[Pacemaker] basic configuration-this time with clones
Frank DiMeo
Frank.DiMeo at bigbandnet.com
Tue Dec 15 20:14:03 UTC 2009
Hi All,
I'm using a two node symmetrical cluster, and want to run stateful clones on them. In the enclosed .xml file, I defined one clone, and I want this clone to run such that one clone's resource is running on each of the two nodes, one of them in the Master state, the other in the Slave state.
I've modified my RA's to return the proper values for all the various states as described in "Configuration Explained", so that the CM can monitor the clone's resources and determine if they are running, and in what state. I also made sure that on startup, the children come up in the slave mode. This is all working fine.
When I run the system though, on instance of the clone's resource (do you call this the "child" of the clone?) is running on each node, but they are both get promoted to the master state by the PE. I realize that in my xml file I define each node to be equally likely to have the Master clone's child running there, but I thought that if a Master child were detected to be running on one node, then the child on the other node would be forced to be slave (or the promote would not happen there).
I'm enclosing the pe-input file for this case, and from it, you can see that both clone's children are promoted to master.
Is this the expected behavior?
-Frank
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/pacemaker/attachments/20091215/b4c78b73/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: out3.png
Type: image/png
Size: 109871 bytes
Desc: out3.png
URL: <https://lists.clusterlabs.org/pipermail/pacemaker/attachments/20091215/b4c78b73/attachment-0003.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 1rsc_basic.xml
Type: text/xml
Size: 2276 bytes
Desc: 1rsc_basic.xml
URL: <https://lists.clusterlabs.org/pipermail/pacemaker/attachments/20091215/b4c78b73/attachment-0003.xml>
More information about the Pacemaker
mailing list