[Samba] [PATCH] Workaround very slow nss_winbind, fix crash on the AD DC (particularly for backups)

[Michael Wood]

Hi

On 24 June 2013 08:56, <Philippe.Simonet at swisscom.com> wrote:

> hi Andrew
>
> the
>         interfaces = eth0, lo
>         bind interfaces only = Yes
>
> doesn't bring anything by me, always crash.
>
> for the 2 other question I need some more time (never used valgrind,
> and have to re-do the bisect ...)
>
>
No need to redo the bisect.  Andrew is just asking you to try reverting the
commit that your previous bisect found and see if you still get the same
crash or a different crash or no crash at all.

So:

$ git revert f77d5d6479c879c8770fbc9a6ca5656ef3e41019

and then save the commit message.

Then test again to see if the crash is the same or different or if the
problem has now disappeared.

Thanks and regards
>
> Philippe
>
>
> > -----Original Message-----
> > From: Andrew Bartlett [mailto:abartlet at samba.org]
> > Sent: Saturday, June 22, 2013 9:09 AM
> > To: Simonet Philippe, ITS-OUS-OP-IFM-NW-IPE
> > Cc: samba at samba.org; samba-technical at samba.org
> > Subject: Re: [Samba] [PATCH] Workaround very slow nss_winbind, fix crash
> > on the AD DC (particularly for backups)
> >
> > On Fri, 2013-06-21 at 08:10 +0000, Philippe.Simonet at swisscom.com wrote:
> > > I tried both, and I get still crashes :
> > > ----
> > > 0001-gensec-work-around-nested-event-loops-by-ensuring-th.patch
> > > 0002-s4-winbind-Add-special-case-for-BUILTIN-domain.patch
> > > ----
> > >
> > > -----
> > > samba version 4.0.6 started.
> > > Copyright Andrew Tridgell and the Samba Team 1992-2012
> > > samba: using 'single' process model
> > > talloc: access after free error - first free may be at
> > > ../source4/kdc/db-glue.c:206 Bad talloc magic value - access after
> > > free
> > > PANIC: Bad talloc magic value - access after free Aborted
> > > ----
> > >
> > > philippe
> >
> > Does setting:
> >
> >         interfaces = virbr0:0 lo
> >         bind interfaces only = yes
> >
> > help?
> >
> > Also, does reverting (with 'git revert HASH', where HASH is the commit id
> > your bisect identified) help?
> >
> > Finally, can you run Samba under valgrind again?  The error you show
> above
> > doesn't seem quite right in the context, and I want to be sure we are not
> > chasing an unrelated issue.
> >
> > Thanks,
> >
> > Andrew Bartlett
> >
> > --
> > Andrew Bartlett
> http://samba.org/~abartlet/
> > Authentication Developer, Samba Team           http://samba.org
> >
>

-- 
Michael Wood <esiotrot at gmail.com>