vSAN Host partition with hosts able to ping other vSAN hosts.

Products

• VMware vSAN 7.x

Issue/Introduction

Symptoms :

• After Network or Power restore, one or more vSAN Host remains in Partitioned state
• Cluster had recent Networking or Power outage which has been restored
• No ping failure on vSAN network
• No Port connectivity issues
• Host remains in partitioned state even after esxcli vsan cluster leave or join command
• Unicastagent table doesn't have any missing entry

Environment

vSAN 7.x

Cause

This happens when the current Master host fails to get the updated information after Power or Network outage.

Under /var/run/log vmkernel logs will show the below events showing hosts entries getting added and then removed again and again:

Resolution

Force re-election in the cluster by running below command on current Master Host. This will refresh the information and cluster partion status will get clear  :

- (ESXi 7.0U2 and higher version) LEADER abdicate
# vsish -e set /vmkModules/cmmds/forceTransition abdicateLeader

 
- (before ESXi 7.0U2) MASTER abdicate
# vsish -e set /vmkModules/cmmds/forceTransition abdicateMaster