>
>
> You can use the built-in streaming replication for what you describe but
> if you want unplanned failover it should be the SYNCHRONOUS replication.
>
> asynchronous replication will work fine for a controlled switchover (stop
> the slons, let your 'DR Master' get caught up with streaming replication'
> then bring it up as a read-write node and restart the slons).
>
> However if you have an unplanned failure of your 'master' node the 'slave'
> node might be further ahead via slony than your 'dr master' node via async
> streaming replication.  If this happens and you promote the 'DR master'
> then you might have a very bad day.
>
>
>
Got it. Thanks Steve for all  your help.
I could able to fix this by storepath() function and as you said, I have
taken care that 'slave' node never pass master events ahead.

--Raghav
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.slony.info/pipermail/slony1-general/attachments/20120622/b04ab9fd/attachment.htm