Dear All,
After a successful c. 3 month trial period of using Multi-Master (*4)
replication (RHEL4) on our test networks we are planning to progress to live
operation in the near term.
Since this trial went a little too well (.healthy paranoia!?) I would
welcome any feedback with respect to other users' experiences with live
operation, for example:
1) Your preferred approach to replacing a Master DS after underlying
hardware failure, in terms of ensuring that recovery from the other three
Masters is indeed achieved in a very timely manner (e.g. usage of any
variants on forced replication updates?).
2) Replication degradation and eventual recovery where the
intermediate network links degrade appreciably or are completely lost for a
significant period, before returning and allowing the replication to
recover.
Thanks in advance!
Regards,
Kevin
Show replies by date