All of lore.kernel.org
 help / color / mirror / Atom feed
* messaging/IO/radosbench results
@ 2012-09-10 20:15 Mike Ryan
  2012-09-10 20:39 ` Mark Nelson
  0 siblings, 1 reply; 8+ messages in thread
From: Mike Ryan @ 2012-09-10 20:15 UTC (permalink / raw)
  To: ceph-devel

*Disclaimer*: these results are an investigation into potential
bottlenecks in RADOS. The test setup is wholly unrealistic, and these
numbers SHOULD NOT be used as an indication of the performance of OSDs,
messaging, RADOS, or ceph in general.


Executive summary: rados bench has some internal bottleneck. Once that's
cleared up, we're still having some issues saturating a single
connection to an OSD. Having 2-3 connection in parallel alleviates that
(either by having > 1 OSD or having multiple bencher clients).


I've run three separate tests: msbench, smalliobench, and rados bench.
In all cases I was trying to determine where bottleneck(s) exist. All
the tests were run on a machine with 192 GB of RAM. The backing stores
for all OSDs and journals are RAMdisks. The stores are running XFS.

smalliobench: I ran tests varying the number of OSDs and bencher
clients. In all cases, the number of PG's per OSD is 100.

OSD     Bencher     Throughput (mbyte/sec)
1       1           510
1       2           800
1       3           850
2       1           640
2       2           660
2       3           670
3       1           780
3       2           820
3       3           870
4       1           850
4       2           970
4       3           990

Note: these numbers are fairly fuzzy. I eyeballed them and they're only
really accurate to about 10 mbyte/sec. The small IO bencher was run with
100 ops in flight, 4 mbyte io's, 4 mbyte files.

msbench: ran tests trying to determine max throughput of raw messaging
layer. Varied the number of concurrently connected msbench clients and
measured aggregate throughput. Take-away: a messaging client can very
consistently push 400-500 mbytes/sec through a single socket.

Clients     Throughput (mbyte/sec)
1           520
2           880
3           1300
4           1900

Finally, rados bench, which seems to have its own bottleneck. Running
varying numbers of these, each client seems to get 250 mbyte/sec up till
the aggregate rate is around 1000 mbyte/sec (appx line speed as measured
by iperf). These were run on a pool with 100 PGs/OSD.

Clients     Throughput (mbyte/sec)
1           250
2           500
3           750
4           1000 (very fuzzy, probably 1000 +/- 75)
5           1000, seems to level out here

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: messaging/IO/radosbench results
  2012-09-10 20:15 messaging/IO/radosbench results Mike Ryan
@ 2012-09-10 20:39 ` Mark Nelson
  2012-09-12 20:08   ` Dieter Kasper
  0 siblings, 1 reply; 8+ messages in thread
From: Mark Nelson @ 2012-09-10 20:39 UTC (permalink / raw)
  To: Mike Ryan; +Cc: ceph-devel

On 09/10/2012 03:15 PM, Mike Ryan wrote:
> *Disclaimer*: these results are an investigation into potential
> bottlenecks in RADOS. The test setup is wholly unrealistic, and these
> numbers SHOULD NOT be used as an indication of the performance of OSDs,
> messaging, RADOS, or ceph in general.
>
>
> Executive summary: rados bench has some internal bottleneck. Once that's
> cleared up, we're still having some issues saturating a single
> connection to an OSD. Having 2-3 connection in parallel alleviates that
> (either by having>  1 OSD or having multiple bencher clients).
>
>
> I've run three separate tests: msbench, smalliobench, and rados bench.
> In all cases I was trying to determine where bottleneck(s) exist. All
> the tests were run on a machine with 192 GB of RAM. The backing stores
> for all OSDs and journals are RAMdisks. The stores are running XFS.
>
> smalliobench: I ran tests varying the number of OSDs and bencher
> clients. In all cases, the number of PG's per OSD is 100.
>
> OSD     Bencher     Throughput (mbyte/sec)
> 1       1           510
> 1       2           800
> 1       3           850
> 2       1           640
> 2       2           660
> 2       3           670
> 3       1           780
> 3       2           820
> 3       3           870
> 4       1           850
> 4       2           970
> 4       3           990
>
> Note: these numbers are fairly fuzzy. I eyeballed them and they're only
> really accurate to about 10 mbyte/sec. The small IO bencher was run with
> 100 ops in flight, 4 mbyte io's, 4 mbyte files.
>
> msbench: ran tests trying to determine max throughput of raw messaging
> layer. Varied the number of concurrently connected msbench clients and
> measured aggregate throughput. Take-away: a messaging client can very
> consistently push 400-500 mbytes/sec through a single socket.
>
> Clients     Throughput (mbyte/sec)
> 1           520
> 2           880
> 3           1300
> 4           1900
>
> Finally, rados bench, which seems to have its own bottleneck. Running
> varying numbers of these, each client seems to get 250 mbyte/sec up till
> the aggregate rate is around 1000 mbyte/sec (appx line speed as measured
> by iperf). These were run on a pool with 100 PGs/OSD.
>
> Clients     Throughput (mbyte/sec)
> 1           250
> 2           500
> 3           750
> 4           1000 (very fuzzy, probably 1000 +/- 75)
> 5           1000, seems to level out here
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Hi guys,

Some background on all of this:

We've been doing some performance testing at Inktank and noticed that 
performance with a single rados bench instance was plateauing at between 
600-700MB/s.  Running multiple concurrent rados bench instances improves 
performance, but only to a certain extent. The fastest throughput we've 
seen so far is around 1160MB/s with 8 rados bench instances, 12 spinning 
disks, and journals on SSDs.  This is true regardless of the underlying 
filesystem on the OSDs, though some hit the limits faster than others. 
Some of the raw data is available here:

https://docs.google.com/a/inktank.com/spreadsheet/ccc?key=0AnmmfpoQ1_94dDlmTHhvM19zd19tb05zbFVqZ2xSYXc#gid=0

To understand why we are plateauing, we wanted to investigate what 
bottlenecks were present in rados bench and if there also were any 
bottlenecks in the messenger code that might be limiting throughput. 
Soon we should have a 36 drive setup in our SC847a chassis where we can 
try to push things even further. :)

Mark

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: messaging/IO/radosbench results
  2012-09-10 20:39 ` Mark Nelson
@ 2012-09-12 20:08   ` Dieter Kasper
  2012-09-12 22:25     ` Mark Nelson
  0 siblings, 1 reply; 8+ messages in thread
From: Dieter Kasper @ 2012-09-12 20:08 UTC (permalink / raw)
  To: Mark Nelson; +Cc: Mike Ryan, ceph-devel@vger.kernel.org

On Mon, Sep 10, 2012 at 10:39:58PM +0200, Mark Nelson wrote:
> On 09/10/2012 03:15 PM, Mike Ryan wrote:
> > *Disclaimer*: these results are an investigation into potential
> > bottlenecks in RADOS. 
I appreciate this investigation very much !

> > The test setup is wholly unrealistic, and these
> > numbers SHOULD NOT be used as an indication of the performance of OSDs,
> > messaging, RADOS, or ceph in general.
> >
> >
> > Executive summary: rados bench has some internal bottleneck. Once that's
> > cleared up, we're still having some issues saturating a single
> > connection to an OSD. Having 2-3 connection in parallel alleviates that
> > (either by having>  1 OSD or having multiple bencher clients).
> >
> >
> > I've run three separate tests: msbench, smalliobench, and rados bench.
> > In all cases I was trying to determine where bottleneck(s) exist. All
> > the tests were run on a machine with 192 GB of RAM. The backing stores
> > for all OSDs and journals are RAMdisks. The stores are running XFS.
> >
> > smalliobench: I ran tests varying the number of OSDs and bencher
> > clients. In all cases, the number of PG's per OSD is 100.
> >
> > OSD     Bencher     Throughput (mbyte/sec)
> > 1       1           510
> > 1       2           800
> > 1       3           850
> > 2       1           640
> > 2       2           660
> > 2       3           670
> > 3       1           780
> > 3       2           820
> > 3       3           870
> > 4       1           850
> > 4       2           970
> > 4       3           990
> >
> > Note: these numbers are fairly fuzzy. I eyeballed them and they're only
> > really accurate to about 10 mbyte/sec. The small IO bencher was run with
> > 100 ops in flight, 4 mbyte io's, 4 mbyte files.
> >
> > msbench: ran tests trying to determine max throughput of raw messaging
> > layer. Varied the number of concurrently connected msbench clients and
> > measured aggregate throughput. Take-away: a messaging client can very
> > consistently push 400-500 mbytes/sec through a single socket.
> >
> > Clients     Throughput (mbyte/sec)
> > 1           520
> > 2           880
> > 3           1300
> > 4           1900
> >
> > Finally, rados bench, which seems to have its own bottleneck. Running
> > varying numbers of these, each client seems to get 250 mbyte/sec up till
> > the aggregate rate is around 1000 mbyte/sec (appx line speed as measured
> > by iperf). These were run on a pool with 100 PGs/OSD.
> >
> > Clients     Throughput (mbyte/sec)
> > 1           250
> > 2           500
> > 3           750
> > 4           1000 (very fuzzy, probably 1000 +/- 75)
> > 5           1000, seems to level out here
> > --
> > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> Hi guys,
> 
> Some background on all of this:
> 
> We've been doing some performance testing at Inktank and noticed that 
> performance with a single rados bench instance was plateauing at between 
> 600-700MB/s.  

4-nodes with 10GbE interconnect; journals in RAM-Disk; replica=2

# rados bench -p pbench 20 write
 Maintaining 16 concurrent writes of 4194304 bytes for at least 20 seconds.
   sec Cur ops   started  finished  avg MB/s  cur MB/s  last lat   avg lat
     0       0         0         0         0         0         -         0
     1      16       288       272   1087.81      1088  0.051123 0.0571643
     2      16       579       563   1125.85      1164  0.045729 0.0561784
     3      16       863       847   1129.19      1136  0.042012 0.0560869
     4      16      1150      1134   1133.87      1148   0.05466 0.0559281
     5      16      1441      1425   1139.87      1164  0.036852 0.0556809
     6      16      1733      1717   1144.54      1168  0.054594 0.0556124
     7      16      2007      1991   1137.59      1096   0.04454 0.0556698
     8      16      2290      2274   1136.88      1132  0.046777 0.0560103
     9      16      2580      2564   1139.44      1160  0.073328 0.0559353
    10      16      2871      2855   1141.88      1164  0.034091 0.0558576
    11      16      3158      3142   1142.43      1148  0.250688 0.0558404
    12      16      3445      3429   1142.88      1148  0.046941 0.0558071
    13      16      3726      3710   1141.42      1124  0.054092    0.0559
    14      16      4014      3998   1142.17      1152   0.03531 0.0558533
    15      16      4298      4282   1141.75      1136  0.040005 0.0559383
    16      16      4582      4566   1141.39      1136  0.048431 0.0559162
    17      16      4859      4843   1139.42      1108  0.045805 0.0559891
    18      16      5145      5129   1139.66      1144  0.046805 0.0560177
    19      16      5422      5406   1137.99      1108  0.037295 0.0561341
2012-09-08 14:36:32.460311min lat: 0.029503 max lat: 0.47757 avg lat: 0.0561424
   sec Cur ops   started  finished  avg MB/s  cur MB/s  last lat   avg lat
    20      16      5701      5685   1136.89      1116  0.041493 0.0561424
 Total time run:         20.197129
Total writes made:      5702
Write size:             4194304
Bandwidth (MB/sec):     1129.269

Stddev Bandwidth:       23.7487
Max bandwidth (MB/sec): 1168
Min bandwidth (MB/sec): 1088
Average Latency:        0.0564675
Stddev Latency:         0.0327582
Max latency:            0.47757
Min latency:            0.029503


Best Regards,
-Dieter

> Running multiple concurrent rados bench instances improves 
> performance, but only to a certain extent. The fastest throughput we've 
> seen so far is around 1160MB/s with 8 rados bench instances, 12 spinning 
> disks, and journals on SSDs.  This is true regardless of the underlying 
> filesystem on the OSDs, though some hit the limits faster than others. 
> Some of the raw data is available here:
> 
> https://docs.google.com/a/inktank.com/spreadsheet/ccc?key=0AnmmfpoQ1_94dDlmTHhvM19zd19tb05zbFVqZ2xSYXc#gid=0
> 
> To understand why we are plateauing, we wanted to investigate what 
> bottlenecks were present in rados bench and if there also were any 
> bottlenecks in the messenger code that might be limiting throughput. 
> Soon we should have a 36 drive setup in our SC847a chassis where we can 
> try to push things even further. :)
> 
> Mark
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html


^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: messaging/IO/radosbench results
  2012-09-12 20:08   ` Dieter Kasper
@ 2012-09-12 22:25     ` Mark Nelson
  2012-09-12 23:24       ` Joseph Glanville
  2012-09-13  7:24       ` Dieter Kasper
  0 siblings, 2 replies; 8+ messages in thread
From: Mark Nelson @ 2012-09-12 22:25 UTC (permalink / raw)
  To: Dieter Kasper; +Cc: Mike Ryan, ceph-devel@vger.kernel.org

On 09/12/2012 03:08 PM, Dieter Kasper wrote:
> On Mon, Sep 10, 2012 at 10:39:58PM +0200, Mark Nelson wrote:
>> On 09/10/2012 03:15 PM, Mike Ryan wrote:
>>> *Disclaimer*: these results are an investigation into potential
>>> bottlenecks in RADOS.
> I appreciate this investigation very much !
>
>>> The test setup is wholly unrealistic, and these
>>> numbers SHOULD NOT be used as an indication of the performance of OSDs,
>>> messaging, RADOS, or ceph in general.
>>>
>>>
>>> Executive summary: rados bench has some internal bottleneck. Once that's
>>> cleared up, we're still having some issues saturating a single
>>> connection to an OSD. Having 2-3 connection in parallel alleviates that
>>> (either by having>   1 OSD or having multiple bencher clients).
>>>
>>>
>>> I've run three separate tests: msbench, smalliobench, and rados bench.
>>> In all cases I was trying to determine where bottleneck(s) exist. All
>>> the tests were run on a machine with 192 GB of RAM. The backing stores
>>> for all OSDs and journals are RAMdisks. The stores are running XFS.
>>>
>>> smalliobench: I ran tests varying the number of OSDs and bencher
>>> clients. In all cases, the number of PG's per OSD is 100.
>>>
>>> OSD     Bencher     Throughput (mbyte/sec)
>>> 1       1           510
>>> 1       2           800
>>> 1       3           850
>>> 2       1           640
>>> 2       2           660
>>> 2       3           670
>>> 3       1           780
>>> 3       2           820
>>> 3       3           870
>>> 4       1           850
>>> 4       2           970
>>> 4       3           990
>>>
>>> Note: these numbers are fairly fuzzy. I eyeballed them and they're only
>>> really accurate to about 10 mbyte/sec. The small IO bencher was run with
>>> 100 ops in flight, 4 mbyte io's, 4 mbyte files.
>>>
>>> msbench: ran tests trying to determine max throughput of raw messaging
>>> layer. Varied the number of concurrently connected msbench clients and
>>> measured aggregate throughput. Take-away: a messaging client can very
>>> consistently push 400-500 mbytes/sec through a single socket.
>>>
>>> Clients     Throughput (mbyte/sec)
>>> 1           520
>>> 2           880
>>> 3           1300
>>> 4           1900
>>>
>>> Finally, rados bench, which seems to have its own bottleneck. Running
>>> varying numbers of these, each client seems to get 250 mbyte/sec up till
>>> the aggregate rate is around 1000 mbyte/sec (appx line speed as measured
>>> by iperf). These were run on a pool with 100 PGs/OSD.
>>>
>>> Clients     Throughput (mbyte/sec)
>>> 1           250
>>> 2           500
>>> 3           750
>>> 4           1000 (very fuzzy, probably 1000 +/- 75)
>>> 5           1000, seems to level out here
>>> --
>>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>>> the body of a message to majordomo@vger.kernel.org
>>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>> Hi guys,
>>
>> Some background on all of this:
>>
>> We've been doing some performance testing at Inktank and noticed that
>> performance with a single rados bench instance was plateauing at between
>> 600-700MB/s.
>
> 4-nodes with 10GbE interconnect; journals in RAM-Disk; replica=2
>
> # rados bench -p pbench 20 write
>   Maintaining 16 concurrent writes of 4194304 bytes for at least 20 seconds.
>     sec Cur ops   started  finished  avg MB/s  cur MB/s  last lat   avg lat
>       0       0         0         0         0         0         -         0
>       1      16       288       272   1087.81      1088  0.051123 0.0571643
>       2      16       579       563   1125.85      1164  0.045729 0.0561784
>       3      16       863       847   1129.19      1136  0.042012 0.0560869
>       4      16      1150      1134   1133.87      1148   0.05466 0.0559281
>       5      16      1441      1425   1139.87      1164  0.036852 0.0556809
>       6      16      1733      1717   1144.54      1168  0.054594 0.0556124
>       7      16      2007      1991   1137.59      1096   0.04454 0.0556698
>       8      16      2290      2274   1136.88      1132  0.046777 0.0560103
>       9      16      2580      2564   1139.44      1160  0.073328 0.0559353
>      10      16      2871      2855   1141.88      1164  0.034091 0.0558576
>      11      16      3158      3142   1142.43      1148  0.250688 0.0558404
>      12      16      3445      3429   1142.88      1148  0.046941 0.0558071
>      13      16      3726      3710   1141.42      1124  0.054092    0.0559
>      14      16      4014      3998   1142.17      1152   0.03531 0.0558533
>      15      16      4298      4282   1141.75      1136  0.040005 0.0559383
>      16      16      4582      4566   1141.39      1136  0.048431 0.0559162
>      17      16      4859      4843   1139.42      1108  0.045805 0.0559891
>      18      16      5145      5129   1139.66      1144  0.046805 0.0560177
>      19      16      5422      5406   1137.99      1108  0.037295 0.0561341
> 2012-09-08 14:36:32.460311min lat: 0.029503 max lat: 0.47757 avg lat: 0.0561424
>     sec Cur ops   started  finished  avg MB/s  cur MB/s  last lat   avg lat
>      20      16      5701      5685   1136.89      1116  0.041493 0.0561424
>   Total time run:         20.197129
> Total writes made:      5702
> Write size:             4194304
> Bandwidth (MB/sec):     1129.269
>
> Stddev Bandwidth:       23.7487
> Max bandwidth (MB/sec): 1168
> Min bandwidth (MB/sec): 1088
> Average Latency:        0.0564675
> Stddev Latency:         0.0327582
> Max latency:            0.47757
> Min latency:            0.029503
>
>
> Best Regards,
> -Dieter
>

Well look at that! :)  Now I've gotta figure out what the difference is. 
  How fast are the CPUs in your rados bench machine there?

Also, I should mention that at these speeds, we noticed that crc32c 
calculations were actually having a pretty big effect.  Turning them off 
gave us a 10% performance boost.  We're looking at faster 
implementations now.

Mark



^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: messaging/IO/radosbench results
  2012-09-12 22:25     ` Mark Nelson
@ 2012-09-12 23:24       ` Joseph Glanville
  2012-09-13  0:39         ` Mark Nelson
  2012-09-13  7:24       ` Dieter Kasper
  1 sibling, 1 reply; 8+ messages in thread
From: Joseph Glanville @ 2012-09-12 23:24 UTC (permalink / raw)
  To: Mark Nelson; +Cc: Dieter Kasper, Mike Ryan, ceph-devel@vger.kernel.org

On 13 September 2012 08:25, Mark Nelson <mark.nelson@inktank.com> wrote:
> On 09/12/2012 03:08 PM, Dieter Kasper wrote:
>>
>> On Mon, Sep 10, 2012 at 10:39:58PM +0200, Mark Nelson wrote:
>>>
>>> On 09/10/2012 03:15 PM, Mike Ryan wrote:
>>>>
>>>> *Disclaimer*: these results are an investigation into potential
>>>> bottlenecks in RADOS.
>>
>> I appreciate this investigation very much !
>>
>>>> The test setup is wholly unrealistic, and these
>>>> numbers SHOULD NOT be used as an indication of the performance of OSDs,
>>>> messaging, RADOS, or ceph in general.
>>>>
>>>>
>>>> Executive summary: rados bench has some internal bottleneck. Once that's
>>>> cleared up, we're still having some issues saturating a single
>>>> connection to an OSD. Having 2-3 connection in parallel alleviates that
>>>> (either by having>   1 OSD or having multiple bencher clients).
>>>>
>>>>
>>>> I've run three separate tests: msbench, smalliobench, and rados bench.
>>>> In all cases I was trying to determine where bottleneck(s) exist. All
>>>> the tests were run on a machine with 192 GB of RAM. The backing stores
>>>> for all OSDs and journals are RAMdisks. The stores are running XFS.
>>>>
>>>> smalliobench: I ran tests varying the number of OSDs and bencher
>>>> clients. In all cases, the number of PG's per OSD is 100.
>>>>
>>>> OSD     Bencher     Throughput (mbyte/sec)
>>>> 1       1           510
>>>> 1       2           800
>>>> 1       3           850
>>>> 2       1           640
>>>> 2       2           660
>>>> 2       3           670
>>>> 3       1           780
>>>> 3       2           820
>>>> 3       3           870
>>>> 4       1           850
>>>> 4       2           970
>>>> 4       3           990
>>>>
>>>> Note: these numbers are fairly fuzzy. I eyeballed them and they're only
>>>> really accurate to about 10 mbyte/sec. The small IO bencher was run with
>>>> 100 ops in flight, 4 mbyte io's, 4 mbyte files.
>>>>
>>>> msbench: ran tests trying to determine max throughput of raw messaging
>>>> layer. Varied the number of concurrently connected msbench clients and
>>>> measured aggregate throughput. Take-away: a messaging client can very
>>>> consistently push 400-500 mbytes/sec through a single socket.
>>>>
>>>> Clients     Throughput (mbyte/sec)
>>>> 1           520
>>>> 2           880
>>>> 3           1300
>>>> 4           1900
>>>>
>>>> Finally, rados bench, which seems to have its own bottleneck. Running
>>>> varying numbers of these, each client seems to get 250 mbyte/sec up till
>>>> the aggregate rate is around 1000 mbyte/sec (appx line speed as measured
>>>> by iperf). These were run on a pool with 100 PGs/OSD.
>>>>
>>>> Clients     Throughput (mbyte/sec)
>>>> 1           250
>>>> 2           500
>>>> 3           750
>>>> 4           1000 (very fuzzy, probably 1000 +/- 75)
>>>> 5           1000, seems to level out here
>>>> --
>>>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>>>> the body of a message to majordomo@vger.kernel.org
>>>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>>
>>>
>>> Hi guys,
>>>
>>> Some background on all of this:
>>>
>>> We've been doing some performance testing at Inktank and noticed that
>>> performance with a single rados bench instance was plateauing at between
>>> 600-700MB/s.
>>
>>
>> 4-nodes with 10GbE interconnect; journals in RAM-Disk; replica=2
>>
>> # rados bench -p pbench 20 write
>>   Maintaining 16 concurrent writes of 4194304 bytes for at least 20
>> seconds.
>>     sec Cur ops   started  finished  avg MB/s  cur MB/s  last lat   avg
>> lat
>>       0       0         0         0         0         0         -
>> 0
>>       1      16       288       272   1087.81      1088  0.051123
>> 0.0571643
>>       2      16       579       563   1125.85      1164  0.045729
>> 0.0561784
>>       3      16       863       847   1129.19      1136  0.042012
>> 0.0560869
>>       4      16      1150      1134   1133.87      1148   0.05466
>> 0.0559281
>>       5      16      1441      1425   1139.87      1164  0.036852
>> 0.0556809
>>       6      16      1733      1717   1144.54      1168  0.054594
>> 0.0556124
>>       7      16      2007      1991   1137.59      1096   0.04454
>> 0.0556698
>>       8      16      2290      2274   1136.88      1132  0.046777
>> 0.0560103
>>       9      16      2580      2564   1139.44      1160  0.073328
>> 0.0559353
>>      10      16      2871      2855   1141.88      1164  0.034091
>> 0.0558576
>>      11      16      3158      3142   1142.43      1148  0.250688
>> 0.0558404
>>      12      16      3445      3429   1142.88      1148  0.046941
>> 0.0558071
>>      13      16      3726      3710   1141.42      1124  0.054092
>> 0.0559
>>      14      16      4014      3998   1142.17      1152   0.03531
>> 0.0558533
>>      15      16      4298      4282   1141.75      1136  0.040005
>> 0.0559383
>>      16      16      4582      4566   1141.39      1136  0.048431
>> 0.0559162
>>      17      16      4859      4843   1139.42      1108  0.045805
>> 0.0559891
>>      18      16      5145      5129   1139.66      1144  0.046805
>> 0.0560177
>>      19      16      5422      5406   1137.99      1108  0.037295
>> 0.0561341
>> 2012-09-08 14:36:32.460311min lat: 0.029503 max lat: 0.47757 avg lat:
>> 0.0561424
>>     sec Cur ops   started  finished  avg MB/s  cur MB/s  last lat   avg
>> lat
>>      20      16      5701      5685   1136.89      1116  0.041493
>> 0.0561424
>>   Total time run:         20.197129
>> Total writes made:      5702
>> Write size:             4194304
>> Bandwidth (MB/sec):     1129.269
>>
>> Stddev Bandwidth:       23.7487
>> Max bandwidth (MB/sec): 1168
>> Min bandwidth (MB/sec): 1088
>> Average Latency:        0.0564675
>> Stddev Latency:         0.0327582
>> Max latency:            0.47757
>> Min latency:            0.029503
>>
>>
>> Best Regards,
>> -Dieter
>>
>
> Well look at that! :)  Now I've gotta figure out what the difference is.
> How fast are the CPUs in your rados bench machine there?
>
> Also, I should mention that at these speeds, we noticed that crc32c
> calculations were actually having a pretty big effect.  Turning them off
> gave us a 10% performance boost.  We're looking at faster implementations
> now.
>
>
> Mark
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Hi Mark

If using primarily Intel machines that are Nahalem or better (I would
imagine most boxes running Ceph would fit this category) then consider
using the Intel CRC32 instructions.
Most of the work to use them is laid out here:
http://www.drdobbs.com/parallel/fast-parallelized-crc-computation-using/229401411

-- 
CTO | Orion Virtualisation Solutions | www.orionvm.com.au
Phone: 1300 56 99 52 | Mobile: 0428 754 846

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: messaging/IO/radosbench results
  2012-09-12 23:24       ` Joseph Glanville
@ 2012-09-13  0:39         ` Mark Nelson
  0 siblings, 0 replies; 8+ messages in thread
From: Mark Nelson @ 2012-09-13  0:39 UTC (permalink / raw)
  To: Joseph Glanville; +Cc: Dieter Kasper, Mike Ryan, ceph-devel@vger.kernel.org

On 09/12/2012 06:24 PM, Joseph Glanville wrote:
> On 13 September 2012 08:25, Mark Nelson<mark.nelson@inktank.com>  wrote:
>> On 09/12/2012 03:08 PM, Dieter Kasper wrote:
>>>
>>> On Mon, Sep 10, 2012 at 10:39:58PM +0200, Mark Nelson wrote:
>>>>
>>>> On 09/10/2012 03:15 PM, Mike Ryan wrote:
>>>>>
>>>>> *Disclaimer*: these results are an investigation into potential
>>>>> bottlenecks in RADOS.
>>>
>>> I appreciate this investigation very much !
>>>
>>>>> The test setup is wholly unrealistic, and these
>>>>> numbers SHOULD NOT be used as an indication of the performance of OSDs,
>>>>> messaging, RADOS, or ceph in general.
>>>>>
>>>>>
>>>>> Executive summary: rados bench has some internal bottleneck. Once that's
>>>>> cleared up, we're still having some issues saturating a single
>>>>> connection to an OSD. Having 2-3 connection in parallel alleviates that
>>>>> (either by having>    1 OSD or having multiple bencher clients).
>>>>>
>>>>>
>>>>> I've run three separate tests: msbench, smalliobench, and rados bench.
>>>>> In all cases I was trying to determine where bottleneck(s) exist. All
>>>>> the tests were run on a machine with 192 GB of RAM. The backing stores
>>>>> for all OSDs and journals are RAMdisks. The stores are running XFS.
>>>>>
>>>>> smalliobench: I ran tests varying the number of OSDs and bencher
>>>>> clients. In all cases, the number of PG's per OSD is 100.
>>>>>
>>>>> OSD     Bencher     Throughput (mbyte/sec)
>>>>> 1       1           510
>>>>> 1       2           800
>>>>> 1       3           850
>>>>> 2       1           640
>>>>> 2       2           660
>>>>> 2       3           670
>>>>> 3       1           780
>>>>> 3       2           820
>>>>> 3       3           870
>>>>> 4       1           850
>>>>> 4       2           970
>>>>> 4       3           990
>>>>>
>>>>> Note: these numbers are fairly fuzzy. I eyeballed them and they're only
>>>>> really accurate to about 10 mbyte/sec. The small IO bencher was run with
>>>>> 100 ops in flight, 4 mbyte io's, 4 mbyte files.
>>>>>
>>>>> msbench: ran tests trying to determine max throughput of raw messaging
>>>>> layer. Varied the number of concurrently connected msbench clients and
>>>>> measured aggregate throughput. Take-away: a messaging client can very
>>>>> consistently push 400-500 mbytes/sec through a single socket.
>>>>>
>>>>> Clients     Throughput (mbyte/sec)
>>>>> 1           520
>>>>> 2           880
>>>>> 3           1300
>>>>> 4           1900
>>>>>
>>>>> Finally, rados bench, which seems to have its own bottleneck. Running
>>>>> varying numbers of these, each client seems to get 250 mbyte/sec up till
>>>>> the aggregate rate is around 1000 mbyte/sec (appx line speed as measured
>>>>> by iperf). These were run on a pool with 100 PGs/OSD.
>>>>>
>>>>> Clients     Throughput (mbyte/sec)
>>>>> 1           250
>>>>> 2           500
>>>>> 3           750
>>>>> 4           1000 (very fuzzy, probably 1000 +/- 75)
>>>>> 5           1000, seems to level out here
>>>>> --
>>>>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>>>>> the body of a message to majordomo@vger.kernel.org
>>>>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>>>
>>>>
>>>> Hi guys,
>>>>
>>>> Some background on all of this:
>>>>
>>>> We've been doing some performance testing at Inktank and noticed that
>>>> performance with a single rados bench instance was plateauing at between
>>>> 600-700MB/s.
>>>
>>>
>>> 4-nodes with 10GbE interconnect; journals in RAM-Disk; replica=2
>>>
>>> # rados bench -p pbench 20 write
>>>    Maintaining 16 concurrent writes of 4194304 bytes for at least 20
>>> seconds.
>>>      sec Cur ops   started  finished  avg MB/s  cur MB/s  last lat   avg
>>> lat
>>>        0       0         0         0         0         0         -
>>> 0
>>>        1      16       288       272   1087.81      1088  0.051123
>>> 0.0571643
>>>        2      16       579       563   1125.85      1164  0.045729
>>> 0.0561784
>>>        3      16       863       847   1129.19      1136  0.042012
>>> 0.0560869
>>>        4      16      1150      1134   1133.87      1148   0.05466
>>> 0.0559281
>>>        5      16      1441      1425   1139.87      1164  0.036852
>>> 0.0556809
>>>        6      16      1733      1717   1144.54      1168  0.054594
>>> 0.0556124
>>>        7      16      2007      1991   1137.59      1096   0.04454
>>> 0.0556698
>>>        8      16      2290      2274   1136.88      1132  0.046777
>>> 0.0560103
>>>        9      16      2580      2564   1139.44      1160  0.073328
>>> 0.0559353
>>>       10      16      2871      2855   1141.88      1164  0.034091
>>> 0.0558576
>>>       11      16      3158      3142   1142.43      1148  0.250688
>>> 0.0558404
>>>       12      16      3445      3429   1142.88      1148  0.046941
>>> 0.0558071
>>>       13      16      3726      3710   1141.42      1124  0.054092
>>> 0.0559
>>>       14      16      4014      3998   1142.17      1152   0.03531
>>> 0.0558533
>>>       15      16      4298      4282   1141.75      1136  0.040005
>>> 0.0559383
>>>       16      16      4582      4566   1141.39      1136  0.048431
>>> 0.0559162
>>>       17      16      4859      4843   1139.42      1108  0.045805
>>> 0.0559891
>>>       18      16      5145      5129   1139.66      1144  0.046805
>>> 0.0560177
>>>       19      16      5422      5406   1137.99      1108  0.037295
>>> 0.0561341
>>> 2012-09-08 14:36:32.460311min lat: 0.029503 max lat: 0.47757 avg lat:
>>> 0.0561424
>>>      sec Cur ops   started  finished  avg MB/s  cur MB/s  last lat   avg
>>> lat
>>>       20      16      5701      5685   1136.89      1116  0.041493
>>> 0.0561424
>>>    Total time run:         20.197129
>>> Total writes made:      5702
>>> Write size:             4194304
>>> Bandwidth (MB/sec):     1129.269
>>>
>>> Stddev Bandwidth:       23.7487
>>> Max bandwidth (MB/sec): 1168
>>> Min bandwidth (MB/sec): 1088
>>> Average Latency:        0.0564675
>>> Stddev Latency:         0.0327582
>>> Max latency:            0.47757
>>> Min latency:            0.029503
>>>
>>>
>>> Best Regards,
>>> -Dieter
>>>
>>
>> Well look at that! :)  Now I've gotta figure out what the difference is.
>> How fast are the CPUs in your rados bench machine there?
>>
>> Also, I should mention that at these speeds, we noticed that crc32c
>> calculations were actually having a pretty big effect.  Turning them off
>> gave us a 10% performance boost.  We're looking at faster implementations
>> now.
>>
>>
>> Mark
>>
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
> Hi Mark
>
> If using primarily Intel machines that are Nahalem or better (I would
> imagine most boxes running Ceph would fit this category) then consider
> using the Intel CRC32 instructions.
> Most of the work to use them is laid out here:
> http://www.drdobbs.com/parallel/fast-parallelized-crc-computation-using/229401411
>

Hi Dieter,

Yes, I've been looking at for Nehalem.  We actually have a number of 
machines using last gen AMD processors so we'll need to consider options 
for that as well.  Earlier today I was reading through the whitepaper here:

http://code.google.com/p/crcutil/

Mark

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: messaging/IO/radosbench results
  2012-09-12 22:25     ` Mark Nelson
  2012-09-12 23:24       ` Joseph Glanville
@ 2012-09-13  7:24       ` Dieter Kasper
  2012-09-13 11:08         ` Mark Nelson
  1 sibling, 1 reply; 8+ messages in thread
From: Dieter Kasper @ 2012-09-13  7:24 UTC (permalink / raw)
  To: Mark Nelson; +Cc: Mike Ryan, ceph-devel@vger.kernel.org

On Thu, Sep 13, 2012 at 12:25:36AM +0200, Mark Nelson wrote:
> On 09/12/2012 03:08 PM, Dieter Kasper wrote:
> > On Mon, Sep 10, 2012 at 10:39:58PM +0200, Mark Nelson wrote:
> >> On 09/10/2012 03:15 PM, Mike Ryan wrote:
> >>> *Disclaimer*: these results are an investigation into potential
> >>> bottlenecks in RADOS.
> > I appreciate this investigation very much !
> >
> >>> The test setup is wholly unrealistic, and these
> >>> numbers SHOULD NOT be used as an indication of the performance of OSDs,
> >>> messaging, RADOS, or ceph in general.
> >>>
> >>>
> >>> Executive summary: rados bench has some internal bottleneck. Once that's
> >>> cleared up, we're still having some issues saturating a single
> >>> connection to an OSD. Having 2-3 connection in parallel alleviates that
> >>> (either by having>   1 OSD or having multiple bencher clients).
> >>>
> >>>
> >>> I've run three separate tests: msbench, smalliobench, and rados bench.
> >>> In all cases I was trying to determine where bottleneck(s) exist. All
> >>> the tests were run on a machine with 192 GB of RAM. The backing stores
> >>> for all OSDs and journals are RAMdisks. The stores are running XFS.
> >>>
> >>> smalliobench: I ran tests varying the number of OSDs and bencher
> >>> clients. In all cases, the number of PG's per OSD is 100.
> >>>
> >>> OSD     Bencher     Throughput (mbyte/sec)
> >>> 1       1           510
> >>> 1       2           800
> >>> 1       3           850
> >>> 2       1           640
> >>> 2       2           660
> >>> 2       3           670
> >>> 3       1           780
> >>> 3       2           820
> >>> 3       3           870
> >>> 4       1           850
> >>> 4       2           970
> >>> 4       3           990
> >>>
> >>> Note: these numbers are fairly fuzzy. I eyeballed them and they're only
> >>> really accurate to about 10 mbyte/sec. The small IO bencher was run with
> >>> 100 ops in flight, 4 mbyte io's, 4 mbyte files.
> >>>
> >>> msbench: ran tests trying to determine max throughput of raw messaging
> >>> layer. Varied the number of concurrently connected msbench clients and
> >>> measured aggregate throughput. Take-away: a messaging client can very
> >>> consistently push 400-500 mbytes/sec through a single socket.
> >>>
> >>> Clients     Throughput (mbyte/sec)
> >>> 1           520
> >>> 2           880
> >>> 3           1300
> >>> 4           1900
> >>>
> >>> Finally, rados bench, which seems to have its own bottleneck. Running
> >>> varying numbers of these, each client seems to get 250 mbyte/sec up till
> >>> the aggregate rate is around 1000 mbyte/sec (appx line speed as measured
> >>> by iperf). These were run on a pool with 100 PGs/OSD.
> >>>
> >>> Clients     Throughput (mbyte/sec)
> >>> 1           250
> >>> 2           500
> >>> 3           750
> >>> 4           1000 (very fuzzy, probably 1000 +/- 75)
> >>> 5           1000, seems to level out here
> >>> --
> >>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> >>> the body of a message to majordomo@vger.kernel.org
> >>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >>
> >> Hi guys,
> >>
> >> Some background on all of this:
> >>
> >> We've been doing some performance testing at Inktank and noticed that
> >> performance with a single rados bench instance was plateauing at between
> >> 600-700MB/s.
> >
> > 4-nodes with 10GbE interconnect; journals in RAM-Disk; replica=2
> >
> > # rados bench -p pbench 20 write
> >   Maintaining 16 concurrent writes of 4194304 bytes for at least 20 seconds.
> >     sec Cur ops   started  finished  avg MB/s  cur MB/s  last lat   avg lat
> >       0       0         0         0         0         0         -         0
> >       1      16       288       272   1087.81      1088  0.051123 0.0571643
> >       2      16       579       563   1125.85      1164  0.045729 0.0561784
> >       3      16       863       847   1129.19      1136  0.042012 0.0560869
> >       4      16      1150      1134   1133.87      1148   0.05466 0.0559281
> >       5      16      1441      1425   1139.87      1164  0.036852 0.0556809
> >       6      16      1733      1717   1144.54      1168  0.054594 0.0556124
> >       7      16      2007      1991   1137.59      1096   0.04454 0.0556698
> >       8      16      2290      2274   1136.88      1132  0.046777 0.0560103
> >       9      16      2580      2564   1139.44      1160  0.073328 0.0559353
> >      10      16      2871      2855   1141.88      1164  0.034091 0.0558576
> >      11      16      3158      3142   1142.43      1148  0.250688 0.0558404
> >      12      16      3445      3429   1142.88      1148  0.046941 0.0558071
> >      13      16      3726      3710   1141.42      1124  0.054092    0.0559
> >      14      16      4014      3998   1142.17      1152   0.03531 0.0558533
> >      15      16      4298      4282   1141.75      1136  0.040005 0.0559383
> >      16      16      4582      4566   1141.39      1136  0.048431 0.0559162
> >      17      16      4859      4843   1139.42      1108  0.045805 0.0559891
> >      18      16      5145      5129   1139.66      1144  0.046805 0.0560177
> >      19      16      5422      5406   1137.99      1108  0.037295 0.0561341
> > 2012-09-08 14:36:32.460311min lat: 0.029503 max lat: 0.47757 avg lat: 0.0561424
> >     sec Cur ops   started  finished  avg MB/s  cur MB/s  last lat   avg lat
> >      20      16      5701      5685   1136.89      1116  0.041493 0.0561424
> >   Total time run:         20.197129
> > Total writes made:      5702
> > Write size:             4194304
> > Bandwidth (MB/sec):     1129.269
> >
> > Stddev Bandwidth:       23.7487
> > Max bandwidth (MB/sec): 1168
> > Min bandwidth (MB/sec): 1088
> > Average Latency:        0.0564675
> > Stddev Latency:         0.0327582
> > Max latency:            0.47757
> > Min latency:            0.029503
> >
> >
> > Best Regards,
> > -Dieter
> >
> 
> Well look at that! :)  Now I've gotta figure out what the difference is. 
>   How fast are the CPUs in your rados bench machine there?

One CPU socket in each node:
model name      : Intel(R) Xeon(R) CPU E5-2630 0 @ 2.30GHz
Logial CPUs: 12
MemTotal:       32856332 kB

> 
> Also, I should mention that at these speeds, we noticed that crc32c 
> calculations were actually having a pretty big effect.  

perf report

Events: 39K cycles
+     26.29%         ceph-osd  ceph-osd                    [.] 0x45e60b                                     
+      4.74%         ceph-osd  [kernel.kallsyms]           [k] copy_user_generic_string                    
+      3.37%         ceph-mon  ceph-mon                    [.] MHeartbeat::decode_payload()               
+      2.88%         ceph-osd  [kernel.kallsyms]           [k] futex_wake                                
+      2.61%          swapper  [kernel.kallsyms]           [k] intel_idle                               
+      2.34%         ceph-osd  [kernel.kallsyms]           [k] __memcpy                                
+      1.71%         ceph-osd  libc-2.11.3.so              [.] memcpy                                 
+      1.70%         ceph-osd  [kernel.kallsyms]           [k] __copy_user_nocache                   
+      1.66%         ceph-osd  [kernel.kallsyms]           [k] futex_requeue                        
+      1.33%         ceph-mon  ceph-mon                    [.] MOSDOpReply::~MOSDOpReply()         
+      1.18%         ceph-mon  libc-2.11.3.so              [.] memcpy                             
+      1.16%         ceph-mon  ceph-mon                    [.] MOSDPGInfo::decode_payload()      
+      0.97%         ceph-osd  [kernel.kallsyms]           [k] futex_wake_op                    
+      0.86%         ceph-mon  ceph-mon                    [.] MExportDirDiscoverAck::print(std::ostream&) const 
+      0.79%         ceph-osd  [kernel.kallsyms]           [k] _raw_spin_lock                                   
+      0.74%         ceph-mon  ceph-mon                    [.] MOSDPing::decode_payload()                      
+      0.52%         ceph-osd  libtcmalloc.so.0.3.0        [.] operator new(unsigned long)                    
+      0.51%         ceph-mon  ceph-mon                    [.] MDiscover::print(std::ostream&) const         
+      0.48%         ceph-osd  [xfs]                       [k] xfs_bmap_add_extent                          
+      0.43%         ceph-mon  [kernel.kallsyms]           [k] copy_user_generic_string                    
+      0.39%         ceph-osd  [kernel.kallsyms]           [k] iov_iter_fault_in_readable                 

Regards,
-Dieter


> Turning them off 
> gave us a 10% performance boost.  We're looking at faster 
> implementations now.
> 
> Mark
> 
> 
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html


^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: messaging/IO/radosbench results
  2012-09-13  7:24       ` Dieter Kasper
@ 2012-09-13 11:08         ` Mark Nelson
  0 siblings, 0 replies; 8+ messages in thread
From: Mark Nelson @ 2012-09-13 11:08 UTC (permalink / raw)
  To: Dieter Kasper; +Cc: Mike Ryan, ceph-devel@vger.kernel.org

On 09/13/2012 02:24 AM, Dieter Kasper wrote:
> On Thu, Sep 13, 2012 at 12:25:36AM +0200, Mark Nelson wrote:
>> On 09/12/2012 03:08 PM, Dieter Kasper wrote:
>>> On Mon, Sep 10, 2012 at 10:39:58PM +0200, Mark Nelson wrote:
>>>> On 09/10/2012 03:15 PM, Mike Ryan wrote:
>>>>> *Disclaimer*: these results are an investigation into potential
>>>>> bottlenecks in RADOS.
>>> I appreciate this investigation very much !
>>>
>>>>> The test setup is wholly unrealistic, and these
>>>>> numbers SHOULD NOT be used as an indication of the performance of OSDs,
>>>>> messaging, RADOS, or ceph in general.
>>>>>
>>>>>
>>>>> Executive summary: rados bench has some internal bottleneck. Once that's
>>>>> cleared up, we're still having some issues saturating a single
>>>>> connection to an OSD. Having 2-3 connection in parallel alleviates that
>>>>> (either by having>    1 OSD or having multiple bencher clients).
>>>>>
>>>>>
>>>>> I've run three separate tests: msbench, smalliobench, and rados bench.
>>>>> In all cases I was trying to determine where bottleneck(s) exist. All
>>>>> the tests were run on a machine with 192 GB of RAM. The backing stores
>>>>> for all OSDs and journals are RAMdisks. The stores are running XFS.
>>>>>
>>>>> smalliobench: I ran tests varying the number of OSDs and bencher
>>>>> clients. In all cases, the number of PG's per OSD is 100.
>>>>>
>>>>> OSD     Bencher     Throughput (mbyte/sec)
>>>>> 1       1           510
>>>>> 1       2           800
>>>>> 1       3           850
>>>>> 2       1           640
>>>>> 2       2           660
>>>>> 2       3           670
>>>>> 3       1           780
>>>>> 3       2           820
>>>>> 3       3           870
>>>>> 4       1           850
>>>>> 4       2           970
>>>>> 4       3           990
>>>>>
>>>>> Note: these numbers are fairly fuzzy. I eyeballed them and they're only
>>>>> really accurate to about 10 mbyte/sec. The small IO bencher was run with
>>>>> 100 ops in flight, 4 mbyte io's, 4 mbyte files.
>>>>>
>>>>> msbench: ran tests trying to determine max throughput of raw messaging
>>>>> layer. Varied the number of concurrently connected msbench clients and
>>>>> measured aggregate throughput. Take-away: a messaging client can very
>>>>> consistently push 400-500 mbytes/sec through a single socket.
>>>>>
>>>>> Clients     Throughput (mbyte/sec)
>>>>> 1           520
>>>>> 2           880
>>>>> 3           1300
>>>>> 4           1900
>>>>>
>>>>> Finally, rados bench, which seems to have its own bottleneck. Running
>>>>> varying numbers of these, each client seems to get 250 mbyte/sec up till
>>>>> the aggregate rate is around 1000 mbyte/sec (appx line speed as measured
>>>>> by iperf). These were run on a pool with 100 PGs/OSD.
>>>>>
>>>>> Clients     Throughput (mbyte/sec)
>>>>> 1           250
>>>>> 2           500
>>>>> 3           750
>>>>> 4           1000 (very fuzzy, probably 1000 +/- 75)
>>>>> 5           1000, seems to level out here
>>>>> --
>>>>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>>>>> the body of a message to majordomo@vger.kernel.org
>>>>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>>>
>>>> Hi guys,
>>>>
>>>> Some background on all of this:
>>>>
>>>> We've been doing some performance testing at Inktank and noticed that
>>>> performance with a single rados bench instance was plateauing at between
>>>> 600-700MB/s.
>>>
>>> 4-nodes with 10GbE interconnect; journals in RAM-Disk; replica=2
>>>
>>> # rados bench -p pbench 20 write
>>>    Maintaining 16 concurrent writes of 4194304 bytes for at least 20 seconds.
>>>      sec Cur ops   started  finished  avg MB/s  cur MB/s  last lat   avg lat
>>>        0       0         0         0         0         0         -         0
>>>        1      16       288       272   1087.81      1088  0.051123 0.0571643
>>>        2      16       579       563   1125.85      1164  0.045729 0.0561784
>>>        3      16       863       847   1129.19      1136  0.042012 0.0560869
>>>        4      16      1150      1134   1133.87      1148   0.05466 0.0559281
>>>        5      16      1441      1425   1139.87      1164  0.036852 0.0556809
>>>        6      16      1733      1717   1144.54      1168  0.054594 0.0556124
>>>        7      16      2007      1991   1137.59      1096   0.04454 0.0556698
>>>        8      16      2290      2274   1136.88      1132  0.046777 0.0560103
>>>        9      16      2580      2564   1139.44      1160  0.073328 0.0559353
>>>       10      16      2871      2855   1141.88      1164  0.034091 0.0558576
>>>       11      16      3158      3142   1142.43      1148  0.250688 0.0558404
>>>       12      16      3445      3429   1142.88      1148  0.046941 0.0558071
>>>       13      16      3726      3710   1141.42      1124  0.054092    0.0559
>>>       14      16      4014      3998   1142.17      1152   0.03531 0.0558533
>>>       15      16      4298      4282   1141.75      1136  0.040005 0.0559383
>>>       16      16      4582      4566   1141.39      1136  0.048431 0.0559162
>>>       17      16      4859      4843   1139.42      1108  0.045805 0.0559891
>>>       18      16      5145      5129   1139.66      1144  0.046805 0.0560177
>>>       19      16      5422      5406   1137.99      1108  0.037295 0.0561341
>>> 2012-09-08 14:36:32.460311min lat: 0.029503 max lat: 0.47757 avg lat: 0.0561424
>>>      sec Cur ops   started  finished  avg MB/s  cur MB/s  last lat   avg lat
>>>       20      16      5701      5685   1136.89      1116  0.041493 0.0561424
>>>    Total time run:         20.197129
>>> Total writes made:      5702
>>> Write size:             4194304
>>> Bandwidth (MB/sec):     1129.269
>>>
>>> Stddev Bandwidth:       23.7487
>>> Max bandwidth (MB/sec): 1168
>>> Min bandwidth (MB/sec): 1088
>>> Average Latency:        0.0564675
>>> Stddev Latency:         0.0327582
>>> Max latency:            0.47757
>>> Min latency:            0.029503
>>>
>>>
>>> Best Regards,
>>> -Dieter
>>>
>>
>> Well look at that! :)  Now I've gotta figure out what the difference is.
>>    How fast are the CPUs in your rados bench machine there?
>
> One CPU socket in each node:
> model name      : Intel(R) Xeon(R) CPU E5-2630 0 @ 2.30GHz
> Logial CPUs: 12
> MemTotal:       32856332 kB

I'm using 2x E5-2360L at 2.0GHz.  So yours are slightly faster, but not 
significantly so.  I am running the tests on localhost though, so 
perhaps that is having a negative effect rather than a positive one. 
Soon I will be testing on 10GbE and bonded 10GbE.

>
>>
>> Also, I should mention that at these speeds, we noticed that crc32c
>> calculations were actually having a pretty big effect.
>
> perf report
>
> Events: 39K cycles
> +     26.29%         ceph-osd  ceph-osd                    [.] 0x45e60b
> +      4.74%         ceph-osd  [kernel.kallsyms]           [k] copy_user_generic_string
> +      3.37%         ceph-mon  ceph-mon                    [.] MHeartbeat::decode_payload()
> +      2.88%         ceph-osd  [kernel.kallsyms]           [k] futex_wake
> +      2.61%          swapper  [kernel.kallsyms]           [k] intel_idle
> +      2.34%         ceph-osd  [kernel.kallsyms]           [k] __memcpy
> +      1.71%         ceph-osd  libc-2.11.3.so              [.] memcpy
> +      1.70%         ceph-osd  [kernel.kallsyms]           [k] __copy_user_nocache
> +      1.66%         ceph-osd  [kernel.kallsyms]           [k] futex_requeue
> +      1.33%         ceph-mon  ceph-mon                    [.] MOSDOpReply::~MOSDOpReply()
> +      1.18%         ceph-mon  libc-2.11.3.so              [.] memcpy
> +      1.16%         ceph-mon  ceph-mon                    [.] MOSDPGInfo::decode_payload()
> +      0.97%         ceph-osd  [kernel.kallsyms]           [k] futex_wake_op
> +      0.86%         ceph-mon  ceph-mon                    [.] MExportDirDiscoverAck::print(std::ostream&) const
> +      0.79%         ceph-osd  [kernel.kallsyms]           [k] _raw_spin_lock
> +      0.74%         ceph-mon  ceph-mon                    [.] MOSDPing::decode_payload()
> +      0.52%         ceph-osd  libtcmalloc.so.0.3.0        [.] operator new(unsigned long)
> +      0.51%         ceph-mon  ceph-mon                    [.] MDiscover::print(std::ostream&) const
> +      0.48%         ceph-osd  [xfs]                       [k] xfs_bmap_add_extent
> +      0.43%         ceph-mon  [kernel.kallsyms]           [k] copy_user_generic_string
> +      0.39%         ceph-osd  [kernel.kallsyms]           [k] iov_iter_fault_in_readable

Looks like you are having the same issues I do with user symbols in 
ceph-osd not showing up in perf.  They show up fine in sysprof for me. 
I bet a good chunk of the 26.29% at the top is crc32c calculation.

>
> Regards,
> -Dieter
>
>
>> Turning them off
>> gave us a 10% performance boost.  We're looking at faster
>> implementations now.
>>
>> Mark
>>
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>


^ permalink raw reply	[flat|nested] 8+ messages in thread

end of thread, other threads:[~2012-09-13 11:08 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-09-10 20:15 messaging/IO/radosbench results Mike Ryan
2012-09-10 20:39 ` Mark Nelson
2012-09-12 20:08   ` Dieter Kasper
2012-09-12 22:25     ` Mark Nelson
2012-09-12 23:24       ` Joseph Glanville
2012-09-13  0:39         ` Mark Nelson
2012-09-13  7:24       ` Dieter Kasper
2012-09-13 11:08         ` Mark Nelson

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.