vdisk Disappeared

[jwellnitz]

I woke up this morning to a dead VM. When I go into Unraid to start my VM it gives me this error "Cannot access storage file '/mnt/user/domains/Gaming VM/vdisk1.img': No such file or directory".  Unraid appears to have taken it upon itself to delete my vdisk over night.  Everything was working just fine yesterday.  I had a warning saying that my cache drive was 75% full but that shouldn't delete my vdisk.  Anyone have a similar issue or anyone know if I can recover my vdisk?

[itimpi]

Never seen Unraid just delete a file.    It might be worth posting your system diagnostics zip file (obtained via Tools->Diagnostics) to see if anyone can spot a reason this might have happened.

[jwellnitz]

Yeah I thought it was strange.  VM worked fine last night and this morning its missing the vdisk.

 

I have attached my diagnostics file.

gamingnas-diagnostics-20190626-0705.zip

[tonyfjr]

I had a power outage and have seen this as well. All my other data is intact just not the domains folder. I am running a parity check as we speak to determine anything crucial.

tower-diagnostics-20190821-1107.zip

Edited August 21, 2019 by tonyfjr

[trurl]

Don't know if your Unassigned Device has anything to do with your VM or not but there is this in syslog:

Aug 21 06:57:09 Tower unassigned.devices: Disk with serial 'Hitachi_HUS724030ALE641_PAJ2DRBT', mountpoint 'Hitachi_HUS724030ALE641_PAJ2DRBT' is not set to auto mount and will not be mounted...
Aug 21 06:57:09 Tower unassigned.devices: Error: ISO file '/mnt/user/isos/ISOS/virtio-win-0.1.141-1.iso' is missing and cannot be mounted.

Also lots of this:

Aug 21 07:07:53 Tower kernel: DMAR: DRHD: handling fault status reg 2
Aug 21 07:07:53 Tower kernel: DMAR: [DMA Write] Request device [05:00.0] fault addr 0 [fault reason 05] PTE Write access is not set
Aug 21 07:07:55 Tower kernel: DMAR: DRHD: handling fault status reg 102

Not sure what that's about, maybe someone else will.

 

You also have this FCP Warning:

Aug 21 04:40:02 Tower root: Fix Common Problems: Warning: Share Software_Applications is set for both included (disk3,disk4) and excluded (disk1,disk2) disks

Include means ONLY these disks. Exclude means EXCEPT for these disks. There is no reason to use both and you should only use include or only use exclude, not both.

[jonp]

Please post a screenshot of your Main tab.  Do you not have a cache device?

[Teth]

I just had this happen to me! My windows VM was working fine and Unraid server running fine. I rebooted the server (cleanly) since the plex docker was not letting me log into the webui. When the server rebooted the VM would not run. I got the same message Vdisk file not found'. To my surprise, the whole vdisk is gone! This is the first time something like this has happened

Edited February 21, 2020 by Teth

[Squid]

You should post your diagnostics

[Amorbellum]

this just happened to me

i was reorienting my drives, so maybe thats relevant? i was experimenting with different cache configurations. 

tower-diagnostics-20200320-2130.zip

 

fuck, am i screwed? shouldnt it have parity?

my cache is empty. fuck

Edited March 21, 2020 by Amorbellum

[trurl]

1 hour ago, Amorbellum said:

i was reorienting my drives, so maybe thats relevant?

Can you elaborate?

[Amorbellum]

i am not going to lie, i really dont know what im doing

i was originally going to use an unassigned disk for downloads, but in the end i created a downloads share and used that instead, so the larger drive that was unassigned, i decided to use that as a cache. at some point, if memory serves, i mightve removed all cache disks. i dont THINK so but its possible. i was adding a disk, and removing another, and then it wouldnt let me change muber of disks from 1 to 2, for some reason, no idea why (my array disks are set to 10 and i have pro)

[trurl]

You have 2 very different sized disks in cache pool. The default btrfs raid1 will give you a mirror, but with total capacity only equal to the smaller of the 2 disks, so you only have 250G capacity in cache the way it is currently configured.

 

Cache pool does appear to be mostly empty now. Before going much further with setting things up again, you might want to reconsider your cache pool configuration and how you would like it to work.

 

See the Cache Pool section of the FAQ for more:

 

https://forums.unraid.net/forums/topic/46802-faq-for-unraid-v6/?page=2&tab=comments#comment-554741

 

[Amorbellum]

right but, is my vm gone?

[Amorbellum]

and my plex docker has been wiped.

 

please, wtf is happening. this is crazy.

[trurl]

The usual configuration is to have VMs and dockers living on cache. So if you wiped your cache that explains both the VM and the docker problem.

[Amorbellum]

so is there no redundancy for vm's?

[JorgeB]

Depends on where you stored them and/or if the cache pool is redundant, but even with redundancy you should have backups of anything important, redundancy it's not a backup.

[Stu]

Adding a "me too" to this issue, which came about following a power outage. Unraid 6.8.3 diags attached.

 

I am running a UPS, and as far as i can tell the server shut down cleanly (parity was still valid when it came back up).

 

Everything seemed normal, shares and docker all ok, except for when my VM didn't autostart.  It flagged an error when i manually started it, that I was missing my second vdisk image that i have on an array only share (for games). The vdisk Windows is installed on, is on cache. 

 

I commented that second vdisk out from the xml and the VM booted fine, though windows did have to do some disk checking.

 

Let me know if i can provide any more info.

 

tower-diagnostics-20200430-2128.zip

[Kamero4K]

Same.. Me too

I really dont know how, but.. i lost like.. 2VMs (vdisks) now..

Both when i added new disks to the array.

A couple weeks ago i bought 2 new SSDs. Server was running fine and i stopped all VMs and stopped the array.

I added my new 2 Disks and startet the array again but when i turnt back to my VMs. Every VM was working fine execpt one of them. I tried to start it, but i got the Error that the VDisk coulnd't be found. I went into the terminal and looked for the VDisk in the folder.. but it was just gone.. I remember that i renamed the VM at somepoint in the past but it shouldn't really matter anyway.

 

Thats how i lost one VM.. the other one i lost a couple days ago.. at Chrismas 24th.

 

Backstory> I once created a Nextcloud - VM ( Disklocation = auto )

The Problem was, the SSD where the NextcloudVM was created on was completely full.

So what i did was, i went into the terminal again and created a folder under /mnt/disk4/domains/ named Nextcloud ( so fullpath was /mnt/disk4/domains/Nextcloud/ )

Then i copyd ( not moved ) the VDisk of my Nextcloud into /mnt/disk4/domains/Nextcloud/vdisk.img.

I changed the Path in the VM GUI and startet the VM.. everything was running fine and so i deleted the old VDisk.img

The VM was running fine for Weeks until Christmas. I got 2 new 1 TB ssds.

So again i stopped the array, plugged in the 2 SSDs and startet the array again but when i returned to the VMs.. the vDisk of my Nextcloud VM was just gone..

I also checked the usage of the SSD where the VDisk should have been : Usage before : 600GB used / Usage after / 200GB used (another vm that was located on the drive)

 

PS : I dont know if this does mean anything but i had a cache SSD installed.. for some reason it was everytime filled with 20GB data.. mover did nothing..

could it have something to too with a error with my cache drive?

 

I included a diagnostics but i dont know if the error is listed there.

I dont know if i can trust unraid anymore.. i have unraid installed on 2 Server.. one Basic and one Pro license.. i got in contact with the unraid Support but they said i should first create a Forum-Post and if nobody can help me i shoud get in contact again.. great support..

dellr720-diagnostics-20211224-1915.zip

[3x3q]

I believe this is happening to me right now. My Ubuntu VM is currently running just fine, but by pure chance, I noticed that I could not find the vdisk.img for this VM. Google'd "disappearing vdisk" and found this thread full of people who've had this happen. Now I'm terrified to shut down the VM for fear that it'll be lost into the abyss.

 

Ubuntu vdisk was installed to /mnt/user/domains/Ubuntu/ on cache pool. The domains share is set to prefer the cache.

 

Is there any situation where the vdisk still exists, but it just isn't displayed in Unraid? Is it just not showing up because it's currently running?

 

Thanks!

 

EDIT: I should also mention that this VM .img file is around 1.6TB in size, so it's very large. I wonder if that has something to do with it?

3x3q-mainframe-diagnostics-20221116-1726.zip

Edited November 17, 2022 by 3x3q

[3x3q]

Thought I'd follow up with some more info on my setup.

 

Like I said, the VM is currently RUNNING, with no apparent vdisk on any of my drives.

 

The last screenshot is of HTOP showing the VM processes, and also of its perceived location on my disk.

 

Someone please just tell me that the .img file will appear after shutting the VM down and the data consolidates... or something like that. You may have noticed that Ubuntu is under heavy CPU/SSD load, and that's because it's processing things that are very special and important to me. 

 

EDIT: Following up with Ubuntu VM log file, in case it helps with anything.

 

Ubuntu VM Logs.txt

Edited November 17, 2022 by 3x3q

[JorgeB]

Go to shares and click on "compute" for the domains share, post a screenshot of that together with the complete diagnostics.

[3x3q]

Here are my shares computed, along with a fresh diagnostics file. What's interesting is that my cache is reporting that there's roughly 0.5 of 2TB free, which would be consistent with a 1.5TB .img file existing on the cache somewhere.

 

Obviously the domains share only reporting anything on Disk1 is a problem. I've just gone through my drives again without detecting the vdisk, but the VM is still running and cranking away.

 

Thanks for your reply Jorge.

3x3q-mainframe-diagnostics-20221117-0943.zip

[JorgeB]

Post output of:

ls -lR /mnt/disk1/domains/

 

[3x3q]

 

EDIT: For reference, this is not the vdisk I'm talking about. I'm concerned about the Ubuntu vdisk that should be located at /domains/Ubuntu/.

Edited November 17, 2022 by 3x3q