[ClusterLabs] Two-node Pacemaker cluster with "fence_aws" fence agent

[Valentin Vidić]

On Fri, Sep 04, 2020 at 05:24:00PM -0400, Digimer wrote:
> It would depend on AWS, and I don't believe it's a good idea to design a
> solution that depends on a third party's behaviour.

It would be strange if AWS control API for node1 would be serialized with
control for node2. In fact fence_aws suggests that both can happen in parallel
and tries to be smart about it:

	if (get_self_power_status(conn,my_instance) == "ok"):
		conn.instances.filter(InstanceIds=[options["--plug"]]).stop(Force=True)
		logger.info("Called StopInstance API call for %s", options["--plug"])
	else:
		logger.info("Skipping fencing as instance is not in running status")

Even with this in place, some delay seems like a good idea because control API
is probably not that fast (a manual fence test can show how long it takes to kill
a node).

-- 
Valentin