[Cluster-devel] [PATCH 5/5] gfs2: dlm based recovery coordination

[Steven Whitehouse <swhiteho@redhat.com>]

Hi,

On Thu, 2011-12-22 at 16:23 -0500, David Teigland wrote:
> On Mon, Dec 19, 2011 at 12:47:38PM -0500, David Teigland wrote:
> > On Mon, Dec 19, 2011 at 01:07:38PM +0000, Steven Whitehouse wrote:
> > > >  struct lm_lockstruct {
> > > >  	int ls_jid;
> > > >  	unsigned int ls_first;
> > > > -	unsigned int ls_first_done;
> > > >  	unsigned int ls_nodir;
> > > Since ls_flags and ls_first also also only boolean flags, they could
> > > potentially be moved into the flags, though we can always do that later.
> > 
> > yes, I can use a flag in place of ls_first.
> 
> I went back to ls_first after finding the flag broke the old code path for
> gfs_controld, and making it work involved changing more of the old code
> that I wanted to in this patch.  We can go back and reorganize how some of
> that old code works (and remove ls_first), in a subsequent patch.
> 
> > > > +	/*
> > > > +	 * Other nodes need to do some work in dlm recovery and gfs2_control
> > > > +	 * before the recover_done and control_lock will be ready for us below.
> > > > +	 * A delay here is not required but often avoids having to retry.
> > > > +	 */
> > > > +
> > > > +	msleep(500);
> 
> This is now msleep_interruptible(500), I couldn't get around adding some
> sort of delay here.  If we think of another way to delay this I'll be
> happy to try it.
> 
> I've finished and tested the rest of the changes.
> https://github.com/teigland/linux-dlm/commits/devel11
> http://git.kernel.org/gitweb.cgi?p=linux/kernel/git/teigland/linux-dlm.git;a=shortlog;h=refs/heads/next
> 

Ok. I think we are going very much in the right direction. Let me think
about this over Christmas and lets catch up again in the New Year,

Steve.