Can't change sync type or failover after tenant upgrade.

I know I can push the configuration and try to sync again, but I've done too much already and I'm a bit wary now. Doing that might get the pair in sync, but that doesn't guarantee a failover would work if I still can't manually fail them over.

Hi! A couple questions to help me understand this process better. 

1. You mention "another pair" of tenants. So you did the same process (upgrading to 17.1.3) to a different unit? Or is it the same? What models are we talking about? Are these virtual vCMP tenants? Are they hosted on the same vCMP Host that succesfully upgraded (itself, or one different guest i presume) ?
2. You mention you installed software on the Active partition. I've personally never tried this but I'm pretty sure this would fail, it's impossible to select it in the GUI and I strogly believe the cli command would return an error.
   Do you mean that you installed 17.1.3 on a partition where software already existed? This isn't a problem in most scenarios -- when F5 performs this operation, it unpacks the .iso on that partition and it creates a new configuration schema using the current config files from your currently Active partition. So configuration will be up-to-date and with new schema. If process completed with no errors, it's pretty safe to assume you can boot it. The only possible error you'll find is license check date, but it's very easy to see and you can resolve it without requiring to activate the old partition back. 
   The only other time I had an issue with an old schema not being wiped, i simply rolled back to previously active location, then deleted the whole partition from the GUI "disk" menu, and created it again. I believe this dated back to BIG-IP v12 tho.  
3. After you upgraded the unit, GUI show the "Disconnected" state and sync was disabled. This is expected behavior while you run different software version on Active and Standby. Also, it's best practice to manually turn off auto-sync before performing cluster upgrade, and I believe the reason auto-sync was disabled was the same reason - you shouldn't sync configuration (and rightfully shouldn't even be able to do so) while config schema is different.
   Nothing to worry about! As soon as version is the same, full functionality will be restored. 
   
   
4. Regarding your current state. Again, nothing to be worried about. F5 keeps track of configuration changes using an increment counter. So MCPD deamons talk to each other and they say "hey my config version is 101, dated 07.11.2025@17:30, what's yours" "mine's version 102 dated 07.11.2025@17:35" "oh well guess I should warn the user to perform a sync". 
   
   What happened is you rebooted one unit, and that reverts this counter back to 1. Again, this is expected! Nothing to worry about at all.
   They're in this state because BOTH units now have different values for version counter (old one will be higher) and date (rebooted one will be more recent). 
5. What you should do -->
   step 1- revert the upgraded unit to previous schema. both units should be running same software version. 
   step 2 - make sure you can sync. perform a MANUAL sync from the node that stayed active this time ( == it was keepinguser traffic alive and no errors were seen ) to the other node. at this point, the top left indicators in the GUI should be green. 
   step 3 - make sure the unit you need to upgrade is Standby.
   step 4 - DISABLE all auto-sync  
   step 5 - delete the 17.1.3 software partition you created from the standby node. check if you need to reactivate license. if you do, do it on the standby node. then, create a new disk partition in version 17.1.3 with current configuration (the automatic process will do it)
   step 6 - upgrade the standby node. there's no need to check the "install configuration" tick box unless you made changes to this single node since the 17.1.3 partition was created. 
   stap 7 - don't panic when it says you can't sync them, it's expected. now perform a cluster failover so that new node starts to manage user traffic. check for problems and eventually rollback. 
   step 8 - repeat step 5 and 6 on the other node. 
   step 9 - after unit completes upgrade, you should see again a manual sync to perform. it's expected again, because both units have same config version counter "1" and different dates. sync from ACTIVE to STANDBY node -- again, i'm assuming that users that connect to Active node haven't raised issues. 
   step 10 - restore auto sync. 

I was not very clear with my language in my original message. When I said active pair, I meant a pair that is currentIy in production and fully upgraded. 

Instead of upgrading the intended pair, I upgraded the one that I had already done but on the other partition while left me with 17.1.3 or both HD1.1 and HD1.2. 

I can't explain what the issue is/was, but after some time, I am able to proceed as I typically would as all the previously mentioned grayed out options are now available to me.

When I ran tmsh load sys config verify, I did not see any errors. I changed the sync type to Auto-Sync and then ran the sync. They are now synced with no pending changes. 

The only thing I saw was that "Force to Standby" wad grayed out on the active member and the sync. This lasted for a while and then became active. 

I forced the active to standby and verified that it was working. It seemed to work fine. However again, the Force to Standby button was grayed out on the newly active member.