On 7/2/07, Michael Schwartzkopff <[EMAIL PROTECTED]> wrote:
Am Montag, 2. Juli 2007 10:56 schrieb Lars Marowsky-Bree:
> On 2007-07-02T10:08:51, Michael Schwartzkopff <[EMAIL PROTECTED]> wrote:
> > I have defined a clone resource (IPaddr2) on two nodes. If one node
> > failes the resource that was located on one node moves to the
> > surviving. OK.
> >
> > But if the dead node comes back again, the resource does not move back
> > to the other node.
> >
> > Since this is a cloned resource with two instances I want to run one
> > instance on every node in case of a healthy system.
>
> Clones are "sticky" by default, so that the cluster doesn't shuffle them
> around all the time. You seem to want exactly that though, and for good
> reasons ;-) So, set resource_stickiness for the clones to 0.
Thanks. That hit the point. I did not know that cloned resources are sticky by
default.
they weren't, but people complained a lot :-)
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
