Date: Thu, 10 Aug 2006 14:28:53 +0200
From: Jens Axboe <axboe@suse.de>
To: Matthias Dahl <mlkernel@mortal-soul.de>
Cc: linux-kernel@vger.kernel.org
Subject: Re: sluggish system responsiveness under higher IO load
Message-ID: <20060810122853.GS11829@suse.de>
References: <200608061200.37701.mlkernel@mortal-soul.de> <200608061554.42992.mlkernel@mortal-soul.de> <20060807134841.GC10444@suse.de> <200608081944.12630.mlkernel@mortal-soul.de> <20060808190241.GB11829@suse.de>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20060808190241.GB11829@suse.de>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 989
Lines: 29

On Tue, Aug 08 2006, Jens Axboe wrote:
> On Tue, Aug 08 2006, Matthias Dahl wrote:
> > > Can you see if this makes any difference whatsoever?
> > 
> > Sorry but no change for the better with this patch.
> > 
> > Are there any more informations I can provide to help?
> 
> Ok, it was worth a shot. I got another report that the patch fixes this
> behaviour, so maybe we are dealing with two seperate issues. It would be
> nice if you could gather vmstat 1 info during a problematic period.
> blktrace info could also be very useful:
> 
> http://brick.kernel.dk/snaps/blktrace-git-20060807122505.tar.gz

Some more things to try/questions:

- Did 2.6.16 work well for you?

- Does disabling preemtion (CONFIG_PREEMPT_NONE=y) help?

-- 
Jens Axboe

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/