If the Production_Server resource group fails it will not automatically move to the recovery node because it is anti-collocated, low priority with the Recovery_Server resource group.
Not doing a automatic Failover allows the creation of a snapshot to validate the Replica data and to restore the Replica to a point in time if required. The Recovery_Server resource group will have to be brought offline before the Production_Server resource group can be brought online on the recovery server.
When the Production_Server resource group is brought online on the recovery node it executes the rtdr -C <Primary Context ID> failover command. When the Production_Server resource group is brought offline (on the recovery server), after the recovery node becomes available, it will send the rtdr -C <Primary Context ID> resync command to the recovery node before stopping.
There may be special cases where you would delay bringing the Production_Server resource group offline on the recovery server. Normally, this would mean that no data is being synchronized to the production server. In that case, you can manually execute the rtdr -C <Primary Context ID> resync command on the production server. This starts the ABA but it will not be monitored by PowerHA for AIX.
If the Recovery_Server resource group fails it will not move, because there is only one node in the participating node list.
Ensure that you have the Recovery_Server resource group online to prevent an automatic failover of the Production_Server resource group to the recovery server.