Forgive me if this is a bit harsh but I am fed up with unraid dropping disks (yes, I know the dropping disks is a hardware issue, how it deals with it is my issue) and forcing rebuilds because exactly things like this can happen. I just lost 12TB of data, luckily it is mostly recoverable but it is a few weeks of work to get it back.

 

Why can't the array simply lock itself and tell me to fix the issue instead of ejecting a disk at the first sign of an issue and forcing a full 24 hour rebuild??? I have had to rebuild 6 or 7 times now due to this. Not everyone can afford proper server gear and have to hobble together old parts on the cheap.

 

/end rant

 

With that out of the way, Once again a disk dropped and it ejected it from the array. I was planning a server upgrade this weekend anyways, so I just wanted to start up the cache and get my data off it since part of my upgrade was a new cache.

 

After some trial and error I finally got the cache moved to another drive and swapped everything to the new setup.

 

I then let it rebuild the disk but now it just says

Unmountable: not mounted

and gives me the option to format it.

 

The only thing I can think of is that I removed the failed drive from the array when starting it as I didn't want it to start a parity sync on it, I was planning on swapping it for another drive in the new case.

 

Did this erase the drive??? That needs to be made WAY more clear.

 

Now it did the rebuild, so I am assuming that what I am dealing with is a deleted partition. Is it possible to recover the BTRFS partition? It was encrypted.

33 minutes ago, TexasUnraid said:

I didn't want it to start a parity sync on it, I was planning on swapping it for another drive in the new case.

Depends on how you put all disk ( normal and the drop one ) to new case and the target drop disk also in correct emulate state. ( Or simple say, create array state same as all disks in old case )

 

If emulate state already unmountable then rebuild also unmountable as expected.

Edited December 6, 20205 yr by Vr2Io

Yeah, I noticed it said unmountable in the emulated state but was more focused on the cache and figured it was just because there was no disk assigned to that slot at the moment.

 

So am I to understand that if a drive actually died and you removed it from the server but did not have another drive to replace it, simply starting the array in this state wipes the drive? I actually mounted to the dropped drive in ubuntu briefly to verify that the drive was ok, it was ok at that point, so it was something unraid did to wipe it, As after checking it was ok in ubuntu I decided to just go ahead and reuse it.

 

Thankfully I have a script that saves a list of the files/folders on each disk just for an occasion like this. Although it did make me realize I need to update it to not overwrite the file each night. BTRFS snapshots FTW.

Edited December 6, 20205 yr by TexasUnraid

29 minutes ago, TexasUnraid said:

simply starting the array in this state wipes the drive?

Unraid won't wipe the drive, disk slot keep in emulate state and won't touch the physical drop disk. I believe if emulate disk show unmountable it is because parity invalid or the file system already corrupt during disk drop.

 

As you can mount the drop disk under ubuntu, so I assume unmountable due to parity out of sync. That's why I will monthly perform parity check, and luckily I never have check error.

 

29 minutes ago, TexasUnraid said:

I actually mounted to the dropped drive in ubuntu briefly to verify that the drive was ok, it was ok at that point,

Does disk SMART data record any abnormal ?

Edited December 6, 20205 yr by Vr2Io

7 minutes ago, Vr2Io said:

Unraid won't wipe the drive, disk slot keep in emulate state and won't touch the physical drop disk. I believe if emulate disk show unmountable it is because parity invalid or the file system already corrupt during disk drop.

 

Does disk SMART data record any abnormal ?

Thats what I thought but can't explain why this happened.

 

Smart is fine on all drives besides the expected UDMA errors, ran a bunch of tests and no signs of any issues.

6 minutes ago, TexasUnraid said:

UDMA errors

 

1 hour ago, TexasUnraid said:

I have had to rebuild 6 or 7 times now due to this.

UDMA error not a good sign, I have this kind problem in long time ago and fixed by change other controller. In past ~3 yers, I haven't drop disk, UDMA, SMART error, even lot of different hardware change.

11 minutes ago, Vr2Io said:

 

UDMA error not a good sign, I have this kind problem in long time ago and fixed by change other controller. In past ~3 yers, I haven't drop disk, UDMA, SMART error, even lot of different hardware change.

I forgot if it was UDMA or CRC (server is off right now as I get my mind around having to redo a lot of data I just spent the last 2 months organizing and was just another hour or 2 of clean up to have it all finished), but it is the error that happens when a cable is loose/bad, it always stops when I re-seat the cables or replace a questionable cable. 

 

If you saw the case everything was mounted in and the rats nest of cables, you would understand why cables were an issue lol. Touch one thing and you move the cables for all the other drives as well.

 

Plus most of the cables were old as well since I was saving up to get all new cables with the new HBA. The drive itself is fine.

 

Also, the 6 or 7 times I have had to rebuild have been on different disks, completely random (well, generally the ones closest to where I touched a cable). It is not like I have a single problem drive, then I would agree with you. HBA is also unlikely as once the cabled are seated well, it has run for a month solid without an issue and a LOT of stress. Still, new larger HBA and cables are supposed to arrive Monday.

Edited December 6, 20205 yr by TexasUnraid

12 minutes ago, TexasUnraid said:

6 or 7 times I have had to rebuild have been on different disks, completely random (well, generally the ones closest to where I touched a cable). It is not like I have a single problem drive, then I would agree with you.

I also face such disk failure problem in 4 or 5 times, after that I have nice time in recent years, so it really depends on what disk quality you got.

 

17 minutes ago, TexasUnraid said:

new larger HBA and cables are supposed to arrive Monday.

Good luck.