  000660cd  Pkts:  1423292  Err:  0  PPS:  94886  
Drop %:  0.0%  Eth3RXInt:  306119  Eth2TXInt:  36382
01/19/05 14:56:34 entries:  000697a1  Pkts:  1458930  Err:  0  PPS:  97262  
Drop %:  0.0%  Eth3RXInt:  302765  Eth2TXInt:  37878


On Wednesday 19 January 2005 04:18 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > >  gc_elasticity, gc_interval, gc_thresh etc I would avoid
>  > > gc_min_interval.
>  >
>  > I have done a little tweaking. I now hold at around 520K routes in the
>  > hash. I still drop packets every secret_interval but I've upped that
>  > counter so I don't whack all of my hash entries all that often.
>
>  Sounds you done progress.
>  This with relative conservative setting if RX-buffers?
>
>  secret_interval needs special care as it flushes the cache totally.
>
>  What did you tweak?
>  Your output from rtstat will be very interesting.
>
>
>       --ro

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-20  1:50                                                                                                     ` Jeremy M. Guthrie
@ 2005-01-20 11:30                                                                                                       ` Robert Olsson
  2005-01-20 14:37                                                                                                         ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-20 11:30 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson


Jeremy M. Guthrie writes:
 > @ 14:51:44 we can see the secret_interval kick in.  Otherwise we run very 
 > solid.  
 > 
 > 101Kpps  8)

 Seems traffic increaed a little bit too?

 > 1024 buffers on the RX of eth3.
 > 
 > echo 86400 > /proc/sys/net/ipv4/route/secret_interval
 > echo 524288 > /proc/sys/net/ipv4/route/gc_thresh

 rhash_entries?

 > It appears to be cruising right along now.

 Nice. rtstat's is still interesting so we can see number fib lookups and
 linear search and GC dynamics etc.


						--ro

 

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-20 11:30                                                                                                       ` Robert Olsson
@ 2005-01-20 14:37                                                                                                         ` Jeremy M. Guthrie
  2005-01-20 17:01                                                                                                           ` Robert Olsson
  0 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-20 14:37 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 6859 bytes --]

On Thursday 20 January 2005 05:30 am, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > @ 14:51:44 we can see the secret_interval kick in.  Otherwise we run
>  > very solid.
>  >
>  > 101Kpps  8)
>
>  Seems traffic increaed a little bit too?
>
>  > 1024 buffers on the RX of eth3.
>  >
>  > echo 86400 > /proc/sys/net/ipv4/route/secret_interval
>  > echo 524288 > /proc/sys/net/ipv4/route/gc_thresh
>
>  rhash_entries?
I left that at 2.4 million.

>  > It appears to be cruising right along now.
>
>  Nice. rtstat's is still interesting so we can see number fib lookups and
>  linear search and GC dynamics etc.
 size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
GC: tot ignored goal_miss ovrf HASH: in_search out_search
524291     73651     805     0     0     0     0     0         5       0 
537114631     537       0         0    0           45657          3
524291     74169     779     0     0     0     0     0         2       0 
1610369017     516       0         0    0           43612          0
524287     74830     777     0     0     0     0     0         6       0      
0     498       0         0    0           40504          1
524285     77458     949     0     0     0     0     0         4       0      
0     619       0         0    0           42731          1
524273     75475     801     0     0     0     0     0        10       0      
0     549       0         0    0           40084          7
524292     76740     902     0     0     0     0     0         5       0      
0     574       0         0    0           41702          3
524295     76681     829     0     0     0     0     0         4       0      
0     527       0         0    0           42328          2
524333     76522     756     0     0     0     0     0         9       1      
0     432       0         0    0           43621          3
524286     80323     792     0     0     0     0     0         6       0      
0     498       0         0    0           44619          2
524294     79409     758     0     0     0     0     0         6       0      
0     508       0         0    0           45864          3
524288     80058     756     0     0     0     0     0         7       0      
0     478       0         0    0           44711          5
524291     77128     794     0     0     0     0     0         8       0      
0     516       0         0    0           42301          3
524288     77433     823     0     0     0     0     0         4       0      
0     539       0         0    0           42357          0
524293     79423     761     0     0     0     0     0         7       0      
0     514       0         0    0           43102          3
524293     80940     804     0     0     0     0     0         6       0      
0     530       0         0    0           45457          3
524287     84864     813     0     0     0     0     0         3       0      
0     532       0         0    0           44309          0
524286     78358     804     0     0     0     0     0         6       0      
0     531       0         0    0           43669          2
524293     72760     717     0     0     0     0     0         2       0      
0     480       0         0    0           41611          0
524287     72833     684     0     0     0     0     0         5       0      
0     464       0         0    0           40571          1
524290     77308     726     0     0     0     0     0         6       0      
0     486       0         0    0           43031          2
 size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
GC: tot ignored goal_miss ovrf HASH: in_search out_search
524288     81205     746     0     0     0     0     0         5       0 
537114631     512       0         0    0           42384          3
524291     81744     736     0     0     0     0     0         5       0 
1610369017     494       0         0    0           43980          3
524285     85813     766     0     0     0     0     0         5       0      
0     485       0         0    0           45385          3
524290     80609     810     0     0     0     0     0         5       0      
0     547       0         0    0           42037          2
524281     75730     784     0     0     0     0     0         7       0      
0     515       0         0    0           42966          3
524295     78139     805     0     0     0     0     0         6       0      
0     540       0         0    0           43026          3
524292     77198     724     0     0     0     0     0         7       1      
0     484       0         0    0           44583          1
524301     80243     802     0     0     0     0     0         7       0      
0     524       0         0    0           43462          2
524286     78010     822     0     0     0     0     0         8       0      
0     557       0         0    0           44684          4
524292     76739     814     0     0     0     0     0        11       1      
0     539       0         0    0           45128          5
524288     79226     765     0     0     0     0     0         6       0      
0     518       0         0    0           47706          1
524286     78367     757     0     0     0     0     0         8       1      
0     486       0         0    0           44462          1
524292     79453     774     0     0     0     0     0        10       0      
0     514       0         0    0           43612          5
524291     75541     749     0     0     0     0     0        10       0      
0     492       0         0    0           43462          3
524289     77849     748     0     0     0     0     0        14       3      
0     493       0         0    0           43390          7
524296     77628     773     0     0     0     0     0         8       0      
0     505       0         0    0           43900          3
524295     79695     699     0     0     0     0     0         8       0      
0     451       0         0    0           45529          4
524293     78484     770     0     0     0     0     0         7       0      
0     339       0         0    0           44700          2
524298     79257     732     0     0     0     0     0         8       0      
0     486       0         0    0           45880          4
524287     79434     749     0     0     0     0     0        12       0      
0     496       0         0    0           45081          5

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-20 14:37                                                                                                         ` Jeremy M. Guthrie
@ 2005-01-20 17:01                                                                                                           ` Robert Olsson
  2005-01-20 17:14                                                                                                             ` Jeremy M. Guthrie
  0 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-20 17:01 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson


Jeremy M. Guthrie writes:

 > >  > 1024 buffers on the RX of eth3.
 > >  >
 > >  > echo 86400 > /proc/sys/net/ipv4/route/secret_interval
 > >  > echo 524288 > /proc/sys/net/ipv4/route/gc_thresh
 > >
 > >  rhash_entries?
 > I left that at 2.4 million.
 > 
 > >  > It appears to be cruising right along now.


 >  size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
 > GC: tot ignored goal_miss ovrf HASH: in_search out_search
 > 524291     74169     779     0     0     0     0     0         2       0 
 > 1610369017     516       0         0    0           43612          0
 > 524287     74830     777     0     0     0     0     0         6       0      
 > 0     498       0         0    0           40504          1
 > 524285     77458     949     0     0     0     0     0         4       0      
 > 0     619       0         0    0           42731          1


 Linear search is under control and number of dst entries very high but very
 constant at the cost of calling GC a number of times second. But I don't 
 understand why we do not see any GC ignored. Did you ever write to gc_min_interval 
 in proc? Never seen rtstat's like this but it seems to do the job.

						  --ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-20 17:01                                                                                                           ` Robert Olsson
@ 2005-01-20 17:14                                                                                                             ` Jeremy M. Guthrie
  2005-01-20 21:53                                                                                                               ` Robert Olsson
                                                                                                                                 ` (2 more replies)
  0 siblings, 3 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-20 17:14 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 2470 bytes --]

On Thursday 20 January 2005 11:01 am, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > >  > 1024 buffers on the RX of eth3.
>  > >  >
>  > >  > echo 86400 > /proc/sys/net/ipv4/route/secret_interval
>  > >  > echo 524288 > /proc/sys/net/ipv4/route/gc_thresh
>  > >
>  > >  rhash_entries?
>  >
>  > I left that at 2.4 million.
>  >
>  > >  > It appears to be cruising right along now.
>  >
>  >  size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot 
>  >    mc GC: tot ignored goal_miss ovrf HASH: in_search out_search
>  > 524291     74169     779     0     0     0     0     0         2       0
>  > 1610369017     516       0         0    0           43612          0
>  > 524287     74830     777     0     0     0     0     0         6       0
>  > 0     498       0         0    0           40504          1
>  > 524285     77458     949     0     0     0     0     0         4       0
>  > 0     619       0         0    0           42731          1
>
>  Linear search is under control and number of dst entries very high but
> very constant at the cost of calling GC a number of times second. But I
> don't understand why we do not see any GC ignored. 
When does GC normally ignore?

> Did you ever write to 
> gc_min_interval in proc? 
I left /proc/sys/net/ipv4/route/gc_min_interval at zero.

> Never seen rtstat's like this but it seems to do the job.

More numbers from right now with higher PPS rate.
 size   IN: hit     tot    mc no_rt bcast madst masrc  OUT: hit     tot     mc 
GC: tot ignored goal_miss ovrf HASH: in_search out_search
524291     92192     848     0     0     0     0     0        17       0 
537114631     553       0         0    0           50947         18
524287     95496     846     0     0     0     0     0         4       0 
1610369017     539       0         0    0           52000          4
524293     98503     791     0     0     0     0     0         7       0      
0     525       0         0    0           53119          3
524290     98711     965     0     0     0     0     0         3       0      
0     626       0         0    0           53448          3


-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-20 17:14                                                                                                             ` Jeremy M. Guthrie
@ 2005-01-20 21:53                                                                                                               ` Robert Olsson
  2005-01-21 21:20                                                                                                                 ` Jeremy M. Guthrie
  2005-01-21 15:23                                                                                                               ` Robert Olsson
  2005-01-31 15:37                                                                                                               ` Jeremy M. Guthrie
  2 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-20 21:53 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson


Jeremy M. Guthrie writes:

 > >  >    mc GC: tot ignored goal_miss ovrf HASH: in_search out_search
 > >  > 524287     74830     777     0     0     0     0     0         6       0
 > >  > 0     498       0         0    0           40504          1

 > When does GC normally ignore?

 At a min ip_rt_gc_min_interval       = HZ / 2;

 From what I understand your GC runs more often than that.
 Do your understand? Test is in rt_garbage_collect()

					    --ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-20 17:14                                                                                                             ` Jeremy M. Guthrie
  2005-01-20 21:53                                                                                                               ` Robert Olsson
@ 2005-01-21 15:23                                                                                                               ` Robert Olsson
  2005-01-21 21:24                                                                                                                 ` Jeremy M. Guthrie
  2005-01-31 15:37                                                                                                               ` Jeremy M. Guthrie
  2 siblings, 1 reply; 88+ messages in thread
From: Robert Olsson @ 2005-01-21 15:23 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson




We have to understand the GC details to 100% confident. Some time to digest 
and a may reboot help.

I also mentioned routing without the route hash. No spinning in route
hash and no GC but we trade this for doing a fib lookup for every packet.
Jamal and I started to play with this at last OLS but it's only tested
in lab sofar.

For environments with small number of flows it's probably not a good idea
Length of flows is of course also important. No idea about the result in 
your case.

I have a patch:

ftp:/robur.slu.se:/pub/Linux/net-development/preroute/preroute10.pat

If you feel very brave you can test this. The files to monitor is 
/proc/net/softnet_stat this verify the new packet path.

With rtstat (tot) should show all routed packets. The rate in pps
we are able to achieve.

					--ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-20 21:53                                                                                                               ` Robert Olsson
@ 2005-01-21 21:20                                                                                                                 ` Jeremy M. Guthrie
  0 siblings, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-21 21:20 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

On Thursday 20 January 2005 03:53 pm, Robert Olsson wrote:
> Jeremy M. Guthrie writes:
>  > >  >    mc GC: tot ignored goal_miss ovrf HASH: in_search out_search
>  > >  > 524287     74830     777     0     0     0     0     0         6   
>  > >  >    0 0     498       0         0    0           40504          1
>  >
>  > When does GC normally ignore?
>
>  At a min ip_rt_gc_min_interval       = HZ / 2;
>
>  From what I understand your GC runs more often than that.
>  Do your understand? Test is in rt_garbage_collect()
I'll look in the function and let you know if I have any questions.

Thanks

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-21 15:23                                                                                                               ` Robert Olsson
@ 2005-01-21 21:24                                                                                                                 ` Jeremy M. Guthrie
  0 siblings, 0 replies; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-21 21:24 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

On Friday 21 January 2005 09:23 am, Robert Olsson wrote:
> We have to understand the GC details to 100% confident. Some time to digest
> and a may reboot help.
> I also mentioned routing without the route hash. No spinning in route
> hash and no GC but we trade this for doing a fib lookup for every packet.
> Jamal and I started to play with this at last OLS but it's only tested
> in lab sofar.
I'd like to hold off on that right now.  Things are running pretty good right 
now.  I'm trying to get my hands on a 'test' system to run side-by-side with 
the production system.  We go live Feb 1 with our app so I need to start 
toning down some of the down time.

> For environments with small number of flows it's probably not a good idea
> Length of flows is of course also important. No idea about the result in
> your case.
>
> I have a patch:
>
> ftp:/robur.slu.se:/pub/Linux/net-development/preroute/preroute10.pat
>
> If you feel very brave you can test this. The files to monitor is
> /proc/net/softnet_stat this verify the new packet path.
>
> With rtstat (tot) should show all routed packets. The rate in pps
> we are able to achieve.
I actually think I could have our other test system ready early next week.  
I'll let you know.  

Have a good weekend.

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-20 17:14                                                                                                             ` Jeremy M. Guthrie
  2005-01-20 21:53                                                                                                               ` Robert Olsson
  2005-01-21 15:23                                                                                                               ` Robert Olsson
@ 2005-01-31 15:37                                                                                                               ` Jeremy M. Guthrie
  2005-01-31 18:06                                                                                                                 ` Robert Olsson
  2 siblings, 1 reply; 88+ messages in thread
From: Jeremy M. Guthrie @ 2005-01-31 15:37 UTC (permalink / raw)
  To: netdev; +Cc: Robert Olsson

[-- Attachment #1: Type: text/plain, Size: 532 bytes --]

Just wanted to verify that by the term linear search from rtstat/lnstat, you 
mean a search of the buckets for a particular hash?

We're going live tomorrow with the app.  Afterwards I'll load up the no route 
caching patch and test that.

-- 

--------------------------------------------------
Jeremy M. Guthrie        jeremy.guthrie@berbee.com
Senior Network Engineer        Phone: 608-298-1061
Berbee                           Fax: 608-288-3007
5520 Research Park Drive         NOC: 608-298-1102
Madison, WI 53711

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 88+ messages in thread

* Re: V2.4 policy router operates faster/better than V2.6
  2005-01-31 15:37                                                                                                               ` Jeremy M. Guthrie
@ 2005-01-31 18:06                                                                                                                 ` Robert Olsson
  0 siblings, 0 replies; 88+ messages in thread
From: Robert Olsson @ 2005-01-31 18:06 UTC (permalink / raw)
  To: jeremy.guthrie; +Cc: netdev, Robert Olsson


Jeremy M. Guthrie writes:
 > Just wanted to verify that by the term linear search from rtstat/lnstat, you 
 > mean a search of the buckets for a particular hash?
 
 Yes.

 > We're going live tomorrow with the app.  Afterwards I'll load up the no route 
 > caching patch and test that.

Ok! Look at the patch just posted to davem. We should able view and control 
gc_min_interval better.
 
net.ipv4.route.gc_min_interval_ms = 300
net.ipv4.route.gc_min_interval = 0

						--ro

^ permalink raw reply	[flat|nested] 88+ messages in thread

end of thread, other threads:[~2005-01-31 18:06 UTC | newest]

Thread overview: 88+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-01-03 20:55 V2.4 policy router operates faster/better than V2.6 Jeremy M. Guthrie
2005-01-03 22:51 ` Stephen Hemminger
2005-01-03 22:56   ` Jeremy M. Guthrie
2005-01-05 13:18     ` Robert Olsson
2005-01-05 15:18       ` Jeremy M. Guthrie
2005-01-05 16:30         ` Robert Olsson
2005-01-05 17:35           ` Jeremy M. Guthrie
2005-01-05 19:25             ` Jeremy M. Guthrie
2005-01-05 20:22               ` Robert Olsson
2005-01-05 20:52                 ` Jeremy M. Guthrie
2005-01-06 15:26                 ` Jeremy M. Guthrie
2005-01-06 18:15                   ` Robert Olsson
2005-01-06 19:35                     ` Jeremy M. Guthrie
2005-01-06 20:29                       ` Robert Olsson
2005-01-06 20:54                         ` Jeremy M. Guthrie
2005-01-06 20:55                         ` Jeremy M. Guthrie
2005-01-06 21:19                         ` Jeremy M. Guthrie
2005-01-06 21:36                           ` Robert Olsson
2005-01-06 21:46                             ` Jeremy M. Guthrie
2005-01-06 22:11                               ` Robert Olsson
2005-01-06 22:18                                 ` Jeremy M. Guthrie
2005-01-06 22:35                                   ` Robert Olsson
2005-01-07 16:17                                     ` Jeremy M. Guthrie
2005-01-07 19:18                                       ` Robert Olsson
2005-01-07 19:38                                         ` Jeremy M. Guthrie
2005-01-07 20:07                                           ` Robert Olsson
2005-01-07 20:14                                           ` Jeremy M. Guthrie
2005-01-07 20:40                                             ` Robert Olsson
2005-01-07 21:06                                               ` Jeremy M. Guthrie
2005-01-07 21:30                                                 ` Robert Olsson
2005-01-11 15:11                                                   ` Jeremy M. Guthrie
2005-01-07 22:28                                             ` Jesse Brandeburg
2005-01-07 22:50                                               ` Jeremy M. Guthrie
2005-01-07 22:57                                                 ` Stephen Hemminger
2005-01-11 15:17                                                   ` Jeremy M. Guthrie
2005-01-11 16:40                                                     ` Robert Olsson
2005-01-12  1:27                                                       ` Jeremy M. Guthrie
2005-01-12 15:11                                                         ` Robert Olsson
2005-01-12 16:24                                                           ` Jeremy M. Guthrie
2005-01-12 19:27                                                             ` Robert Olsson
2005-01-12 20:11                                                               ` Jeremy M. Guthrie
2005-01-12 20:21                                                                 ` Robert Olsson
2005-01-12 20:30                                                                   ` Jeremy M. Guthrie
2005-01-12 20:45                                                                   ` Jeremy M. Guthrie
2005-01-12 22:02                                                                     ` Robert Olsson
2005-01-12 22:21                                                                       ` Jeremy M. Guthrie
     [not found]                                                                         ` <16869.42247.126428.508479@robur.slu.se>
2005-01-12 22:42                                                                           ` Jeremy M. Guthrie
2005-01-12 22:47                                                                           ` Jeremy M. Guthrie
2005-01-12 23:19                                                                             ` Robert Olsson
2005-01-12 23:23                                                                               ` Jeremy M. Guthrie
2005-01-13  8:56                                                                                 ` Robert Olsson
2005-01-13 19:28                                                                                   ` Jeremy M. Guthrie
2005-01-13 20:00                                                                                     ` David S. Miller
2005-01-13 20:43                                                                                       ` Jeremy M. Guthrie
2005-01-13 23:13                                                                                         ` David S. Miller
2005-01-13 21:12                                                                                     ` Robert Olsson
2005-01-13 22:27                                                                                       ` Jeremy M. Guthrie
2005-01-14 15:44                                                                                         ` Robert Olsson
2005-01-14 14:59                                                                                       ` Jeremy M. Guthrie
2005-01-14 16:05                                                                                         ` Robert Olsson
2005-01-14 19:00                                                                                           ` Jeremy M. Guthrie
2005-01-14 19:26                                                                                             ` Jeremy M. Guthrie
2005-01-16 12:32                                                                                               ` Robert Olsson
2005-01-16 16:22                                                                                                 ` Jeremy M. Guthrie
2005-01-19 15:03                                                                                                 ` Jeremy M. Guthrie
2005-01-19 22:18                                                                                                   ` Robert Olsson
2005-01-20  1:50                                                                                                     ` Jeremy M. Guthrie
2005-01-20 11:30                                                                                                       ` Robert Olsson
2005-01-20 14:37                                                                                                         ` Jeremy M. Guthrie
2005-01-20 17:01                                                                                                           ` Robert Olsson
2005-01-20 17:14                                                                                                             ` Jeremy M. Guthrie
2005-01-20 21:53                                                                                                               ` Robert Olsson
2005-01-21 21:20                                                                                                                 ` Jeremy M. Guthrie
2005-01-21 15:23                                                                                                               ` Robert Olsson
2005-01-21 21:24                                                                                                                 ` Jeremy M. Guthrie
2005-01-31 15:37                                                                                                               ` Jeremy M. Guthrie
2005-01-31 18:06                                                                                                                 ` Robert Olsson
2005-01-12 22:05                                                                     ` Jeremy M. Guthrie
2005-01-12 22:22                                                                       ` Robert Olsson
2005-01-12 22:30                                                                         ` Jeremy M. Guthrie
2005-01-11 17:17                                                     ` Jeremy M. Guthrie
2005-01-11 18:46                                                       ` Robert Olsson
2005-01-12  1:30                                                         ` Jeremy M. Guthrie
2005-01-12 16:02                                                           ` Robert Olsson
2005-01-04 15:07   ` Jeremy M. Guthrie
     [not found] <200501071619.54566.jeremy.guthrie@berbee.com>
2005-01-07 23:23 ` Jesse Brandeburg
2005-01-10 21:11   ` Jeremy M. Guthrie
     [not found] <C925F8B43D79CC49ACD0601FB68FF50C02D39006@orsmsx408>
2005-01-13 22:55 ` Jeremy M. Guthrie

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).