On Mon, Oct 08, 2007 at 07:18:10PM +0200, Claus Guttesen wrote:
quoted text
> > I was able to reproduce the panic by rsyncing big files and trying
> > bonnie++ test suggested in this thread.
> >
> > Can you guys retry with this patch:
> >
> >         http://people.freebsd.org/~pjd/patches/vm_kern.c.2.patch
> >
> > It's a hack, yes, but allows to mitigate the problem quite well. I'm
> > looking for a solution that can be used for 7.0 before we find a better
> > fix.
>=20
> Congrats Pawel! You made my server survive my rsync of 90 GB. :-)
>=20
> This is on same src as the one that required a reboot except for your
> patch. So this fix does 'alleviate kmem_map too small' in my case.

While we have come across the 'kmem_map too small' panics in the past,
these are now solved - but our rsync processes still hang at random,
just spinning the CPU. This is when running multiple rsyncs to the
same zpool:

  http://www.brixandersen.dk/tmp/zfs-rsync-debug.1.txt

This is reproduceable (but only after running an intensive set of
rsync processes for ~10-12 hours or so) on our i386 RELENG_7 box with
the above mentioned patch, 4GB RAM, KVA_PAGES=3D512 and the following
settings in loader.conf:

vfs.zfs.prefetch_disable=3D"1"
vfs.zfs.zil_disable=3D"1"
vfs.zfs.arc_max=3D"209715200"
vm.kmem_size=3D"1258291200"
vm.kmem_size_max=3D"1258291200"

Regards,
Brix
--=20
Henrik Brix Andersen <henrik@brixandersen.dk>