A cluster with only 2 nodes (master and replica) is a valid and supported configuration. Depending on the count of objects you are monitoring you may need more (data) nodes.
When HA is enabled, the replica can take over all functions that the master provides, were the master to fail for any reason. If the master fails, failover to the replica is automatic and requires only two to three minutes of vRealize Operations Manager downtime to resume operations and restart data collection.
thanks for you reply.
So, if the configuration is supported and the take over doesn't work I presume it's a problem with my installation.
I'll open a SR.