RE: Berkely DB/OpenLDAP Upgrade issues

["Howard Chu" <hyc@highlandsun.com>]

There is a fix in CVS for some back-ldbm indexing problems. These fixes were
released in OpenLDAP 2.1.16. The corresponding patches have been applied to
the OpenLDAP 2.0 CVS stream, but have not been released. To my knowledge, no
one has tested them to confirm that they fix the observed problems.

You can try the patch yourself; if you do please report back whether it
resolves your problem. From CVS get the diff to back-ldbm/idl.c,
	-r OPENLDAP_REL_ENG_2_0_27 -r OPENLDAP_REL_ENG_2
and apply that to your source.

Note that OpenLDAP 2.0 is all old code and has not had any real active
maintenance for quite a while. It will very soon be retired, same as OpenLDAP
1.2.

  -- 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-software@OpenLDAP.org
> [mailto:owner-openldap-software@OpenLDAP.org]On Behalf Of Jong 't, D
(Dennis)

> Hi all,
>
> I have a question regarding the installation and functioning
> of Berkely
> DB/OpenLDAP.
>
> We are currently running a RedHat 7.2 with Berkely DB 3.3.11 and
> OpenLDAP
> 2.0.18. This all runs fine. Now we are in a stage of upgrading our
> corporate
> DS to Berkely DB 4.0.14 and OpenLDAP 2.0.27.
>
> After compliation and installation of both packages on test
> machines, I
> can
> import an LDIF (which is generated by the former configuration) with
> slapadd
> (commented out the inedexes), but when running slapindex (uncommented
> the
> indexes), I get a segmentation fault. (after it has run a while). This
> all
> runs on a Compaq Proliant 1600 with a dual processor config and SMP
> kernel.
> It also produces a seg. fault when running slapadd with the indexes
> turned
> on.
>
> On an other system, where the reindexing succeeded, we got spontanious
> hangups of the slapd when deleting entries from the
> directory. Also the
> ACL
> is interpreted incorrectly. This also runs on RH 7.2, but on
> a "normal"
> Compaq desktop PC.
>
> Does anyone recognize the mentioned problems and are there known
> solutions
> for it?
>
> Thanks in advance for your respone.
>
> With kind regards,
>
> Dennis 't Jong, The Netherlands