[PATCH] ctdb-tests: Make reloadips tests more reliable (bug #13924)

[Amitay Isaacs]

On Tue, May 14, 2019 at 2:46 PM Martin Schwenke via samba-technical
<samba-technical at lists.samba.org> wrote:
>
> On Tue, 14 May 2019 14:43:52 +1000, Martin Schwenke <martin at meltin.net>
> wrote:
>
> > On Fri, 10 May 2019 09:19:43 +1000, Martin Schwenke via samba-technical
> > <samba-technical at lists.samba.org> wrote:
> >
> > > On Wed, 8 May 2019 11:31:14 +1000, Martin Schwenke via samba-technical
> > > <samba-technical at lists.samba.org> wrote:
> > >
> > > > The "ctdb reloadips" can fail for a few reasons, including
> > > > colliding with a takeover run already in progress.  This was a
> > > > simplifying design decision.
> > > >
> > > > The tests try to be careful, but there's a startup edge case where an
> > > > unexpected takeover run can be triggered, though it has been delayed
> > > > for a quite a while after startup.
> > >
> > > Please hold this for a while.  This is still failing because it isn't
> > > retrying.
> > >
> > > Amitay, this is almost certainly an issue with matching the error
> > > message, which you queried the other day...
> >
> > OK, the testcase changes were correct.  There was a bug in a previous
> > patch that fixed try_command_on_node().
> >
> > Additional patch included here.
> >
> > Please review and maybe push...
>
> Sorry, wrong patch.  Correct patch now attached...
>
> peace & happiness,
> martin

Pushed to autobuild.

Amitay.