[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: SLAPD Locksup...problems with slapcat as well

--On Monday, January 05, 2004 10:39 PM -0700 Jayson Henkel <jhenkel@sterlingcrane.ca> wrote:


For no obvious reason (there were no changes to the directory today) I
had slapd stop answering queries today after working flawlessly for the
last 8 or so months. I logged into the box and saw at least 10 slapd
process running. Attempting the simple reboot process didn't rectify the
issue. I attempted to slapcat the entire database out and attempt to
reload it. However slapcat seems to lock or more correctly loop forever.
Please see slapcat.trace at http://xoa.darktech.org/slapcat.strace which
is a strace of the process... the sched_yield()   0 continues looping to
infinity. I also attempted debugging an interaction between the server
(which starts up fine) and a simple client search done via ldapsearch -x
-h localhost, it's located at http://xoa.darktech.org/slapd.strace. I
would appreciate any insight into the problem. I am using Debian
Unstable on a uniprocessor machine with version 2.1.23-1 of
both the slapd server and the ldap-utils and 2.3.2.ds1-10 of libc6.

Saying what backend you are using would be helpful. If you are using BDB, you may wish to try the db_recover command.


Quanah Gibson-Mount
Principal Software Developer
ITSS/TSS/Computing Systems
ITSS/TSS/Infrastructure Operations
Stanford University
GnuPG Public Key: http://www.stanford.edu/~quanah/pgp.html