[Pacemaker] Pacemaker 1.0.8 and -INFINITY master score

Dejan Muhamedagic dejanmm at fastmail.fm
Fri Aug 13 06:03:49 EDT 2010


Hi,

On Thu, Aug 12, 2010 at 12:54:10PM -0700, Bob Schatz wrote:
> I upgraded to Pacemaker 1.0.8 since my application consists of Master/Slave 
> resources and I wanted to pick up the fix for setting negative master scores.

Why not to 1.0.9.1?

> I am now able to set negative master scores when a resources starts and is SLAVE 
> but can't be promoted.  (The reason I want this is the process needs an 
> administrative override and has to be up and running for the administrative 
> override.)
> 
> However, when I test this on a one node cluster I see that the resource loops 
> through the cycle (attempt promote, timeout, stop the resource, start the 
> resource, ...).
> 
> I would have thought that a master score of -INFINITY would have prevented the 
> promotion.

Yes, sounds like it. Where is the score set? Wouldn't "resource
demote" do the right thing?

> My configuration is:
> 
> node $id="f4e5e15c-d06b-4e37-89b9-4621af05128f" mgraid-bob00001-0
> primitive SSJ5AMKP00009010REPK ocf:omneon:ss \
>     params ss_resource="SSJ5AMKP00009010REPK" 
> ssconf="/var/omneon/config/config.J5AMKP00009010REPK" \
>     op monitor interval="3s" role="Master" timeout="7s" \
>     op monitor interval="10s" role="Slave" timeout="7" \
>     op stop interval="0" timeout="100" \
>     op start interval="0" timeout="120" \
>     meta id="SSJ5AMKP00009010REPK-meta_attributes"
> ms ms-SSJ5AMKP00009010REPK SSJ5AMKP00009010REPK \
>     meta clone-max="2" notify="true" globaally-unique="false" 

You have a typo here.

Thanks,

Dejan

> target-role="Master"
> location ms-SSJ5AMKP00009010REPK-master-w1 ms-SSJ5AMKP00009010REPK \
>     rule $id="ms-SSJ5AMKP00009010REPK-master-w1-rule" $role="master" 100: #uname 
> eq mgraid-BOB00001-0
> property $id="cib-bootstrap-options" \
>     dc-version="1.0.8-9881a7350d6182bae9e8e557cf20a3cc5dac3ee7" \
>     cluster-infrastructure="Heartbeat" \
>     stonith-enabled="false"
> 
> Is this the expected behavior?
> 
> 
> Thanks,
> 
> Bob
> 
> 
> 
>       
> 
> 
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> 
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker




More information about the Pacemaker mailing list