Reset-resistant rootfs mount options for btrfs?

Software
1

I’m using debian 13 and the (default?) mount options are:

/dev/mmcblk1p2 on / type btrfs (rw,noatime,compress=zstd:3,ssd,discard=async,space_cache=v2,subvolid=5,subvol=/)

Seems like the needed barrier option is on by default, but maybe I miss something?

The goal is to have it survive random reboots once a while without any intervention (e.g. fsck asking to press ‘y’ on the console after an unfortunate reboot)

Should usebackuproot be added? Or just rescue=all?

The proper way would be to have a read-only / with /var and other stuff on a separate volume, but perhaps with current filesystems this isn’t needed?

Thanks.

2

Btrfs is designed to be crash-resistant with COW so it won’t corrupt. It will just potentially lose a certain period of data. You can set the values below:

  1. commit=10 (or commit=30): Sets the interval (in seconds) for flushing metadata to disk. Lower values improve data safety at the cost of write performance. You lose the last 10 seconds of writes.
  2. flushoncommit: Forces a flush to the storage device on every commit. Recommended for systems without battery-backed caches or when power-loss protection is critical.
  3. discard may also be appropriate.