[SOLVED] [6.10.0-rc1] Dual Parity Array, 1 parity drive failed & 1 data drive failed while doing parity check - can I rebuild?

[metalchad]

I'll start this off by saying that unfortunately this happened on a day that I did not have the bandwidth to deal with the issue. Being a novice to unRAID, I did not get the syslog before stopping the array and shutting down thinking, "I'll deal with this when I get back". Really wish I'd have gotten that syslog... (feature request: email syslog when array errors are encountered, perhaps a plugin or user script already exists that I'm not aware of?)

 

I'm basically looking for the next steps I should take to attempt to recover here.

 

Sequence of events according to email notification I have set up:

 

• [02:01] unRAID started the parity check
• [11:09] unRAID parity disk error
• [11:09] unRAID disk 1 error
• [11:09] unRAID Array has errors
• [13:17] Parity check cancelled (this was my doing, I believe it suspended itself and I cancelled before stopping the array then shutting down)

 

From all that I've read, I believe that an issue with the controller (within an enclosure) caused the drive to drop unexpectedly. But I may have a unique set-up here, open to critiques.

 

ASROCK Z490M PRO4

Intel I9-10850K

EVGA GeForce RTX 3080 FTW

G.SKILL 32GB 2X16 D4 3200 RIPJAWS (non-ECC)

2x 480GB SanDisk SSD - SATA

1x 1TB NVMe - PCIe

5x 4TB HDD - USB 3.2 in a Mediasonic Probox enclosure

 

Perhaps I'm a bit paranoid, as I understand nobody should need dual parity for the amount of disks I'm using, be that as it may, I set it up with dual parity. But, hopefully that saves me because it would seem that the rarest of cases happened with 2 drives failing at the same time.

 

I'm not storing the planet on my rig. I run a few dockers including Plex and have a Windows 10 VM dedicated to gaming. And then of course I store personal files, photos/gopro videos, etc., and have had the intention to set up Backblaze for offsite backup of those, but am not quite there yet. 

 

At the time of the failure, I don't believe there was any disk activity other than the reads from the parity check happening.

 

In all my searching I haven't found another post or article that explains how to rebuild from the state I am currently in, which is that Parity disk 1 and array disk 1 are both in disabled states.

 

Please advise. Diagnostics attached. Thanks in advance.

 

 

 

 

 

 

mimir-diagnostics-20210908-2135.zip

Edited September 10, 2021 by metalchad

[trurl]

2 hours ago, metalchad said:

5x 4TB HDD - USB 3.2

USB not recommended for array or pool disks

[trurl]

SMART for both disks looks OK. You should be able to rebuild both disks at the same time if you could keep them connected, but dropped USB connections is almost certainly the reason they were disabled.

[metalchad]

27 minutes ago, trurl said:

SMART for both disks looks OK. You should be able to rebuild both disks at the same time if you could keep them connected, but dropped USB connections is almost certainly the reason they were disabled.

Seems odd though, that only 2 of the drives dropped, and not all 5, as they are all in the same 8 bay enclosure.. really wishing I had that syslog. Anyway, I'm not entirely sure what the exact procedure is to re-enable the drives and start the the rebuild. Any chance you could point me in the right direction?

[trurl]

https://wiki.unraid.net/Manual/Storage_Management#Rebuilding_a_drive_onto_itself

[trurl]

12 minutes ago, metalchad said:

only 2 of the drives dropped

Only 2 drives disabled because it won't disable more disks than you have parity. Entirely possible they were all dropped.

[metalchad]

3 minutes ago, trurl said:

Only 2 drives disabled because it won't disable more disks than you have parity. Entirely possible they were all dropped.

Ah, that makes sense. Thanks for the link. I'll provide an update when there is one to report.

[metalchad]

Rebuild of both parity and data drive completed successfully. Haven't yet verified my files are not corrupt, but they seem to be there; good enough for me to mark this as solved. Thanks @trurl