[ClusterLabs] Antw: Failover to spare node
Ulrich.Windl at rz.uni-regensburg.de
Fri Oct 23 02:42:33 EDT 2015
>>> Andrei Borzenkov <arvidjaar at gmail.com> schrieb am 22.10.2015 um 18:49 in
Nachricht <562913B5.6070105 at gmail.com>:
> Let's say I have a pool of nodes and multiple services, somehow
> distributed across them. I would like to keep one node as "spare",
> without services by default, and if any of "worker" nodes fail, services
> that were running there should be relocated to spare together.
> This ensures each service keeps the same resources available and does
> not compete with services on other nodes.
> It obviously can be achieved with explicit location constraints on each
> service for "primary" and "secondary" node; but is there some generic
> trick that avoids configuring each and every service?
Maybe "node colors" could do the trick: Makr one of your nodes as "not that good" while all nodes are up; if one node fails, mark that standby node as "good" (automatically, of course).
I never tried it, but I remember having read about it...
> Users mailing list: Users at clusterlabs.org
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
More information about the Users