Re: (ITS#4390) slapd-ldap crashes with failed assertion

[aleksander.adamowski@gmail.com]

On 2/6/06, Pierangelo Masarati <ando@sys-net.it> wrote:
> > slapd: ../../../../servers/slapd/back-ldap/unbind.c:60: ldap_back_conn_destroy:
> > Assertion `lc->lc_refcnt == 0' failed.
> If this is related to ITS#4387, I need to know more about the type and
> sequence of operations that occur.  In fact, I haven't been able to
> reproduce this issue so far, even by running multiple consecutive binds
> on the same connection (see last commits to tests/progs/slapd-bind.c).

This is the same instance, same configuration and workload. I now have
idle-timeout set on all slapd-ldap backends in this instance.

There are 2 backends configured, the first for base DN "o=Altkom", the
second for empty base DN, it rewrites queries to the 1st backend using
slapo-rwm:

database        ldap
suffix          ""
uri           "ldap://localhost:392/ ldaps://ldapreplica"
idle-timeout 120
overlay rwm
rwm-rewriteEngine on
rwm-rewriteContext searchBase
rwm-rewriteRule   "$" "o=Altkom" ":"


Both backends have their idle-timeout set to 120 seconds to workaround ITS#4387.

One time, shortly before slapd-ldap crashed, I were running top and
could see it consume lots of CPU.

I'm keeping a log of last 256 lines of output from each crashed
slapd-ldap session (it runs in foreground, in an infinite loop), I
could send some of those samples to your private e-mail address if you
wish.

Best Regards,
--
    Aleksander Adamowski
        Jabber JID: olo@jabber.altkom.pl
        GG#: 274614
        ICQ UIN: 19780575
        http://olo.ab.altkom.pl