Hosts in a 2 node stretched cluster fail to enter maintenance mode with 'Full Data Migration'

Products

• VMware vSAN

Issue/Introduction

While performing a ESXi hosts upgrade, an attempt to put a host of a 2 node stretched cluster fail to enter maintenance mode with 'Full Data Migration' with the following message:
1 more standalone host is required. Some objects will become inaccessible or non-complaint with their storage policy

Environment

ESXi
VSAN

Cause

'Full Data migration' with RAID 1 FTT=1 will require a 4th host apart from the two hosts and the Witness. This is explained in the article vSAN host fails to enter Maintenance Mode with 'Full data migration' option selected

Resolution

Select Maintenance Mode 'Ensure Accessibility' option - note that all data with components on the host in maintenance mode will be in a reduced redundancy state until the host has exited Maintenance Mode and the data resynced back to compliance.

Note 1: While using 'Ensure Accessibility, when the ESXi Host is upgraded and moved out of Maintenance Mode, a Resync will occur if this took more than 60 minutes which is the default Resync delay timer. If you foresee a reason for the upgrade of each host to take more than 60 minutes, you may extend the timer by following the article Changing the default repair delay time for a host failure in vSAN

Note 2: For stretched vSAN clusters configured with a dedicated witness and target builds of 7.0 or higher, always upgrade the witness host before the ESXi hosts.
When upgrading a stretch cluster configured with a shared witness, always upgrade the witness after all the ESXi hosts in the cluster.
This is as per the VMware vSAN Upgrade Best Practices