[Linux-cluster] CS4 Update 2 / problem with systems dump ?

[David Teigland]

On Tue, Mar 21, 2006 at 06:02:22PM +0100, Alain Moulle wrote:
> Hi
> 
> How do you manage the problem of dumping machines ?
> 
> I mean, if in a HA pair, the 1st node is crashing
> and proceeding to dump, the CS4 on peer node will
> automatically power off the first one, and therefore
> interrupt the dump process.
> 
> Does that mean that we never can have dump when
> the nodes are "under" CS4 ?
> Or is there a way to manage this point ?

You might set a fencing delay that would allow the dump to complete, e.g.
  <fence_daemon post_fail_delay="10">
  </fence_daemon>

Dave