WARNING - OLD ARCHIVES

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/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-devel

Re: [Xen-devel] unfair servicing of DomU vbd requests

To: Keir Fraser <keir.xen@xxxxxxxxx>
Subject: Re: [Xen-devel] unfair servicing of DomU vbd requests
From: Ian Campbell <Ian.Campbell@xxxxxxxxxx>
Date: Thu, 3 Mar 2011 08:22:40 +0000
Cc: James Harper <james.harper@xxxxxxxxxxxxxxxx>, xen devel <xen-devel@xxxxxxxxxxxxxxxxxxx>
Delivery-date: Thu, 03 Mar 2011 00:23:11 -0800
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <C994F3F3.14124%keir.xen@xxxxxxxxx>
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/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
Organization: Citrix Systems, Inc.
References: <C994F3F3.14124%keir.xen@xxxxxxxxx>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
On Thu, 2011-03-03 at 07:29 +0000, Keir Fraser wrote:
> On 03/03/2011 02:25, "James Harper" <james.harper@xxxxxxxxxxxxxxxx> wrote:
> 
> > A user of GPLPV (see thread "blue screen in windows balloon driver") is
> > getting a bug check in Windows under extremely high memory usage and
> > swapfile thrashing tests across multiple DomU's. Responses to my query
> > on the ntdev mailing list say that this would happen if an IO request is
> > not completed after 70 seconds during high memory/pagefile pressure,
> > which is what is happening.
> > 
> > It appears that Dom0 is not servicing vbd requests from DomU's fairly so
> > one or two end up getting stalled while the others are mostly okay. How
> > are vbd requests supposed to be serviced? Is there potential for one to
> > be overlooked for a long period of time? Is there some settings that
> > could be changed to avoid this happening?
> 
> Dom0 does round-robin scanning of pending event channels these days, which
> helps fairness a fair bit.

I have a feeling this isn't true of pvops kernels...

looks like we need to pull 324:7fe1c6d02a2b (and subsequent fixes) out
of 2.6.18-xen.hg into the pvops world.

I'll take a look shortly if no one beats me to it.

>  When a pending event is found, the corresponding
> blkfront has a batch of requests pulled down and submitted into Linux's
> block subsystem at which point we have no more control over scheduling (it
> could generally be configured though -- Linux has an admin mechanism for
> that).
> 
>  -- Keir
> 
> > Thanks
> > 
> > James
> > 
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@xxxxxxxxxxxxxxxxxxx
> > http://lists.xensource.com/xen-devel
> 
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-devel



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