Downgrade from 6.12.4 to 6.12.3 and won't boot

[Mlatx]

The past few days I've been having issues with unraid becoming unresponsive after nightly docker updates. I went to downgrade to 6.12.3, and now it won't boot. I have a backup, but it's on the server. The software I use to copy from the server from my mac has not copied for a while now unfortunately. Rather than remove the drives to find the correct one, how can I get back into unraid? 

 

I would hope the easiest way would be to download 6.12.4 and replace the files on the flash drive that may have been corrupted leaving the system configuration intact.

[JorgeB]

Try recreating the flash drive, backup first, recreate with v6.12.3, restore the /config folder only from the backup

[Mlatx]

Here is what I did.

 

1. Copy the current USB contents somewhere

2. Wipe the usb with a fresh 6.12.3 installation

3. Replace the config folder with the one I copied

 

I still can't boot. I'm getting stuck at the bzfirmware checksum. How can I get around this?

[JorgeB]

Try a different USB with a stock install, no key needed, if the same run memtest.

[Mlatx]

The new usb worked. I'm assuming I put the config folder back on and all should work.

[itimpi]

3 minutes ago, Mlatx said:

The new usb worked. I'm assuming I put the config folder back on and all should work.

You could also try the original USB but with a stock config folder rather than your personalised one.    That could give you an indication as to whether it is the physical flash drive or something in your configuration information.    Not much point in doing a licence transfer to a new flash drive if it is not going to help.

[Mlatx]

That's a good idea. Although if that is the case, this is going to suck to fix, meaning I'll have to find the backup on the drive.

 

[Mlatx]

I copied the stock config files to the old usb. It got stuck in the same place. I copied my configuration files to the new usb, and now it gets stuck at bzmodules checksum Ugh.

 

I'm going to try installing 6.12.4 and copy my files over and see if that works. At least this way I can get to my backup and restore it.

Edited September 21, 2023 by Mlatx

[Mlatx]

That didn't work. Something in the configuration got corrupted in the downgrade. I can either find the backup on the disk (not optimal), or I have a backup of the last 6.11 I can go back to. If I did that, would it affect my dockers in anyway?

[Mlatx]

Update so far. I was trying to use a linux live cd to pull the last 6.12.x backup I had. All the live distros I've tried lock up after a few minutes.  

 

I installed the 6.11.5 backup to a flash drive. It got stuck at the bzfirmware checksum. So I put it into another usb port. It move passed the error but then got stuck on a plugin coral-driver/packages/5.19.17/._Coral... md5 error.  I tried a another usb with this version and was able to boot in after a couple of attempts. The array didn't start because the license wasn't valid on the new disk. The UI eventually locked up, and I had to restart it.

 

I figured the other usb slot might work for getting 6.12.3 working. I got past the checksum but then through an error "eth0" not detected. It went through the process of booting but did not assign IP addresses.  

 

Out of 3 USB's, I had all types of issues. Random restarts. Lockups at different points. 

 

I'm lost at this point. I'm going to run a memory test. Is it possible the motherboard is bad? Here are 2 pictures I took. One was interested because it was saying something about lack of communication between processors. This is the generic named photo. The other was after the UI froze when I got in. 

 

Any other suggestions?

Edited September 22, 2023 by Mlatx
new info

[JorgeB]

5 hours ago, Mlatx said:

All the live distros I've tried lock up after a few minutes.

This and the other issues suggest a hardware problem, start by running memtest.

[Mlatx]

I had to download and run memtest from memtest86.com because I needed a UEFI boot. Unraid's memtest, based on that, needs legacy CSM enabled. I needed my monitor and let it run 2 cycles. No issues were found. I updated the bios and still have boot problems. 

 

I tried each ram stick individually, and the problems persist. I doubt 2 ram sticks would go bad at the same time. 

 

It's either the motherboard or CPU. This is not going to be fun. I figure I'll start with the motherboard as CPU's have a lower failure rate.

 

New mobo arrives tomorrow. 

 

 

Edited September 22, 2023 by Mlatx

[Mlatx]

Hi All, 

 

It was the processor. Replacement came today, and all is booted up.