[Pacemaker] cluster-delay property

[Andrew Daugherity]

> Message: 1
> Date: Thu, 24 Oct 2013 11:07:10 +0200
> From: Karl R??mann <K.Roessmann at fkf.mpg.de>
> To: pacemaker <pacemaker at oss.clusterlabs.org>
> Subject: [Pacemaker] cluster-delay property
> Message-ID:
> 	<20131024110710.Horde.MXWLBLi1JerSN7Rv2744mg9 at multix51.mpi-stuttgart.mpg.de>
> 	
> Content-Type: text/plain; charset=UTF-8; format=flowed; DelSp=Yes
> 
> 
> Hi,
> 
> we have a two node HA cluster using SuSE SlES 11 HA Extension SP3
> 
> For some reason there was heavy I/O load on both nodes yesterday.
> and one of the nodes went down. (Which was a serious problem)
> 
> Maybe we have to change a timeout value ?
> I crm_gui I see  'cluster-delay'. The default is 60s.
> Is it a good idea to change it to 120s or more ?

Others have already addressed the cluster side of things.  In addition, I would highly recommend setting elevator=deadline in your kernel boot parameters.  This will make the machine much more responsive when under heavy I/O load, which should help the cluster maintain communication.


Andrew Daugherity
Systems Analyst
Division of Research, Texas A&M University