Recently I was running some Cloud Foundation upgrades for a customer and encountered a problem trying to execute an ESXi upgrade bundle.
The error was the following:
clicking on the host name that’s all you could see from the GUI
ESX_UPGRADE_VUM_PRECHECK_CLUSTER_SIZE_INVALID
Not super meaningful in all honest. All you can see is that Update Manager (VUM) didn’t even succeed at pre-staging the patch into the host, for some invalid size cluster.
As always logs are you best friend 😉 but /home/vrack/lcm/logs/lcm.log can be very verbose and tricky to follow through.
PRO-TIP: if you’re going to tail the file with the follow option (-f) make sure you’re also filter by a specific pattern, in my example I was expecting some ERROR to show up.
Bingo! This message is now more useful:
1 cluster(s) expected. Found: 2
SDDC Manager is expecting 1 cluster but found 2
This was because was testing happened on said WLD; a second cluster was added to the vCenter Server managing the workload domain (multi-cluster isn’t supported on VCF 2.x) and even though all hosts were removed the cluster object was still present on vCenter Server.
As a result SDDC Manager was getting a total of 2 clusters and wouldn’t continue with whatever LCM operation it’s executing.