Tales from the field; while being at a customer site who needed to reboot their ESX hosts (and thus put their hosts in Maintenance Mode), who found that their 2-host cluster hung at 2% while initiating VMware Maintenance Mode, the fix is below:

• Ensure Distributed Resource Scheduling (DRS) is enabled
• The error is generated because the default HA setting is to support 1 host failure. Well, putting 1 host in Maintenance Mode will leave a 2-host VI Cluster unable to support 1 host failure, thus generating the error.
• The fix is to disable Admission Control, which is enabled by default.
• NOTE: I'm not a huge fan of enabling (default) Admission Control in the first place. I prefer to have all of my designated VMs come back after a host failure at <100% performance than have some of them come back at ~100% performance.
• Enter Maintenance Mode.
• Do your goodness, such as patching the host, drinking some whiskey, or watching Formula 1 replays.
• Exit Maintenance Mode.
• Re-enable Admission Control (if so desired).

NOTE: This could also happen in a VI cluster that can not support the amount of hosts in Maintenance Mode. The steps for remediation should roughly be the same.