[Linux-HA] Re: forced failover again

[John R Mocho]

Sorry, but I am jumping in mid-thread (hopefully).

Has anyone been able to clearly define the equation for resources in 
resource groups / resource groups?

> > >
> > > Linux-HA moved a group to the backup_node after 6 failures. But per
> > > this document: http://wiki.linux-ha.org/v2/faq/forced_failover it had
> > > to be 10:
> > >
> > > (500 - 0 + 500) / abs (-100) = 1000 / 100 = 10
> > >

I am having great success with group based rsc_location constraints based 
on custom node attributes and was hoping to nail down what my 
default_resource_failure_stickiness and default_resource_stickiness values 
should be to allow the cluster to fail over resources in a logical way 
based our our needs.  It currently isn't.

Thanks,
   John.

> beekhof at gmail  -- 13 Jul, 2006
> -----------------------------------------------------------------
> oh, they're in a group! that changes everything.
> i'll take a look tomorrow but that formula only applies to single 
> resources