This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


Re: [Xen-devel] Xen / Dell 2850 PERC 4e/Di lock up

To: xen-devel@xxxxxxxxxxxxxxxxxxx
Subject: Re: [Xen-devel] Xen / Dell 2850 PERC 4e/Di lock up
From: David H <davidh.davidh@xxxxxxxxx>
Date: Wed, 20 Jul 2005 10:24:38 -0700
Delivery-date: Wed, 20 Jul 2005 17:23:45 +0000
Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=P0mfeUeGZmMpT5vHjo/jymlIN3+Y5CQDKc6dP0IymhJ/ZyRU/Kkls5lZNKrOFi0mdHKDXKYSZH0Tvtteacl8nStXmR9UZABnHcBi49iB0p638Jf3kqPs3875B/P3qb3OjbAHBPSIeoOba4X9o2VDVHO+1i/oopypnsXsa6czSgc=
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <32bd0f116d592e65b6136e3642b0b356@xxxxxxxxxxxx>
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
References: <A95E2296287EAD4EB592B5DEEFCE0E9D2826E3@xxxxxxxxxxxxxxxxxxxxxxxxxxx> <4a51ce970507191439709cc197@xxxxxxxxxxxxxx> <32bd0f116d592e65b6136e3642b0b356@xxxxxxxxxxxx>
Reply-to: David H <davidh.davidh@xxxxxxxxx>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
I am seeing the same behavior on a similar system.  As previously
discussed, on these severs domain0 hangs under I/O load.  The results
of my testing with 2.0-testing and the latest unstable are as follows.

basic xen:              Hangs under load, never* comes back  *(waited 12 hours)
with noirqbalance:  Seems to fix the problem (network throughput ~5%
lower then with "nousb")
with nousb:            Reliably running for weeks

basic xen:              Hangs under load
with noirqbalance:  Long delays under load but does not hang (peak
network throughput greater than 2x 2.0-testing but delays lower
average throughput)
with nousb:            Reliably running for one day (peak network
throughput greater than 2.5x 2.0-testing with average throughput
slightly greater than 2x)

I am testing by coping a 1.5GB file using scp.  This will reliably,
and quickly hang both xen versions without noirqbalance or nousb.  The
same file is being copied each time between the same two systems.  For
each test a script copies the file 10 times or untill the domain
hangs.  The systems are connected by a GigE switch.  The xen version
are from yesterdays tar balls.

This system will be available for testing for the remainder of the
week.  Please let me know if there is anything I can do to help
resolve this problem.

I would also like to take a moment to thank everyone involved in this
project.  Xen is truly amazing and getting better all the time.


On 7/20/05, Keir Fraser <Keir.Fraser@xxxxxxxxxxxx> wrote:
> On 19 Jul 2005, at 22:39, Shane Chen wrote:
> > However, I'm not completely
> > sure I managed to get it working because I never saw "XEN: Platform
> > quirk -- Disabling IRQ balancing/affinity" when I did `xm dmesg`.  So
> > then the curious question for me is then why it didn't completely hang
> > when I tried the same thing (multiple times).
> Only the unstable tree prints that message, and only then when it
> automatically detects and applies fix to a buggy chipset (not if you
> manually specify noirqbalance)
>   -- Keir
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-devel

Xen-devel mailing list