PoorMan's San

PoorMan's San

[Robert L. Harris]

  Back on the Kernel list after a long hiatis.  Need some experienced
optinions.

Current Setup:

Linux Legato Backup server  (E1000 nic)   Linux-2.4.19-pre4
5xLinux NFS servers         (E1000 nic)   Linux-2.4.19-pre4

Legato mounts the NFS shares over a Gig ether and uses them as backup
media for non-offsite/non-longterm.  This works very well when NFS
behaves.

The probems is that every now and then the network/nfs just hang up for
a while and gets VERY slow.  Some research has been done and it seems
that the PCI latency is part of the problem, especially when IDE Raid
hits the disks hard, this can overrun the buffers on the GigE card.

I've got a test NFS server coming in and will look at upgrading the
kernel to 2.4.19-pre11 as there were some bugs associated with E1000 and
NFS I saw in pre-[7-9] area. I'm also considering trying to use
Network Block Device instead of NFS.  Has anyone tried this?

Any suggestions that doesn't involve hitting up a non-existant budget?


Robert



:wq!
---------------------------------------------------------------------------
Robert L. Harris                
                               
DISCLAIMER:
      These are MY OPINIONS ALONE.  I speak for no-one else.
FYI:
 perl -e 'print $i=pack(c5,(41*2),sqrt(7056),(unpack(c,H)-2),oct(115),10);'

Re: PoorMan's San

[Ben Greear]

Robert L. Harris wrote:
> 
>   Back on the Kernel list after a long hiatis.  Need some experienced
> optinions.
> 
> Current Setup:
> 
> Linux Legato Backup server  (E1000 nic)   Linux-2.4.19-pre4
> 5xLinux NFS servers         (E1000 nic)   Linux-2.4.19-pre4
> 
> Legato mounts the NFS shares over a Gig ether and uses them as backup
> media for non-offsite/non-longterm.  This works very well when NFS
> behaves.
> 
> The probems is that every now and then the network/nfs just hang up for
> a while and gets VERY slow.  Some research has been done and it seems
> that the PCI latency is part of the problem, especially when IDE Raid
> hits the disks hard, this can overrun the buffers on the GigE card.
> 
> I've got a test NFS server coming in and will look at upgrading the
> kernel to 2.4.19-pre11 as there were some bugs associated with E1000 and
> NFS I saw in pre-[7-9] area. I'm also considering trying to use
> Network Block Device instead of NFS.  Has anyone tried this?
> 
> Any suggestions that doesn't involve hitting up a non-existant budget?
> 
> 
> Robert

Have you increased the e1000 Rx and Tx descriptors?

Could try the e1000 NAPI patch as well...

Ben



-- 
Ben Greear <greearb@candelatech.com>       <Ben_Greear AT excite.com>
President of Candela Technologies Inc      http://www.candelatech.com
ScryMUD:  http://scry.wanfear.com     http://scry.wanfear.com/~greear

D-Link dfe-580tx (nic of doom)

[Robin Darby]

......So did anyone ever get this 4 port nic to go under linux?
I booted it up with a fresh copy of RH 8 this morning and still no joy. 
I even tried to compile the newest version of the sundance code, but no 
fun (couldn't even get the darn thing to compile).
Can someone point me in the right direction (working binary rpm would 
be nice) or give me a quick run down on how to get this thing going????

Ta
   Robin...

Re: D-Link dfe-580tx (nic of doom)

[Gianni Tedesco]

[-- Attachment #1.1: Type: text/plain, Size: 706 bytes --]

On Mon, 2002-10-21 at 19:48, Robin Darby wrote:
> ......So did anyone ever get this 4 port nic to go under linux?
> I booted it up with a fresh copy of RH 8 this morning and still no joy. 
> I even tried to compile the newest version of the sundance code, but no 
> fun (couldn't even get the darn thing to compile).
> Can someone point me in the right direction (working binary rpm would 
> be nice) or give me a quick run down on how to get this thing going????

You need the latest driver, patch attached.

-- 
// Gianni Tedesco (gianni at ecsc dot co dot uk)
lynx --source www.scaramanga.co.uk/gianni-at-ecsc.asc | gpg --import
8646BE7D: 6D9F 2287 870E A2C9 8F60 3A3C 91B5 7669 8646 BE7D

[-- Attachment #1.2: 16_sundance2.diff.gz --]
[-- Type: application/x-gzip, Size: 26542 bytes --]

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 232 bytes --]