[ClusterLabs] pcs create master/slave resource doesn't work

[Jan Pokorný]

On 23/11/17 23:52 +0800, Hui Xiang wrote:
> I am working on HA with 3-nodes, which has below configurations:
> 
> """
> pcs resource create ovndb_servers ocf:ovn:ovndb-servers \
>   master_ip=168.254.101.2 \
>   op monitor interval="10s" \
>   op monitor interval="11s" role=Master
> pcs resource master ovndb_servers-master ovndb_servers \
>   meta notify="true" master-max="1" master-node-max="1" clone-max="3"
> clone-node-max="1"
> pcs resource create VirtualIP ocf:heartbeat:IPaddr2 ip=168.254.101.2 \
>     op monitor interval=10s
> pcs constraint order promote ovndb_servers-master then VirtualIP
> pcs constraint colocation add VirtualIP with master ovndb_servers-master \
>   score=INFINITY
> """

(Out of curiosity, this looks like a mix of output from 
pcs config export pcs-commands [or clufter cib2pcscmd -s]
and manual editing.  Is this a good guess?)

>   However, after setting it as above, the master is not being selected, all
> are stopped, from pacemaker log, node-1 has been chosen as the master, I am
> confuse where is wrong, can anybody give a help, it would be very
> appreciated.
> 
> 
>  Master/Slave Set: ovndb_servers-master [ovndb_servers]
>      Stopped: [ node-1.domain.tld node-2.domain.tld node-3.domain.tld ]
>  VirtualIP (ocf::heartbeat:IPaddr2): Stopped
> 
> 
> # pacemaker log
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++ /cib/configuration/resources:  <primitive class="ocf"
> id="ovndb_servers" provider="ovn" type="ovndb-servers"/>
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++                                  <instance_attributes
> id="ovndb_servers-instance_attributes">
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++                                    <nvpair
> id="ovndb_servers-instance_attributes-master_ip" name="master_ip"
> value="168.254.101.2"/>
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++                                  </instance_attributes>
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++                                  <operations>
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++                                    <op
> id="ovndb_servers-start-timeout-30s" interval="0s" name="start"
> timeout="30s"/>
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++                                    <op
> id="ovndb_servers-stop-timeout-20s" interval="0s" name="stop"
> timeout="20s"/>
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++                                    <op
> id="ovndb_servers-promote-timeout-50s" interval="0s" name="promote"
> timeout="50s"/>
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++                                    <op
> id="ovndb_servers-demote-timeout-50s" interval="0s" name="demote"
> timeout="50s"/>
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++                                    <op
> id="ovndb_servers-monitor-interval-10s" interval="10s" name="monitor"/>
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++                                    <op
> id="ovndb_servers-monitor-interval-11s-role-Master" interval="11s"
> name="monitor" role="Master"/>
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++                                  </operations>
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++                                </primitive>
> 
> Nov 23 23:06:03 [665249] node-1.domain.tld      attrd:     info:
> attrd_peer_update: Setting master-ovndb_servers[node-1.domain.tld]: (null)
> -> 5 from node-1.domain.tld

If it's probable your ocf:ovn:ovndb-servers agent in master mode can
run something like "attrd_updater -n master-ovndb_servers -U 5", then
it was indeed launched OK, and if it does not continue to run as
expected, there may be a problem with the agent itself.

You can try running "pcs resource debug-promote ovndb_servers --full"
to examine the executation details (assuming the agent responds to
OCF_TRACE_RA=1 environment variable, which is what shell-based
agents built on top ocf-shellfuncs sourcable shell library from
resource-agents project, hence incl. also agents it ships,
customarily do).

> Nov 23 23:06:03 [665251] node-1.domain.tld       crmd:   notice:
> process_lrm_event: Operation ovndb_servers_monitor_0: ok
> (node=node-1.domain.tld, call=185, rc=0, cib-update=88, confirmed=true)
> <29>Nov 23 23:06:03 node-1 crmd[665251]:   notice: process_lrm_event:
> Operation ovndb_servers_monitor_0: ok (node=node-1.domain.tld, call=185,
> rc=0, cib-update=88, confirmed=true)
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: Diff: --- 0.630.2 2
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: Diff: +++ 0.630.3 (null)
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: +  /cib:  @num_updates=3
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_perform_op: ++
> /cib/status/node_state[@id='1']/transient_attributes[@id='1']/instance_attributes[@id='status-1']:
> <nvpair id="status-1-master-ovndb_servers" name="master-ovndb_servers"
> value="5"/>
> Nov 23 23:06:03 [665246] node-1.domain.tld        cib:     info:
> cib_process_request: Completed cib_modify operation for section status: OK
> (rc=0, origin=node-3.domain.tld/attrd/80, version=0.630.3)

Also depends if there's anything interesting after this point...

-- 
Jan (Poki)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.clusterlabs.org/pipermail/users/attachments/20171123/f6a82fca/attachment.sig>