77.30 to 80.20 - Cluster connectivity upgrade question

[Josh_B]

I have two 77.30 gateways using ClusterXL in high availability mode. Looking at doing a connectivity upgrade from 77.30 to 80.20.

After I upgrade the standby member I'm looking to fail over after hours and test connectivity, but then fail back and leave the 77.30 member running for a few more days. Is this supported? I guess my question is, can you fail back over to the 77.30 member once you fail over to the upgraded 80.20 cluster member? I've read over the R80.20 "Upgrading a Security Gateways and Cluster" section in the Installation and Upgrade Guide, but this part isn't clear.

[PhoneBoy]

I believe the sync is only "one-way" in a Connectivity Upgrade.
Meaning once you fail over to R80.20, connections aren't synced to R77.30.
I'll see if I can get a definitive answer, though.

[Josh_B]

Thanks, much appreciated.  

That is what I believe I was understanding. If I failed back over to R77.30, I would lose all connections. I would be OK with this as it I'm doing this during a planed maintenance window. I wanted to be sure I wouldn't be causing any issues when I failed back over to R80.20  during the second maintenance window a few days later, where I planned to keep the R80.20 cluster member as the active. 

[PhoneBoy · Accepted solution]

Confirmed with R&D that Connectivity Upgrade is a one-way sync.
It is possible to fail back to R77.30 by performing a cpstart on that cluster member (assuming R80.20 upgrade was done on the standby member), but you lose all connections and redundancy is obviously limited.
We do plan to improve this in later releases.

[HeikoAnkenbrand]

This is a very nice information that we can use FCU to sync to a lower version in the future.

[PhoneBoy]

It will most likely be a different, but similar mechanism.
Stay tuned 🙂