Hi everyone,

 

I'm having a serious issue with my Unraid server and need some expert help. Yesterday, my network shares became inaccessible. When I checked on the server today, I found that the parity drive was disabled, and the system was showing multiple read errors.

 

I suspect the problem might be related to my USB enclosure. I noticed the enclosure's power was off, I had not turned if off, so I suspect a power interruption to the enclosure. I have not experienced any issues with this setup before.

 

Here's what I've done so far (in order):

1. Found the server on, but the USB enclosure was off.
2. I downloaded the diagnostics from the Unraid WebGUI before doing anything else.
3. I rebooted UnRaid cleanly, and turned on the enclosure at the same time.
4. I started the array successfully and was able to see the file structure on individual data drives but, it reported the Parity drive was disabled. The two data drives mounted fine.
5. I've started an extended SMART test on all three drives (still running). I also quickly ran the Short SMART self-test on all drives, which all passed.
6. I've since shut down the array for safety, keeping SMART going.

 

Here's my setup:

• Unraid Version: 6.12.6 (replace with your version)
• Server: Intel NUC model NUC8BEB
• USB Enclosure: Mediasonic HFR2-SU3S2 PRORAID 4 Bay 3.5" SATA Hard Drive Enclosure - USB 3.0 & eSATA
• Parity Drive:  WDC WD100EMAZ-00WJTA0

• Data Drives:

  • Disk 1: WDC_WD80_EMAZ-00WJTA0
  • Disk 2: WDC_WD80_EMAZ-00WJTA0
• Cache Drive: 250GB Samsung 970 EVO Plus NVMe M.2
• Connection: The enclosure is connected via USB 3.0 to the rear USB-A port on the NUC.

 

 

I've attached the full diagnostics files below - both after error and before reboot, and since rebooting and starting the array.

 

The first syslog shows multiple md: diskX read error messages for all my data drives, all happening at the same time. This, combined with the enclosure power issue, makes me think the USB connection was interrupted.

 

I also saw these concerning messages in the syslog:

• md: recovery thread: stopped logging
• XFS (md1p1): metadata I/O error

 

I'm worried about data loss, as I have a large collection of media on the array, as well as some important family history documents (which are backed up elsewhere, thankfully).

 

Is it safe to try running xfs_repair on the array? Or should I take other steps first? I'm hesitant to do anything without expert guidance, as I don't want to make things worse.

 

Thank you in advance for any help you can provide!

 

EDIT: Removing logs for privacy post fix

 

Edited March 5, 20251 yr by JohnCaldwell
Information redaction

Can't see the start of the problem in the first diags since the syslog had already rotated, but all disks were offline, so likely enclosure issue.

 

Latest diags look fine, you just need to re-enable parity, I would recommend doing a new config and trusting parity, then run a correcting check.

 

Also, keep in mind that USB is not recommended for array/pool devices.

Thanks Jorge - appreciate it. And yes, I’ve just finished speccing out a new build with internal storage to avoid this in the future - I knew I was playing with fire. 

 

Appreciate your advice! Once the SMART checks finish I’ll go ahead and do as you suggest.