Sorry I only have partial information so far, but I figured I start someplace and see where it leads.

 

A well liked unRAID system was damaged 12.26, likely bumped or jarred. But nearly all at once it reported a parity drive disabled, disk 1 disabled, and 6 drives with errors. Reported to be running unRAID 6.7.2. During shutdown, md3, md4, and md6 were reporting xfs errors. It was delivered for repair yesterday. So, no for the syslog or diag.

 

The drives have been pulled and inventoried, but I have not yet mapped serial numbers to errors or failures. 

 

The server will not boot, likely due to old SuperMicro BIOS, suspect the date issue as it is not running 2.3a.

 

My first few questions;

1) it has 8T, 6T, 5T, and 4T drives. Pretty sure at least some new drives will be needed. Normally I go like for like, but it seems with unRAID you can replace with larger.  Which way to go?

2) Where is the drive mapping stored on the USB? I am told there are "extra drives" in the system and I would like to focus on the active ones.

3) Any pointer on a action plan? Currently thinking of ddrescue each to new drive. From question 1, I could use new 10+T drives if the space eventually becomes usable.

 

Thanks for all your help!

Edited December 29, 20223 yr by PaulS

Unlikely BIOS would prevent it booting from flash drive if it worked before. 

 

Would have been better to not pull any drives before getting it booted again.

 

Backup the flash drive then see if you can get it booting. It should tell you which disk it expects where. If that doesn't work then we can try other things. 

 

No point in thinking about replacing any drives yet until we can see what you have and come up with a data recovery plan. 

These boards, Supermicro X9 E3, seem to have a BIOS bug which is triggered by 2021. 

 

A bad port was found on a SAS card. Not sure that covers errors on 6 drives, as a port is only 4 drives.

 

All the disk serial numbers are mapped to assignment.

 

But the USB seems to lock the machine up.

Just really odd behavior with this USB device. If it is plugged in during POST, POST seems to stop. 

Really it is just very slow or long pause, because eventually it gets to EFI Shell. It does not boot from the USB.

 

The board will boot by using the boot menu, F11. But reaching the boot menu with the original USB installed hasn't happened. Boot IPMI media and other devices works.

 

Hence the need to update the BIOS. Hopefully doing that wont ruin the recovery.

 

Is the USB hard to rebuild? I am tempted to run ddrescue, or some phison util to attempt recovery.

You can try rebuilding the flash drive or using a different one, backup current one, recreate manually or with the USB tool then restore the config folder, try first with the current flash drive, a new one will require a key change, but you can try first to see if it boots.

On 12/29/2022 at 11:11 AM, trurl said:

Backup the flash drive

Did you?