[ClusterLabs] corosync totem.token too long may cause pacemaker(cluster) unstable?

[Jan Friesse]

cys napsal(a):
> At 2017-03-08 17:10:36, "Jan Friesse" <jfriesse at redhat.com> wrote:
>>> Hi,
>>> We changed totem.token from 3s to 60s. Then something strange were observed, such as unexpected node offline.
>>> I read corosync.conf manpage, but still don't understand the reason.
>>> Can anyone explain this? or maybe our conf is broken?
>>
>> What corosync version are you using? Your config file is not exactly
>> "broken", but I'm not sure how exactly all the settings play together,
>> so maybe minimize it. Setting high token timeout shouldn't cause
>> problems, only failure detection takes long time.
>>
>>
>
> Corosync Cluster Engine, version '2.4.0'
> Copyright (c) 2006-2009 Red Hat, Inc.
>
> Sometimes corosync took very long time to be in operational state. I'm not sure whether it's related with totem.token.

It can be caused by token timeout + consensus timeout.

> Could you tell me how to diagnose it?

Basically try following config file:

---- CUT HERE ---

quorum {
          provider: corosync_votequorum
          two_node: 0
}
totem {
	version: 2

	crypto_cipher: none
         crypto_hash: none

	transport: udpu
	cluster_name: ClusterName
}

nodelist {
          node {
                  ring0_addr: 39.39.0.4
                  nodeid: 1
          }
          node {
                  ring0_addr: 39.39.0.5
                  nodeid: 2
          }
          node {
                  ring0_addr: 39.39.0.6
                  nodeid: 3
          }
}

--- CUT HERE ---

Replace ClusterName with something more clever and give it a try. This 
is almost minimal corosync.conf with default values and it usually works 
quite well. If you observe too much lost tokens, try increase token 
(timeout) but 60 seconds is too much. 10 should do the job.

Regards,
   Honza

>
> Thanks.
> _______________________________________________
> Users mailing list: Users at clusterlabs.org
> http://lists.clusterlabs.org/mailman/listinfo/users
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
>