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

Re: slapd dying ...

I have strange crashes as well, but under much heavier load and a
different setup.  ;)
Is it possible that your indexes are messed up, or something along those lines?
What's the behaviour of the crash?  Do you get core files, or does it
just shutdown?

Did you try truss/ktrace/strace-ing the search and/or server during
these hangs to see what was happening?  Are there some db-specific
monitorings you could try?

On 5/4/05, Samuel Tran <stran@amnh.org> wrote:
> Hi all,
> I run several OpenLDAP server 2.2.23 + Berkeley 4.2.52 (patched) on Debian
> Linux Woody. I have more than 2,000 entries in the directory. I have a LVS
> load balancer in front of the slaves. Every slave processes about 50
> req/sec.
> Once in a while slapd would die. When I restart it everything looks fine
> until I do a ldapsearch of the whole directory and find out that the
> output of the ldapsearch stops at a specific entry. When I try to stop
> slapd, it wouldn't kill the processes. I have to manually kill each one of
> them. In order to fix the problem I have to run a db_recover, then restart
> slapd.
> This problem occured on the master and on the slaves.
> I am not sure what is going on.
> Would an upgrade to OpenLDAP 2.2.26 fix the problem?
> Has anyone experienced the same problem?
> I would appreciate any input.
> Thanks in advance.
> Sam