Replace drive question (array order/infos get lost on array stop)

[JorgeB]

Please post new diags.

[Darkguy]

On 6/18/2022 at 11:16 AM, JorgeB said:

Please post new diags.

New diags attached, thanks

darkrack-diagnostics-20220619-1146.zip

[JorgeB]

Jun 18 09:41:01 DarkRack kernel: sd 8:0:1:0: Power-on or device reset occurred

Multiple errors like the above, this is usually a power/cable problem.

[Darkguy]

Seems I fixed the error, unplugged all SATA cables to the bay in question and reattached them. The error seems to be gone now, I copied the whole contents from the ddrescue-image of the original disk back over to the new disk and no more UDMA CRC errors.

 

There still seems to be a problem with the super.dat file though - all drives lose assignments on a reboot (but no longer on a stop/start array action).

 

There also is no new super.dat created in /boot - there was the first time I started off the new flash drive, but I still had the error about it in the syslog, deleted it and it doesn't get created now.

 

Current diagnostics attached.

darkrack-diagnostics-20220622-1925.zip

[JorgeB]

Didn't re-read the entire thread, but if you didn't yet try recreating the flash drive, backup current one, re-do it manually or by using the USB tool, then restore the config folder.

[Darkguy]

8 minutes ago, JorgeB said:

Didn't re-read the entire thread, but if you didn't yet try recreating the flash drive, backup current one, re-do it manually or by using the USB tool, then restore the config folder.

 

That's basically what I did to create the current flash drive a few days ago.

[JorgeB]

Then I would try with a different flash drive, you can use a trial key to make sure it works before transferring the current one.

[Darkguy]

Restarting this thread:

 

I had to replace my flash drive again this weekend (the previous one I used showed read/write errors). I opted for a new, factory sealed SanDisk USB 2.0 drive this time around.

 

There still seems to be some problem with the super.dat file/storage of disk slots:

 

• if I merely stop and restart the array, everything is file; disks keep their assignments and I can start the array without problems
• as soon as I change anything in the list of disks (right now, I'd like to replace disk5 as described here and set disk5 to "no device"), all disk assignments are lost; there is a corresponding entry in the log:

Sep 26 08:49:53 Tower kernel: read_file: error 2 opening /boot/config/super.dat

Sep 26 08:49:53 Tower kernel: md: could not read superblock from /boot/config/super.dat

 

The super.dat file is fine up to this point, and can also be read and restored from backup.

 

What I am trying to achieve is:

• replace disk5 by disabling the array, setting disk5 to be emulated by parity, insert a new disk, preclear the new disk using the Unassigned Devices Preclear script, then use the new disk in the disk5 slot
• repeat the whole procedure for disk3

Log attached, anything else I can try? Slowly running out of space and I have disks lying around to increase storage space by a total of 4.5 TB which I currently cannot use.

darkrack-diagnostics-20220926-1006.zip

[JorgeB]

Strange one, do you have another PC you could use the flash drive with just to test if you have the same issue when changing assignemnts?

[Darkguy]

5 hours ago, JorgeB said:

Strange one, do you have another PC you could use the flash drive with just to test if you have the same issue when changing assignemnts?

You mean another PC to create the drive from?

[JorgeB]

Just to see if you can assign and unassign drives without losing the array config.