[Linux-HA] Score Calculation

Dejan Muhamedagic dejanmm at fastmail.fm
Thu Sep 13 07:44:37 MDT 2007


On Thu, Sep 13, 2007 at 11:53:46AM +0200, Dominik Klein wrote:
> Hi
> I did some tests on score calculation as I could not find a good web 
> resource on this. Now I would like to get this confirmed or corrected.
> Suppose I have a resource with a resource location constraint wit score 
> 300 for node A, score 250 for node B, resource stickiness = 200, failure 
> stickiness = -100.
> Startup. Resource goes to node A with score 300, gets the stickiness 
> addition and scores end up with A=500, B=250
> Now say the monitor action on the resource reports an error. Failcount 
> on A is increased, scores are recalculated. Is the following correct?:
> score + stickiness + (failcount*failurestickiness) = actualscore
> So in my exmaple
> 300+200+(1*(-100))=400
> This would make it A=400, B=250, so the resource would get stop/started 
> on A, no more calculation is done. Correct?
> Another question:
> Back to the top first, A=500, B=250, Resource running on A. All 
> failcounts=0.
> Now say the monitor action reports the resource is not running. 
> Failcount on A is increased and scores are recalculated. Is the 
> following difference (to the first calculation) correct?
> score+(failcount*failurestickiness)=actualscore
> So in my example
> 300+(1*(-100))=200
> So A=200, B=250, Resource gets stop on A, start on B. *Afterwards*, B 
> gets the stickiness Bonus of +200, so it ends up with
> A=200, B=450.
> Is this correct?
> Is there any website that describes these calculations?


That's were you'll find one example of calculating scores. Check
also the list archives---quite a few times this issue came up.



> Regards
> Dominik
> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems

More information about the Linux-HA mailing list