In case it helps, the DAS isn't really a single point of failure for the
cluster. In GMS terms, if the DAS fails, shuts down, or restarts, then
another group member will become the master. As long as the DAS is restarted
before you want to do any kind of administrative work, you're ok.
I have an example in my "asadmin get-health" blog where I restart the DAS,
and when it comes back up it rejoins the group and still has valid
information about the state of the running instance(s) -- for example, it can
restart and still tell you that an instance has been up since some time
before the DAS restarted.
http://blogs.sun.com/bobby/entry/state_of_the_cluster_with
Cheers,
Bobby
--
[Message sent by forum member 'bbissett']
View Post: http://forums.java.net/node/796262