So you didn't answer anything about the old version of OpenLDAP versus
the current one. I would note that according to the db_stat output,
you've experienced zero deadlocks. That seems somewhat in conflict with
your earlier report. Did you run db_recover (resetting the stats)?
Sorry; I missed that one.
We were running 2.4.33 before the upgrade.
I have not run db_recover in between the original report and now. It
is worth noting that the deadlock errors have ceased in the log file,
but I can't speak to whether or not we're more stable than we were
before; as I mentioned earlier, it seemed kinda random when it would
crash.
I was thinking that doing a slapcat and then cleaning out the database
files and then slapadd'ing everything back might be helpful. Do you
agree?