This issue has been plaguing me for a month. I noticed it initially because my Plex would start failing, but navigating my array filesystem was completely fine. After a bunch of looking into it to find the problem I came across these error messages from my cache drive when fixCommonProblems reported my Cache drive was read-only or full (and it's not full)

 

Looking into BTRFS errors I tried a whole slew of things to rebuild the file system safely, but ultimately nothing worked. I ended up completely wiping and rebuilding the drive including my docker image and my app data. Not the biggest deal, just a little bit annoying.

 

However, this didn't fix the issue at all. Within a few hours I was back to getting these errors. So I thought perhaps the entire drive is just dying and went ahead and bought a new 1TB cache. After a few days of working great, it's back. Same symptoms as before.

 

When I start my server everything will work fine and I won't get any errors. Somewhere between a few hours and even a few days my Plex will stop working and my drive will start returning these filesystem corrupted errors. If I restart my server it will start the loop over. Working for some amount of time and then failing again.

 

Jan 25 14:23:38 Server kernel: BTRFS error (device sdd1): bdev /dev/sdd1 errs: wr 0, rd 0, flush 0, corrupt 71, gen 0
Jan 25 16:46:17 Server kernel: BTRFS: error (device sdd1: state A) in __btrfs_free_extent:3110: errno=-117 Filesystem corrupted
Jan 25 16:46:17 Server kernel: BTRFS info (device sdd1: state EA): forced readonly
Jan 25 16:46:17 Server kernel: BTRFS info (device sdd1: state EA): leaf 419787161600 gen 1781 total ptrs 178 free space 5959 owner 2
Jan 25 16:46:17 Server kernel: BTRFS warning (device sdd1: state EA): bad eb member start: ptr 0x403be0 start 419787161600 member offset 4209648 size 8

 

My drives don't show any SMART errors and the cache is brand new. The drives are fairly full, but I have about 1TB free on the array.

My Docker Image is 64GB and is only 11% full. Plenty of RAM is free, only 1% for the log filesystem.

 

Since I rebuilt my cache and appData from scratch, I don't have that many dockers doing anything. I just set up the bare necessities to get Plex back online.

 

Prior to all these issues I was on 6.12.13 and have since upgraded to 7.0.0. I included my system logs from right after a crash.

android13-syslog-20250125-2304.zip

Since btrfs is detecting data corruption, recommend running memtest first, then backup and reformat the pool