I thought raid0 "striped" the data across two or more devices to
increase total capacity, for example when adding a new device to an
existing filesystem. But that is not apparently what I ended up with.
Before:
btrfs device usage /mnt/backup/
/dev/sdc1, ID: 1
Device size: 300.00GiB
Device slack: 0.00B
Data,single: 226.01GiB
Metadata,DUP: 8.00GiB
System,DUP: 64.00MiB
Unallocated: 65.93GiB
/dev/sdc2, ID: 2
Device size: 300.00GiB
Device slack: 0.00B
Data,single: 1.00GiB
Unallocated: 299.00GiB
Then, I ran command:
btrfs balance start -dconvert=raid0 -mconvert=raid1 /mnt/backup
And what I ended up with seems to be double the amount of data used,
like what I think would happen with raid1, not raid0:
btrfs device usage /mnt/backup/
/dev/sdc1, ID: 1
Device size: 300.00GiB
Device slack: 0.00B
Data,RAID0: 228.00GiB
Metadata,RAID1: 5.00GiB
System,RAID1: 64.00MiB
Unallocated: 66.94GiB
/dev/sdc2, ID: 2
Device size: 300.00GiB
Device slack: 0.00B
Data,RAID0: 228.00GiB
Metadata,RAID1: 5.00GiB
System,RAID1: 64.00MiB
Unallocated: 66.94GiB
Or, am I misinterpreting what I am seeing? Thank you.
# uname -a
Linux tux 5.10.7-arch1-1 #1 SMP PREEMPT Wed, 13 Jan 2021 12:02:01
+0000 x86_64 GNU/Linux
# btrfs --version
btrfs-progs v5.9
# btrfs fi show
Label: none uuid: c0f4c8e2-b580-4c0d-9562-abdb933b9625
Total devices 1 FS bytes used 13.11GiB
devid 1 size 449.51GiB used 14.01GiB path /dev/sda3
Label: none uuid: 4fe39403-7ba1-4f22-972f-5041e3b6ff6f
Total devices 1 FS bytes used 37.36GiB
devid 1 size 600.00GiB used 40.02GiB path /dev/sdb1
Label: none uuid: 1751eeca-c1a2-47bb-906b-c7199b09eb6d
Total devices 2 FS bytes used 229.57GiB
devid 1 size 300.00GiB used 233.06GiB path /dev/sdc1
devid 2 size 300.00GiB used 233.06GiB path /dev/sdc2
# btrfs fi df /mnt/backup
Data, RAID0: total=456.00GiB, used=226.65GiB
System, RAID1: total=64.00MiB, used=64.00KiB
Metadata, RAID1: total=5.00GiB, used=2.92GiB
GlobalReserve, single: total=401.84MiB, used=0.00B
