In our environment we have multiple data engines in use (4 in total).
we have 1 primary hub server, 1 secondary primary hub server and failover servers for each server.
the primary hub runs everything except the discovery server, which we had to move off to separate server due to queue limits and the qos_processor. the secondary primary hub server has all the alarm and qos queues on it and has the qos_processor on it, this server writes all the qos to the DB and passes the alarms over to the primary hub server. the failover servers config matches the primary and secondary primary hub with various queues stopped and some probes stopped as well.
we have configured the data_engines with data_engine id's so that the primary hub is primary and runs the DB maintenance job.
we found that if the primary hub is unavailable then none of the other servers pickup this fact and the maintenance job doesn't run.
Has anyone else encountered this issue or have any suggestions on changes that could be made to the configuration/setup?