BTRFS Read Only Mystery

Issues & Assistance
1

Hi, this is a weird one.

System: 2TB NVME, 8TB HDD, 10TB HDD
Partitions: half the NVME is NTFS (Windows), the other half is BTRFS (CachyOS root/home), 8TB is BTRFS, 10TB is BTRFS. Both of the larger hard drives are used only with Cachy and not with Windows.

System clocks in Windows and Cachy were set to correct time.

The issue: While downloading for half an hour or more, the 8TB drive will go read-only and give a disk-write error. This has occurred when downloading games from steam (in Cachy, directly to the BTRFS 8TB drive) or from my NAS (Synology) via Rsync.

What I’ve done so far:
sudo btrfs property set / ro false →KDE Partition Manager Unmount then Mount this partition removes the read-only status and use can resume, but long downloads will cause it to re-occur.

sudo btrfs scrub start /dev/sda2 (the affected drive) returns no errors found. Disk must be mounted for this operation

sudo btrfs check /dev/sda2 (the affected drive) returns no error found and correct byte count. Disk must be unmounted for this operation

/etc/fstab lists all drives by UUID. The 8TB and 10TB drives have nofail and 0 0.

What should I try next?

2

Checking for hardware issues should always be top of mind when a checksumming filesystem flags errors.

3

I suppose, but i’m fairly convinced there is no hardware issue and that this is some software flag that keeps resetting inadvertently. What else

4

This is a classic symptom of thermal throttling or overheating in the hard drive’s controller or firmware.

2 sets of eyes here and agree on hardware related. :frowning:

5

Check logs right after drive goes read only:

sudo dmesg | tail -n 50

Look for “I/O error”, “resetting link”, or Btrfs-specific “critical space allocation failed” or “forced readonly” messages. This will confirm if the kernel is losing the connection to the hardware.

Check smart data:

sudo smartctl --all /dev/nvme0 #or whatever your drive is /dev/sda etc......

check fstab doesn’t have any conflicting mount points

most likely failing drive or drive heat

6

by just vibes? what have you checked to rule hardware issues (which match your problem description…) out?

I understand the desire to have your hardware working perfectly…but these are almost always symptoms that it isn’t.

7

not by just vibes, but this development is unexpected. I pinpoint it to a thermal issue. The hardware is fine as I believed, it’s just that now, the position in the case is no good. Swapping drive bays resolved it.

8

To be clear and explicit - this is a hardware problem

“Hardware problem” doesn’t necessarily mean “you need to replace something”.