Hello, recently went to download some things and noticed it kept failing. looked into fix common problems and it said my cache drive for downloads is read only, probably full capacity is used. i checked and its only sitting at half the space used. I don't have any of my docker containers on here only new data which gets moved as the drive is nearing capacity. I have removed the cache as the primary storage to check if things still work with my main array and they do however when i add the cache back in the same thing happens. I don't want to unmount the cache yet before getting some advice from here first as best steps to take. The data on it isn't precious but id like to try and not lose anything if possible.

I have attached my diagnostics

any advice or pointers would be helpful.

sith-diagnostics-20250729-1536.zip

Your sysl;og is full of errors of the form:

Jul 29 14:02:01 Sith kernel: BTRFS error (device nvme0n1p1: state EA): bdev /dev/nvme0n1p1 errs: wr 0, rd 270833, flush 0, corrupt 0, gen 0

Jul 29 14:02:01 Sith kernel: BTRFS error (device nvme0n1p1: state EA): bdev /dev/nvme0n1p1 errs: wr 0, rd 270834, flush 0, corrupt 0, gen 0

Jul 29 14:02:01 Sith kernel: BTRFS error (device nvme0n1p1: state EA): bdev /dev/nvme0n1p1 errs: wr 0, rd 270835, flush 0, corrupt 0, gen 0

Jul 29 14:02:24 Sith kernel: BTRFS error (device nvme0n1p1: state EA): bdev /dev/nvme0n1p1 errs: wr 0, rd 270836, flush 0, corrupt 0, gen 0

Jul 29 14:02:24 Sith kernel: BTRFS error (device nvme0n1p1: state EA): bdev /dev/nvme0n1p1 errs: wr 0, rd 270837, flush 0, corrupt 0, gen 0

Jul 29 14:02:24 Sith kernel: BTRFS error (device nvme0n1p1: state EA): bdev /dev/nvme0n1p1 errs: wr 0, rd 270838, flush 0, corrupt 0, gen 0

which shows that the cache_downloads pool is corrupt. @JorgeB is probably the best person for giving recovery advice.

Since corruption of btrfs file systems is frequently the result of RAM issues have you recently run memtest from the Unraid boot menu?

That NVMe device appears to have dropped, but the syslog already rotated, reboot and post new diags after array start.

7 minutes ago, itimpi said:

Your sysl;og is full of errors of the form:

Jul 29 14:02:01 Sith kernel: BTRFS error (device nvme0n1p1: state EA): bdev /dev/nvme0n1p1 errs: wr 0, rd 270833, flush 0, corrupt 0, gen 0

Jul 29 14:02:01 Sith kernel: BTRFS error (device nvme0n1p1: state EA): bdev /dev/nvme0n1p1 errs: wr 0, rd 270834, flush 0, corrupt 0, gen 0

Jul 29 14:02:01 Sith kernel: BTRFS error (device nvme0n1p1: state EA): bdev /dev/nvme0n1p1 errs: wr 0, rd 270835, flush 0, corrupt 0, gen 0

Jul 29 14:02:24 Sith kernel: BTRFS error (device nvme0n1p1: state EA): bdev /dev/nvme0n1p1 errs: wr 0, rd 270836, flush 0, corrupt 0, gen 0

Jul 29 14:02:24 Sith kernel: BTRFS error (device nvme0n1p1: state EA): bdev /dev/nvme0n1p1 errs: wr 0, rd 270837, flush 0, corrupt 0, gen 0

Jul 29 14:02:24 Sith kernel: BTRFS error (device nvme0n1p1: state EA): bdev /dev/nvme0n1p1 errs: wr 0, rd 270838, flush 0, corrupt 0, gen 0

which shows that the cache_downloads pool is corrupt. @JorgeB is probably the best person for giving recovery advice.

Since corruption of btrfs file systems is frequently the result of RAM issues have you recently run memtest from the Unraid boot menu?

No I haven’t run any mem tests, the server has been running fine for months with no input from me for a while, it does its thing. I have rebooted the server and it all came back ok. The cache is working again. Currently moving everything to the array in case it drops again.

4 minutes ago, JorgeB said:

That NVMe device appears to have dropped, but the syslog already rotated, reboot and post new diags after array start.

I’ve no idea what has caused this. I’ve seen some posts else where saying btrfs might be the cause of this ? Should I clean the drive out and reformat to xfs? After a reboot things have come back to normal and I’m currently moving the data to the array.

No, Btrfs doesn't make devices drop, it's typically a hardware/firmaware issue, or a combination of the hardware and the kernel.

1 minute ago, JorgeB said:

No, Btrfs doesn't make devices drop, it's typically a hardware/firmaware issue, or a combination of the hardware and the kernel.

So possibly the drive is going bad. I’ll need to look at buying a new one. As for a posibility of hardware and kernel would updating to the latest stable version be advisable to eliminate that as a factor?

2 minutes ago, Darth-Sidious said:

So possibly the drive is going bad

Not necessarily; it would be good to see the error when it drops. In some cases adding some kernel options can help, if it's the more common error.