Am 2015-08-06 um 13:18 schrieb Rich Freeman: > It isn't necessarily essential, but btrfs fi df /mnt/gentoo will show > you that before the balance there are still some chunks in single mode > - it seems like mkfs creates the first device and adds the second one, > leaving some residual non-RAID chunks (that hopefully will never have > data written to them). The balance of an empty filesystem is really > fast and completely converts it to raid1, so I figured it would be > cleaner to do it this way. I have no idea what happens if those > single chunks remain and you degrade the array.
This reminded me of doing a balance-run on the 2-hdd btrfs-RAID1 in my desktop machine. Runs now. The machine runs and boots on btrfs only (as well as my 2 thinkpads), I know that btrfs still isn't as well tested as extX or XFS, for example ... but I am quite happy so far (doing backups is essential for everyone, right?) - Regarding the topic of this thread ... I am off-topic here ;-) sorry -> GPT on a single SSD (containing / and the OS, the hdds hold data), and systemd ... I run 2 systems (desktop and one laptop) with 2 distros installed in parallel, Fedora and Gentoo, and btrfs helps to share storage nicely here. It even works to share the EFI-boot-partition etc ... the only issue is that having multiple kernels for each distro frequently leads to manually remove one older kernel to be able to add another -> Yes, that partition was sized too small and isn't so easy to grow right now. No big problem. As I mentioned in another btrfs-related thread here a few months ago I really appreciate the move from partitions/LVM/RAID/filesystems to this new concept where all these layers are somehow integrated and interacting. Sorry for OT-ing here, regards, Stefan