My cache SSD was running hot and throwing thermal alerts, even when the rest of the drives and server were running cool.  The logs were showing some errors, and the drive is nearly six years old so I decided to replace it and move to a cache pool in RAID1.

 

Have added the two new drives (one brand new, one that was lying spare) and tried to balance in the Cache settings but it aborts.  Another forum post said it might be down to errors and the existing SSD should be scrubbed, and then the balance would work so I tried that.  Scrub also aborts, so I'm a little stuck.

 

Any help anyone could offer would be appreciated.  Happy to work in the terminal if given commands to paste.

 

Diagnostics attached.

kelvin-diagnostics-20230626-1317.zip

Log is being spammed with these:

Jun 26 11:55:37 Kelvin nginx: 2023/06/26 11:55:37 [error] 5612#5612: *18071 limiting requests, excess: 20.785 by zone "authlimit", client: 192.168.123.101, server: , request: "PROPFIND /login HTTP/1.1", host: "192.168.123.77"

Do you know what 192.168.123.101 is?

192.168.123.101 is my Windows workstation.

 

On the last reboot I had to wait to login as the webpage was giving a bad page error.  Left it alone for a while, then it worked fine.  

Try rebooting to see if they clear, due to that spam it's not possible to see anything else.

New diags attached.

kelvin-diagnostics-20230626-1425.zip

Jun 26 14:24:32 Kelvin kernel: BTRFS info (device sdb1): bdev (efault) errs: wr 211785, rd 51347, flush 2014, corrupt 1767, gen 0
Jun 26 14:24:32 Kelvin kernel: BTRFS info (device sdb1): bdev /dev/sdb1 errs: wr 206663196, rd 981210, flush 1771262, corrupt 1014082, gen 467

 

Current pool is kind of a mess, there's a missing device and many read/write errors logged for both, suggest you copy what you can form the current pool and create a new one, also see here for better pool monitoring for the future.

Thank you.  Have created a new pool, and have everything up and running again.