Hi,
Recently I testing out Veeam to backup my VMs, and activated it's direct storage access feature - meaning Veeam directly accesses the VMFS store instead of going via a ESX host.
However, during a power outage, afterwards one of the VMs that was being backed up showed as inaccessible. Looking on the store, the folder was still present, but it was empty inside.
In vCenter, I could see that Veeam had put the VM into snapshot mode first, so it is very strange that suddenly all the files are missing.
I looked via SSH and browsed to the folder, but was very surprised to see a strange message once I landed on the affected folder:
Instead of showing the directory name it showed "unknown".
At the same moment in the vmkernel.log file it shows this:
2020-04-05T14:29:30.184Z cpu0:2099304)Fil3: 7897: DirEntry corruption (type 0, gen 0, addrType 0)
2020-04-05T14:29:30.184Z cpu0:2099304)WARNING: Fil3: 209: Unknown object type 0
2020-04-05T14:29:30.184Z cpu0:2099304)WARNING: Fil3: 209: Unknown object type 0
2020-04-05T14:29:30.184Z cpu0:2099304)ALERT: DC: 783: Duplicate name '' entry in cache.
2020-04-05T14:29:30.184Z cpu0:2099304)WARNING: Fil3: 209: Unknown object type 0
This also showed up when I ran the VOMA check.
I have found that VMFS Recovery from Diskinternals can see the vm files, but without purchasing it, I have no hope of recovery. However, since that application can see the files (I can tell from the date, and size which are VMDK etc) I am certain there must be some way to still recover them.
I can only imagine that they and/or the folder is in some sort of limbo "locked" state.
Does anyone have any ideas please?
Environment:
4 x ESXi 5.5 Hosts, 1 x vCenter, 1 x iSCSI Datastore, 1 x NAS (for Veeam to put backups on to).