[DRBD-user] Out of memory error when invoking fence-handler

Lars Ellenberg lars.ellenberg at linbit.com
Tue Nov 11 10:29:34 CET 2014

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

On Mon, Nov 10, 2014 at 12:21:01PM -0500, Digimer wrote:
> On 10/11/14 10:42 AM, Lars Ellenberg wrote:
> >On Mon, Nov 10, 2014 at 09:32:32AM -0500, Digimer wrote:
> >>So to summarize, I only add:
> >>
> >>====
> >>          flush_signals(current);
> >>====
> >
> >No, you upgrade to 8.4 ;-) ^a^k Yes.
> Oh you.
> Soon, just not yet, promise!
> >>I'll brush off my old RPM notes and see if I can sort out a patch. Thanks!
> >
> >No need, just fetch from git.
> >
> >	Lars
> I'd like to keep it as an RPM, and iirc, the git checkout doesn't
> have an RPM builder does it?

Sure has. What did you think we build our packages from.

> In either case, is there a reproducer for this bug? However I fix
> it, I'd like to have a way to verify the fix is in place before I
> push it out to production.

If you have fencing configured,
and the replication link goes down,
that code path is triggered.

If it spawns the fence handler just fine, your are good.

If it gives you that nonsens error message about ENOMEM
(where it should report EINTR, if anything),
and fails to even start the handler script,
it is broken.

: Lars Ellenberg
: http://www.LINBIT.com | Your Way to High Availability
: DRBD, Linux-HA  and  Pacemaker support and consulting

DRBD® and LINBIT® are registered trademarks of LINBIT, Austria.
please don't Cc me, but send to list   --   I'm subscribed

More information about the drbd-user mailing list