On 17.05.2015 22:48, Nuno Magalhães wrote: > On Sun, May 17, 2015 at 8:36 PM, Stefan G. Weichinger <li...@xunil.at> wrote: >> tl;dr ... maybe you listed some reason to stick with mdadm/lvm2/xfs etc >> ... sorry in that case > > I didn't. 2 disks with RAID1/LVM, 2 disks (maybe) with ZFS. Pairs > because by board has 2 SATA channels, otherwise i'd go RAID5 and gain > an extra TB. > BTRFS seems a bit unstable at the moment (i could be wrong).
not for your use-case = RAID1 -> google for Chris Mason, btrfs, stable ... Ok, it is not as tested as XFS or ext4 ... but stable enough for desktops with a few disks (avoid RAID5/6 for now, yes .. but lev1 should be fine already). btrfs is in development still, sure (all other fs are in development as well ;) ). - So if you decide for it, keep your kernel and btrfs-progs updated ... and an eye on the btrfs-mailing-list(s). And don't overdo with the snapshots ( ... 10, ok ... hundreds, hmm ... ). backups ... no comment needed, right? There were problems with btrfs and the kernel a few months ago (Rich Freeman was hit by that, maybe he chimes in here), but in general for me it is still a very positive experience. I run it on 2 servers with RAID1-setups, a desktop with an SSD and a RAID1 for 2 hdds ... and on two laptops with dualboot-setups ... after the initial learning phase I am basically just using it and happy. No more re-partitioning hassle ... checksums overall ... snapshots if you want ... etc etc For example I cloned my working rootfs into a separate btrfs-subvolume, and rebuilt everything in there (using systemd-nspawn ... off-topic here) ... and then switched over to that rootfs by simply adding a new gummiboot entry. As you ordered 2 ssds right now this seems a perfect opportunity to start over and test something "new" (btrfs is in the linux kernel since 2009). good luck, have fun, Stefan