Announcement

Collapse
No announcement yet.

Btrfs In Linux 4.1 Has Fixes For File-Systems Of 20 Terabytes & Up

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

  • Btrfs In Linux 4.1 Has Fixes For File-Systems Of 20 Terabytes & Up

    Phoronix: Btrfs In Linux 4.1 Has Fixes For File-Systems Of 20 Terabytes & Up

    It's nearing the end of the Linux 4.1 kernel and Chris Mason has now sent in his pull request of Btrfs file-system updates for this next kernel update...

    http://www.phoronix.com/scan.php?pag...trfs-Linux-4.1

  • #2
    I think I just learned something about Facebook. I'm not exactly sure what it is that I learned, but learn I did.

    Comment


    • #3
      Who uses Btrfs?

      Who uses Btrfs in production?
      Does Facebook use Btrfs?

      Comment


      • #4
        Originally posted by uid313 View Post
        Who uses Btrfs in production?
        Does Facebook use Btrfs?
        Yes, for some things. They also employ some of the main developers.

        Comment


        • #5
          Originally posted by uid313 View Post
          Who uses Btrfs in production?
          AFAIR they do at Facebook. We use btrfs in our mirror servers at our university department, yet for smaller filesystems of only about 6-10 TiB oft size. They do a nightly rsync copy oft the Main fileserver and snapshot that copy. All the snapshots are reexported via NFS. Works fine since about two years.

          Before, we used rsync to hardlink the unchanged files during sync. That worked, too, however was much slower than the snapshot variant. Speed is important, as the backup needs to be done on every server in 8h.

          Comment


          • #6
            Originally posted by oleid View Post
            AFAIR they do at Facebook. We use btrfs in our mirror servers at our university department, yet for smaller filesystems of only about 6-10 TiB oft size. They do a nightly rsync copy oft the Main fileserver and snapshot that copy. All the snapshots are reexported via NFS. Works fine since about two years.

            Before, we used rsync to hardlink the unchanged files during sync. That worked, too, however was much slower than the snapshot variant. Speed is important, as the backup needs to be done on every server in 8h.
            Have you considering using send+receive instead of rsync?

            Comment


            • #7
              Originally posted by nanonyme View Post
              Have you considering using send+receive instead of rsync?
              Not really. You have to run btrfs on both ends, right? We are (still) using ext4 on the main file server. The last time I checked (and that was some time ago) btrfs didn't support quota (via NFS).

              But I'll try it at home for my backup needs...

              Comment

              Working...
              X