Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 06 Oct 2007 04:09:35 +0200
From:      =?utf-8?Q?Dag-Erling_Sm=C3=B8rgrav?= <des@des.no>
To:        Pawel Jakub Dawidek <pjd@FreeBSD.org>
Cc:        freebsd-fs@FreeBSD.org, freebsd-current@FreeBSD.org
Subject:   Re: ZFS kmem_map too small.
Message-ID:  <86r6k9ouxs.fsf@ds4.des.no>
In-Reply-To: <20071005000046.GC92272@garage.freebsd.pl> (Pawel Jakub Dawidek's message of "Fri\, 5 Oct 2007 02\:00\:46 %2B0200")
References:  <20071005000046.GC92272@garage.freebsd.pl>

next in thread | previous in thread | raw e-mail | index | archive | help
Pawel Jakub Dawidek <pjd@FreeBSD.org> writes:
> We'are about to branch RELENG_7 and I'd like to start discussion with
> folks that experience 'kmem_map too small' panic with the latest HEAD.

I've been getting it recently (twice in the last ten days).  Intel E6600
with 4 GB RAM running an amd64 kernel from September 25.  After the
first panic, I increased kmem_size_max to 2G so that I now have:

des@ds4 ~% sysctl -h vm | grep kmem
vm.kmem_size_scale: 3
vm.kmem_size_max: 2,147,483,648
vm.kmem_size_min: 0
vm.kmem_size: 1,381,412,864

After the second panic, I set kmem_size_max to 4G and kmem_size to 2G in
loader.conf, but I haven't rebooted yet.

No backtraces because the watchdog kicked in and rebooted the machine.
I now have tace_on_panic=3D1 and debugger_on_panic=3D0, so next time I
should get at least a backtrace, and a dump if I'm lucky.

Both panics occurred during the nightly rsync of my imap spool from my
mail server to my file server, where it is kept in a compressed dataset
which I snapshot every day.  The spool currently contains about 1.5 M
files with an average size of slightly less than 8 kB.

DES
--=20
Dag-Erling Sm=C3=B8rgrav - des@des.no



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?86r6k9ouxs.fsf>