[ClusterLabs] Corosync and pacemaker token thresholds with Juniper CoS

[Brad Zynda]

I guess to simplify this, what is the max latency of corosync multipath?
I have seen a few articles pointing to 2ms is this still the case?

Thanks,
Brad

On 12/04/2017 08:25 AM, Brad Zynda wrote:
> Hello,
> 
> I was hoping someone could explain the use of thresholds in
> communication between HA/Cluster members, more specifically their use of
> ping and latency.
> 
> We have noticed issues with ICMP latency gathering in other services and
> hardware and are wondering if this could be the case with the token
> timeouts we are seeing from corosync log events.
> 
> We have increased the token amount from 1200 to 2000.
> 
> We are still seeing issues with token timeout increase warnings and
> 
> corosync[1256]: [TOTEM ] A processor failed, forming new configuration.
> 
> 
> Could this all be related to
> https://kb.juniper.net/InfoCenter/index?page=content&id=KB28157
> 
> Now just to add some more info specific to the article we are seeing 100
> to 1000 not just 10 (maybe downplayed via the article)
> 
> and if so can corosync/pacemaker be used with tcping vs ping?
> 
> Thanks,
> Brad
> 
> _______________________________________________
> 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
>