autobuild: intermittent test failure detected

[Andrew Bartlett]

On Thu, 2015-07-09 at 08:27 +0200, Volker Lendecke wrote:
> On Thu, Jul 09, 2015 at 09:16:48AM +1200, Andrew Bartlett wrote:
> > On Wed, 2015-07-08 at 18:39 +0200, Stefan (metze) Metzmacher wrote:
> > > Hi,
> > > 
> > > it seems the new notify code is more timing sensitive than the old one..
> > > 
> > > I also saw this a few times today.
> > > 
> > >   [366(1357)/1863 at 33m53s] samba3.raw.notify(ad_dc)
> > >   TESTING CHANGE NOTIFY WITH RECURSION
> > >   UNEXPECTED(failure): samba3.raw.notify.recursive(ad_dc)
> > >   REASON: Exception: Exception: ../source4/torture/raw/notify.c:512:
> > > notify.nttrans.out.num_changes was 2 (0x2), expected 3  (0x3): wrong
> > > number of changes
> > > 
> > >   FAILED (1 failures, 0 errors and 0 unexpected successes in 0 testsuites)
> > > 
> > > Should we add this to the flapping tests or is there an easy way to fix
> > > this?
> > 
> > This reproduces 100% of the time on the Catalyst Cloud.  I've had to add
> > a knownfail for it ever since we started doing builds on our cloud.
> > 
> > http://git.catalyst.net.nz/gw?p=samba.git;a=commitdiff;h=66d33fab08e0962c817966b51cd3f44fe6cfc83d
> > 
> > I can get you a login on one of the hosts if that would help.
> 
> Is this flapping only with notifyd or is the flappiness older?

As metze points out, on the Catalyst Cloud it failed reliably since
March.

Andrew Bartlett

-- 
Andrew Bartlett                       http://samba.org/~abartlet/
Authentication Developer, Samba Team  http://samba.org
Samba Developer, Catalyst IT          http://catalyst.net.nz/services/samba