Re: 2.6.13.4 - "lockd: weird return 1 for CANCEL call"

["Jesper Juhl" <jesper.juhl@gmail.com>]

On 21/02/07, Neil Brown <neilb@suse.de> wrote:
> On Wednesday February 21, jesper.juhl@gmail.com wrote:
> > Greetings,
> >
> > I just got a unusual message from lockd on a webserver running a
> > custom compiled 2.6.13.4 (yes it is ancient, I know).
> >
> > # uname -a
> > Linux webserv.somedomain.example 2.6.13.4 #1 SMP Wed Nov 16 10:03:05
> > CET 2005 i686 unknown
> >
> > # dmesg | tail -n 3
> > do_vfs_lock: VFS is out of sync with lock manager!
> > lockd: weird return 1 for CANCEL call
> > lockd: weird return 1 for CANCEL call
> >
> > The server seems to be running fine, so I'm not really worried, but I
> > thought maybe someone would like to know. :-)
>
> Thanks.
> Don't be worried.  They are both warning about perfectly acceptable
> (though uncommon) behaviour.  These warnings have since been removed.
>
I know the "VFS is out of sync with lock manager" warning has been
removed (I pushed quite a bit to get that done myself ;) I was just
providing context.  The one I had not seen before was the "weird
return ..." one.

> A process was interrupted while trying to get a lock, so the NFS
> client didn't know if it actually had the lock or not.
> When the process exited, the tried to cancel the lock just to be on
> the safe side and was told that it didn't own any such lock.
>
Thanks a lot for that very nice explanation.

-- 
Jesper Juhl <jesper.juhl@gmail.com>
Don't top-post  http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please      http://www.expita.com/nomime.html