Well the idea is to automate system administration while letting the
administrator be the commander in chief :)
Setting up disaster recovery sites is difficult and errorprone.
DR site is not a supported scenario so far. There's so much more we need
to get right until then ...
You are right. And why is this - basically because it is REALLY hard to
reliably tell the current status the nodes (as it really is and not as
the cluster sees it, split brain etc.). This is due to the really
unreliable heartbeat infrastructure in DR / Split Site / Wan
environments, isn't is ? Well with a manually confirming administrator
your have a reliable source of information. the admin can contact other
departments prior to doing the failover and ask for the real status of
the data center. That's better than any cluster can eventually get.
Robert
_______________________________________________
Pacemaker mailing list
[email protected]
http://list.clusterlabs.org/mailman/listinfo/pacemaker
