[Date Prev][Date Next]
Re: OpenLDAP 2.2.10 locked up
--On Tuesday, May 18, 2004 10:01 AM -0500 Digant Kasundra <firstname.lastname@example.org>
In the middle of the night, while doing a batch of adds, slapd seems to
have locked up. I can't query it and the slapadd is still sitting there
in locked state. When I do try to query the directory, I do see a
connection operation logged in the ldap log on the server, but the
nothing else after that, and no results on searches or adds or modifies.
slapadd or ldapadd? Generally you don't use slapadd on a running server...
For the moment, the server and client at both just sitting there. I was
wondering if anyone knew of somethings I might want to check to see if I
can discover the problem.
CPU load is 0 on the server, and OpenLDAP seems to be consuming 350M of
memory. The DB_CONFIG is setup to create 2GB share.
The adds were all people and account entries. They are all of similar
format. It successfully added several thousand before stopping in the
J's and I see nothing particularly wrong with the entry its stuck on.
The log shows the ADD operation starting but nothing about its result.
The log also shows my ldap monitor being able to successfully query the
directory for an hour or two after that (while that add was still
locked)and then all I see is connect operations from the ldap monitor,
but no binds or searchs, etc.
Any help would be appreciated.
Generally if my system gets in that state, I kill it, run db_recover in the
db area, restart slapd, and then verify that the database is up, and the
last entry to successfully make it in.
I've seen this same problem using OpenLDAP 2.2.11 with syncRepl(ication) on
my master. For me though, the master locks up immediately on the first ADD.
Principal Software Developer
GnuPG Public Key: http://www.stanford.edu/~quanah/pgp.html