Unable to enumerate all disks. One of the parameters supplied is invalid.

[Cubea]

I had some issues with one of my two nodes on my ESXI Cluster with vSAN. It wasn't working properly with networking, refusing to connect, and brought vSAN down with a bunch of VMs running. So, after hours of troubleshooting and trying to fix it, I gave up and simply reinstalled it and the other node for good measure. I recreated the vSAN and Witness, and everything came back up as if it never happened. But, then none of my 20 or so VMs will start with the error linked in the subject, even the ones that were offline when vSAN went down. There is also now 14 Non-availability related incompliance objects among 111 health objects. On top of that, vSAN says 2TBs of storage is being used, but every virtual machine is saying it's 1GB of size or less, with vSAN reporting almost all the storage used are User Objects instead of VMDKs. I'm also unable to put either machines into maintaince mode, with it reporting:

The host cannot enter maintenance mode. Resource check failed - Internal failure happened during the resource check. Please check clomd.log on host node2.REDACTED.com for more details.

If you have any ideas, or other information you'd like me to retrieve please let me know. You can contact me via this page, or directly via Discord.

Cubea#1884

[TheBobkin]

@Cubea, "Unable to enumerate all disks. One of the parameters supplied is invalid." - this likely means it cannot access the vmdk descriptors (and thus the data Object these point to) with the current registered VM/vmx information (or some Object tied to the VM is inaccessible).

If I were to hazard an educated guess, I reckon you may have a new vsanDatastore and your VMs are all still pointing to the old one (e.g. VM registered with /vmfs/volumes/vsan:old-info/VMName/VMName.vmx but the datastore 'vsan:old-info' is no longer present and replaced with vsan:new-info and thus everything below it is not present).

First step is to validate what you see in either datastore browser (e.g. do you see the VM namespace folders or is it empty?) and via SSH
# cd /vmfs/volumes/vsanDatastore (or whatever you have named this to)
# ls -lah
If all the namespaces are gone then your vsanDatastore is basically a new one and not where everything is referenced from - this isn't a dead-end, there are numerous ways this can be addressed but IMO I wouldn't advise randomers on the internet to be joining you via Discord to poke and prod at it, open a ticket with VMware vSAN GS if the data and/or prodcution state of this cluster is in any way important.

It *may* be possible to join the nodes back to the 'old' vSAN cluster and thus vsanDatastore if you have the Sub-Cluster UUID, alternative workarounds would be to script changing all Objects paths to replace old vsanDatastore name/ID with new one, but this isn't trivial and someone that knows exact details and what they are doing would need to be applied (AKA VMware vSAN GS).

Assuming I am not mis-guessing here are some previous discussions relating to this:
https://communities.vmware.com/t5/VMware-vSAN-Discussions/vSAN-Expert-Challenge-2-Update-1/td-p/2752... 
https://communities.vmware.com/t5/VMware-vSAN-Discussions/I-have-lost-my-vSAN-Could-you-help/td-p/51...