Rackspace CI and password lockout tests (was: Re: [PATCH] Fix several mem leaks in ldb_index.c, ldb_tdb.c, ldb_search.c)

Andrew Bartlett abartlet at samba.org
Sat Jun 16 06:09:19 UTC 2018


On Fri, 2018-06-15 at 20:15 +0200, Andrew Bartlett via samba-technical
wrote:
> On Fri, 2018-06-15 at 10:25 -0700, Jeremy Allison via samba-technical
> wrote:
> > On Fri, Jun 15, 2018 at 08:57:28AM +0000, Andrej Gessel via samba-technical wrote:
> > > Hello list,
> > > 
> > > 
> > > I have found and fixed some mem leaks, at least i think, that these are mem leaks.
> > > 
> > > 
> > > Pipeline[1] on Gitlab is not successful and I don't know why. If i run this test "samba4.ldap.password_lockout.python" locally, it runs successful.
> > > 
> > > So please review.
> > > 
> > > [1] https://gitlab.com/samba-team/devel/samba/pipelines/23846671
> > 
> > Hi Andrej,
> > 
> > All of these patches look completely correct to me.
> > 
> > RB+.
> > 
> > Can I get a second Team reviewer please ?
> > 
> > The samba4.ldap.password_lockout.python error may be
> > transient. Once I've gotten a second Team reviewer,
> > if it fails the autobuild with the same error I'll
> > look into the problem specifically, but right now
> > many of the python tests are just unreliable.
> 
> We are having some teething trouble with the new Rackspace based build
> hosts (they have different timing properties to the Catalyst Cloud
> hosts, local worktations and importantly sn-devel).  To make things
> even more challenging the PSO support brought in more runs of timing-
> dependent tests. 
> 
> sn-devel will likely be fine, and the timing will be investigated to
> see either if it can be make more tolerant or if the VM can be tweaked.
> 
> I'm running a CI on the Catalyst Cloud (the catalyst-samba repo is
> hooked up to different runner VMs) via this link to see if we can break
> the deadlock:
> 
> https://gitlab.com/catalyst-samba/samba/pipelines/23955774

It passed there.

> In the meantime I've also reviewed it and pushed it to autobuild.

It is now in master.  I've done some research on ways we can reduce the
storage latency (by forcing nobarrier to the filesystem mount).  It is
tricky getting options all the way from gitlab-runner to docker-machine 
to the host, but Joe Guo and I will work to see if we can update the
Rackspace runners.  Hopefully that will help.

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




More information about the samba-technical mailing list