[Linux-cluster] lm_dlm_cancel during quota operations

[David Teigland]

On Fri, Apr 20, 2007 at 04:31:17PM +0100, Robert Clark wrote:
>   I have a script which runs gfs_quota to set quotas for all the users
> on my GFS filesystem. When it's run simultaneously on two nodes, errors
> like the following begin to appear:
> 
>     lock_dlm: lm_dlm_cancel 2,34 flags 80
>     lock_dlm: lm_dlm_cancel rv 0 2,34 flags 40080
>     lock_dlm: complete dlm cancel 2,34 flags 40000

Writes to a hidden/internal file (like the quota file) set the PRIORITY
flag in the glock request, which cause the cancels.  I don't understand
why gfs would want to do cancels in this case, though, I was very
surprised to see that.  So, first job is to figure out why these writes
need to do cancels (or if that's the bug right there), and the second job
is to figure out why those cancels, if legitimate, are causing problems.

Dave