2014-12-03 08:58:38 +03:00
= Proxmox HA Manager =
2014-11-29 13:14:59 +03:00
2014-12-03 08:58:38 +03:00
== Motivation ==
The current HA manager has a bunch of drawbacks:
- no more development (redhat moved to pacemaker)
- highly depend on corosync (old version)
- complicated code (cause by compatibility layer with
older cluster stack (cman)
- no self-fencing
In future, we want to make HA easier for our users, and it should
be possible to move to newest corosync, or even a totally different
cluster stack. So we want:
- possible to run with any distributed key/value store which provides
some kind of locking (with timeouts).
- self fencing using linux watchdog device
- implemented in perl, so thatw e can use PVE framework
2014-11-29 13:14:59 +03:00
- only works with simply resources like VMs
2014-12-03 08:58:38 +03:00
= Architecture =
== Cluster requirements ==
=== Cluster wide locks with timeouts ===
The cluster stack must provide cluster wide locks with timeouts.
The Proxmox 'pmxcfs' implements this on top of corosync.
== Self fencing ==
A node needs to aquire a special 'agent_lock' (one separate lock for
each node) before starting HA resources, and the node updates the
watchdog device once it get that lock. If the node loose quorum, or is
unable to get the 'agent_lock', the watchdog is no longer updated. The
node can release the lock if there are no running HA resources.
This makes sure that the node holds the 'agent_lock' as long as there
are running services on that node.
The HA manger can assume that the watchdog triggered a reboot when he
is able to aquire the 'agent_lock' for that node.