--On Tuesday, August 10, 2004 5:28 PM -0400 James Courtney
<Jcourtney@inphonic.com> wrote:
I just wanted to ask my question again about anyone seeing OpenLDAP
2.1.25 backed by BDB 4.2.52 on RedHat 3 ES (or some similar combination)
silently dying under load. My thread and conn_max* settings are the
defaults. I'm going to mine my logs a bit to see how many nearly
concurrent requests the server is seeing when it passes away as I know
that I get spikes of requests (maildrop delivering to a distribution
list). I suppose these *COULD* maybe exceed some internal capacity of
slapd (surely not the conn_max_pending_auth of 1000 though) and result in
some misfortune.
Also, should I consider a more recent version of OpenLDAP and/or BDB? If
so, which are most stable?
What are your DB_CONFIG settings?
What is your cachesize setting?
What is your idlcachesize setting?
What backend are you using? (bdb, ldbm, etc).