On Mon, 2007-10-01 at 07:14 -0700, Howard Chu wrote:
Stelios Grigoriadis wrote:
On Fri, 2007-09-28 at 13:44 -0700, William R. Somsky wrote:
I've had a similar problem here. I don't really have much to add in the way of info for tracking it down, but here is an additional question:
Once a replica server gets out of sync w/ the master in this way, how does one get the replica back in sync? I've not been able to figure that out, short of deleting the entire database on the replica and rebuilding it from scratch, which can take a very long time.
IMO it's not worth answering this question until you know why things got out of sync in the first place.
To my knowledge, you can't. Since the replica contextCSN is timestamped at a later time than the out-of-sync entries, they can't be synced.
You obviously haven't read the slapd(8) manpage carefully enough. See the "-c" option.
No I didn't, my bad. And it works like a charm.
Some additional info. It seems to me that it is easier to reproduce this problem with the current stable version than the previous one (2.3.32).
/Stelios