[Linux-cluster] gfs_glock_is_locked_by_me assertion failures

[Bob Peterson]

On Tue, 2008-02-26 at 15:25 +0100, Kadlecsik Jozsef wrote:
> Hi,
> 
> In our five node cluster running the same software version on all 
> machines, two of them have started to emit assertion failure messages:
> 
> Feb 26 13:54:47 web0 kernel: GFS: fsid=kfki:services.4: warning: assertion
> "gfs_glock_is_locked_by_me(ip->i_gl)" failed
> Feb 26 13:54:47 web0 kernel: GFS: fsid=kfki:services.4:   function = gfs_readpage
> Feb 26 13:54:47 web0 kernel: GFS: fsid=kfki:services.4:   file =
> /services/src/2.6.23.16/cluster-2.01.00/gfs-kernel/src/gfs/ops_address.c, 
> line = 279
> Feb 26 13:54:47 web0 kernel: GFS: fsid=kfki:services.4:   time = 1204030487
> 
> One of the machines in question has been rebooted today, the other one has 
> been running for several days.
> 
> How could I debug what is the reason of the problem?
> 
> Best regards,
> Jozsef

Hi Jozsef,

We already have a bugzilla 429343 open for this problem.  I believe it
was reported by a customer, so the bug record was automatically locked
to protect the customer's privacy.  We are working on a fix.
If you have a test or scenario that will recreate the failure, it may
help us with testing any fix we design.

Regards,

Bob Peterson
Red Hat GFS