PE6650 Just Hangs
Anthony Porcano
anthony_porcano at kaplan.com
Wed Jan 21 09:29:01 CST 2004
Thanks for writing Timothy. I have been running the tg3 driver on the onboard nics of one of our 6650's for just about 24 hours now, and the server is still online and working well. Since we have gone several weeks without a lockup in the past I am not inclined to say this is a magic bullet yet, but at this point I think I would lean more towards RedHat's solution. The developer who maintains the tg3 driver works directly for RedHat and has been vigilant in answering user issues on this list and addressing them quickly. (Thanks Jeff)
Again, I am not trying to flame Dell on these issues. I have a ton of respect for the Dell reps on this list who spend their free time answering questions and helping users. They have been great about listening to our issues and escalating them up the ladder. However, thus far running a Dell certified configuration has not gotten us anywhere. So my only suggestion to someone new to this issue would be don't overlook non-standard configurations. From the sounds of it, the folks who have had the best luck with these lockups have opted out of using the onboard nics completely, and replaced them with PCI intel cards (which both Dell and RedHat will support AFAIK).
It should probably also be noted that although I strongly suspect the nics as a common point of failure, I have not done load testing on the interfaces as other users have. Jay has been able to reproduce the lockup using a packet blaster which seems evident of at least some issue with the broadcom drivers/chipsets; however, it is possible that the lockups others are experiencing may not be caused by the same issue that Jay is having.
Best luck to you on these issues.
--Anthony
-----Original Message-----
From: Timothy Bruess [mailto:netmgr at learningresources.com]
Sent: Wed 1/21/2004 8:12 AM
To: 'Paul Anderson'; Anthony Porcano; Jay Haque; linux-poweredge at dell.com; linux-poweredge-admin at dell.com
Cc:
Subject: RE: PE6650 Just Hangs
We have just bought several PE 6650's preloaded with Red Hat AS 2.1 and the
Broadcomm onboard NIC. Being a complete newbie to Linux (1 week and
counting) we have had to pester both Dell and Red Hat for technical support
numerous times.
During the course of these support calls we have learned that not only does
Dell not support the tg3 driver, Red Hat won't support the Broadcomm driver,
which made for some very interesting discussions as we tried to configure
these servers.
Trying to learn a new OS on the fly is challenging enough without having
these "Catch 22" events occur.
My $.02 worth.
-----Original Message-----
From: Paul Anderson [mailto:Paul.Anderson at priorityhealthcare.com]
Sent: Wednesday, January 21, 2004 3:13 AM
To: Anthony Porcano; Jay Haque; linux-poweredge at dell.com;
linux-poweredge-admin at dell.com
Subject: RE: PE6650 Just Hangs
The only problem with the tg3 driver is that it is not officially supported
by Dell. If anything goes wrong the Dell techs will tell you that you are
not compliant. The Broadcomm crap cards have a long history of proprietary
problems, like sharing the code with those writing the drivers.
In our shop, and in several locations, we replaced the on board crap with
Intel Pro 1000's and they run great. No more system lockups due to the
network.
**HINT HINT HINT Dell people**
Dell, IMHO, needs to start selling boxes designed for Linux and stop selling
windows boxes that hobble running Linux. Oh yeah, and try building more than
two servers (6650, 2650) while you are at it.
Paul
-----Original Message-----
From: Anthony Porcano [mailto:anthony_porcano at kaplan.com]
Sent: Tuesday, January 20, 2004 1:59 PM
To: Jay Haque; linux-poweredge at dell.com; linux-poweredge-admin at dell.com
Subject: RE: PE6650 Just Hangs
Jay,
Our 6650 issues may be unrelated, but FWIW I am running version 6.2.11 of
the bcm5700 driver and we are still getting frequent lockups with our
systems. I just switched one of our 6650's over to tg3 this morning, and
will let you know if this makes a difference with the lockups.
I have yet to try load testing the interfaces, but based on your ability to
reproduce the issue I am now very inclined to run such a test. Thanks for
posting back your results.
--Anthony
-----Original Message-----
From: linux-poweredge-admin at dell.com on behalf of Jay Haque
Sent: Tue 1/20/2004 12:22 PM
To: linux-poweredge at dell.com; linux-poweredge-admin at dell.com
Cc:
Subject: Re: PE6650 Just Hangs
Thanks for the info.
We used Netperf to hammer the Broadcom gig port. In about ten
minutes the
server went bye-bye... no keyboard, no video, no network.
I feel a lot better now that we can reproduce the problem. We had a
conversation with a Dell tech earlier. He suggested installing the
Broadcom
BCM5700 Linux Driver V. 6.0.5.
We'll give this a shot and see how it goes.
If anyone else has tried this please share your experience.
Everything I've
read so far leads me to believe this solution is too easy.
Thank You,
Jay
Xose Vazquez Perez
<xose at wanadoo.es> To: Jay Haque
<jhaque at nypl.org>
Sent by: cc:
linux-poweredge at dell.com
linux-poweredge-adm Subject: Re:
PE6650 Just Hangs
in at dell.com
01/16/2004 10:36 AM
Jay Haque wrote:
> Found this thread at Bugzilla:
> http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=79997. Most of
you
have
> probablly seen it already. It discuss the tg3 driver mentioned by
Jim.
> It looks like the tg3 is the culprit. Anyone know of any good
torture
> testing tools I can use to reproduce the problem? I'd hate to
upgrade the
> Kernel and then have the problem reappear.
tg3 in latest RHL errata kernel should be stable.
(deja vu) There are lots of NET tools:
NetPIPE http://www.scl.ameslab.gov/netpipe/
Netperf http://www.netperf.org/
Iperf http://dast.nlanr.net/Projects/Iperf/
gensink http://jes.home.cern.ch/jes/gensink/
TTCP http://ftp.arl.mil/~mike/ttcp.html
New TTCP http://www.leo.org/~elmar/nttcp/
nuttcp ftp://ftp.lcp.nrl.navy.mil/pub/nuttcp/
net-stress - http://colorfullife.com/~manfred/TestApps/
pktgen -
ftp://robur.slu.se/pub/Linux/net-development/pktgen-testing/
I like NetPIPE and nuttcp(latest son of ttcp).
_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge at dell.com
http://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq or search the
list
archives at http://lists.us.dell.com/htdig/
_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge at dell.com
http://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq or search the
list archives at http://lists.us.dell.com/htdig/
_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge at dell.com
http://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq or search the list
archives at http://lists.us.dell.com/htdig/
_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge at dell.com
http://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq or search the list
archives at http://lists.us.dell.com/htdig/
More information about the Linux-PowerEdge
mailing list