Dominik Klein wrote:
My first HA setup is for a squid proxy where all I need is to move an IP
address to a backup server if the primary fails (and the cache can just
rebuild on its own). This seems to work, but will only fail if the
machine goes down completely or the primary IP is unreachable. Is that
typical or are there monitors for the service itself so failover would
happen if the squid process is not running or stops accepting connections?
Second question (unrelated): Can heartbeat be set up so one or two
spare machines could automatically take over the IP address of any of a
much larger pool of machines that might fail?
Heartbeat in v1 mode (haresources configuration) cannot do any resource
level monitoring itself. You'd need to do that externally by any means.
If you're just starting out learning now, I'd suggest going with openais
and pacemaker instead of heartbeat right away. Check out the
documentation on www.clusterlabs.org/wiki/install and
www.clusterlabs.org/wiki/Documentation
Is there some documentation available for openais? I can't even find a
good description of what it does or why you would use it. Also, will
this help with my 2nd question: having a few spares for a large number
of servers? While my objective with the squid cache is to proxy
everything through one server to maximize the cache hits, I may switch
to memcached on a group of machines and would like to have a standby or
2 that could take over for any failing machine.
--
Les Mikesell
lesmikes...@gmail.com
_______________________________________________
Linux-HA mailing list
Linux-HA@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
