White Papers

Dell HPC NFS Storage Solution - High Availability Configurations
Page 47
A.8. Quick test of HA set-up
1) Test fencing:
 First disable the cluster service:
[root@active ~]# clusvcadm –d HA_service
 From the active server run the command fence_node passive. This should power cycle
the passive server via DRAC. Check /var/log/messages on active.
 From passive run the command fence_node active. This should power cycle the active
server via DRAC. Check /var/log/messages on passive
 Disconnect the DRAC cable on passive. From “active” run the command fence_node
passive. This should power cycle the passive server via the switched PDU.
 Disconnect the DRAC cable on active. From passive run the command fence_node
active. This should power cycle the “active” server via the switched PDU.
2) Simple HA failover test:
 Start the service on the “active” server
[root@active ~]# clusvcadm –e HA_service –m active
 Power down the active server by holding down the power button/pulling out both power
cords.
 Service should migrate to passive. Watch /var/log/messages and check clustat
status on passive.
A.9. Useful commands and references
1) clustat
2) cman_tool status
3) If the /etc/cluster/cluster.conf file is edited manually, make the changes only on one
server and increment the version number field. Use the following command to propagate the
changes to both servers. ccs_tool update /etc/cluster/cluster.conf
4) service cman status
5) service rgmanager status
6) clusvcadm to manually stop, start, disable and relocate the cluster service.
7) group_tool
8) rg_test is a debugging tool. Use it with care. Ensure the cluster service is stopped and disabled
before you use this tool.
 clusvcadm –s HA_service
 clusvcadm –d HA_service
 rg_test test /etc/cluster/cluster.conf start service HA_service
- mount will show the file system mounted.
- ip addr list will show the floating resource IP configured.