Host is partitioned after a reboot when vCenter is on vSAN or is shutdown
Article ID: 326907
Updated On:
Products
Issue/Introduction
Host is partitioned after a reboot when vCenter is on vSAN or is shut down.
Follow the below steps to confirm if you've hit this issue:
1. Using configstorecli list the vSAN network-related fields.
configstorecli config current get -c vsan -g system -k vsan|grep network_address -B 2 -A 6 (use vsan_witness when running the command on a witness node)
2. Check the above output. The issue matches if it has duplicate entries of the same address and port.
For example:
"network_address": [
{
"address": "x.x.198.222",
"port": 12321,
"vmknic": "vmk1"
},
{
"address": "x.x.198.222",
"port": 12321,
"vmknic": ""
}
Note: The command will not include the info for the node the command is run on.
Environment
VMware vSAN 7.0.x
Resolution
Upgrade vCenter/ESXi to 7.0U2c or higher
Workaround:
Perform one of these steps to workaround this issue:
1. If vCenter Server is active, go to vSAN Cluster > Monitor > vSAN > Skyline Health > Cluster > vCenter state is authoritative and click on "Update ESXi Configuration"
2. If vCenter Server is inactive,
a. Run this command to remove entries from the host that have duplicate configuration store entries:
esxcli vsan cluster unicastagent remove
For example: esxcli vsan cluster unicastagent remove -a x.x.198.222 -c 5283e04b-xxxx-xxxx-xxxx-16695223e964 -p 12321 -t node -u 60d28a08-xxxx-xxxx-xxxx-0200362fd9aa
Note: The information can be obtained using the command esxcli vsan cluster unicastagent list
b. Run this command to add it back without specifying "-i":
esxcli vsan cluster unicastagent add
For example: esxcli vsan cluster unicastagent add -a x.x.198.222 -c 5283e04b-xxxx-xxxx-xxxx-16695223e964 -p 12321 -t node -u >60d28a08-xxxx-xxxx-xxxx-0200362fd9aa -U true
c. Ensure that there are no duplicate configuration store entries.
d. Reboot the hosts.
Additional Information
Feedback
