My apologies for not grabbing the diagnostics before rebooting

 

A few weeks ago, I set up a Win10 VM for my dad to manage some financials on ancient accounting software.  I left the VM running after we set it up so he could practice VPNing in and connecting to the VM (which he did manage to do successfully without any problems).  I didn't shut the VM down, so it sat there running for a few weeks.  We tried to connect to the VM over Christmas ... even though it was listed as "started" in the VM tab, the VM no longer had an IP address and wasn't responding to pings.  I tried to stop the VM from the VM tab, but it was more or less unresponsive.  So, I resorted to the "force stop" option with the bomb icon (not sure what the wording actually was), and left it to troubleshoot when I returned home (now).

 

Trying to restart the VM resulted in an execution error (`cannot read header for vdisk1.img` and `Input/output error`).  Analyzing 'fix common problems' revealed that the drive containing my VM was either full or read only.  As far as I could tell, this only happened after force stopping the VM.  Its status on the main page indicated it was far from full (~49GiB used on a 1TiB nvme drive), so I elected to reboot the server (before collecting diagnostics -- my bad).

 

The drive in question is a WD 1TB Blue SN550 Gen3 M.2 SSD.  It no longer shows up as even being available to assign to `vms_nvme` which was originally in my pool devices.

 

I suspect this drive has actually failed since unraid can no longer see it, but I'm curious if that could actually be the issue (the drive was less than a year old when I found the problem).  

 

I'm hesitant to pull it from the system without further advice, as I'm no longer sure that my vdisk1.img is even safe (I'm hoping the image file was saved and I can just replace with a new SSD to recreate).

 

Any advice at this point is welcome ... I know this is probably a very basic problem.  I've been so happy with unraid since abandoning Synology

 

galileo-diagnostics-20230102-0924.zip

14 minutes ago, gizmo000 said:

WD 1TB Blue SN550 Gen3 M.2 SSD

I do see a disk like that in pool cache_nmve, but I think you must instead be talking about the single disk pool named vms_nvme. I don't see any disk for that and the pool doesn't currently exist. You might try replugging it.

 

16 minutes ago, gizmo000 said:

I'm no longer sure that my vdisk1.img is even safe (I'm hoping the image file was saved and I can just replace with a new SSD to recreate).

If it was on that pool, unless you backed it up yourself, perhaps with a plugin, then there is no copy.

Your domains share is configured to prefer the pool vms_nvme, it has files on disk3 (and no files on that pool).

 

What do you get from command line with this?

ls -lah /mnt/disk3/domains

 

root@galileo:~# ls -lah /mnt/disk3/domains/
total 0
drwxrwxrwx  2 root   root    6 Jan  2 09:23 ./
drwxrwxrwx 18 nobody users 286 Jan  2 09:33 ../

 

12 minutes ago, trurl said:

I do see a disk like that in pool cache_nmve, but I think you must instead be talking about the single disk pool named vms_nvme. I don't see any disk for that and the pool doesn't currently exist. You might try replugging it.

 

If it was on that pool, unless you backed it up yourself, perhaps with a plugin, then there is no copy.

 

I did in fact mean `vms_nvme` ... they are (were) both on the same type (and size) NVME SSD.

 

To clarify ... `cache_nvme` has its own 1TiB SSD, and `vms_nvme` has its own 1TiB SSD.

Edited January 2, 20233 yr by gizmo000
clarification

4 minutes ago, gizmo000 said:

root@galileo:~# ls -lah /mnt/disk3/domains/
total 0
drwxrwxrwx  2 root   root    6 Jan  2 09:23 ./
drwxrwxrwx 18 nobody users 286 Jan  2 09:33 ../

 

Empty folder, probably freshly created by VM Manager when the pool went missing.

Well the good news is that I have pulled the affected NVME and plugged it into another computer ... and the vdisk image is still there.  

 

So now, I just need to find out why unraid won't see the drive.

1 minute ago, gizmo000 said:

Well the good news is that I have pulled the affected NVME and plugged it into another computer ... and the vdisk image is still there.  

 

So now, I just need to find out why unraid won't see the drive.

40 minutes ago, trurl said:

You might try replugging it.

If unraid was seeing it before then likely a hardware problem as simple as poor connection.

1 hour ago, trurl said:

You might try replugging it.

This appears to have solved my issue ... I'm back up with the correct drive assigned!  And the VM boots up just fine.

 

1 hour ago, trurl said:

If it was on that pool, unless you backed it up yourself, perhaps with a plugin, then there is no copy.

And since I've averted disaster, do you have any recommendations on how to cleanly or safely backup my VM disk images to the array?  For example, a recommended plugin?  I assume the VM would have to be shut down in order to back it up as well...

 

Sorry for posting such an elementary issue, but you've really helped

 

galileo-diagnostics-20230102-1135.zip

2 minutes ago, gizmo000 said:

recommended plugin

Search your Apps page

Thanks!

 

Found JTok's vmbackup plugin in CA ... also his github page is very informative.