| To: | Christoph Hellwig <hch@xxxxxxxxxxxxx>, linux-kernel@xxxxxxxxxxxxxxx, linux-xfs@xxxxxxxxxxx |
|---|---|
| Subject: | Re: XFS lstat() _very_ slow on SMP |
| From: | Jan Kasprzak <kas@xxxxxxxxxx> |
| Date: | Sun, 29 May 2005 20:37:50 +0200 |
| In-reply-to: | <20050528091123.GA19330@infradead.org> |
| References: | <20050516162506.GB19415@fi.muni.cz> <20050518140258.GA22587@infradead.org> <20050518174530.GF19173@fi.muni.cz> <20050528091123.GA19330@infradead.org> |
| Sender: | linux-xfs-bounce@xxxxxxxxxxx |
| User-agent: | Mutt/1.4.1i |
Christoph Hellwig wrote:
: > ("time ./bigtree.pl /new-volume 3 128" for 128*128*128 files), and then
: > "strace -c find /new-volume -type f -mtime +1000 -print" (the numbers
: > without strace are almost the same, so strace is not a problem here).
:
: I couldn't reproduce the odd case here. Could you try to get some profiling
: data with oprofile for the odd and one of the normal cases?
I have solved this by adding "ihashsize=65537" to /etc/fstab.
Now my find(1) is limited by the disk speed instead of system time.
-Y.
--
| Jan "Yenya" Kasprzak <kas at {fi.muni.cz - work | yenya.net - private}> |
| GPG: ID 1024/D3498839 Fingerprint 0D99A7FB206605D7 8B35FCDE05B18A5E |
| http://www.fi.muni.cz/~kas/ Czech Linux Homepage: http://www.linux.cz/ |
-- Yes. CVS is much denser. --
-- CVS is also total crap. So your point is? --Linus Torvalds --
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: XFS lstat() _very_ slow on SMP, Christoph Hellwig |
|---|---|
| Next by Date: | XFS & cryptoloop problems, Udo Wolter |
| Previous by Thread: | Re: XFS lstat() _very_ slow on SMP, Christoph Hellwig |
| Next by Thread: | XFS & cryptoloop problems, Udo Wolter |
| Indexes: | [Date] [Thread] [Top] [All Lists] |