On Thu, Jul 14, 2011 at 5:10 AM, JB <jb.1234abcd@xxxxxxxxx> wrote: > Well, then you have to read the thread more carefully before you bark back - > right in the first OP's post you have references, e.g. > > https://lkml.org/lkml/2010/6/18/144 Heh - now that you provide links it's better... anyway, that's a completely hilarious email you're linking to. "All such modifications must be reviewed by specialists in the theory of algorithms." I love it. Can you get some of those specialists to review the progress of all the other filesystems in use? ext2, 3, xfs, ReiserFS? > Here is one more: > http://www.ilsistemista.net/index.php/linux-a-unix/6-linux-filesystems-benchmarked-ext3-vs-ext4-vs-xfs-vs-btrfs.html Well, that's interesting. It is a pretty old benchmark, and it shows btrfs has is slower ins some operations -- on the other hand, it has some fantastic new features that the other FSs don't have. It's a tradeoff. > They all confirm the same picture ! The only picture here is that people who don't understand filesystems are talking nonsense. Myself, I use btrfs on some of my F14/F15 machines, and it's been rock-solid. Some things are slower - yep - but competing in performance with ext3/4, which has had decades of fine-tuning is... quite unfair. All FSs have some use-cases where they are... inappropiate. For example, ext3 is very badly behaved (performance-wise) in Maildir-based email servers due to directory fragmentation issues; similarly, any journalled filesystem will be inefficient tracking the transaction log of a database (pg_xlog). So far, it looks that having VM "disks" as files on BTRFS partition sucks in performance -- ok, not a big deal. Just use a different FS for your VM disks _or_ take the slowdown _or_ be helpful to people trying to address the problem. cheers, m -- martin.langhoff@xxxxxxxxx martin@xxxxxxxxxx -- Software Architect - OLPC - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff -- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/devel
