btrfs corruption

[modest hare]

So 24 hours ago I got out of memory and had to force a poweroff and now btrfs gave up and remounted in read-only. (probably because I had swap partition)
After booting up into an iso I have following, what do I do? btrfs check --repair?

[mild carbon]

@modest hare are you in the chroot?

[modest hare]

no

right now I mounted spare ssd and successfully mounted falling apart btrfs filesystem and just copying my home to the spare ssd

but after I made the backup what do I do?

well first what I do is shutdown and flash fresh iso because btrfs on falling apart system is v6.7 vs v6.5.2 and doubt it is a good idea to repair with this old version

[modest hare]

so should I do btrfs check --repair in chroot?

[modest hare]

so I just did the repair without chroot or anything it said it deleted something but system boots just fine so far I did not notice anything vital missing

but how do I avoid such things in the future? (aside from setting up oom killer)

[dull sail]

btrfs is usually very force-poweroff-resistant, this is a rare case

you dont even need to fsck it after a force poweroff, unlike smth like ext4

[modest hare]

How do I (possible?) extract filenames or paths of corrupted stuff from such a state?

[spring arrow]

Mounting it read only

[modest hare]

of files that cause the corrupted state, like when btrfs check tells me data extent and all that other information how do I convert it into a path to a file on the filesystem

I did mount the filesystem in read only and copied everything over but I still don't know what files were broken/missing

[long ledge]

AFAIK there's no way to know what files are broken/missing unless you compare them to a known working state. EI: You would know if any are missing by comparing old file list to new file list. You could know which are corrupt by comparing old checksum and new checksum.