Neil Bothwick <[email protected]> writes: > On Tue, 05 Jan 2016 23:16:48 +0100, lee wrote: > >> > I would run btrfs on bare partitions and use btrfs's raid1 >> > capabilities. You're almost certainly going to get better >> > performance, and you get more data integrity features. >> >> That would require me to set up software raid with mdadm as well, for >> the swap partition. > > There's no need to use RAID for swap, it's not like it contains anything > of permanent importance. Create a swap partition on each disk and let > the kernel use the space as it wants.
When a disk fails a swap partition is on, the system is likely to go down. Raid is not a replacement for backups. >> The relevant advantage of btrfs is being able to make snapshots. Is >> that worth all the (potential) trouble? Snapshots are worthless when >> the file system destroys them with the rest of the data. > > You forgot the data checksumming. Not at all, I'm seeing it as an advantage, especially when you want to store large amounts of data. Since I don't trust btrfs with that, I'm using ZFS. A system partition of 50 or 60GB --- of which about 10GB are used --- is not exactly storing large amounts of data, and the data on it doesn't change much. In this application, checksums would still be a benefit, yet a rather small one. So as I said, the /relevant/ advantage of btrfs is being able to make snapshots. And that isn't worth the trouble. > If you use hardware RAID then btrfs > only sees a single disk. It can still warn you of corrupt data but it > cannot fix it because it only has the one copy. or it corrupts the data itself ;) >> Well, then they need to make special provisions for swap files in btrfs >> so that we can finally get rid of the swap partitions. > > I think there are more important priorities, its not like having a swap > partition or two is a hardship or limitation. Still needing swap partitions and removing the option to use swap files instead simply defeats the purpose of btrfs and makes it significantly harder to use.
