Re: How to convert a directory to a subvolume

[John Center <jlcenter15@gmail.com>]

root@Mariposa:/# mount | grep btrfs

/dev/sda3 on / type btrfs 
(rw,relatime,space_cache=v2,autodefrag,subvolid=256,subvol=/@)
/dev/sda3 on /home type btrfs 
(rw,relatime,space_cache=v2,autodefrag,subvolid=257,subvol=/@home)
/dev/sda3 on /opt type btrfs 
(rw,relatime,space_cache=v2,autodefrag,subvolid=268,subvol=/@opt)

Thanks, Hugo & Chris!

     -John


On 5/2/22 12:15 PM, Chris Murphy wrote:
> On Mon, May 2, 2022 at 11:37 AM John Center <jlcenter15@gmail.com> wrote:
>> Hi Chris,
>>
>> Thanks for responding!  Well, I made some progress:
>>
>> root@Mariposa:/# mount /dev/sda3 /mnt
>>
>> root@Mariposa:/# btrfs subvolume create /mnt/@opt
>> Create subvolume '/mnt/@opt'
>>
>> root@Mariposa:/# btrfs subvolume list /
>> ID 256 gen 11732 top level 5 path @
>> ID 257 gen 11733 top level 5 path @home
>> ID 268 gen 11733 top level 5 path @opt
>>
>> Once I unmounted /mnt, I had to create the /opt directory again:
>>
>> root@Mariposa:/# umount /mnt
>>
>> root@Mariposa:/# mkdir opt
>>
>> Then I mounted /opt & it didn't complain:
>>
>> root@Mariposa:/# mount /opt
>> root@Mariposa:/# btrfs subvolume list /
>> ID 256 gen 11751 top level 5 path @
>> ID 257 gen 11751 top level 5 path @home
>> ID 268 gen 11733 top level 5 path @opt
>>
>> So, is this correct?
> Looks correct. You need to include something like 'mount | grep btrfs'
> to know for sure, that should show subvol and subvolid for each mount
> point.
>
>
>
>> It would have been good if there was a direct way
>> to add the level that you wanted the subvolume created & mounted with
>> the btrfs command.  So if I take a snapshot of @, /opt won't be
>> included?
> Btrfs snapshots via the btrfs-progs CLI tool are not recursive. So
> whether the subvolumes are arranged in a nested or flat layout doesn't
> matter. There is a non-atomic recursive snapshot (create and remove)
> option in libbtrfs both for C API and Python bindings.
>
>
>> That'd my goal.  I've been using btrfs at a very basic level,
>> added on top of an md raid1.  It worked well for a number of years, but
>> then I needed to replace the disks, I thought I'd go all in on btrfs.
>>
>