Announcement

Collapse
No announcement yet.

Linux 5.4 EXT4 / XFS / Btrfs RAID Performance On Four HDDs

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

  • Linux 5.4 EXT4 / XFS / Btrfs RAID Performance On Four HDDs

    Phoronix: Linux 5.4 EXT4 / XFS / Btrfs RAID Performance On Four HDDs

    Recently a Phoronix reader inquired about seeing some fresh hard drive RAID benchmarks on the current kernel release and using Btrfs / EXT4 / XFS. While we don't often look at HDD RAID performance these days compared to speedier SSD testing, since the reader was a generous Phoronix Premium member I was happy to oblige to his test request. Here is a look at the Linux 5.4 HDD RAID performance per his request with Btrfs, EXT4, and XFS while using consumer HDDs and an AMD Ryzen APU setup that could work out for a NAS type low-power system for anyone else that may be interested.

    http://www.phoronix.com/vr.php?view=28708

  • #2
    Pretty good improvements on Btrfs part compared to a year ago.

    I think you should run these benchmarks with noatime as default though. atime hurts performance, especially for COW filesystems, and is not really used by anything these days, unless you use mutt.

    Comment


    • #3
      i cure myself of Seagate a few years ago. i had 2 hdds of 3 tb each bought at 1 year difference. and both crashed at 1 year difference. i was lucky that both still had warranty. and these drivers was mostly for movies or storage. i mostly do only read after both of them were full. i didn't even bother to delete what i have on it once it was full. its strange to have hdd failure from 90% of usage time for reading. and to make it clear none of them were use for 2-3 hours/day. probably not even 1 min/week sometimes

      Comment


      • #4
        What a coincidence , I was actually planning on asking about a BTRFS RAID test too, but here it is. That being said I am a bit disappointed as I think the benchmark would have been fairer if i twas tested with BTRFS on top of MDRAID as with the other filesystems as well. It would be interesting to see BTRFS on top of MDRAID vs it's own built in "RAID"-like capabilities.

        A word of warning, BTRFS does NOT yet (/still) distribute reads to the fastest/least busy storage devices in the pool neither for "RAID1" or "RAID10" so it could in theory perform much better for parallel read operations.

        http://www.dirtcellar.net

        Comment


        • #5
          BTRFS or ZFS on top of MDRAID is half useless. Don't use it that way.

          Errors can be detected but never repaired and you'll have to rebuild the filesystem to fix the problem.

          Comment


          • #6
            With 4 drives it is possible to get RAID5, would be possible to test it is as well?

            Comment


            • #7
              Pretty good improvements on Btrfs part compared to a year ago.
              Maybe it's because Red Hat is no longer boycotting?

              Comment


              • #8
                Originally posted by Zan Lynx View Post
                BTRFS or ZFS on top of MDRAID is half useless. Don't use it that way.

                Errors can be detected but never repaired and you'll have to rebuild the filesystem to fix the problem.
                It is not as useless as you might think. BTRFS on top of MDRAID is just as useful as any other filesystem on top of MDRAID. With BTRFS you will get data errors detected (or if you store withe DUP profile you can get them corrected as well). Metadata will be corrected by default and you have nice little features like snapshots, compression etc which you don't always get with other filesystems.

                http://www.dirtcellar.net

                Comment


                • #9
                  Originally posted by waxhead View Post

                  It is not as useless as you might think. BTRFS on top of MDRAID is just as useful as any other filesystem on top of MDRAID. With BTRFS you will get data errors detected (or if you store withe DUP profile you can get them corrected as well). Metadata will be corrected by default and you have nice little features like snapshots, compression etc which you don't always get with other filesystems.
                  with md underneah, btrfs is not able to perform seal-healing in case of bad blocks. Which from my point of view is the killer feature of both ZFS and btrfs. So yes, in this case using btrfs is useless.

                  Comment


                  • #10
                    Originally posted by waxhead View Post

                    With BTRFS you will get data errors detected (or if you store withe DUP profile you can get them corrected as well). Metadata will be corrected by default
                    Maybe yes, maybe no. With a faulty device, btrfs will detect errors and will try to correct it with the dupe. But it has no knowledge about the RAID and the RAID-layout. It doesn't know how to find dupes on non-error devices. This will induce new error-scenarios. I don't know about btrfs and caches, but because of this, ZFS must not run hard- or software RAID. Correctness is only guaranteed when ZFS has full control about all caches.

                    Comment

                    Working...
                    X