[Pacemaker] Resource is running only on one node, even if the node is not available anymore
bt047265 at gmail.com
Tue Mar 1 04:55:16 EST 2011
thanks a lot for your hint. Double negation is the smartest way ;)
On Mon, Feb 28, 2011 at 9:36 PM, Lars Ellenberg
<lars.ellenberg at linbit.com> wrote:
> On Mon, Feb 28, 2011 at 07:59:30PM +0100, Thomas Baumann wrote:
>> I tried to stick a group fixed to one node, but if the service
>> corosync is stopped the group is jumping to the next node.
>> Here is my configuration:
>> <rsc_location id="loc_lb1" rsc="group_lb1" node="m-lab-prx-lb-1"
>> How I can configure that behavior that my group is running only in
>> that specific host ?
> I really like the crm shell.
> And I like the double negation trick:
> crm configure location X-on-and-ONLY-on-Y X rule -inf: \#uname ne Y
> (the \ is for the shell, not for the crm shell,
> but it does not hurt in the crm shell either).
> The xml produced by this is
> <rsc_location id="X-on-and-ONLY-on-Y" rsc="X">
> <rule id="X-on-and-ONLY-on-Y-rule" score="-INFINITY">
> <expression attribute="#uname" id="X-on-and-ONLY-on-Y-expression" operation="ne" value="Y"/>
> The idea is to absolutely forbid (-inf) the resource X to run,
> unless it runs on #uname Y (rule is only valid for #uname NOT EQUAL Y).
> : Lars Ellenberg
> : LINBIT | Your Way to High Availability
> : DRBD/HA support and consulting http://www.linbit.com
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> 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