Please refer to the attached screenshot.
After running some VMs for a while on VSAN 6.7U1, they stop showing their proper VM Display names under Cluster > Monitor > Virtual Objects.
When this happens to a VM, it gets grouped to the "Others" object category as shown in the image.
These VMs are still VSAN-healthy and still run fine, but we'd still like to get this 'abnormal condition' fixed if possible, and get these objects back to displaying their VM names properly. (Like the two other normal VMs in the image)
We already checked this link: https://www.virtuallyghetto.com/2017/11/translating-vsan-vm-object-ids-uuid-to-vm-and-vm-to-uuid.htm...
It is the same issue, but unlike in his example, we don't use the Content Library.
Furthermore, he only provided a way to lookup a VM's name from its VSAN UUID, but not a way to make things to back to normal.
Interestingly, there is no need to use his script to lookup the VM's name anyway, if you check my attached screenshot, you can see that the VM's name is already revealed in the UUID path.
Still, we're just looking for a way to get these VM objects back to displaying normally in this menu.
Thanks for your help.
Are the namespace symlinks still present? In a healthy cluster with the VMs registered it should show both namespace UUID (e.g. 1e73525c-f793-7db9-2cd7-ac1f6b6c98) and namespace symlink (e.g. ws01):
# ls -lah /vmfs/volumes/cluster1_vsanDatastore/
Have you tried reloading and/or re-registering the VMs?
# vim-cmd vmsvc/reload <vmID>
or via RVC:
> vsan.check_state -r <pathToCluster>
Looks like they haven't been renamed with last-known path so unlikely vsan.fix_renamed_vms <pathToVM> will work.
(Screenshot) ls -lah /vmfs/volumes/cluster1_vsanDatastore/
The two VMs in my original image do show symlink > UUID, but strangely, each ALSO have another entry where the same UUID does not have a symlink, as highlighted in the screenshot.
Not sure if normal.
Also tried reloading with vim-cmd vmsvc/reload <correct respective VMID>, no change.
/vc1/dc1/computers/cluster1> vsan.check_state -r ./
2019-03-10 08:30:17 -0800: Step 1: Check for inaccessible vSAN objects
2019-03-10 08:30:22 -0800: Step 1b: Check for inaccessible vSAN objects, again
2019-03-10 08:30:22 -0800: Step 2: Check for invalid/inaccessible VMs
2019-03-10 08:30:22 -0800: Step 2b: Check for invalid/inaccessible VMs again
2019-03-10 08:30:22 -0800: Step 3: Check for VMs for which VC/hostd/vmx are out of sync
Did not find VMs for which VC/hostd/vmx are out of sync
... looks fine I guess?
Still don't know what's going on. VMs are still healthy. Can vMotion, Storage vMotion, remove from inventory then re-register. No problem, they still boot and run fine.
The only problem still is that they are only showing UUIDs instead of names in Cluster > Monitor > Virtual Objects.
Are ALL hosts in the cluster upgraded to ESXi 6.7 U1 same as the vCenter?
I ask as we changed how VM IDs are handled in 6.5 U2/6.7 U1 and Objects appearing as 'Other' can occur if there are older hosts in the cluster.
It's a new deployment, so they're all 6.7U1. vCenter and hosts.
Just want to mention that they were displaying normally to begin with, but at some point in time they switched to UUID only and appeared in "Others".
We have no idea why, and it doesn't happen to all VMs.