[Pacemaker] cman+pacemaker - way to join/leave cluster?
Yuriy Demchenko
demchenko.ya at gmail.com
Tue Apr 16 08:42:16 UTC 2013
yea, that's what i meant by edit CIB&cluster.conf by hands
however, two problems with it:
1. it requires restart of cman on each node to "forget" expelled node
after cluster.conf edit
2. after editing CIB (removing all entries belonging to expelled node)
and replacing CIB via "pcs cluster push cib.xml" (or cibadmin --replace;
doesn't matter) cluster DC still automaticly inserts "<node>" and
"<node_status>" entries for expelled node which I cannot cleanup.
> <node_state id="node-2" uname="node-2" in_ccm="false"
> crmd="offline" join="down" expected="down"
> crm-debug-origin="do_cib_replaced"/>
Yuriy Demchenko
On 04/16/2013 12:16 PM, Daniel Black wrote:
> http://clusterlabs.org/doc/en-US/Pacemaker/1.1-crmsh/html-single/Pacemaker_Explained/index.html#_options_2
>
> example 2.4
>
> # cibadmin --query > tmp.xml
> # vi tmp.xml
> # cibadmin --replace --xml-file tmp.xml
>
>
> ----- Original Message -----
>> Hi,
>>
>> Can anyone point me to the correct procedure of adding/removing node
>> to/from cluster in cman+pacemaker stack, preferably without stopping
>> whole cluster?
>> "pcs cluster node add|remove" doesn't work, as there's no "pcsd"
>> daemon
>> in centos6.4; cman tools (cman_tool, ccs_tool, ccs_sync) - doesn't
>> work
>> either, as it expects configured pki infrastructure as "ricci" service
>> does...
>>
>> I can just edit CIB & cluster.conf and remove/add node by hands, but
>> it
>> seems kind of crude way - requires restart of cman&pacemaker on all
>> nodes
>>
>> --
>> Yuriy Demchenko
>>
>>
>> _______________________________________________
>> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
>> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>>
>> Project Home: http://www.clusterlabs.org
>> Getting started:
>> http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
>> Bugs: http://bugs.clusterlabs.org
More information about the Pacemaker
mailing list