On 2005-12-12T12:19:23, Alan Robertson <[EMAIL PROTECTED]> wrote:
> >I'd simply run the management daemon on each and every node. One could
> >then connect either to the admin IP of each node (maybe via the
> >management net only based on ACLs), or to a "Cluster IP" as soon as one
> >is added to the cluster (probably co-located with DC status)...
>
> If it didn't require the administrator to be smarter, and work harder, I
> like the idea of just running it once. But, it _does_ require them to
> be smarter and work harder - so I guess I've changed my mind.
I agree, I also would prefer that... Doesn't seem possible easily though
:-(
BTW, this is also the way how FailSafe worked ;-) You could connect the
management client to any node, if you wanted that to be HA, you added a
virtual IP. This probably means that the client should support
reconnecting "smoothly" (in case the IP fails over).
Sincerely,
Lars Marowsky-Brée
--
High Availability & Clustering
SUSE Labs, Research and Development
SUSE LINUX Products GmbH - A Novell Business -- Charles Darwin
"Ignorance more frequently begets confidence than does knowledge"
_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
