autobuild failure due to python replication tests - why ?

[Jeremy Allison]

On Wed, Aug 03, 2016 at 08:33:50PM +0200, Ralph Böhme wrote:
> On Wed, Aug 03, 2016 at 11:29:15AM -0700, Jeremy Allison wrote:
> > On Wed, Aug 03, 2016 at 08:10:37PM +0200, Ralph Böhme wrote:
> > > On Wed, Aug 03, 2016 at 10:52:06AM -0700, Jeremy Allison wrote:
> > > > On Wed, Aug 03, 2016 at 08:43:55PM +0300, Uri Simchoni wrote:
> > > > > 
> > > > > I tried a non-committing background buildnice autobuild on sn-devel,
> > > > > skipping the tests mentioned in those commits (the ones in knownfail in
> > > > > the commits put in skip) and it failed twice, but in places that don't
> > > > > seem related.
> > > > 
> > > > Yes, that's what we seem to be getting a lot - hence the
> > > > sledgehammer :-(.
> > > > 
> > > > On another note - can you post how to do a "non-committing
> > > > background buildnice autobuild on sn-devel" ?
> > > > 
> > > > I don't remember saving off a receipe on how to do that,
> > > > and it might help lower my blood pressure when watching
> > > > autobuilds via html :-).
> > > 
> > > thanks to metze since a few days that's now just a:
> > > 
> > > $ cd <git checkout>
> > > $ autobuild-private.sh
> > > or
> > > $ autobuild-private.sh samba
> > > for just the Samba target.
> > 
> > On what machine ?
> > 
> > ssh jra at git.samba.org
> > 
> > $ which autobuild-private.sh
> > $
> 
> sn-devel. Sorry, I thought that was obvious. :)

ssh jra at git.samba.org
$ ssh jra at sn-devel
ssh_exchange_identification: Connection closed by remote host

Can't get there from here :-(.