[Fwd: [DRBD-user] heartbeat and drbd / Failover / Failback]

Rois Cannon rois at cobiz.com
Tue Dec 4 23:54:22 CET 2007

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


Removing "fencing  resource-only" from drbd.conf seems to do the trick
but then I'm back to secondary failure rolling back to an out of date
(but marked as uptodate) primary (node1.)  

You may have already figured out my problem from looking at my logs so
ignore my rambling if that is so and I'm out in left field, but is there
a way to get ha to outdate the currently primary node if ipfail shuts it
down for a ping node failure.  Actually, having ha outdate any node that
fails a ping node test would work well I would think.  That would
effectively keep that node from taking over resources without human
intervention.

Thoughts?
Rois



On Tue, 2007-12-04 at 14:05 -0800, Rois Cannon wrote:
> I attached the logs for each.  node2 is svr92_syslog.  Failure of node1
> shows up first at 13:22:07 when the serial shows dead.
> 
> Looks like I'm getting a drbd0 error at 13:22:13
> -----------------------------------------------------------------
> Dec  4 13:22:13 svr92 kernel: drbd0: outdate-peer helper broken,
> returned 255
> -----------------------------------------------------------------
> and then a bunch of refused to become primary stuff.
> 
> Anyway, have a look and tell me what you think.
> 
> It's greatly appreciated.
> Rois
> 
> 
> On Tue, 2007-12-04 at 17:27 +0100, Florian Haas wrote:
> > Rois,
> > 
> > can you provide a syslog snippet to include any drbd/dopd message from node2, 
> > around the time you pulled the plug on node1?
> > 
> > Thanks.
> > 
> > Cheers,
> > Florian
> > 
> > On Tuesday 04 December 2007 01:55:03 Rois Cannon wrote:
> > > Florian,
> > > I'm sure I'm just missing something.  Probably a timing thing.  I added
> > > the lines to drbd.conf and ha.cf per the instructions on your
> > > blog (see below for for full file.)  Brought up the system and made sure
> > > it was correctly primary on node1 and secondary on node2.  On node1, if
> > > I do a "halt" on the machine or restart heartbeat it correctly brings up
> > > node2 as primary.  If I pull the plug on node1, then node2 is being set
> > > to outdated so heartbeat can't bring it up.  Can you tell me what I'm
> > > missing?  Just FYI (in case it makes a difference) I'm running this in 2
> > > VMServer's as a test bed.
> > >
> > > [...]
> > 
> _______________________________________________
> drbd-user mailing list
> drbd-user at lists.linbit.com
> http://lists.linbit.com/mailman/listinfo/drbd-user




More information about the drbd-user mailing list