(SOLVED) Read Check found 553322 errors - changing case and replacing (at least) 1 HDD to fix

[RollinDolan]

Hi all,

 

My server is on Unraid Version 6.8.0 2019-12-10.  

 

Plugins / Addons:

• LSIO Nvidia drivers. I'm not clear if I'm on vanilla unraid or Nvidia'd unraid at this point. I may have updated to 6.8.0 via Unraid by mistake.
• Unassigned Devices
• Preclear Disks
• CA
• CA Backup / Restore Appdata
• Several docker containers (plex, delugeVPN, sonarr, radarr... but they are all turned off for the time being)

 

Here is my hardware part list

 

Note - I've always had seating issues with the hotswap cages in my Rosewill case. Sometimes they need an extra push in to be seated fully. With this issue, I decided to switch to an ATX tower case where all drives will be connected directly (no backplane / hotswap).

 

The problem:

My server has been disconnected for over a month, as we just moved. Before that, I think it had been existing on one emulated disk for a while. i figured I just needed to re-seat the disk as usual, and I would wait until after I moved (stupid decision, but my server was not in a service-able area anyway while we were in our old home).

 

We get to the new home, and I plug in the server last night. One drive (Disk 2) has a red X, "device disabled, contents emulated".

Ok. I decide to 1. Stop Array > 2. Set disk to uninstalled > 3. Start array > 4. Stop array > 5. set disk to installed in right spot > 6. start array > 7. rebuild

 

Except after step 6, unraid kicks off a read check (not a rebuild of the emulated disk). I'm not sure why. So I say OK, I'll wait for the read check to finish.

 

Today, I see the check has finished, it's found 553322 errors, and I still don't see an option to rebuild. In fact, that disk is again not found. It's not in unassigned devices or in my array devices. It could be the finnicky backplane.

I plan to remedy the backplane issue by switching over to my new ATX case where everything will be directly connected to the MoBo or SAS card.

 

However, is that my only problem? I checked out my system diagnostics (attached) briefly, and the SMART data seems to be crazy bad. Almost all of my drives have at least a few "P"s in the "Prefailure Warning" section, and many have values of 100 in attributes where 100 is the "WORST" possible value. Am I reading this right? We did just drive across the country with the server in our car. We did try to take precautions with padding though, and some of these drives were relatively new. Is it possible I'm misunderstanding these results?

 

Note - I've disabled all my docker / plex stuff for now. I'm not as interested in getting that back up yet. That should be much easier once my server's disks are healthy and happy again.

tower-diagnostics-20210217-1622.zip

Edited February 18, 2021 by RollinDolan

[JorgeB]

9 hours ago, RollinDolan said:

Except after step 6, unraid kicks off a read check (not a rebuild of the emulated disk).

It started rebuilding disk2, but then there were errors on disk1 and then disk2, disks look fine, and multiple disk errors suggest a power, controller or connection problem.

 

 

[RollinDolan]

4 hours ago, JorgeB said:

It started rebuilding disk2, but then there were errors on disk1 and then disk2, disks look fine, and multiple disk errors suggest a power, controller or connection problem.

 

 

Thanks for the help! That would make sense, since the case is inconsistent.

 

Do you mind sharing what you looked for to determine the disks look OK? I’d like to be able to read the SMART reports in the future, but they looked negative to me this time.

[JorgeB]

29 minutes ago, RollinDolan said:

I’d like to be able to read the SMART reports in the future

Several places already exist with info, easily found with google, e.g.:

https://ma.juii.net/blog/interpret-smart-attributes