Running into a really frustrating problem following an unexpected shutdown. When the server was brought back online one of the drives was disabled/emulated. I stopped the array, set the disk to No device, started the array in maintenance mode, stopped the array, set the disk back to its original option, and restarted the array. Now any time I try a data rebuild (normal mode), sync (maintenance mode), or an unbalance to move everything off of the emulated disk after a few seconds one, sometimes two, of the drives goes into standby (which drive seems to change with each attempt) and starts racking up errors (the standby disk also then appears in unassigned devices even though it is shown as assigned to the array).

I've been digging for a good part of the day but can't find a reason or solution. Has anyone run into a similar problem or have any suggestions? Running 7.1.0 and haven't had any problems until now. The diagnostics are attached

bigboy-diagnostics-20251004-1911.zip

Replace both cables for that disk and try again, if the same, post new diags.

1 hour ago, JorgeB said:

Replace both cables for that disk and try again, if the same, post new diags.

By both cables I am assuming you mean the data and power? I'll see what I can do but there are 18 disks on the backplane with the disks that are going into standby being sometimes on the same cable and sometimes not (four drives per cable). I'll try moving them to a different bay in the enclosure that is on different connections as I don't currently have any spare SAS cables.

3 hours ago, JorgeB said:

Replace both cables for that disk and try again, if the same, post new diags.

I moved the disks that were going to standby to other cables where available and to other bays where not. Results are roughly the same with the addition of another drive going to standby that previously had not. New diagnostics are attached.

I have ordered new cables and am looking for an alternate HBA to try while I continue to look for a solution.

bigboy-diagnostics-20251006-0934.zip

Edited October 6, 2025Oct 6 by scott.harris

That's worth a try, could also be a backplane issue.

25 minutes ago, JorgeB said:

That's worth a try, could also be a backplane issue.

That's one of the things I really fear. Getting a replacement backplane is going to be a massive PITA if I can even get it on its own. New cables will be here tomorrow and I will update with the results then.

All new cables, same problem. Moved the HBA to a different slot, no change. New diagnostics are attached if anyone is interested.

I am working on getting a replacement backplane and HBA but the first I won't have a quote on until next week and it probably will be another month+ before it gets here and I didn't pull the trigger fast enough on a replacement HBA and someone else snapped it up so am waiting for another to pop up.

bigboy-diagnostics-20251007-1858.zip

Still looks like a power/connection issue, if everything else has been changed, it would be good to test with a different or without the backplane.

I'm curious; if it was losing power or data connection would it not switch to offline or disabled instead of standby? The connections are still seemingly good. The drive goes to standby in the array but then gets listed to Unassigned Devices where it is mountable and the data can be accessed form there which would seem to indicate all of the connections were good.

What do you mean by standby? There are multiple disks with read errors logged, you can check the syslog to see which ones.

On the "Main" tab the indicator for the drives with the read errors are a light green dot and when the cursor is over it the description is, roughly from memory as the server is currently offline, the disk is in standby and I have the option to spin up the drive. Selecting that option does nothing and my spin down setting is set to never on all drives.

Unraid is losing contact with the drives, hence why they appear as in standby, but that's not the problem.

13 minutes ago, JorgeB said:

Unraid is losing contact with the drives, hence why they appear as in standby, but that's not the problem.

But it is still mountable and useable in Unassigned Devices so the connection is still present. That's what is throwing me off. If Unraid has lost connection why is it still connected?

26 minutes ago, scott.harris said:

But it is still mountable and useable in Unassigned Devices so the connection is still present.

Because the drives are dropping offline and reconnecting with a different ID, UD doesn't show drives that are currently assigned to array or pools.

2 hours ago, JorgeB said:

Because the drives are dropping offline and reconnecting with a different ID, UD doesn't show drives that are currently assigned to array or pools.

Interesting. Sometimes the "unassigned" drive would show up in UD with either "Array" or "InArray", can't recall exactly which, and if I remember correctly those might not have been mountable? But they connected and Unraid recognized them as supposed to be in the array. It is all very odd and I'll admit I'm not doing a great job of communicating all of the details. I appreciate the insight, though. I didn't realize that if a drive lost connection and reconnected it could be with a different ID. That helps.

I am in the process now of moving all of the drives back to my old server so I can at least get the original down drive rebuilt, if possible, and to at least confirm the drives are all fine. A new HBA and another set of cables will be here some time in the next week and hopefully that will get things resolved. Otherwise I will need to wait until I can get a replacement backplane from China. Fingers crossed and thanks for the help.

Damn. So the move back to the old hardware went smoothly. I lost the data on the "bad" drive but that can be replaced. No read errors at all. Gave it a few days to settle in and everything was still running smoothly. That brings us to today. Got a new HBA and cables in moved the drives back to the new hardware. Started the array in maintenance mode without any problems. Started a read check and BAM, 2048 errors immediately followed by the drive dropping to disabled then popping up in UA. I guess the next step is the backplane which will likely not be available for a couple of months. I guess we're sticking with the old server for a while.

Damn it. Everything moved again. Started it all back up and now the drive with the read errors on the previous attempt is now showing as "Unmountable: wrong or no file system". This is a full 24TB disk. Doesn't even come up as emulated. I can stop the array, set that slot to no device, and start again and then mount via UA and everything is still present. No complaints about the file system at all. This is becoming extremely frustrating.

Is there a way to add this disk back into the array? I don't want to go poking around too much as doing that previously cost me everything on another drive.

If the emulated disk doesn't mount, but the actual disk is fine, and you don't suspect the disk is the problem, you can do a new config to re-enable it, but make sure the disk is not the actual issue.

The disk is not emulated. It shows up with an "Unmountable: wrong or no file system" notice and an option to format. If I remove it from the array (set that slot to No Device) and then start the array it pops up in UA and is mountable and the data is present. The disk itself appears to be fine and now that it is again in the old server there shouldn't be any problems with the HBA, cables, or backplane. It ran there fine for days during testing over the previous weekend.

I am hoping there is a way to get this disk recognized as part of the array again but without having to reformat and with no data loss.

1 hour ago, scott.harris said:

The disk is not emulated. It shows up with an "Unmountable: wrong or no file system" notice and an option to format. If I remove it from the array (set that slot to No Device) and then start the array it pops up in UA and is mountable and the data is present.

If the actual disk mounts in UD, it should also mount in the array, unless it's being emulated, please post new diags.

New diags attached. Here is a screenshot of what happens when I assign the disk to its slot in the array now:

And when I don't select it the option is either one of my unassigned disks or "unassigned" where previously it was a disk or "No device".

bigboy-diagnostics-20251016-1119.zip

Disk16 is disabled (being emulated), so

9 hours ago, JorgeB said:

If the emulated disk doesn't mount, but the actual disk is fine, and you don't suspect the disk is the problem, you can do a new config to re-enable it, but make sure the disk is not the actual issue.

You can also see if the filesystem for the emulated disk can be repaired and then rebuild; just don't rebuild on top without repairing the filesystem first, if you still have doubts, please ask before proceeding.

I think I am going to move everything off of disk16 and on to other array drives via UA and then add it and format. Everything I see in the options leaves me with the fear that I will lose the data like I did with the previous drive (tried repairing the file system and adding the drive back in and lost 12TB). When I start the array with nothing assigned to that slot it does say it is disabled and emulated but it cannot be browsed like a disabled disk usually can, only that it is unmountable. Frankly, I lack the confidence in my knowledge of the process of adding a full drive to the array and having it remain intact to go with that.

I will work on that while I wait for the replacement backplane to arrive. I truly appreciate your time with this.

4 minutes ago, scott.harris said:

When I start the array with nothing assigned to that slot it does say it is disabled and emulated but it cannot be browsed like a disabled disk usually can, only that it is unmountable.

To try and correct that, you just need to check filesystem, and it's worth a try, since it won't cause any more damage; just leave the disk unassigned for now.

Ok. I completed the file check and repair. Prior to starting the array after the repair it no longer says it is unmountable but it does warn that starting the array will begin data rebuild. I mounted it in UA and the data appears intact. Starting the array with the disk unassigned and the emulated files are again browseable. Thank you!

Now...is there a way to not rebuild the data but instead just use what is already present and only recalculate parity? I think that may be what the "new config" option is for but the wording on it is unclear, referring to the drives as new.

Edited October 16, 2025Oct 16 by scott.harris