[ClusterLabs] Using the max_network_delay option of the corosync

Jan Friesse jfriesse at redhat.com
Fri Aug 17 08:11:45 EDT 2018


Vladislav,
> Hello!
> I have the cluster based on pacemaker and corosync. The cluster nodes are
> located in several data centers. And there is up to 450 ms ping latency
> between some of the nodes. Sometimes these delays lead to split brains.
> 

Yep, that's quite a lot. Enlarge token timeout so something like:

totem {
...
	token: 5000
...
}

will help

> I have found themax_network_delay option for the corosync.

This options does nothing until heartbeat_failures_allowed is set. I 
don't know if you have it enabled, but I would recommend to not enable it.

Regards,
   Honza

> 
> But there is a warning:
> It is not recommended to override this value without guidance from the
> corosync
> community.
> 
> Can anyone help me with this?
> 
> PS:
> https://serverfault.com/questions/926548/using-the-max-network-delay-option-of-the-corosync
> 
> 
> 
> _______________________________________________
> Users mailing list: Users at clusterlabs.org
> https://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
> 




More information about the Users mailing list