Drive "Not installed" (Odd Drive Issue)

[falconexe]

Welp, I've gone a very long time without a single drive issue/failure, and I guess I was due.

 

Here is the Order of Events:

 

1. Stopped the Array (Was going to Power Off Due to a House Move)

2. Disk 7 Threw an Error and Said "Not installed" Immediately After the Array Stopped. (Red Banner Notification Stating Disk Was Disabled)

3. Rebooted UNRAID (Hoping the Problem Would Just Go Away)

4. Upon Array Start, Disk 7 Still Showed "Not Installed" (Was Not Displaying Anywhere in GUI)

5. Stopped the Array and Powered Off (Moved Chassis to New House)

6. Once in New House, I Opened the Chassis and Reseated Every Cable on HBA cards

7. Powered On Array

8. Disk 7 Still Showed "Not installed", But is Now Showing Up and Powered on Under Unassigned Devices (UD) with XFS (correct) Partition

9. Ran S.M.A.R.T Test on Disk 7 in UD and Came Back Clean

10. Stopped Array

11. Assigned UD Disk 7 to Array Disk 7

12. Disk 7 Now Shows Blue Square (New Device) & UNRAID Warned That All Content on Drive 7 Will Be Lost (Assuming Due to Rebuild)

13. Did NOT Proceed

 

 

Questions:

 

Q.1. How come UNRAID could not just start the array since the previous drive was placed back into the same exact drive assignment? The drive literally blinked out like a lightbulb, and there had been no signs of trouble prior to this. S.M.A.R.T. comes back clean. I thought if the same drive was assigned in the same spot, the array would just start right up.

 

Q.2. Assuming I proceed with the format, the drive would be rebuilt from the Dual Parity (emulated contents of Disk 7). Is this a correct assumption? I also assume that NO DATA will be Lost that Was On Drive 7? I'd hate to proceed and find out that Drive 7 formats correctly, but that all of the previous data (emulated via parity) is now gone. From past experience, this would NOT be the case and once the disk is rebuilt from parity, it would be an exact duplicate of the pre-issue Drive 7.

 

Q.3. Is there any way to get this drive to just start without a data rebuilt/format?

 

I'm not touching this until some more seasoned experts chime in. I've been doing UNRAID since 2014, and have had drive failures/successful recoveries in the past, but I have never seen this particular set of circumstances and/or series of events (workflow). I did some searching on the forums and did not come across this exact scenario.

 

I have everything properly backed up, but I would like to understand completely what just happened. I may just buy a new drive and move on, but that would be the easy road 😂

 

@SSD Any thoughts? Maybe I'm just overthinking this. Been a very long 2 days with no sleep due to a pack/move/unpack for a new house... I just put this thing into its new air conditioned server room (not final, still building out the rack...)

 

 

Thanks for the help guys!

 

 

 

 

Edited November 1, 2020 by falconexe

[itimpi]

Q1:  As far as Unraid was concerned there WAS a problem.

 

Q2: Do NOT format unless you want to lose the drives contents.  A format creates an empty file system on the drive (currently being emulated) and updates parity accordingly.   A rebuild always puts onto the drive exactly what you see on the emulated drive and is NOT preceded by a format.

 

Q3: Not sure about the best way to handle this.   Are you certain that nothing has been written to the array since the drive played up?   

[falconexe]

28 minutes ago, itimpi said:

Q1:  As far as Unraid was concerned there WAS a problem.

 

Q2: Do NOT format unless you want to lose the drives contents.  A format creates an empty file system on the drive (currently being emulated) and updates parity accordingly.   A rebuild always puts onto the drive exactly what you see on the emulated drive and is NOT preceded by a format.

 

Q3: Not sure about the best way to handle this.   Are you certain that nothing has been written to the array since the drive played up?   

 

Q1. Yeah, shoulda grabbed the logs/diags, but I was in the middle of a time sensitive move out...

 

Q2. THANKS for confirming that I SHOULD NOT do that.

 

Q3. YES, no writes to the array WHATSOEVER. The array was only started for about 2 minutes once it was physically moved to the new house (after I checked all cables), and I saw that drive 7 was not assigned, but was showing up in UD. I then stopped the array. Nothing writes to Drive 7 like a VM or anything docker/plugin related. All shares are set to READ ONLY.

 

 

Here is my current plan. MOVE all emulated data (4.55 TB) via unBalance plugin to other drives. Once data is secure, perform the format (new device) of Disk 7 (same physical drive) and see how it responds. That way no data is actually lost/risked. Or I can do that same plan, but trash drive 7, and replace with a new larger (because why not?) precleared drive.

 

 

Thoughts?

 

Edited November 1, 2020 by falconexe

[itimpi]

personally I would not bother copying the data to another drive.
 

If you are interested in getting a larger drive into the array (assuming it is not larger than parity) then now is probably a good time to do it.    In other words simply rebuild the emulated drive onto the larger drive. You can keep the original drive intact just in case.    In fact as long as it is not assigned to the array you could try mounting it as a Unassigned Device) to see if its contents look intact.    If the contents look good then it is a fallback in the off-chance something goes wrong in rebuilding onto a larger disk.

 

The section of the online documentation (accessible via the Manual link at the bottom of the Unraid GUI) that covers replacing disks under several different scenarios is can be found here.

[falconexe]

I am well versed with how to rebuild data drives after a failure and a the wiki. I've done this a few times in the past without any issues.

 

Can you please clarify the following?

 

Since UNRAID will not recognize Disk 7 as the true original disk, and it is trying to format it, won't it just rebuild the same Disk 7 (physical drive) with the contents from the emulated Disk 7? Yes, the physical drive will be wiped, but it would be rebuilt from the contents of itself. You said I should not do this in Q2...

 

 

Doesn't that mean that it will be OVERWRITTEN with the contents of the EMULATED Disk 7?

 

I don't care if Disk 7 gets wiped as long as it gets replaced with itself. I just didn't want the emulated Disk 7 to disappear.

 

In the end, it is just very odd that since the same disk that became disabled, was put back, and it won't just recognize that and start the array.

 

Edited November 1, 2020 by falconexe

[falconexe]

10 hours ago, itimpi said:

Q2: Do NOT format unless you want to lose the drives contents.  A format creates an empty file system on the drive (currently being emulated) and updates parity accordingly.   A rebuild always puts onto the drive exactly what you see on the emulated drive and is NOT preceded by a format.

 

Maybe the issue is that IT IS trying to format. And that IS NOT a normal part of the rebuild process. I wonder of the XFS file system became corrupted and that is why it is not being recognized as the missing disk it actually is.

 

So the real question is, is the GUI indicating that the disk will be REBUILT here? Or is it indicating that it will be FORMATTED?

Edited November 1, 2020 by falconexe

[falconexe]

Welp, there is something seriously wrong with this drive. It just disappeared again, no where to be found in the GUI with the array started or stopped. Seems like it is going DEAD intermittently. Not a power issue, and not a cabling issue. I'm going to pull it and replace it with a new 14TB drive. I'll throw this old one into my backup server and run some pre-clears on it (if it can even handle that). Sucks because it is less than a year old. I'm not RMAing it either because of security concerns. No worries. Thanks.

 

 

 

Edited November 1, 2020 by falconexe