[Lustre-discuss] slow direct_io , slow journal .. in OST log

[Brian J. Murrell]

On Tue, 2010-01-26 at 14:12 +0700, Lex wrote:
> There was a problem with our raid controller this morning, raid array
> was degraded ( I reinstalled the hard drive and its state was
> rebuilding right after that )  and from messages log, i saw many
> warnings like this: 
> 
> Jan 26 13:59:13 OST6 kernel: Lustre:
> 9589:0:(filter.c:1363:filter_parent_lock()) lustre-OST0006: slow
> parent lock 49s
> Jan 26 13:59:13 OST6 kernel: Lustre:
> 9589:0:(filter.c:1363:filter_parent_lock()) Skipped 10 previous
> similar messages
> Jan 26 13:59:13 OST6 kernel: Lustre:
> 9638:0:(filter_io.c:398:filter_preprw_read()) lustre-OST0006: slow
> preprw_read setup 34s

This is not surprising during a rebuild.

> I have never seen slow parent lock or preprw-read  before. Maybe
> everything is going down fast and severely, Do all you guys have any
> ideas for my situation ? 

About this and your previous message, I think somebody has already
suggested that you are oversubscribing your disks.  You need to tune the
number of OST threads your OSSes are using so that they don't
oversubscribe your disks.  Details are in the manual.

b.

P.S.  Please don't put the mailing list address in the BCC list.  Doing
that makes it less intuitive for people to reply back to the list.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20100126/f1d6568c/attachment.pgp>