How to correct, shared folders going N/A for device after ZFS unmount/remount? Cannot edit Shared folder...

vulcan4d · 26. September 2017

OK so I'm experiencing a chain of events that is driving me a little nuts. My shared folders go N/A for device and I cannot edit them to correct, only option I have is to delete every shared folder and recreate. This is caused because I'm running OMV under HyperV and once in a while Windows decides to make the drives "Online", causing the drives to be unmounted from the HyperV guest machine after reboot. They can be easily made offline and readded, however when OMV boots the ZFS drives disapear and need to be imported again. The drives disappearing causes Shared folders to N/A under device. When I try to edit the Shared folders it gives me a lovely error and unable to edit. The only option is to delete and recreate. I've done this a number of times but I think it's time to figure this one out :). If there is a way to quickly fix these Shared folders without recreating them I'm willing to live with this silly behavior. Any ideas?

Code

Failed to execute XPath query '//system/fstab/mntent[uuid='822316bc-a9ef-4f40-89d8-96eeccf16bcd']'.

Code

Error #0: exception 'OMV\Config\DatabaseException' with message 'Failed to execute XPath query '//system/fstab/mntent[uuid='822316bc-a9ef-4f40-89d8-96eeccf16bcd']'.' in /usr/share/php/openmediavault/config/database.inc:78 Stack trace: #0 /usr/share/openmediavault/engined/rpc/sharemgmt.inc(231): OMV\Config\Database->get('conf.system.fil...', '822316bc-a9ef-4...') #1 [internal function]: OMVRpcServiceShareMgmt->get(Array, Array) #2 /usr/share/php/openmediavault/rpc/serviceabstract.inc(124): call_user_func_array(Array, Array) #3 /usr/share/php/openmediavault/rpc/rpc.inc(86): OMV\Rpc\ServiceAbstract->callMethod('get', Array, Array) #4 /usr/sbin/omv-engined(536): OMV\Rpc\Rpc::call('ShareMgmt', 'get', Array, Array, 1) #5 {main}

vulcan4d · 26. September 2017

Another look shows that the loss of the ZFS pools and re-import caused the drives to change the type of drive identity. I believe it normally is the UUID, how it looks like the path (dev/sdx). If drives get moved or HyperV does it's thing again it looks like this is probably why. Any suggestions, I'm at a loss.

hoppel118 · 28. September 2017

Which Windows Server OS are you using?

I don’t use Windows as the host, but it would be interesting to know.

Thanks and greetings Hoppel

henkall · 28. September 2017

The drives should not just go online. What version of Windows are you using? I am running omv in hyper-v on a Windows Server 2016 Datacenter and I have not yet seen any off my drives suddenly go online yet.

cabrio_leo · 2. November 2017

Hi folks,

I have exactly the same problem, but not in a VM environment! My OMV installation is running directly on real hardware. In Shared Folders I get N/A for Device and when I try to edit it I get the same error message as written in post no. 1. All my Shared Folders are located on a ZFS pool.

It seems that this happend after a ZFS pool export / import by the "Export Pool" and "Import Pool" button of the ZFS plugin which I have made for a while.

I have seen this behavior while I was searching for a different problem with USBbackup. From one day to another the wrong content was backuped by USBbackup. It didn´t backup the configured shared folder any more but the root folder of the system drive. Because I have set the "delete files"-option in the USBbackup job definition, my backup was corrupted therefore. Not really funny.

In the USBbackup email then I saw:
Please wait, syncing '//' to '/srv/f144d955d8d7fd1a199ea0dc188c38c9/Video' ...
sending incremental file list

This is the email message when the USbbackup was executed correctly (some weeks ago).
Please wait, syncing '/naspool/FilmeFS//' to '/srv/95330157ea8b6b143b13cf8cd279a999/Video' ...
sending incremental file list

So for me the corruption of the USBbackup is somehow related to the ZFS pool export/import.

First question: Is there a possibility to correct the shared folder error without deleting all references (Samba shares, ...) and without recreating all shared folder and SMB shares?

Second question: Somebody should look for the initial error.

henkall · 6. November 2017

I would also really like to know the answer to this one because I have had the same issue with my kernel was updated and then my zfs did not mount and then my shares suddenly was showing the device n/a and when I got my zfs up and running again I had to remake all the shares again. That is really not optimal.

cabrio_leo · 6. November 2017

Hi @'henkall you are expressing exactly what I feel.

Here in another thread that I wrote with the opinion of the developer of OMV: https://forum.openmediavault.o…?postID=158505#post158505

Jetzt mitmachen!