Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 26 Feb 2010 22:43:20 +0100
From:      Gerrit =?ISO-8859-1?Q?K=FChn?= <gerrit@pmp.uni-hannover.de>
To:        Daniel Braniss <danny@cs.huji.ac.il>
Cc:        stable@freebsd.org, Willem Jan Withagen <wjw@digiware.nl>, Jack Vogel <jfvogel@gmail.com>, Jeremy Chadwick <freebsd@jdc.parodius.com>
Subject:   Re: mbuf leakage with nfs/zfs? (was: em0 freezes on ZFS server)
Message-ID:  <20100226224320.8c4259bf.gerrit@pmp.uni-hannover.de>
In-Reply-To: <E1Nl6VA-000557-D9@kabab.cs.huji.ac.il>
References:  <4B86F384.3010308@digiware.nl> <2a41acea1002251459v40e8c6ddxd0437decbada4594@mail.gmail.com> <4B8795B1.4020006@digiware.nl> <20100226120339.GB17798@icarus.home.lan> <20100226133138.d47dd080.gerrit@pmp.uni-hannover.de> <20100226134429.041ea6f2.gerrit@pmp.uni-hannover.de> <E1Nkzry-0000AC-0S@kabab.cs.huji.ac.il> <20100226141754.86ae5a3f.gerrit@pmp.uni-hannover.de> <E1Nl1mb-0002Mx-M9@kabab.cs.huji.ac.il> <E1Nl2JK-00033U-Fw@kabab.cs.huji.ac.il> <20100226174021.8feadad9.gerrit@pmp.uni-hannover.de> <E1Nl6VA-000557-D9@kabab.cs.huji.ac.il>

next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 26 Feb 2010 22:09:32 +0200 Daniel Braniss <danny@cs.huji.ac.il>
wrote about Re: mbuf leakage with nfs/zfs? (was: em0 freezes on ZFS
server) :

DB> > Furthermore I found this via Google:
DB> > <http://lists.freebsd.org/pipermail/freebsd-current/2009-December/014062.html>.

This did not help, I still see the same problem.

DB> I'll have to do some packet snooping to check if it's TCP or UDP nfs
DB> traffic, since some of the clients are Linux ...

I have Linux clients, too. Some use tcp, some udp.

DB> > 2264/1321/3585 mbufs in use (current/cache/total)
DB> > 1239/1017/2256/65000 mbuf clusters in use (current/cache/total/max)
DB> > 1239/809 mbuf+clusters out of packet secondary zone in use
DB> > (current/cache)

DB> > but the uptime is only 12min so far. In some hours I'll know for
DB> > certain if this patch has anything to do with the problem.

It did not help. In the meantime the values read

20555/1465/22020 mbufs in use (current/cache/total)
19529/1029/20558/65000 mbuf clusters in use (current/cache/total/max)
19529/823 mbuf+clusters out of packet secondary zone in use (current/cache)


I created a little graph here:
<http://www.pmp.uni-hannover.de/test/Mitarbeiter/g_kuehn/data/mbuf.pdf>.

y-axis are the total mbuf clusters, x-axis in minutes. The flat part in
the upper right corner is a 10min-interval when I had stopped nfsd.

DB> at the moment there is not much activity, but if you check the latest
DB> plot.ps you will see that the bottom is slowly increasing, so my bet
DB> is that there must be some leakage!

There certainly is. I wonder when this came in and why it has gone
unnoticed so far. Probably not all people serving nfs from zfs see this,
or this would have popped up earlier. Maybe the Linux clients are somehow
triggering the issue? Or did it start with the import of zvol version 14?
Unfortunately I have upgraded my pool, so I cannot easily go back to 8-REL
to test this (otoh, I need a stable server quite urgently).


cu
  Gerrit



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20100226224320.8c4259bf.gerrit>