Indeed that was the first thing on my mind too, on those systems it has
happened on, we have told it to syncrepl from scratch into an empty
directory. This does not help the master though, but we could always
promote a slave to master if required.
I have just been told that we have disabled the testing script due to the
frequent on-call troubles. So I will have to replicate it on test
environment to debug further.
Quanah Gibson-Mount wrote:
--On Tuesday, September 26, 2017 9:56 AM +0900 Jorgen Lundman
> Unfortunately, this is not our experience - although it is possible the
> message is just a distraction.
> slapd stops all syncrepl, and responding, until they call us to restart
Interesting, I wonder if the BDB database has corrupted. Have you tried
rebuilding the database(s) on the system via slapcat/slapadd?
Packaged, certified, and supported LDAP solutions powered by OpenLDAP:
Jorgen Lundman | <lundman(a)lundman.net>
Unix Administrator | +81 (0)90-5578-8500
Shibuya-ku, Tokyo | Japan