RE: slapd stuck in mutex_lock() in BDB (ITS#2972)

[hyc@highlandsun.com]

You should upgrade to the current release (OpenLDAP 2.1.27) before spending
any more time on this. Also, look at db_stat -m when the hang occurs.
BerkeleyDB 4.1.25 is known to hang when it runs out of cache.

  -- Howard Chu
  Chief Architect, Symas Corp.       Director, Highland Sun
  http://www.symas.com               http://highlandsun.com/hyc
  Symas: Premier OpenSource Development and Support

> -----Original Message-----
> From: owner-openldap-bugs@OpenLDAP.org
> [mailto:owner-openldap-bugs@OpenLDAP.org]On Behalf Of tr@tng.de

>  > Was the server running continuously before this problem
> developed, or had it
>  > been restarted recently?
>
> No, it was running for weeks and started from a clean state.
>
>  > Well, there's definitely something wrong there, but to
> really track it down
>  > would require a much greater level of detail - enabling
> full debug symbols in
>  > the BerkeleyDB library so we can examine individual
> variables in each thread
>  > stack to try to identify which thread owns the offending read lock.
>
> Unfortunately, the problem occurs infrequently. If I can
> compile the BerkeleyDB
> library with full debug symbols, I will give it a try. Do I
> also need to compile
> OpenLDAP with full debug symbols?
>
> Thanks for looking into this.
>