On 2006-01-26T16:50:53, Alan Robertson <[EMAIL PROTECTED]> wrote:
> No I'm not. You're using the term in two different ways in the same
> sentence to give the appearance of contradiction - one as "the heartbeat
> processes" and one as "all of Linux-HA".
I'm not sure I follow this one, but it's late ;-)
> But, even then _from heartbeat's point of view_ it's a cluster - just a
> really small one. And, it's not a larval stage. It's a full-fledged
> cluster.
>
> And, the idea of merging clusters is a useful one - not just for this
> case. But, this is, of course, the most obvious case.
Well, OK. So a one-node "cluster" might start up even without
communication channels to other nodes, which is I think what this thread
refered to.
But, you _do_ bring up an excellent point and one I had hoped to avoid
;-) Because a "merge" (ie, first join ever) operation is, in a subtle
way, different from a re-join of a member which already was there once.
Namely we can't just supersede one CIB with the other, but would have to
merge them too.
Uhm.
Next year, I think. ;-)
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/
