On 2006-06-26T15:27:26, Huang Zhen <[EMAIL PROTECTED]> wrote:
> Sorry for the mistake :)
> I mean support more than two nodes.
> One is in primary and others in secondary.
It's on the list of features the drbd project is interested in, but it
is a difficult problem - resynchronization can get realllly complex,
write & read quorum begins to matter et cetera...
Much more interesting would be to have "floating peers", where the
two end-points of any given replication pair can move among different
nodes (which would of course need to share a storage backend then). This
would be very useful for multi-site clusters, where one peer would be on
some node at site A and the other peer on some node at site B.
This isn't too hard to add to the RA (basically instead of hard-coded
hostnames, use a globally unique clone and the clone number as the
hostname), but nobody has stepped forward to do it yet ;-)
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/
