Re: slapd memory leak (ITS#249)

[dnehring@telemedia.de]

On Thu, Aug 19, 1999 at 07:04:01AM -0700, Kurt D. Zeilenga wrote:
> Dirk Nehring wrote:
> >
> > On Wed, Aug 18, 1999 at 08:10:55AM -0700, Kurt D. Zeilenga wrote:
> > > At 11:55 AM 8/18/99 GMT, dnehring@telemedia.de wrote:
> > > >Unfortunatly, it doesn't help. slapd is eating memory for lunch when
> > > >modifying many entries. I have no idea.
> > >
> > > Without decent tools, one has to resort to a process of elimination.
> > > I suggest now trying --without-threads --with-ldbm-api=gdbm.
> >
> > Hmmh, that seems to be very much work to test different backends. Is
> > there no change to reproduce the problem (10000 entries in a database
> > and write an LDIF file for updating 1 attribute of each entry)? If not,
> > I'll try to set up a test enviroment to track the memory leak further
> > down.
>
> Dirk, I hacked up the devel test008-concurrency test such that I
> could plug in different slapd (including non-threaded/gdbm).  The
> test uses add/delete, not modify, but I do see memory growth
> similiar to what you describe.  We are digging, but the more
> eyes the better...

OK, not problem. Unfortunately (for you, not me ;-), I'm going to
vacation for 2 weeks. After that, I'll try some of my leak detector
libraries...

Dirk

--
Dirk Nehring                | Phone: +49 5241 80-1560
Telemedia GmbH & Co. KG     | Fax:   +49 5241 80-9518
Carl-Bertelsmann-Str. 161 I | EMail: dnehring@telemedia.de
33311 Gütersloh             |