UnRAID 7.2.0

The system locked up. I couldn't SSH into it, or get to it through the ui. I pushed the power button, and it looked like it got the sigterm, and sat there for a while. I looked like it hung on Stopping the UNRAID API. I just sat there for several minutes. I forced shutdown with holding the power button. After I tried to start it it said it was missing the USB drive. I ended up pull out the drive and reseating it. Then it found the drive. I have had that happen before. The difference this time is that when I started the array back up, (needed to do parity sync obviously). But when it tried to start I started getting a ton of errors in the logs about drive read errors.

Nov 22 23:52:42 palazzo kernel: md: disk5 read error, sector=7897400

Nov 22 23:52:42 palazzo kernel: md: disk6 read error, sector=7897400

Nov 22 23:52:42 palazzo kernel: md: disk3 read error, sector=7897408

I have stopped docker process, and tried to stop the vm process. The UI is now unresponsive. I did have it generate a diagnostics while this was going on. I couldn't find an earlier diagnostics in /boot/logs from the last day or two.

Am I good to just power off the computer, and try to re-seat the drives again? Or what should my next steps be here?

ls -alh /var/log/syslog* && uptime

-rw-r--r-- 1 root root 128M Nov 23 00:18 /var/log/syslog

00:18:09 up 29 min, 1 user, load average: 67.74, 67.20, 54.19

palazzo-diagnostics-20251122-2358.zip

Edited November 23, 2025Nov 23 by openam
Adding details about syslog size and uptime.

The syslog just stops growing at 128M. I don't see any syslog.1, syslog.2, etc getting created.

Nov 22 23:52:28 palazzo kernel: mpt2sas_cm0: SAS host is non-operational !!!!

HAB problem make sure it's well seated and sufficiently cooled; you can also try a different PCIe slot.

Turned off the machine re-seated the card (only one pci slot). When I brought everything back up it show disk 5 is disabled? Ran SMART short self-test said completed without error. Took diagnostics, figured I should run the SMART extended self-test. Probably a bad idea. I guess that might take a day or two to complete. When it started yesterday it started to run parity check. Would that mess up the parity drive with the SAS device going offline like it did? Wondering what the best plan of action is here. Attached are the diagnostics from before the extended test being started.

palazzo-diagnostics-20251123-1006.zip

5 hours ago, openam said:

When it started yesterday it started to run parity check. Would that mess up the parity drive with the SAS device going offline like it did?

If it is the Parity Check that runs automatically when starting the array after (what Unraid thinks is) an unclean shutdown, that is a read-only check - parity is not updated.

If you click the button on the Main tab to start a parity check AND the box that says "Write corrections to parity" is checked (it defaults to being checked) then that will write to the Parity drive.

Now it doesn't say "Write corrections to parity" now. I don't remember what it said. But if it defaults to being checked like you said then I guess that's what I clicked. I guess the question is would it have been writing corrections still when it lost 3 drives with the HBA/SAS lost connection? I'm still waiting on the extended SMART test on that drive it currently says 50%.

What should my next steps be?

I just re-read your message. Sounds like 2 different ways to start parity check.

1 - Start array after unclean shutdown

2 - Explicitly start parity check (I assume this is after array is started)

Is that what you mean? If so I originally just did the 1st option.

If the array was shut down uncleanly, Unraid will always automatically start a non-writing parity check when the array starts.

Well the SMART extended test is about 90% complete. What's the next steps assuming it finishes and says the disk is healthy, or bad? If it's bad I assume I pull it out and put a new one in and rebuild from parity. Do I also rebuild the same one from parity if it comes back as good?

Post new diags after array start to see if the emulated disk is mounting.

Started the array, attaching diagnostics. Looking at disk5 it appears that there is stuff in there.

palazzo-diagnostics-20251124-0853.zip

The emulated disk is mounting and has plenty of data. Assuming the contents look correct, you can rebuild on top:

https://docs.unraid.net/unraid-os/using-unraid-to/manage-storage/array/replacing-disks-in-array/#re-enabling-a-disabled-disk-rebuilding-onto-itself

I started in maintenance mode. How do I know if it's being re-built. It feels like I'm getting conflicting messages.

This says it's being re-constructed

But this says Sync to start Data-Rebuild

Is there a difference between reconstructed and Data-Rebuild? Is this just a UI issue? Or do I need to hit the Sync button to actually get this to fix it?

I ended up hitting the Sync button because there was no progress showing that the rebuild was started. It appears to be running now.

In maintenance mode, you need to click "sync", in normal mode, the rebuild starts automatically.

Thanks everyone. Everything is back up and running. Sync completed. Then I stopped the array, and restarted it not in maintenance mode.