[Linux-HA] forced failover, node score

[Andrew Beekhof]

On 2/12/07, Andreas Kurz <andreas.kurz at gmail.com> wrote:
> On 2/12/07, Andrew Beekhof <beekhof at gmail.com> wrote:
> > On 2/10/07, Anton Lewur <alewur at web.de> wrote:
> > > Hello,
> > >
> > > let's assume following example
> > >
> > > A 2 node cluster, myNode_A myNode_B, heartbeat version 2.0.8.
> > >
> > > is the following calculation the equivalent for the example configuration below?
> >
> > seems correct
>
> And what about groups? .... I now this a favourite topic of Andrew ;-)
>  While testing 2.0.8 I saw that the score calculation is now not the
> addition of all resources in a group any more, but for the
> resource_stickiness it is the 'old' behaviour.  So the stickiness of a
> group is at the moment (default_resource_stickiness * sum(resources in
> group)) and this leads to a higher local restart count as most would
> expect.

ah, yes.  annoying but true :-(

i'll see if there is anything i can do for this in the next release

>
> >
> > >
> > >
> > > (MyNode_A - MyNode_B + default_resource_stickiness) / abs(default_resource_failure_stickiness)
> > >
> > > (100 - 0 + 500) / abs(-200)
> > >
> > > 600 / 200
> > >
> > > 3
> > >
> > >.....
> _______________________________________________
> 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
>