xfs
[Top] [All Lists]

Re: Read performance issues with 2.6.0test5?

To: Seth Mos <knuffie@xxxxxxxxx>
Subject: Re: Read performance issues with 2.6.0test5?
From: Frank Hellmann <frank@xxxxxxxxxxxxx>
Date: Tue, 23 Sep 2003 14:44:04 +0200
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <4.3.2.7.2.20030923092739.038dee40@xxxxxxxxxxxxx>
Organization: Optical Art Film- und Special-Effects GmbH
References: <4.3.2.7.2.20030923092739.038dee40@xxxxxxxxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030512
Hi Seth!

It seems that this only applies to the 2.4 kernel, does it? On my 2.6 installation there is no such entry in the /proc fs.

        Cheers,
                        Frank...

Seth Mos wrote:
At 15:55 22-9-2003 +0200, Frank Hellmann wrote:

Hi!

That is a bit weird result, isn't it? On the hand-patched 2.4.20-18.9 Kernel with XFS 1.3.0pre4 I'll get much better read performance and about similar write performance. I checked if there are any issues regarding the internal log and tried an external one, but not much of a change with that.


The readahead defaults are skewed. Which comes from the Red Hat Errata kernel. It changed from 2.4.18-27 to 2.4.20-18.

see /proc/sys/vm/max-readahead.
Default went from 127 to 31.

Since you seem to have some decent hardware *koff* I suggest setting max-readahead to something like 511 or so.

Cheers

--
Seth
It might just be your lucky day, if you only knew.





--
--------------------------------------------------------------------------
Frank Hellmann          Optical Art GmbH           Waterloohain 7a
Digital Cinema          http://www.opticalart.de   22769 Hamburg
frank@xxxxxxxxxxxxx     Tel: ++49 40 5111051       Fax: ++49 40 43169199


<Prev in Thread] Current Thread [Next in Thread>