From opHA 3.3.2, redundancy in primary servers is supported, following the schema:
The primaries don't talk to each other, each of them pulls the pollers information but, the Main primary will also push the nodes to the peers. The secondary primary will just pull or read the nodes data from the pollers.
That means, the secondary primary could have a delay in updating the information from the Main primary. This delay shouldn't be bigger than the frequency where the cron job is configured, by default, 5 minutes. It can be updated in /etc/cron.d/opha
Here you can find more detailed information about each role functions.
From previous versions, it is important to change the opha cron job:
# this cron schedule controls opHA data collection # # m h dom month dow user command # # pull info every five minutes #*/5 * * * * root /usr/local/omk/bin/opha-cli.pl act=pull # push nodes every five minutes #*/5 * * * * root /usr/local/omk/bin/opha-cli.pl act=sync-all-nodes # push processed nodes every minute #*/1 * * * * root /usr/local/omk/bin/opha-cli.pl act=sync-processed-nodes */5 * * * * root /usr/local/omk/bin/opha-cli.pl act=pull data_types=nodes,inventory,latest_data,events,status,registry,selftest |
It is also important, not edit the nodes from the secondary primary server.