When adding a host to an existing vSAN cluster, you may encounter a vSAN Unicast Cluster Partition error. This issue typically stems from either foundational network configuration errors (such as missing VLANs or switch port misconfigurations) or, less frequently, from stale unicast agent list entries. This article provides steps to diagnose and resolve partition errors triggered by host addition.
• VMware vSAN 7.x, 8.x, 9.x
• Newly added host shows "partitioned" or "isolated" in the vSAN cluster.
• vSAN Skyline Health reports a "vSAN Cluster Partition" alert.
• esxcli vsan cluster get shows a Sub-Cluster Member Count lower than the expected total number of hosts.
• VMs may show as "invalid" or have reduced availability.
1. Verify Foundational Network Configuration: Before modifying vSAN cluster settings, ensure foundational network connectivity is verified.
• Verify that the switch ports associated with the new host have the correct VLANs configured.
• Perform a connectivity test using vmkping to verify communication between the new host and existing cluster members over the vSAN VMkernel interfaces:
vmkping -I vmkX <Peer_IP_Address> -d -s 1472
• Review the vSAN Network Troubleshooting Guide for detailed MTU and port protocol requirements.
2. Verify IgnoreClusterMemberListUpdates Parameter: If foundational network configuration is verified, the vCenter Server may fail to update the unicast agent list. Check if the IgnoreClusterMemberListUpdates parameter is set to 1, which prevents automatic updates.
• Run:
esxcfg-advcfg -g /VSAN/IgnoreClusterMemberListUpdates
• If set to 1, reset to the default value of 0:
esxcfg-advcfg -s 0 /VSAN/IgnoreClusterMemberListUpdates
3. Rebuild Unicast Agent List (If Required): If the partition persists after network and parameter verification, manually rebuild the unicast agent list. Refer to Configuring vSAN Unicast networking from the command line.