(6.12.3) USB Boot Drive Read Only after every reboot/shutdown

[air_marshall]

Almost everytime I reboot/shutdown my server comes back-up with no GUI, no shares mounted. SSH is available and 'diagnostics' command results is a 'directory' doesn't exist error. USB drive chkdisk repair results in a successful boot.

 

Diagnostics file is attached - I believe this one was captured automatically during post Tip and Tweaks install.

 

TL;DR

 

Server was off for c12 months during our house renovation. Upon reboot I've had multiple issues. First seemed to be related to docker not being terminated on array stop/reboot/shutdown. I found the forum post, used the commands to stop and then was able to clean reboot/shutdown. Server upgraded to latest stable.

 

I set-about reformatting my cache to ZFS using spaceinvader video. Created a pool, but on copying back to the cache pool 1 of the drives had significant ATA issues, that drive eventually disappeared (I suspect went bad, but I've now pulled the SATA cable on it). So back to a single drive cache.

 

Current situation is as above. I've carried out most suggestions from the 'Unclean Shutdown Thread", docker is disabled, no VMs, Tips & Tweaks plugin is installed to capture diagnostics on shutdown and time-outs set at 7 mins. There isn't a delay in unmounting drives however I feel like something is hanging on reboot/shutdown that results in the usb drive being mounted "read-only" on almost every reboot. Fixed by pulling the usb and doing a repair in windows but obviously not sustainable.

 

Any/All help is much appreciated. I'm a long time unraid user and these are the first stability issues i've experienced.

 

TIA

tower-diagnostics-20230827-1032.zip

Edited August 28, 2023 by air_marshall

[JorgeB]

Make sure the flash drive is using a USB 2.0 port, also try recreating the flash drive, backup current one, format and re-create the flash drive manually or by using the USB tool, then restore the /config folder from the backup.

[air_marshall]

Good thinking. Done.

 

First reboot from freshly prepared and restored stick has given the same result. Server GUI not accessible. Shares not mounted. SSH available. Looks like USB is read-only again.

 

root@Tower:~# diagnostics
mkdir: cannot create directory ‘/boot/logs’: Input/output error
Starting diagnostics collection... tail: cannot open '/boot/bz*.sha256' for reading: No such file or directory
sed: can't read /tower-diagnostics-20230829-0942/config/go.txt: No such file or directory
done.
ZIP file '/boot/logs/tower-diagnostics-20230829-0942.zip' created.

 

Last lines of syslog using: 

 

Aug 29 09:42:42 Tower kernel: FAT-fs (sda1): Directory bread(block 30620) failed
Aug 29 09:42:42 Tower kernel: FAT-fs (sda1): Directory bread(block 30621) failed
Aug 29 09:42:42 Tower kernel: FAT-fs (sda1): Directory bread(block 30622) failed
Aug 29 09:42:42 Tower kernel: FAT-fs (sda1): Directory bread(block 30623) failed
Aug 29 09:42:42 Tower kernel: FAT-fs (sda1): Directory bread(block 30624) failed
Aug 29 09:42:42 Tower kernel: FAT-fs (sda1): Directory bread(block 30625) failed
Aug 29 09:42:42 Tower kernel: FAT-fs (sda1): Directory bread(block 30626) failed
Aug 29 09:42:42 Tower kernel: FAT-fs (sda1): Directory bread(block 30627) failed

 

Does this mean the USB is bad OR files in the USB config folder are corrupted?

 

Diagnostics file attached from last shutdown before erasing and reinstalling usb.

 

USB 2 port is in use.

 

Best path forward at this point?

 

 

tower-diagnostics-20230829-0914-shutdown.zip

[JorgeB]

Try a new flash drive.

[itimpi]

1 hour ago, air_marshall said:

Does this mean the USB is bad

Suggests this as it indicates that blocks on the flash drive are failing when trying to read them.   

 

Sometimes manually rewriting all the bz* type files on the flash drive from a zip of the release works to help with reading borderline sectors but not if the drive is really going bad.