Rebooting loop on Unraid versions newer than 6.9.2

[malcolmkza]

Hi,

 

My original Unraid server that I setup in 2013 with an MSI FM2-A85XA-G43 motherboard, AMD A4-5300 APU and 4 x 2GB DDR3-1333 ram sticks, goes on a rebooting loop whenever I upgrade beyond Unraid 6.9.2. So anything from 6.10.0 to 6.12.6 loops the reboot.

 

This machine runs Unraid Pro and I have 11 x 4Tb hdd's and it's for long term storage that gets booted up every few weeks receiving data from my other Unraid server that I use more frequently.

 

I've had intermittent issues with booting into the original 2013 server and I've finally sought to resolve them this week.

 

I've disconnected all the hard drives and removed the one daughter card so the issue is not the hard drives.

 

There's a Corsair 550W PSU in the server too.

 

I upgraded the BIOS today from 2.4 to 2.6 (https://www.msi.com/Motherboard/FM2-A85XA-G43/support#bios) but the rebooting persists.

 

I also backed up the contents of the boot USB stick to my PC and installed a fresh copy of 6.9.2 to remove the possibility that it was a configuration issue. So all the boot testing I've done today is with fresh installs, albeit nothing has been configured beyond that.

 

On some of the reboots, I do recall seeing a passage of text where the phrase "kernal panic" was displayed. I have read through other posts on these here forums and nothing really stands out.

 

The following warning / error from the Tools > System Log section seem important:

 

Jan 19 09:46:09 Tower kernel: ACPI: Early table checksum verification disabled

Jan 19 09:46:09 Tower kernel: ACPI BIOS Warning (bug): Optional FADT field Pm2ControlBlock has valid Length but zero Address: 0x0000000000000000/0x1 (20200925/tbfadt-615)

Jan 19 09:46:09 Tower kernel: acpi PNP0A03:00: _OSC failed (AE_SUPPORT); disabling ASPM

Jan 19 09:46:09 Tower kernel: floppy0: no floppy controllers found

Jan 19 09:46:09 Tower kernel: random: 6 urandom warning(s) missed due to ratelimiting

Jan 19 09:46:09 Tower kernel: EDAC amd64: Error: F1 not found: device 0x1601 (broken BIOS?)

Jan 19 09:46:09 Tower kernel: EDAC amd64: Error: F1 not found: device 0x1601 (broken BIOS?)

Jan 19 09:46:09 Tower kernel: EDAC amd64: Error: F1 not found: device 0x1601 (broken BIOS?)

Jan 19 09:46:09 Tower kernel: EDAC amd64: Error: F1 not found: device 0x1601 (broken BIOS?)

Jan 19 09:46:21 Tower rpc.statd[1505]: Failed to read /var/lib/nfs/state: Success

Jan 19 09:46:26 Tower kernel: read_file: error 2 opening /boot/config/super.dat

 

Any suggestions on what the issue could be?

 

If I can avoid upgrading motherboard, cpu and ram, that would be great. I'm fine with staying on 6.9.2 if that's the only alternative I have.

 

However, I'd like to keep up to date on the Unraid OS if possible.

 

Thanks

tower-diagnostics-20240119-2018.zip

[JorgeB]

Try blacklisting the Radeon driver:

 

https://docs.unraid.net/unraid-os/release-notes/6.10.0#linux-kernel

[malcolmkza]

Thank you for your quick reply JorgeB!

 

So, I did this on the System Devices page:

 

 

Then I updated to 6.12.6 and I've rebooted and Unraid has loaded successfully.

 

I just want to confirm that I've done, what you suggested, correctly?

 

Thank you

 

 

[malcolmkza]

Thinking through the steps here, I realised that I'd be restoring the config folder from the backup I made earlier yesterday, to the currently working fresh install of 6.12.6 and that this might override the settings I've changed above.

 

I did see in the config folder on the working 6.12.6 install that there's a file called vfio-pci.cfg and it contains:

 

"BIND=0000:00:01.0|1002:9993 0000:00:01.1|1002:9902"

 

There isn't a similar file on the backed up config folder so I wouldn't be overwriting anything.

 

Am I good to go and copy the backed up folder to the working usb install?

 

Thanks

[JorgeB]

Binding the GPU will also work, but you can just blacklist the driver as the link I posted above explains.

[malcolmkza]

Thanks again @JorgeB, my 2013 Unraid Pro server is running again! Greatly appreciated.