Date: Tue, 20 Dec 2005 15:02:26 +1100
From: Nathan Scott <nathans@sgi.com>
To: Aaron Kulbe <akulbe@gmail.com>
Cc: linux-kernel@vger.kernel.org, xfs-masters@oss.sgi.com,
       linux-xfs@oss.sgi.com, hch@xfs.org
Subject: Re: PROBLEM: Page allocation failure
Message-ID: <20051220040226.GE1355@frodo>
References: <4a10e1350512191355t5c0abe18k4269e0ebbabc147e@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <4a10e1350512191355t5c0abe18k4269e0ebbabc147e@mail.gmail.com>
User-Agent: Mutt/1.5.3i
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1070
Lines: 27

On Mon, Dec 19, 2005 at 03:55:56PM -0600, Aaron Kulbe wrote:
> [1.] SUMMARY:  During rsync and heavy I/O operations, we experienced
> abnormally slow response times on the file server.
> ...
> [4.] Kernel version (from /proc/version): Linux version 2.6.8-24.14-bigsmp (

Try a recent kernel, both XFS and the VM have moved on in leaps
and bounds since 2.6.8 timeframe.

> I tried Christoph Hellwig's suggestion from
> http://oss.sgi.com/bugzilla/show_bug.cgi?id=345 of increasing the value of
> /proc/sys/vm/min_free_kbytes.  I set the value to 1024.  The results were
> the same.  No change in performance, and errors persisted.

There were older kernel versions where that parameter was not
being honoured particularly well, I can't recall anymore if it
was 2.6.8 or not though - it may well have been.

cheers.

-- 
Nathan
-
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/