Hi all,

Today I swapped my unraid setup into a new case. Booted everything up and I’m getting read errors on each disk and one of the disks was disabled. Diagnostics attached. Any help is appreciated. Thanks!

tower-diagnostics-20230204-1938.zip

All disks dropped offline at the same time, check power and cables.

I checked the cords and connections, and even swapped the PSU. Still the same issue. I notice it crashes as soon as I start my VM.
 

New diagnostics attached. Thank you!

tower-diagnostics-20230205-1004.zip

17 hours ago, LakersFan said:

I notice it crashes as soon as I start my VM.

That's a big clue, you are binding the HBA to VFIO-PCI, delete/rename /boot/config/vfio-pci.cfg and reboot.

 

Note that the addresses for bound devices can change if you add or remove any hardware.

Ugh. I let my parity check / data repair for all drives. (Left my vm off the entire time.) I then backed up the vfio-pci.cfg file and rebooted. Everything looked and worked normal with the vm stopped. The moment I pressed start on my VM, all disks had read errors simultaneously and all have “unable to write to disk XX” messages. Latest diagnostic files attached. 

tower-diagnostics-20230206-1901.zip

Edited February 7, 20233 yr by LakersFan

I also have problems with SAS2008 based cards sometimes.

 

What have changed every small detail can give clues?

 

What have changed while moving to the new case?

 

How do you handle loosening connectors and SATA power cabling?

 

Any OC ?

 

Try replacing the SATA cables.

I just took everything out of the old case and placed it in the new one. One thing I may have done differently is swapped the location of my pci-e Sata card with my graphics card. Would that make a difference?

Ok! So based off the question of what changed and my answer on the last post, I swapped my pci-e Sata card with my graphics card so they’re back to their original locations on the MB. Everything boot up fine and the vm now starts without any errors! Soooo… I think I got past that hurdle. However new problem is the vm doesn’t output to my monitor. I’ll post the latest diagnostics here, but mods let me know if I should call this thread solved and start a new one in the vm section…

 

tower-diagnostics-20230206-2140.zip

Did your revert that CFG file change from earlier? Or do that again to let it readjust it. It may be that as well.

 

And the monitor is connected at the same port at the graphics card ?

 

Sometimes BIOS is weird with graphics output selection and remembering it on reboot.

 

If it's the same hdmi port, Try the following - power down, remove power cable, push power On button to clear capacitors.

 

Power back On.

 

If it didn't work, try this - get into bios and re-adjust the graphics card selection. Switch to onboard, save, start the system, then back to PCI-e.

Edited February 7, 20233 yr by GRRRRRRR

Thanks for the suggestions. I tried reverting back to the prior cfg file, same behavior (no video on my vm). I tried to power down, pull the power cord, pressed power to clear the capacitors (it flashed on for a brief second), plugged back in and powered on, same thing, no video.  Went to the bios, changed graphics card back and forth, still the same thing - no video on my vm. I appreciate the suggestions but anything else?

 

I’ll add that I don’t think it’s a connection issue, as I see all the Unraid boot up print when I start it. When I click vm, my monitor then says no signal and goes auto-off. 

Edited February 7, 20233 yr by LakersFan
More info added

The problem was that the VM was still passing-through that device, you'd need to edit the VM template/XML and correct that.

4 hours ago, JorgeB said:

The problem was that the VM was still passing-through that device, you'd need to edit the VM template/XML and correct that.

Thank you. Just to clarify, where in the XML do I need to change?

As you had the devices originally, you need to remove this part:

 

</hostdev>
    <hostdev mode='subsystem' type='pci' managed='yes'>
      <driver name='vfio'/>
      <source>
        <address domain='0x0000' bus='0x02' slot='0x00' function='0x0'/>
      </source>
      <address type='pci' domain='0x0000' bus='0x06' slot='0x00' function='0x0'/>
    </hostdev>

 

14 hours ago, JorgeB said:

As you had the devices originally, you need to remove this part:

 

</hostdev>
    <hostdev mode='subsystem' type='pci' managed='yes'>
      <driver name='vfio'/>
      <source>
        <address domain='0x0000' bus='0x02' slot='0x00' function='0x0'/>
      </source>
      <address type='pci' domain='0x0000' bus='0x06' slot='0x00' function='0x0'/>
    </hostdev>

 

I think that part was from the first post with the diagnostics (xml) that was already messed up?
I deleted the hostdev from the xml and re-added the graphics card and sound from the form view. It still not displaying on my monitor. 
 

Question - do I need to delete the vfio-pci.cfg file still? I restored it from the backup I made initially when the suggestion in the second post of this thread ended up not being the issue. 

If it's working with vfio-pci.cfg restored no need to delete it.

It’s starting but it’s not displaying anything on the screen… so it’s not working on that way 

Unraid or the VM?

It was the VM. I ended up recreating the VM based off my previous vdisk.img. Everything seems to be working now. Thank you @JorgeB and @GRRRRRRR for your help!!