Some vSAN Objects inaccessible after vSAN kernel communication is restored.

Products

• VMware vSAN

Issue/Introduction

After a vSAN outage, some objects are still inaccessible even after the following has been verified:

1) vSAN kernels on all hosts are present and vSAN is enabled on them.

2) Communication between the vSAN kernels has been verified using vmkping -I and testing for Jumbo Frames where applicable.

3) Unicast tables on all participating hosts are correct, in that they include all other hosts in the sub-cluster and NO entries for their own kernels.

 

One or more hosts displays that it is in a sub-cluster of '1' and/or it has a mismatched UUID from the other ESXi hosts in the vSAN cluster/sub-cluster.

Environment

vSAN 8.x

vSAN 9.x

Resolution

1)Run the following to ensure that the value is set to 0:esxcfg-advcfg -s 0 /VSAN/IgnoreClusterMemberListupdates

2) With the host still in maint mode, click and drag it out of the current cluster into the Datacenter object in vCenter's inventory.

3) Wait about 5 min, then click and drag it back into the host cluster.This should have the effect of wiping and rewriting its Unicast Table as well as rejoining the vSAN sub-cluster.

4) Run esxcli vsan cluster get to verify that it's rejoined the rest of the cluster.

5) After 5-30 minutes, recheck for resynching objects in Skyline Health.