On Wed, 2017-09-13 at 09:38 +0000, Roberto Muñoz Gomez wrote:
> Hello,
> 
> I have a two node cluster with 56cores and 512Gb RAM each.
> 
> In this cluster I have configured 964 resources in 153 groups.
> 
> The first problem I had was regarding PCMK_ipc_buffer. Which is now
> configured as 9562060.
> 
> But I still have problems when try to "cleanup" several resources.
> 
> Is there another parameters I have to tune to deal with this amount
> of resources?

You may need the new cluster-ipc-limit cluster property available in
Pacemaker 1.1.17. The recommended value for large clusters is the
number of nodes times the number of resources. The default is 500. If
you see "Evicting client" log messages after cleanups or node joins,
you definitely need this.

I'm not aware of any other tuning needed, but let us know if you
encounter anything else.

> 
> Best Regards.
> ·····················································
> Roberto Muñoz
> 
> 
> 
> P Antes de imprimir, piensa en el MEDIO AMBIENTE
> AVISO LEGAL/DISCLAIMER
-- 
Ken Gaillot <[email protected]>




_______________________________________________
Users mailing list: [email protected]
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

Reply via email to