On Fri, Sep 10, 2010 at 12:48 AM, Phillips, William G (BPHILLIP)
<[email protected]> wrote:
> Hi,
>
> I'm running a Pacemaker 0.6 two-node active/passive cluster (plan to migrate
> to latest in the next couple of months when my management will allow it). I
> have a group resource that runs on the active node and a clone resource that
> runs on both nodes. I'm using OCF resource agents. I want to implement the
> following location constraint assuming I have Group Resource "GrpApps"
> running on the active node and Clone Resource "CloneApp":
>
> If CloneApp stops .and. CloneApp node == GrpApps node, Then move GrpApps to
> other node

You'll need to colocate GrpApps with CloneApp.  Not sure about the
syntax of colocation constraints in 0.6 anymore though.

> I've tried a rsc_location constraint of the following form:
>
> <rsc_location id="move-apps-grp-CloneApp:0-failure" rsc="GrpApps">
>   <rule id="CloneApp:0-failure-rule" score="-INFINITY" boolean_op="and">
>      <expression id="CloneApp:0-failure-check"
> attribute="fail-count-CloneApp:0" operation="gt" value="0"/>
>      <expression id="CloneApp:0-on-primary-check" attribute="#uname"
> operation="eq" value="node1"/>
>   </rule>
> </rsc_location>
>
> There's similar rsc_location constraints for the CloneApp:1 version of the
> clone and for the other node. Also, I reset the failcount for the CloneApp
> resource on restart.
>
> This rsc_location constraint works but when the location constraint
> executes, I've now restricted my GrpApps resource from running on node1. I
> haven't been able to figure out a way to reset the GrpApps resource to allow
> it to run on node1 again. I tried using the 'crm_resource -U -r "GrpApps"'
> command in the start section of the CloneApps OCF script, but that didn't do
> anything.
>
> Any ideas on how I can resolve the restriction to allow GrpApps to run on
> node1 again? Is there a better way to setup the rsc_location constraint I'm
> trying to use?
>
> Any ideas or point me in the right direction, it would be greatly
> appreciated.
>
> Thanks,
> Bill Phillips
>
> ---------------------
> This e-mail (including any attachments) is intended only for the use of the
> individual or entity named above and may contain privileged, proprietary, or
> confidential information.  The information may also contain technical data
> subject to export control laws.
> ---------------------
>
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to