Replication test flapping

[Andrew Bartlett]

On Tue, 2016-06-07 at 08:35 +0200, Andreas Schneider wrote:
> On Tuesday, 7 June 2016 01:46:03 CEST Andrew Bartlett wrote:
> > commit fa2a94ae4c073be267f79ba16592c444cbbb0c84
> > Author: Andrew Bartlett <abartlet at samba.org>
> > Date:   Tue Apr 19 17:28:59 2016 +1200
> > 
> >     selftest: These replication tests are now OK after we fixed all
> > the
> > replication bugs
> > 
> >     We have run this test a number of times, and these tests do not
> > flap
> >     in those tests
> > 
> >     Signed-off-by: Andrew Bartlett <abartlet at samba.org>
> >     Reviewed-by: Garming Sam <garming at catalyst.net.nz
> 
> Awesome! Great work

We had lots of runs on the Catalyst cloud.  I ended up up running up to 10 jobs in parallel isolated VMs to try and get some confidence. 

Now, sn-devel isn't the same (much as we tried to emulate it), and there may still be issues, but I will watch it closely.

Thankfully my latest autobuild not only shortens the test time
(trimming down the krb5.kdc test runs), it also includes a fix for a
use-after-free we found with AddressSanitizer, in the schema
replication code.  This should fix some of the remaining flaps (such as
the repl_schema one in particular). 
Hopefully I wasn't too bold in declaring 'all' the replication bugs
fixed, but I really think Samba 4.5 will be a big improvement in this
space.
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