[prev in list] [next in list] [prev in thread] [next in thread] 

List:       e1000-devel
Subject:    Re: [E1000-devel] IGB 1.3.8.6 under stress -- kernel  errors
From:       David Lawless <david_lawless () flumedata ! com>
Date:       2008-11-18 18:49:46
Message-ID: 6.2.5.6.2.20081118134534.01d7d830 () flumedata ! com
[Download RAW message or body]

At 10:39 11/18/2008 -0800, Duyck, Alexander H wrote:
>David Lawless wrote:
>> igb-1.3.8.6 compiled with DCA enabled
>> 
>> F9 distribution
>> 
>> kernel.org 2.6.27.1
>> 
>> -----
>> 
>> high load on two of four 82575 ports
>> 
>> heavy disk cache pressure
>> 
>> -----
>> 
>> dmesg attached
>
>Other than these messages where there any other issues?  Also 
>how often are the messages occurring?

Alex,

Yes, large amounts of UPD packets were lost during the
events.  This happened once during an unusually intense
period of activity.  2P 8-core Opteron 2354 was at
something like 90% overall CPU utilization along with
the heavy writing of dirty file cache pages.

I've been planning to switch the disk activity to direct IO
to eliminate the cache pressure even before this happened.

>What you are describing sounds like a known issue that has been 
>seen with multiple drivers in the Linux kernel.  If you Google 
>"GFP_ATOMIC page allocation failures" you should find an email 
>thread from the linux-kernel mailing list that goes into detail 
>on the issue.

Yes it does look like this.

David


-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
[prev in list] [next in list] [prev in thread] [next in thread]