strange file system error after host freeze

[robvoi3]

One of our hosts completely freezed. We still don't kow what exactly happened.

After a cold reset of the host one VM didn't start anymore. The disks for the VM couldn't be found. We move dthe flat disks to a new VM. The VM is running all fine.

If the folder of the old VM there are now some "leftovers" which can't be deleted because they aren't really there. Also when executing a command while in this folder stange things happen in the ssh console.

Please see the screenshot for both effects.

Any idea how to fix this and remove the folder?

Thanks

Robert

[continuum]

Interesting ....
I have seen the same before on my own system - for me those issues disappeared after a reboot of the host that runs Putty.
Maybe also a reboot of the ESXi was required - dont remember.
Like you I first suspected VMFS corruption - but never actually found any VMFS problems when investigating

If you see the same problems after a reboot of both systems plus a cleanup of putty-temp-files then you can assume a problem - the few times I run into this it it was just anoying but harmless.

[robvoi3]

Unfortunately neither a reboot of the ESXi Host not using a different client for the putty session lets the problem disappear.

It still looks that strange. But only once when I list the content of this specific folder.

[jagdish_rana]

Hi There,

Could you please check performance of datastore by esxtop command. It could be latency of disk group.

Please check with storage team as well for the duration. When problem  was reported.

Thanks

[robvoi3]

No performance problem. Queue is at 0/0.01. Latency is below 25 all the time. The problem persists.

[FritzBrause]

Move off all valid data/VMs from this datastore. Make sure there is only data left on the problem datastore which is NOT needed.

Then delete it from the ESXi hosts.

Unpresent it on the array and delete the logical volume on the array.

Create a new logical volume, present it to the ESXi hosts and create a new datastore.

I sometimes have seen those strange issues and a recreation of the VMFS datastore was the best solution.

[robvoi3]

Thanks. This is the way to go then. There is a 10TB VM on the local storage. With 1Gbit only network this will be a long downtime. 🙂