Wow i'm scared at the moment. I'm trying to replace my remaining 12TB with 18TB. I unassigned a 12TB and assigned a precleared 18TB, all fine no issues, rebuild starts. 4TB in I get a notification that disk 2 has read errors and is disabled, 2048 errors. Ok i'll let it rebuild and replace that drive maybe its dying after. I wake up today and parity 2 has 563,197,701 errors and increasing by the second?! What is going on I attached my diagnostics. I've never had issues with unraid till the last few weeks in the years Ive been using it! Current rebuild has 5 hrs left. Looks like syslog is full cause of "Mar  2 09:37:00 Toblerone kernel: md: disk29 read error, sector=30864609368" I just increased log size to get more info. The temp on parity 2 has a star idk why but its still marked as green, "normal operation"?

toblerone-diagnostics-20240302-0932.zip

Biggest question is since one drive is rebuilding, one is disabled and one parity drive is having a few million read errors, is my data fine at the moment?

Under attributes it says Smartctl open device/dev/sdi failed for parity 2

 

All user shares on the Array do say "SOME OR ALL FILES UNPROTECTED"

Edited March 2, 20242 yr by Johann

I found this similar situation but didn't really identify the fix here:

Looks like the rebuild thinks is continuing and reading from drives but its not writing to the rebuilding drive. I'll let it finish its course. I think the next step before doing anything would be to wait for its to finish "rebuilding" and running xfs_repair -v /dev/mdX on each disk in maintenance mode from https://docs.unraid.net/unraid-os/manual/storage-management/#xfs-and-reiserfs to verify everything all filesystems are working. Once confirmed all the filesystems are good to go, i think I would keep disk 2 disabled since the data should be intact on that drive but I'm not sure why it had read errors. But next step to rebuilding disk 3 on top of itself? The disk it is replacing is still in unassigned drives with data intact and untouched.

 

What would be the best course of action here, I'm gonna wait for a response before I break everything...

 

I have a precleared drive sitting in unassigned drives and I'm not sure if this is related, but on the precleared drives, running fdisk -l, I see this error "Partition 1 does not start on physical sector boundary". I saw this before on the drives I precleared.

Here's the problems right now. Disk 2 is disabled. Disk 3 is rebuilt, but I'm uncertain whether or not the data is correct since there weren't writes to the disk for a while but other disks were reading. Parity 2 has dropped itself from the array and currently its "unassigned.

Looks more like a power/connection problem, check/replace cables and post new diags after arrays start, don't write anything to the array for now.