Am Samstag 19 M=E4rz 2011, 14:37:46 schrieb dieter@dkluenter.de:
While modifying a database, either subordinate or not, slapd crashes. Debugging and configuration information can be found here: http://pastebin.de/16349 (debugging) http://pastebin.de/16350 (slapd.conf)
I wonder if such a configuration is even supposed to work. The accesslog=20 overlay on the master is configured for the glued database=20 (dc=3Ddkluenter,dc=3Dde). On the slave however you have individual syncrepl= =20 configuration for each subordinate database. Wouldn't it be correct to=20 have separate accesslog databases for the subordinates?
Additionally I think that at least the "overlay syncprov" for the=20 subordinate databases on the master is wrong (as syncprov is already=20 configured for the glued database). Not sure if it is harmful though.
A more selfcontained and stripped down example configuration with actual=20 data and information on which operation exactly fails could also help=20 debugging this (assuming that your config turns out to be valid).
Ralf