When my master node is in trouble and begins to hang.....
control is indeed passed to the second node, the resources load
and it assumes being the Master node .....that part works fine.
why the failed node doesn't 'Auto Restart after an Abend'
but rather waits for a 'Prompt for coredump/abend.log"
before it is brought back online is in IMHO limited.
I think the workflow should have a timeout on the prompt
so that the administrator can decide whether its time for
'deep troubleshooting' or 'just get that damn server back up'
On Fri, 22 May 2009 01:06:02 GMT, ataubman
<ataubman@no-mx.forums.novell.com> wrote:
>
>Not in clustering, no. The server must go completely so clustering fails
>resources over to another node. If the abending server tries to limp
>along the resources it hosts will not fail over but may still be
>unavailable to users, obviating the whole point of having clustering at
>all.