Recently upgraded to 6.12.14.

 

One of my disks has been disabled post upgrade. I figured I'd test clearing it and re-adding it to the array, as I could still see and read the disk as an unassigned device. However, after doing this it was disabled again, I assume while rebuilding. 

 

• I can see the drive in the unassigned devices, which I find unusual, the drive ID was re-assigned from SDI (the disabled disk) and the unassigned device is SDK. I can mount the partition and read the data within it, indicating data was copied to it at least partially. I can't see any obvious errors in the SMART logs, no CRC errors, etc...
  • In the Disk Log Information I can see repeating warnings for : sd 7:0:7:0: [sdk] Synchronize Cache(10) failed: Result: hostbyte=0x01 driverbyte=DRIVER_OK
  • As well as an error: hotplug_devices, 1709: No such file or directory (2): tagged device WDC_WD40EZAZ-..... was (sdi) is now (sdk)

 

This leads me to wonder if the culprit may simply be a cable going bad, or something to do with the mobo. 

 

Side note this is the second time I upgrade to 6.12.14, I had to downgrade initially due to RTL8111 NIC issue, it did not like the drivers no matter what I tried. I've since added a secondary NIC I'm using which is an Intel 82574L and seems to be working fine. I noted there's an error complaining about eth0 not having an IP (it's physically unplugged) I may look to stub the onboard NIC just so that Unraid doesn't have to deal with it, but that involves a few additional settings I do not expect/suspect are impacting the drive issue. That is unless there's potentially something wrong with the mobo itself. The majority of the drives  run off an HBA but 2 are connected directly to the mobo. 

 

Diag file attached. 

scylla-unraid-diagnostics-20241220-1404.zip

Dec 20 03:41:59 Scylla-Unraid kernel: md: disk1 read error, sector=6750709872
Dec 20 03:41:59 Scylla-Unraid kernel: md: disk0 read error, sector=6750711544
Dec 20 03:41:59 Scylla-Unraid kernel: md: disk2 read error, sector=6750710528

 

Errors with multiple disks, possible a power/connection issue.

Thanks for the insight into the logs. I'm going to say you're likely on the right track with potential power issues, considering that this morning it was completely unresponsive but still powered on. It didn't even respond to the power button being pressed at all.

 

• Tried to see if it would just gracefully shutdown from me pushing the power button, usually will start the shutdown procedure, but there was zero activity after waiting for a while. Usually see some writes happening to USB (has led activity light).
• Then I tried to just hold the power button 4seconds to force shutdown, but even after 10 seconds of holding it didn't do anything. I'm going to guess the PSU levels may be dropping to low on some of the voltage rails and causing issues.
• I had to forcibly shut it down using the power switch on the PSU to be able to get it to power off. 

 

I'll order a new PSU before trying anything else with it, just to be on the safe side. 

 

If the issues persist with a new PSU, then the next likely culprit in my mind is the motherboard. Luckily I run a second Unraid box that just replicates the data from this first one, and I can stand up most of the dockers I use off of this one as well. So not completely dead in the water so to speak, and not overly worried about data loss. 

 

Happy holidays to you and your family.

Brand new PSU installed, was able to finish parity check successfully this time. Upgraded to 850W instead of 750, also not needing any extra adapters to swap from molex to sata power. I may see if I can test the old PSU in an older PC. beyond that seems to be stable for the most part.