Full_Name: Nikolay Bogdanov Version: 2.4.23 OS: Debian/Ubuntu URL: ftp://ftp.openldap.org/incoming/sysinfo.txt Submission from: (NULL) (95.221.182.198) Dear �olleagues, We used open ldap for three years. But some weeks ago we disided to restructurate our LDAP server. We did it, but there are a lot of users in the company. So we enabled the "rwm" overlay and added "oldDn" field in a schema. Futhermore, we use "memberof" overlay. As a result users can bind with their old dn. But after that slapd stops with deadlock everyday. What can you advice for us? And what information should we provide to solve this problem? (strace, gdb log, db_stat or something else?) With best regards, Nikolay Bogdanov. CJSC Flant.
nikolay.bogdanov@flant.ru writes: > Version: 2.4.23 > (...) What can you advice for us? Upgrade, building from the source at www.openldap.org. There are several possibly-related bugfixes since 2.4.23. Then report back whether nor not that fixed the problem. If it is: > And what information should we provide to solve this problem? (strace, gdb > log, db_stat or something else?) Server log from slapd, and gdb 'thread apply all backtrace'. Unfortunately OpenLDAP installation strips debugging symbols by default, which makes backtraces nearly useless. You can prevent that by configuring with ./configure --enable-memberof ac_cv_prog_STRIP=: -- Hallvard
--On Saturday, December 10, 2011 7:07 PM +0000 h.b.furuseth@usit.uio.no wrote: > Unfortunately OpenLDAP installation strips debugging symbols by default, > which makes backtraces nearly useless. You can prevent that by > configuring with > ./configure --enable-memberof ac_cv_prog_STRIP=: You can also do it at install time by simply typing: make install STRIP="" --Quanah -- Quanah Gibson-Mount Sr. Member of Technical Staff Zimbra, Inc A Division of VMware, Inc. -------------------- Zimbra :: the leader in open source messaging and collaboration
Dear Colleagues, we compiled new ldap (2.4.28) on weekend and started in production, but problem does not solved. Now i trying to collecl debug info. best regards, Nikolay Bogdanov. CJSC Flant.
I wrote: >> And what information should we provide to solve this problem? (strace, gdb >> log, db_stat or something else?) > > Server log from slapd, and gdb 'thread apply all backtrace'. Eh, I forgot the obvious one: slapd's configuration, minus any passwords. -- Hallvard
Hello. This is some debug info, that we collect. Also, now we have two ldap servers. Stable without rwm for some servicies, like a e-mail server and current (2.4.28) with rwm for users. So we can collect debuginfo easily. With best regards, Nikolay Bogdanov. CJSC Flant.
moved from Incoming to Software Bugs
Multiple fixes made to slapo-rwm in the 2.5 release. If this still occurs with that release please follow up.