--On Tuesday, August 27, 2013 8:06 AM +0200 Ulrich Windl Ulrich.Windl@rz.uni-regensburg.de wrote:
Quanah Gibson-Mount quanah@zimbra.com schrieb am 26.08.2013 um 17:21 in
Nachricht <564FB11DDFEE88059AA204C7@[192.168.1.22]>:
--On Monday, August 26, 2013 4:16 PM +0200 Ulrich Windl Ulrich.Windl@rz.uni-regensburg.de wrote:
Hi!
I have a related question: How is the updateref expected to be for a multi-master server with replicated config database? None, one, or one for each server?
Why would a master have an updateref statement? Then it wouldn't be a master.
I haven't fully investigated, but in SLES11 SP2 YaST seems to refuse updates (users and groups) on multi-master if either "mirrormode: true" is missing, or "updateref". It's a bug in YaST, yes, but I have to live with it. I could do updates with LDIF, but other people want to use YaST...
Sounds like there is something seriously broken with YaST. How does it talk to a regular master if it doesn't have updateref incorrectly set? I.e., how does it "know" the difference between a single or multi-master configuration?
In any case, I always advise avoiding SLES. They can't release SP's without breaking backwards compatibility. SP2 broke SP1 backwards compat. SP3 breaks SP2 backwards compat. It's insane. How that can be considered "enterprise" is beyond me.
--Quanah
--
Quanah Gibson-Mount Lead Engineer Zimbra, Inc -------------------- Zimbra :: the leader in open source messaging and collaboration