/ 12.2. Debugging a HA Cluster
	Chapter 12. Troubleshooting		12.3. FAQs

“Debugging” a HA Cluster

If something is not working for your cluster, try the following first. It displays you the resource operation history (option -o) and inactive resources (-r):

crm_mon -o -r

The display is refreshed every 10 seconds (you can cancel it with Ctrl+C.) An example could look like:

Example 12.1. Stopped Resources

Refresh in 10s...

============
Last updated: Mon Jan 19 08:56:14 2009
Current DC: d42 (d42)
3 Nodes configured.
3 Resources configured.
============

Online: [ d230 d42 ]
OFFLINE: [ clusternode-1 ]

Full list of resources:

Clone Set: o2cb-clone
         Stopped: [  o2cb:0 o2cb:1o2cb:2 ]
Clone Set: dlm-clone
         Stopped [ dlm:0 dlm:1 dlm:2 ]
mySecondIP      (ocf::heartbeat:IPaddr):        Stopped

Operations:
* Node d230:
   aa: migration-threshold=1000000
    + (5) probe: rc=0 (ok)
    + (37) stop: rc=0 (ok)
    + (38) start: rc=0 (ok)
    + (39) monitor: interval=15000ms rc=0 (ok)
* Node d42:
   aa: migration-threshold=1000000
    + (3) probe: rc=0 (ok)
    + (12) stop: rc=0 (ok)

First get your node online (see Section 12.3). After that, check your resources and operations.

The Configuration Explained PDF under http://clusterlabs.org/wiki/Documentation explains in section How Does the Cluster Interpret the OCF Return Codes? three different recovery types: