Upgrade R81.10 Cluster to R82 issues

[PingMe]

I have 2 pairs of gateways both running R81.10 latest recommended take. I have upgraded my management and log servers to R82. When I do an upgrade on a member of either pair (5600's and 6600's) I cannot fail over to the upgraded member to update the other member. I am using CPUSE to perform the updates via the GAIA interface. After the upgrade the only blades that come up are the Firewall, Application Control, IPSec VPN and  URL filtering. All others are coming up as 'not responding'. The ClusterXL says inactive or machine is down. Am I missing something with this upgrade? Been over the installation guide and I don't seem to see anything that jumps out. The log server had to be a clean install, and that was not obvious in the documentation, so I feel like I am missing a step here. 

[_Val_]

Check the licenses and also remember to push the policy again after updating the version in SmartConsole. Allow installation in case of a version mismatch.

[the_rock]

Make sure on UPGRADED member that MVC is on, you can check by running cphaprob mvc, to turn it on, its cphaconf mvc on.

This is what I would always follow, never had an issue.

Andy

https://sc1.checkpoint.com/documents/R81/WebAdminGuides/EN/CP_R81_Installation_and_Upgrade_Guide/Top...

[PingMe · Accepted solution]

Thanks for the responses - My issue was using CPUSE in the GAiA - I needed to release my control freak issues, and allow the upgrade to happen in the SmartConsole. After I just went in and allowed it to update on its own, full cluster, latest release, the upgrade went thru. I had no luck using SmartConsole trying to just upgrade one member. It looks like I will be giving more trust to the SmartConsole upgrade/updates in the future. 

The 'install on all cluster members' option does what I normally do. Install on the standby member, reboots it, checks to see if/when it comes back up, then proceeds to make it active, upgrades the other member, and switches it back to active after everything is done.

Thanks again for the respones!

[Alex-]

Well, the CPUSE upgrade should still work, especially when you need to run a slew of verification before going to the other node as part of some change management processes, which go further than just checking cluster state.

[the_rock]

Good job!