[Linux-cluster] Freeze with cluster-2.03.11

Kadlecsik Jozsef kadlec at mail.kfki.hu
Thu Apr 2 19:29:45 UTC 2009


Hi,

On Thu, 2 Apr 2009, Wendy Cheng wrote:

> > If you have any idea what to do next, please write it.
> >   
> Do you have your kernel source somewhere (in tar ball format) so people can
> look into it ?

I have created the tarballs, you can find them at 
http://www.kfki.hu/~kadlec/gfs/:

- Kernel is vanilla 2.6.27.21, the '.config' file is preserved in the 
  tarball as 'config'.

- On top of that I installed vanilla e1000-8.0.6, e1000e-0.5.8.2 and
  aoe6-69. The same e1000-8.0.6 and e1000e-0.5.8.2 are used with the 
  working cluster-2.01.00 to which the earlier aoe6-59 was added.

- The cluster-2.03.11 is also the vanilla version, except that since
  this thread started I have added two small corrections:
  - fence/fenced/agent.c fixed, see
    https://www.redhat.com/archives/linux-cluster/2009-March/msg00222.html
  - gfs2/mount/umount.gfs2.c, '-l' flag support added, see
    https://www.redhat.com/archives/cluster-devel/2009-April/msg00000.html

  The configure options are in 'configure-options', the locally used init
  scripts can be found under deb/DEBIAN/etc ;-)

The GFS volumes are mounted with noatime, quota is enabled (can't leave 
that off). The volumes are tuned with the values:

statfs_slots 128
statfs_fast 1
demote_secs 30
glock_purge 50
scand_secs 3

Those are mostly remnants of the time when Maildir was in use instead 
of plain mailbox format and we tried to cure the terrible performance.

Probably it's worth to note that 'statfs_fast 1' takes a lot of time to 
complete (usually around 15-20 seconds) which is, at least for me, 
surprising.

I think that's all. 

Best regards,
Jozsef
--
E-mail : kadlec at mail.kfki.hu, kadlec at blackhole.kfki.hu
PGP key: http://www.kfki.hu/~kadlec/pgp_public_key.txt
Address: KFKI Research Institute for Particle and Nuclear Physics
         H-1525 Budapest 114, POB. 49, Hungary




More information about the Linux-cluster mailing list