Just wondering, when deploying this OVF, did you pick Thin provisioning by any chance?
I had the same issue as you describe, but as soon as I redeployed it from scratch with Thick provisioning Eager zeroed - this issue was gone.
I have the same problem with version 6.7 and 6.7u1.
Where does the virtual disk type touch the root password policy?
It would appear that the vSAN Witness Appliance is not being deployed properly.
The recommendation would be to deploy a new appliance.
I have the same problem with a 2 nodes cluster with the appliance 6.7 u1 or u2 . I tryed the u3 and it works well but my vcenter is u2 and cannot talk properly with the appliance.
What is the solution ? I tryed thin provisionning, thick, different datastore, different sizing, nothing works.
"Cannot talk properly" - Is this networking, or simply won't properly join the cluster.
It isn't until VCSA 6.7 U3 that we better support mixed versions in a cluster.
"Cannot talk properly": many problems in fact:
- error message on the vcenter console
- do not see the capacity disk, event with the command line to force it
- hazardous behavious of the console in general
All those problems were gone after reinstalling the esx on 6.7u2 instead of 6.7u3
For my problem, i could solve it by installing the witness on another cluster and then do a storage vmotion, i still don't know why.
It isn't a supported combination to use vSAN 6.7 U3 (Witness or otherwise) and vCenter 6.7 U2:
Multiple time I have seen (in GSS and here) issues such as you say with disk claim/management and general issues with Health - these are resolved by upgrading the vCenter so they can function as intended, is there a reason you cannot/won't upgrade the vCenter?
upgrading the vcenter implies upgrading Veeam, deep security, Citrix (MCS), and so on...
Many things talk to the vcenter now.