openSUSE Conference 2018 - btrfs is awesome, except when it isn't

แชร์
ฝัง
  • เผยแพร่เมื่อ 19 พ.ย. 2024

ความคิดเห็น • 11

  • @danceswithglowstix22
    @danceswithglowstix22 6 ปีที่แล้ว +8

    I love the idea of btrfs. I'd rather see it production ready than getting new features and unstable though.

    • @kwagnert
      @kwagnert 6 ปีที่แล้ว +1

      It's already production level. Just some features are not: btrfs.wiki.kernel.org/index.php/Production_Users

    • @ypierro
      @ypierro 3 ปีที่แล้ว +1

      @@kwagnert I had more issues with btrfs in the only year I'd used it than with every other filesystem in 20 years. No, it's not production ready... at the same level as ext4.

  • @galtbarber2640
    @galtbarber2640 3 ปีที่แล้ว

    This command will attempt to relocate data in empty or near-empty data chunks
    (at most 5% used, in this example), allowing the
    space to be reclaimed and reassigned to metadata.

  • @IkarusKommt
    @IkarusKommt 5 ปีที่แล้ว +1

    Just drop it already, like RH did. It endangers home users.

    • @hunibuni
      @hunibuni 4 ปีที่แล้ว

      RH?

    • @IkarusKommt
      @IkarusKommt 4 ปีที่แล้ว +1

      @@hunibuni Redhat

    • @hunibuni
      @hunibuni 4 ปีที่แล้ว

      @@IkarusKommt So what does Redhat use now after dropping it? And how does it endanger home users?

    • @IkarusKommt
      @IkarusKommt 4 ปีที่แล้ว +2

      @@hunibuni RH recommends ext4 or xfs. btrfs is prone to data corruption when employed on the computers without UPS and hard drives without autonomous power, such as home and small business users usually have.

    • @uzefulvideos3440
      @uzefulvideos3440 4 ปีที่แล้ว +4

      ​@@IkarusKommt That's only true for Btrfs RAID5/6, and when you use Btrfs RAID5/6 with RAID1 (at least raid1c3 for RAID6) metadata you can even circumvent that last "bug" Btrfs has.
      I don't say that Btrfs doesn't have any issues, but they have nothing to do with data corruption. It's quite the opposite: Btrfs can detect and repair bit rot, no other mainstream file system that's natively supported by the Linux kernel can do that.