[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [xen-unstable test] 88047: regressions - FAIL

On Fri, 2016-04-01 at 07:44 -0600, Jan Beulich wrote:
> > On 01.04.16 at 15:36, <Ian.Jackson@xxxxxxxxxxxxx> wrote:
> > That message is from libvirt.  I don't know the exact situation.
> > 
> > But we have seen other timeouts with merlot[01].  They seem to
> > sometimes have weird stalls when running under Xen.
> Might it be worth trying whether that's a deep C-state issue (by
> temporarily limiting their use of deep C states)?
IIRC, it's the box(es) with the weird NUMA setup... something about
nodes without any memory:

Apr  1 01:03:31.095559 (XEN) SRAT: Node 0 PXM 0 0-a0000
Apr  1 01:03:31.103546 (XEN) SRAT: Node 0 PXM 0 100000-c0000000
Apr  1 01:03:31.103565 (XEN) SRAT: Node 0 PXM 0 100000000-240000000
Apr  1 01:03:31.111600 (XEN) SRAT: Node 2 PXM 2 240000000-43f000000
Apr  1 01:03:31.111619 (XEN) NUMA: Allocated memnodemap from 43ee16000 - 
Apr  1 01:03:31.119615 (XEN) NUMA: Using 8 for the hash shift.
Apr  1 01:03:31.127613 (XEN) SRAT: Node 1 has no memory. BIOS Bug or 
mis-configured hardware?
Apr  1 01:03:31.135548 (XEN) SRAT: Node 3 has no memory. BIOS Bug or 
mis-configured hardware?

Not that I see how this could cause the long latencies / timeouts, though... :-/

In any case, I think Jan's C states suggestion is worth a shot.

<<This happens because I choose it to happen!>> (Raistlin Majere)
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)

Attachment: signature.asc
Description: This is a digitally signed message part

Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.