Greetings,
We had an unsuccessful CUCM upgrade from 6.1.5 to 8.6.2(22900-9) over the weekend. Publisher failed to upgrade the first time (switch to new version), but worked the second time (booted in 6.1.5).
Subsequently we tried some of the remaining servers in the cluster and had different failures and some successful attempts.
Some servers failed the upgrade without any errors. We could only see this message in CLI "current upgrade status is set to Installing but no upgrade process was detected. This can occur when the Cisco Unified Communications OS Platform CLI window is closed before the upgrade command has completed. Please review the upgrade logs and verify the previous upgrade has completed. Once verified, run the upgrade cancel command to reset the upgrade status.". Servers simply boot with 6.1.5 with nothing in the inactive partition. We cancelled and re-tried with various success.
Other servers failed the upgrade with the following message "the installation has failed and must be restarted to recover from the failure." Servers would actually not boot (tried recovery CD) after that error and we had to rebuild them.
Another two servers failed the upgrade by failing to boot after the second reboot that is part of the upgrade. They had “Kernel Panic: UFS: Unable to mount root fs on 00:00” and had to re-build them.
All in all we had to cancel the upgrade with only 4 servers upgrading successfully in the inactive partition. Cisco TAC watched all this and could not provide any resolution or meaningful explanation. TAC is trying to involve BU now, but nothing yet.
We had the COP file installed in advance. Hardware is MCS7845-I3.
Any ideas? Thank you in advance.
Val
