Mgmt server HA broken - standby not communicating R80.20

[aks_2512]

We have 2x management servers configured in HA. Both the servers were in collision state.
================================================================
Conflict - multiple active servers
mgmt-1: Active. Conflict multiple active servers
mgmt-2: Active. In collision with mgmt-2. Last sync time unavailable.
================================================================
Following Check Point's document (https://sc1.checkpoint.com/documents/R80.20/SmartConsole_OLH/EN/html_frameset.htm?topic=IgYiJZ3l4vRm...), we set the mgmt-2 as Standby. After nearly an hour, the sync completed but now the state has changed to "Failed to communicate with peer "mgmt-2""

================================================================
Failed to communicate with peer "mgmt-2"
mgmt-1: Active. Failed to communicate with peer "mgmt-2"
mgmt-2: No Communication. Communication with mgmt-1 failed since x.y PM.
================================================================

Is a database install required at this stage?

Also, when we view mgmt-2 object, it shows SIC status as "Trust Established" however when we Edit mgmt-2 object it shows SIC status as Uninitialized.
Do we need to reset SIC in this case. If yes, and we do not have the initial SIC (one time passwd) on the primary mgmt server, will setting up a new SIC on the primary mgmt server cause existing trust between pri mgmt server and gateways to break? Or it has no effect on existing trusts. 

Thank you.

[PhoneBoy]

Is there some reason you're still on R80.20, which has been an unsupported release for a while now?

[aks_2512]

Hello PhoneBoy,

Devices were previously managed by third party. We are in the process of replacing the mgmt server and firewalls, one of the prerequisite was to get the existing mgmt server to R81.20 and then move. 

Thank you.

[PhoneBoy · Accepted solution]

As part of the planned upgrade, you're going to have to do a fresh install of the Management HA node and set up synchronization again.
Seems like this would be a better approach than trying to get the existing management HA on R80.20 working.

[aks_2512]

Thank you PhoneBoy, noted. We will proceed as suggested.