Runing bees for block de-duping and avoiding btrfs read only snapshots

[Marc MERLIN <marc@merlins.org>]

I have a backup server that gets btrfs sends from multiple machines.
I've been trying to use bees to dedupe blocks from multiple backup
sources.
Looks something like this:
/usr/local/lib/bees/bees -t -p --timestamps --absolute-paths /run/bees/mnt/a97dec85-a0d5-42ab-a0ef-e9b7479fbe43

The problem is a lot of my filesystem looks like this:
gargamel:~# l -d /mnt/btrfs_bigbackup/DS1/Soft_*
lrwxrwxrwx 1 root root  25 Mar  7 21:57 /mnt/btrfs_bigbackup/DS1/Soft_last_ro -> Soft_ro.20240307_21:56:57/
drwxr-xr-x 1 root root 196 Jan 23  2022 /mnt/btrfs_bigbackup/DS1/Soft_ro.20240101_03:13:05/
drwxr-xr-x 1 root root 196 Jan 23  2022 /mnt/btrfs_bigbackup/DS1/Soft_ro.20240107_23:25:43/
drwxr-xr-x 1 root root 196 Jan 23  2022 /mnt/btrfs_bigbackup/DS1/Soft_ro.20240212_13:26:05/
drwxr-xr-x 1 root root 196 Jan 23  2022 /mnt/btrfs_bigbackup/DS1/Soft_ro.20240307_21:35:59/
drwxr-xr-x 1 root root 196 Jan 23  2022 /mnt/btrfs_bigbackup/DS1/Soft_ro.20240307_21:56:57/

I don't want bees to look in every single RO copy to try do dedupe things since
they are already unique thanks to btrfs send/receive (that's how they were created)

bees does not seem to have a way to exclude some paths, although I could
give it --workaround-btrfs-send and then make a single rw snapshot for each
of my btrfs send targets, so that bees tries to dedupe those blocks once against
other snapshots from other targets.

Is it the least bad way to do things, or is there another way?

Thanks,
Marc
-- 
"A mouse is a device used to point at the xterm you want to type in" - A.S.R.
 
Home page: http://marc.merlins.org/                       | PGP 7F55D5F27AAF9D08