Re: [Xen-devel] slow storage backend kills stubdomains

On Tue, Sep 29, 2009 at 2:38 PM, Pasi Kärkkäinen <pasik@xxxxxx> wrote:
> On Mon, Sep 28, 2009 at 04:43:17PM -0400, Keith Coleman wrote:
>> After deploying stubdomains in a production environment with xen 3.4.1
>> and the official 2.6.18-xen kernel I've had a lot of crashes. Each
>> server hosts several PV and HVM domUs. Whenever the storage backend
>> gets slow (RAID rebuild, domU with heavy disk IO, etc) the stubdomains
>> begin dying off. I have two questions:
>>
>> 1. Has anyone else seen this issue?
>> 2. fs-backend is still running so is there any way to attach a new
>> stubdom to the existing (orphaned) HVM domU?
>>
>
> Did you apply this patch?
>
> http://lists.xensource.com/archives/html/xen-devel/2009-08/msg00335.html
>
> -- Pasi
>
>

This is a different. The patch that you suggested was already in place
since I contributed to the discovery and resolution of that problem.
This particular issue seems to be related to high IO latency.


Keith Coleman

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel

<Prev in Thread]	Current Thread	[Next in Thread>
[Xen-devel] slow storage backend kills stubdomains, Keith Coleman Re: [Xen-devel] slow storage backend kills stubdomains, Pasi Kärkkäinen Re: [Xen-devel] slow storage backend kills stubdomains, Keith Coleman <=

Previous by Date:

RE: [Xen-devel] [RFC] [PATCH] use "reliable" tsc properly when available, but verify, Dan Magenheimer

Next by Date:

RE: [Xen-devel] [PATCH] replace rdtsc emulation-vs-native xen boot option with per-domain (hypervisor part), Dan Magenheimer

Previous by Thread:

Re: [Xen-devel] slow storage backend kills stubdomains, Pasi Kärkkäinen

Next by Thread:

[Xen-devel] default xen "make world" again dependent on non-xenbits repositories and non-http protocol, Dan Magenheimer

Indexes:

[Date] [Thread] [Top] [All Lists]