VMDK disk show zero size after snapshot operation

Products

• VMware vCenter Server
• VMware vSphere ESXi

Issue/Introduction

Symptoms:

• After taking snapshot on a virtual machine and the task completes successfully, however the VMDK disk size shows Zero.
• Consequently, Unable to do any task on the virtual machine.
• Any Backup job will be failed.
• Reviewing hostd.log, you will see error similar to :

2019-12-12T16:02:09.743Z warning hostd[1B4C2B70] [Originator@6876 sub=Vmsvc.vm:/vmfs/volumes/5aa77f4c-db5986e0-695f-aa2dc8900078/x/x.vmx] Snapshot create succeeded, but UID 650 not in list.

2019-12-12T16:02:09.743Z warning hostd[1B4C2B70] [Originator@6876 sub=Vmsvc.vm:/vmfs/volumes/5aa77f4c-db5986e0-695f-aa2dc8900078/x/x.vmx] Failed operation

2019-12-12T16:02:09.744Z info hostd[1B4C2B70] [Originator@6876 sub=Vmsvc.vm:/vmfs/volumes/5aa77f4c-db5986e0-695f-aa2dc8900078/x/x.vmx] State Transition (VM_STATE_CREATE_SNAPSHOT -> VM_STATE_ON)

2019-12-12T16:02:09.744Z info hostd[1B4C2B70] [Originator@6876 sub=Vimsvc.TaskManager] Task Completed : haTask-49-vim.VirtualMachine.createSnapshot-127313764 Status error

2019-12-12T16:02:09.595Z warning hostd[1B4C2B70] [Originator@6876 sub=Vmsvc.vm:/vmfs/volumes/5aa77f4c-db5986e0-695f-aa2dc8900078/x/x.vmx] UpdateStorageAccessibilityStatusInt: The datastore is not accessible ide1:0.fileName = "/vmfs/volumes/5cee6432-4df76266-7681-aa2dc8900054/contentlib-a2a6d08a-0b4f-4c5b-bc5b-69a5c3aadf3e/05667003-4f37-47e9-aa1e-21fe542dd38f/x.iso"

 

 

 

Environment

VMware vCenter Server 6.5.x
VMware vSphere ESXi 6.7
VMware vSphere ESXi 6.5
VMware vCenter Server 6.7.x
VMware vCenter Server 7.0.x
VMware vSphere ESXi 7.0.0

Cause

The main cause of this issue is that there is inaccessible ISO image is connected to the virtual machine cause VMDK file size show Zero after takign snapshot .

Resolution

• This is expected behavior, generally, when any datastore that is relevant to a VM is inaccessible, Hostd will treat it as an all-or-nothing situation, with the ultimate goal of either restoring the accessibility or transitioning the VM to an invalid load state, which would indicate there is an issue.
• When the storage is inaccessible, Hostd does not fill in the disk cache - in this case, default entries with 0 capacity are generated for virtual disks which are not in the cache.
• When the storage is inaccessible, Hostd does not refresh the snapshot list. This is why during the snapshot creation workflow even though the snapshot is actually created, the final check is unable to find the new UUID in the list, so the task fails. Reloading the VM when storage is accessible again will indicate that the new snapshot does indeed exist.

Workaround:
The solution is to make sure all storage is accessible again, either by reconnecting the ISO datastore or un-mount ISO Image.

Additional Information

Impact/Risks:
Unable to do any task on the virtual machine "Including any hot operations"