Esxi 7.0 Bug? (snapshot error if unreachable iso is present in the vm config)

[Dotmax]

Good morning to all, maybe not the right place to post, but i'm just curious on how many have find this bug/strange behaviour on esxi 7.0

Here are the steps of what happens:
Mount an nfs share on a qnap nas
In the share put an iso file (example: os image)
Mount the iso as a cd in a vm (in esxi gui, via edit -> cd drive -> datastore ISO file)
Disconnect the qnap nas from the network (the nfs share therefore becomes inaccessible)
Backups start failing with this error (Failed to create VM snapshot. Error: CreateSnapshot failed, vmRef 16, timeout 1800000, snName VEEAM BACKUP TEMPORARY SNAPSHOT, snDescription Please do not delete this snapshot.It is being used by Veeam Backup., memory False, quiesce False )

Try to manually take a snapshot of the VM (via gui -> actions -> snapshot -> take snapshot)
The snapshot fails (timeout) and no snapshot are present in the manage snapshots window
But inside the vm filesytem folder, a file vmname-0000x-delta.vmdk (delta file) is created (like in a successfully snapshot)
If you take more snapshots, a delta file is created for each snapshot taken (always nothing in the gui)

At this point reconnect the nfs share
or delete from the vm config (via gui or by editing the vmx file) the path to the iso file (that is not accessible)

Take another snapshot (which at this point is done correctly) The complete chain of snapshots appears in the gui (base -> snapshot1 -> snapshot2 -> etc etc -> last snapshot)

TLDR:When the path to an iso file is present in the vile vmx, but the NFS share is not accessible, esxi is unable to take snapshots but delta files are created on the file system. In the gui, however, nothing is shown.
You need to remove the unreachable path or make the share reachable again, take an extra snapshot (to update the gui), and delete all/intermediate unwanted snapshots


Has anyone else noticed this strange behavior?

Thank You
Max

[HannesK]

Hello,
from Veeam side, everything looks "as expected".

I guess it makes most sense to post such things on the VMware forums.

Best regards,
Hannes

[Regnor]

I haven't seen this behaviour so far. But I once had a situation were an inaccessible NFS store for ISOs caused the VMs to go into 'invalid' state.
Thinking about all the possible problems and the security vulnerability from last year, it's probably best to always unmount the ISO after usage.