[Top] [All Lists]

Re: xfs performance problem

To: Linux fs XFS <xfs@xxxxxxxxxxx>
Subject: Re: xfs performance problem
From: pg_xf2@xxxxxxxxxxxxxxxxxx (Peter Grandi)
Date: Sun, 1 May 2011 17:32:28 +0100
In-reply-to: <4DB75C6D.1080901@xxxxxxxxxxx>
References: <4DB72084.8020205@xxxxxxxxxxx> <4DB74331.3030804@xxxxxxxxxxxxxxxxx> <4DB75C6D.1080901@xxxxxxxxxxx>
> [ ... ]

> This is of course with delaylog enabled. I don't think a
> difference of a factor 7 is normal, given that writing to a
> raid-0 (xfs numbers) is supposed to be faster than writing to
> raid-1 (ext3 numbers) [ ... ]

To summarize some previous detailed discussion the actual
"performance" difference is either a factor of around 2 (12m for
'ext3', 24m for XFS) in the regular case or a factor of around 1.2
(27s for 'ext3', 33s for XFS) in the EatMyData (plus 'umount')
case (the one giving you over 1,300 transactions second).

Numbers like a bit over 2m for XFS with 'delaylog' and a bit
over 4m without are for intermediate cases between the regular
and EatMyData case, depending on how infrequently data and
metadata are committed by XFS.

If you want best "performance" with XFS, 'exec eatmydata "$SHELL"'
might be the top solution ;-).

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