19.6.1 Placing NNMi in maintenance mode
The maintenance mode is a function for temporarily disabling failover during NNMi maintenance.
NNMi running in an HA environment is monitored by the HA product. If NNMi stops, the HA product determines that a failure has occurred and fails over from the primary node to the secondary node. This means that failover also occurs when NNMi is stopped for maintenance purposes.
The maintenance mode disables failover processing by suppressing the monitoring of NNMi. This means that you can perform maintenance work by executing ovstop and ovstart on the active cluster node. Make sure that you do not execute ovstop and ovstart on the passive cluster node.
- Important
If a product that requires NNMi is running, placing only NNMi in the maintenance mode will still trigger failover if the related product fails. In such a case, first stop the related product or place it in the mode equivalent to the maintenance mode, and then place NNMi in the maintenance mode.
- Organization of this subsection
(1) Placing NNMi in maintenance mode
When NNMi is placed in the maintenance mode, NNMi's monitoring process is disabled. While NNMi is in the maintenance mode, stopping or starting NNMi for that HA resource group does not trigger failover.
To place NNMi in the maintenance mode, create the following file (which can be empty) on the active cluster node:
Windows
%NnmDataDir%hacluster\resource-group\maintenance
Linux
$NnmDataDir/hacluster/resource-group/maintenance
(2) Removing NNMi from maintenance mode
Taking NNMi out of maintenance mode re-enables NNMi monitoring. If NNMi is stopped, the HA resource group will fail over to the passive cluster node.
To remove an HA resource group from maintenance mode:
Verify that NNMi is running correctly:
ovstatus -c
All NNMi services must show the state RUNNING.
Delete the maintenance file from the node that was the active cluster node before maintenance was initiated.
For details about the maintenance file, see (1) Placing NNMi in maintenance mode.