[Gluster-infra] NetBSD jobs hang on umount

Nigel Babu nigelb at redhat.com
Mon Aug 22 05:16:06 UTC 2016


On Sat, Aug 20, 2016 at 9:33 AM, Emmanuel Dreyfus <manu at netbsd.org> wrote:

> Nigel Babu <nigelb at redhat.com> wrote:
>
> > Atin says we've noticed this in the past and somehow fixed it. Do you
> > recall what we did to fix it?
>
> Is it the same problem? The key test is ps -axl and observe the WCHAN
> column for stuck umount process. If it is tstile then this is the
> ancient bug.
>
> --
> Emmanuel Dreyfus
> http://hcpnet.free.fr/pubz
> manu at netbsd.org
>

Yep, I see tstile:
   0 16295  2404 11509  83  0   3624  1164 wait     I    ?        0:00.02
sh -c /sbin/umount -f -R /mnt/glusterfs/0 /mnt/glusterfs/1 /mnt/glusterfs/2
/mnt/nfs/0 /mnt/nfs/1 /build/install/var/run/gluster/patchy
   0 17824 16295 11509 124  0   1612   652 tstile   D    ?        0:00.00
/sbin/umount -f -R /mnt/glusterfs/0 /mnt/glusterfs/1 /mnt/glusterfs/2
/mnt/nfs/0 /mnt/nfs/1 /build/install/var/run/gluster/patchy
   0 20703 27838 10365 115  0   1612   652 tstile   D    ?        0:00.00
/sbin/umount -f -R /mnt/glusterfs/0 /mnt/glusterfs/1 /mnt/glusterfs/2
/mnt/nfs/0 /mnt/nfs/1 /build/install/var/run/gluster/patchy
   0 27838  6136 10365  83  0   3624  1164 wait     I    ?        0:00.02
sh -c /sbin/umount -f -R /mnt/glusterfs/0 /mnt/glusterfs/1 /mnt/glusterfs/2
/mnt/nfs/0 /mnt/nfs/1 /build/install/var/run/gluster/patchy
   0 21943  9415   682  85  0   3388   992 pipe_rd  S+   pts/0    0:00.00
grep umount

I've added a pkill gluster before the clean up script, which should fix
this up. Let's see how that works out for this week.


-- 
nigelb
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-infra/attachments/20160822/76a50905/attachment.html>


More information about the Gluster-infra mailing list