Btrfs Sees A Number Of Improvements With Linux 5.8

[pal666]

I fail to see how this could possibly happen. Can anyone explain that to me?

if(error) mark_readonly();

[pal666]

First you write new metadata, then you delete the old

you can't delete old when it's used by some snapshot

[pal666]

Btrfs introduced a "GlobalReserve" to make sure that this temporary space is always available.

s/always/usually/

[pal666]

readwrite, showing an error about failing to recover the transaction log.

did you try mount -o nologreplay or mount -o norecovery ?

[archkde]

you can't delete old when it's used by some snapshot

Recursive snapshots don't exist.

[Zan Lynx]

Sometimes (rarely) on unclean shutdown of my laptop, if the power is interrupted due to lack of battery, on next boot the kernel refuses to mount the filesystem readwrite, showing an error about failing to recover the transaction log.

I had that happen once to me. Although in my case the error was unrecoverable (the data was recoverable but the metadata could not be repaired) and more serious.

The SSD in the laptop somehow didn't write some of the data in the middle of its cache. The btrfs metadata sequence numbers were impossible because blocks that had to have been committed were not there.

Anyway, apparently bad SSDs which don't either commit everything or roll back to a fully consistent state really mess up btrfs metadata.