All of lore.kernel.org
 help / color / mirror / Atom feed
From: Bernd Schubert <bernd-schubert@web.de>
To: Olaf Kirch <okir@suse.de>
Cc: nfs@lists.sourceforge.net
Subject: Re: async vs. sync
Date: Wed, 28 Jul 2004 14:35:51 +0200	[thread overview]
Message-ID: <200407281435.54555.bernd-schubert@web.de> (raw)
In-Reply-To: <20040728085644.GD20264@suse.de>

Hello Olaf,

> the way the sync export option affects NFSv3 writes is limited to
> COMMITs, so if you see a slow-down here it must be bottle-necking in
> that part of the code.
>
> Quite possibly, this is a problem of the underlying file system.
> You're using a journaled file system, right? So what seems to happen
> is that on every n-th commit call or so all nfsd processes stall
> as the file system tries to write its journal. Note that the
> VM currently allows dirty data to accumulate for up to 30 seconds
> before it is forcibly written to disk (dirty_expire_centisecs sysctl).
>
> A good way to simulate this is to run several iozone processes on the
> server and tell them to sync() every 1 MB or so.
>
> 	iozone -s 1g -r 1m -o -i 0

I just did some tests:

/home: 10MB/s (mirrored via drbd)
/worka: 60MB/s (not mirrored)

The filesystem is reiserfs in both cases, but it seems drbd has in this case a 
terrible performance problem. I also tested to write to a drbd mirrored ext2 
partition, but it has the same problem, so I think this is filesystem 
independent.

Well, during the testing period of our server, I also tested the drbd 
performance, but unfortunately I did most tests with linux-2.6.7. With 2.6.7 
I got more than 30MB/s over drbd. Well, I also did some tests with 2.4., but 
this was with drbd-0.6.12, now we are using drbd-0.7.0. As far as I remember 
the numbers with drbd-0.6+linux-2.4 were similar or even faster than 
drbd-0.7+linux-2.6.

>
> This takes NFS out of the equation.

It seems you are right, its not a nfs issue.

>
> Maybe it would help to play with the dirty writeback strategy, e.g. by
> lowering /proc/sys/vm/dirty_writeback_centisecs (to e.g. 250), increasing
> dirty_background_ratio or lowering vm_dirty_ratio.

Well, I only see this on 2.6. systems, but not on 2.4., are there similar 
triggers in 2.4.?  With 2.6. we don't have the problem at all.


Thanks a lot for pointing me in the right direction!


Best regards,
	Bernd



-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

  reply	other threads:[~2004-07-28 12:36 UTC|newest]

Thread overview: 63+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <482A3FA0050D21419C269D13989C61130435E51E@lavender-fe.eng.netapp.com>
2004-07-26 17:05 ` async vs. sync Bernd Schubert
2004-07-26 19:47   ` Jan Bruvoll
2004-07-26 22:06     ` Bernd Schubert
2004-07-27 12:00       ` Jan Bruvoll
2004-07-27 13:00         ` Bernd Schubert
2004-07-27 13:56           ` raven
2004-07-27 14:04             ` Jan Bruvoll
2004-07-27 14:11           ` Jan Bruvoll
2004-07-28  8:56       ` Olaf Kirch
2004-07-28 12:35         ` Bernd Schubert [this message]
2004-07-28 12:49           ` Olaf Kirch
2004-11-24 19:05 Lever, Charles
  -- strict thread matches above, loose matches on Subject: below --
2004-11-23 16:36 Lever, Charles
2004-11-23 18:16 ` Dan Stromberg
2004-11-23 14:30 Lever, Charles
2004-11-23 21:46 ` jehan procaccia
2004-11-24 18:45   ` jehan.procaccia
2004-11-24 22:24     ` Neil Brown
2004-11-24 23:14       ` jehan procaccia
2004-11-24 23:34         ` Neil Brown
2004-11-24 22:09   ` Neil Brown
     [not found]   ` <Pine.GSO.4.53.0412010900500.5486@int1.cdc.noaa.gov>
2004-12-01 17:27     ` jehan.procaccia
2004-11-23  3:53 Lever, Charles
2004-11-23 16:33 ` Dan Stromberg
2004-11-22 22:14 Lever, Charles
     [not found] <20041122214605.8E2B31D0FE1@sc8-sf-uberspam1.sourceforge.net>
2004-11-22 21:57 ` Joshua Baker-LePain
2004-11-22 21:50 Lever, Charles
2004-11-22 22:06 ` jehan procaccia
2004-11-23  1:09 ` Dan Stromberg
2004-11-22 19:02 Lever, Charles
2004-11-22 21:25 ` jehan procaccia
2004-11-22 21:45   ` Nicolas.Kowalski
2004-11-22 23:51     ` jehan procaccia
2004-11-22 18:31 Lever, Charles
2004-11-16 18:48 Lever, Charles
2004-11-22 15:36 ` Olaf Kirch
2004-11-22 17:55   ` jehan.procaccia
2004-11-22 18:06     ` Roger Heflin
2004-11-22 18:46       ` jehan.procaccia
2004-11-22 19:10         ` Roger Heflin
2004-11-22 21:44           ` jehan procaccia
2004-11-22 21:52             ` jehan procaccia
2004-11-22 22:20               ` Trond Myklebust
2004-11-22 22:57                 ` jehan procaccia
2004-11-23  9:50                   ` jehan procaccia
2004-11-23 14:57                     ` J. Bruce Fields
2004-11-22 18:08     ` Trond Myklebust
2004-11-22 18:57       ` jehan.procaccia
2004-11-22 19:05         ` Roger Heflin
2004-11-22 20:14         ` Trond Myklebust
2004-11-22 21:04           ` Paul Cunningham
2004-11-22 21:14             ` Trond Myklebust
2004-11-22 22:07               ` Paul Cunningham
2004-11-22 22:26                 ` Trond Myklebust
2004-11-16 18:45 Lever, Charles
2004-11-16 16:15 Lever, Charles
2004-11-16 16:32 ` Trond Myklebust
2004-11-16 17:18   ` jehan.procaccia
2004-11-16 18:08     ` Trond Myklebust
     [not found] <482A3FA0050D21419C269D13989C61130435E530@lavender-fe.eng.netapp.com>
2004-07-27 15:07 ` Bernd Schubert
2004-07-26 23:05 John Roberts
     [not found] <482A3FA0050D21419C269D13989C61130435E523@lavender-fe.eng.netapp.com>
2004-07-26 21:28 ` Bernd Schubert
2004-07-23 16:20 Linux NFS writes to Solaris very, very slow John Roberts
2004-07-26 15:17 ` async vs. sync Bernd Schubert

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200407281435.54555.bernd-schubert@web.de \
    --to=bernd-schubert@web.de \
    --cc=nfs@lists.sourceforge.net \
    --cc=okir@suse.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.