[SOLVED] Drive disabled after moving to a different slot

[gnollo]

tower-diagnostics-20200529-1303.zip

[JorgeB]

1 minute ago, gnollo said:

And now I get read errors from disk7,

Post current diags but if you're still getting errors on multiple disks rebuilding the disable disk will be a problem.

[JorgeB]

Emulated disk2 needs a filesystem check, but first you need to fix the disk7 errors, replace/swap cables on that disk.

[gnollo]

I did move the drive with read errors just now from the third Norco, so perhaps I have cabling issues on that one, so i will swap cables on that drive next.

I am now running 10 drives in two cages, trying to reduce elements of errors and so i could double up on the power connectors. 

[gnollo]

I will swap cables later on, can't access that room for another 3 hours. Why cant I access the emulated data? Is parity also fuxored?

[gnollo]

No folder for me to click on the right side of the WebUI either

[JorgeB]

22 minutes ago, gnollo said:

Why cant I access the emulated data? Is parity also fuxored?

Parity should be fine, but since you have a disk disable with single parity all the other disks need to be working correctly for the emulated disk to work, and disk7 is not.

[gnollo]

Ok turned the server off until I replace the cable.

[gnollo]

Ok replaced the cable, still no mounting and no drive access after starting the array.

Then I actually removed the drive physically from the server.

And rebooted it.

The UI still shows "Disk 2 not installed Unmountable: No file system"

And I cannot see a folder next to it.

And I cannot see data being emulated, or a drive being emulated.

No read errors now.

Diagnostics which I just donwloaded also attached.

 

tower-diagnostics-20200529-1758.zip

Edited May 29, 2020 by gnollo

[JorgeB]

5 hours ago, johnnie.black said:

Emulated disk2 needs a filesystem check

 

[gnollo]

Thanks Joe, do I put the disk back in? How do I peform a filesystem check? Sorry but I am really out of my depth with this issue.

[JorgeB]

2 minutes ago, johnnie.black said:

Emulated disk2 needs a filesystem check

Though not sure it will work since now unlike in the previous diags it's not detecting a valid reiser filesystem, if reiserfsck doesn't work check if the old disk mounts with UD, if yes and no new data was written since it got disabled it might be better to do a new config and re-sync parity instead.

[JorgeB]

8 minutes ago, gnollo said:

How do I peform a filesystem check?

https://wiki.unraid.net/Check_Disk_Filesystems#Checking_and_fixing_drives_in_the_webGui

or

https://wiki.unraid.net/Check_Disk_Filesystems#Drives_formatted_with_ReiserFS_using_unRAID_v5_or_later

[gnollo]

9 minutes ago, johnnie.black said:

Though not sure it will work since now unlike in the previous diags it's not detecting a valid reiser filesystem, if reiserfsck doesn't work check if the old disk mounts with UD, if yes and no new data was written since it got disabled it might be better to do a new config and re-sync parity instead.

What is UD? I have started the array in maintenance mode and clicked on CHECK for drive 2.

It says "running" below the following message "reiserfsck 3.6.27 Will read-only check consistency of the filesystem on /dev/md2 Will put log info to 'stdout'"

[gnollo]

Failed to open the device '/dev/md2': Unknown code er3k 127

[gnollo]

Just loaded the drive in unassigned devices, and I can access folders and download files.

[gnollo]

So I guess next step is new configuration? This time I just calculate parity from scratch? Does this overcome the system file corruption?

Sent from my SM-A520F using Tapatalk

[gnollo]

Started a fresh new config, parity date rebuild in progress, 28 hours to go.

[trurl]

On 5/24/2020 at 5:59 AM, gnollo said:

Last time I checked parity in 2019: 

Event: Unraid Parity check
Subject: Notice [TOWER] - Parity check finished (47 errors)
Description: Duration: 21 hours, 50 minutes, 34 seconds. Average speed: 101.8 MB/s
Importance: warning

I know this is several days old now, but I felt it needed commenting on.

 

Are you saying that your previous parity check had errors? Was that a correcting parity check? 

 

Exactly zero parity errors is the only acceptable result. If you get any errors at all you must correct them, then you should check again to make sure there isn't still some problem.

 

If you have parity errors how can you expect to rebuild a disk? One possible result of rebuilding a disk when you have parity errors is a corrupt filesystem and unmountable disk on the rebuild.

[gnollo]

Hi trurl, thanks for commenting, it' useful because I never really paid attention to what type of parity check is in progress. That particular parity check started upon booting automatically after an unclean shutdown due to a power outage, which type would that be, in your opinion?

I don't regularly perform parity checks, which I guess I should start doing on a regular basis. Which type of parity checks should I perform and how often would you suggest?

This drive that has filesystem issues used to be the parity drive. I bought a new 10TB drive to replace a drive that was showing pending sectors, but as it was the largest drive I have, it ended up becoming a copy of the old parity (bit more history in the thread below). I rebuild the drive with pending sectors using the old parity as replacement (8TB drive). So the drive was rebuild over two months ago. I still have the replaced drive, and I used Checksum Compare, a windows app, to check all the movie files across the old and replacement drive, which gave thumbs up for all the files.

 

 

[trurl]

I think unclean shutdown parity check is noncorrecting. Unclean shutdown often results in a small number of parity errors. You should try to avoid unclean shutdown and in this case it seems an UPS would have prevented it. The large number of parity errors you got on that other check suggests you had done something to invalidate parity. 

 

I do monthly noncorrecting parity checks. Then if parity errors are detected (rare) and there doesn't seem to be any other problem that might have caused them, then a correcting parity check. Finally, after parity errors are corrected, another noncorrecting parity check to confirm there are exactly zero parity errors. If not then you have some other problem which might be diagnosed by examining diagnostics.

 

We would need the complete syslog during those parity checks for comparison so no rebooting. Bad RAM is one possible reason for not being able to get to zero parity errors, for example.

[gnollo]

Thanks trurl, I will incorporate that in my monthly routines, and post a syslog once parity has been constructed in 17 hours time.

[gnollo]

Parity rebuild completed. Duration 1 day, 2 hr, 55 min, 32 sec Speed 103.2 MB/s errors 0 Status OK

[gnollo]

and diagnostics attached

tower-diagnostics-20200531-2335.zip

[gnollo]

Just completed a parity check, 0 errors, everything is looking good.