Announcement

Collapse
No announcement yet.

On-Disk Format Changes Ahead To Improve "Painful" Parts Of Btrfs Design

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • On-Disk Format Changes Ahead To Improve "Painful" Parts Of Btrfs Design

    Phoronix: On-Disk Format Changes Ahead To Improve "Painful" Parts Of Btrfs Design

    Prominent Btrfs file-system developer Josef Bacik is working through a big set of patches that will result in on-disk format changes to Btrfs but address some of "the more painful parts" to the file-system's design...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Oh no!

    Comment


    • #3
      If the ondisk format changes, why isn't btrfs still experimental?
      Are they going to make this btrfs2 or something?

      Comment


      • #4
        Originally posted by baka0815 View Post
        If the ondisk format changes, why isn't btrfs still experimental?
        Are they going to make this btrfs2 or something?
        I guess that changes to filesystem on-disk format is actually OK, XFS has ondisk format changes when it is already in the kernel.

        However the filesystem needs to retain backwards compatibility, meaning it needs to be able to mount all the old on-disk format.

        Comment


        • #5
          Honestly I, as a user, don't like where this is going.

          This work (well, not specifically this work, but the whole "extent tree v2" effort) is basically destroying the single killer feature of btrfs, that is, the ability to relocate extents at will (i. e. the btrfs balance operation, which leads into RAID reshape/restripe).

          When this is done, there will be no practical reason to prefer btrfs to zfs.
          Last edited by intelfx; 11 November 2021, 08:58 AM.

          Comment


          • #6
            intelfx Btrfs still comes native with the kernel, which I see as a big advantage compared to OpenZFS.

            @Thread: Why not combine the planned on-disk format changes with patches that fix raid5/6 and be done with it? Otherwise fixing raid will need another format change…

            Comment


            • #7
              Originally posted by kiffmet View Post
              Why not combine the planned on-disk format changes with patches that fix raid5/6 and be done with it? Otherwise fixing raid will need another format change…
              I guess because nobody so far is interested in writing those patches... They don't exist.

              Comment


              • #8
                Originally posted by kiffmet View Post
                intelfx Btrfs still comes native with the kernel, which I see as a big advantage compared to OpenZFS.
                It's really only an advantage at install time...and even then that's not much of an advantage seeing as how the installer and default distribution environment can include OpenZFS much like they can any other piece of software.

                It's not like we install Linux from the vanilla kernel. For most of us we have installers and frameworks and all sorts of stuff that manages installs for us and most of those don't cover everything the kernel supports. Even with vanilla supported features we have to pick our poison in regards to distributions that support the features we're after or we roll our own with Arch, Gentoo, Debian Minimal, or some other lower level install method.

                The only other time it's an advantage is if there is a critical system failure and the kernel and module somehow get out of sync OR your bootloader decides to go all loopy, possibly FUBAR, after an update (that can happen to any FS and any bootloader). Either way you almost likely need a rescue environment or disk handy with every tool under the sun which makes what vanilla Linux supports moot.

                Comment


                • #9
                  Originally posted by siyia View Post
                  Oh no!
                  Said GRUB to my BTRFS root around 4.19

                  Comment


                  • #10
                    I am perfectly willing to not fight, and say I am an idiot and doing it wrong--however, I have more issues with zfs than btrfs in normal use. I was hit with an openzfs bug in the last few months where docker and snapshots caused runaway writes to disk. The only solution for me was a new install and disabling a zfs feature. I had problems years ago, but it seems unfair to mention them. I'm also partial to the btrfs subvolumes compared to new zfs filesystems and zfs send/receive to move things around. Now, I have to close by saying I was lured to try both btrfs and zfs by all the buzz, but for root fs' I will probably return to ext4 and for RAID I will return to mdraid and xfs. If I'm not forced into upgrading my btrfs filesystems and can wait until I'm certain the changes are stable, I'm OK with that. If not, it will just be time to go back to the boring days of xfs and ext4. I'm not a fan of excitement when it comes to my storage.

                    Comment

                    Working...
                    X