Hi there,

This morning I received the following errors from my (recently rebuilt) server:

Event: Unraid Disk 2 error
Subject: Alert [WILLOW] - Disk 2 in error state (disk dsbl)
Description: TOSHIBA_HDWG180_90G0A17XFAUG (sde)
Importance: alert

 

and

Warning [WILLOW] - array has errors
Array has 1 disk with read errors

 

I immediately shut it down.

I am of the mind that those errors are usually a loose cable, and I only recently built the server (moving my HBA and disks from an old one), so I assumed that was the problem, so I opened up the case, checked the power and SATA cables for the disk and rebooted.

After starting the array, it says that the disk is being emulated.

Though I then received the notifications:

Warning [WILLOW] - udma crc error count is 74
TOSHIBA_HDWG180_90G0A17XFAUG (sde)

and

Notice [WILLOW] - array turned good
Array has 0 disks with read errors

 

As the disk was being emulated, I assumed it must still be a bad connection, so I shut it down and tried a different connector on my HBA card.

It still says it's emulated.

However, I can run a SMART test on the disk (it says it passed - report attached), so it must be connected, right?

 

And if the SMART test passed, how come it still says that the disk is emulated?

 

What do I do now? Do I make Unraid accept the disk? Or is the disk dead and I buy a new one?

 

I've seen topics with a similar problem linking to this help article: https://docs.unraid.net/#Rebuilding_a_drive_onto_itself but the link is no longer working, so your insight is appreciated. EDIT: I have found the new URL for it: https://docs.unraid.net/unraid-os/manual/storage-management/#rebuilding-a-drive-onto-itself I don't want to do the rebuild unless it's the right thing to do, though, so I will wait until someone here can confirm that my disk is ok, and that it was more than likely just a bad cable.

 

Diagnostics attached (I was in a panic after receiving the notification, so I didn't think to download the diagnostics before shutting down the server - sorry), as is the SMART report for the disk in question.

Thank you for your help.

TOSHIBA_HDWG180_90G0A17XFAUG-20250125-0929.txt willow-diagnostics-20250125-0943.zip

Edited January 25, 20251 yr by jademonkee
Found info on rebuilding onto itself

I note that this post:

 

...says that I can just add the disk back in if I know Parity is good.
I last ran a parity check only last week, so Parity should be good. I am as certain as I can be that the disk wasn't being written to this morning (there's no process that would have been writing to it), and the error was a 'read' error, after all (though, TBF, I don't know why anything was reading it at the time, so my confidence that something wouldn't have been writing to it is a little shaken).

Is that a feasible approach? Or should I do the rebuild? (or, is my disk dying?).

Thanks again.

Ah.

I just shutdown and connected it back to the original connector on my SAS - SATA cable, booted up, and now the udma crc error count is 75.

So I think maybe that cable or port on the card (the other connector I used was on the other SAS port) might be dying.

I have a spare SAS > SATA cable, so I will try that one and see if the CRC error count remains stable. If I don't get any more errors, then I will rebuild on top (I now no longer trust the data on the disk, so I think rebuilding will be best).

If I do get more errors, I will try the other connector (on the other SAS port) again. If I get errors on THAT, then I will replace the disk.

If I don't get errors on that, I will replace the HBA card.

 

If anyone wishes to chime in with a sanity check on this process, then that would be appreciated.

Thank you.

Diags are after rebooting so we can't see what happened, but SMART looks OK, the emulated disk is mounting, assuming contents look correct, you can rebuild on top, recommend replacing/swapping the cables first to rule that out.

Thanks for the prompt reply @JorgeB

I've now swapped the SAS > SATA cable. The new SMART diags are attached (since the crc error count rose to 75), in case that's useful.

I will run an 'extended self test' now to see if any more errors kick up.

If not, then I will keep this cable and turf the old one.

Thanks again.

 

EDIT: Ah, I just realised I can't run the extended test on it (presumably while it's emulated). I'll do a rebuild and see what happens.

 

TOSHIBA_HDWG180_90G0A17XFAUG-20250125-1021.txt

Edited January 25, 20251 yr by jademonkee

SMART looks fine, don't see the need to run an extended test for now, just rebuild, and any more errors gran the diags before rebooting.

Just closing the loop on this:

Disk rebuild was succesful.

UDMA CRC error count has not risen since swapping to a new SAS > SATA cable.

Thanks for your help and guidance.