Upgrading with Reduced Redundancy

vSAN on-disk format upgrades require the existing VM storage policies to be satisfied during the upgrade process. For example, in a three node cluster, a Failure To Tolerate =1 policy requires three nodes. Bringing a node offline to perform the upgrade can create reduced redundancy.

By default, the upgrade process does not permit reduced redundancy. Attempts to perform an on-disk format upgrade without sufficient spare resources fail. In cases where the vSAN cluster has insufficient resources to satisfy a VM storage policy, such as a three node cluster with FTT=1 using mirroring, you must set a reduced redundancy flag. You can use the Ruby vSphere Remote Console (RVC) to set the reduced redundancy flag.

You can also set the reduced redundancy flag programmatically. You can set the flag when you initiate the upgrade using the following script:

print 'call PerformVsanUpgradeEx to upgrade disk versions'
task = vsanUpgradeSystem.PerformVsanUpgradeEx(
    cluster=cluster,
    performObjectUpgrade=args.objupgrade,
    allowReducedRedundancy=args.reduceredundancy)