• MangoPenguin@lemmy.blahaj.zone
    link
    fedilink
    English
    arrow-up
    3
    ·
    7 days ago

    That sounds neat!

    I’ve always read the performance hit of a VM is pretty minimal, like 1% or less on most tasks. Is it really that much faster for you?

    • F04118F@feddit.nl
      link
      fedilink
      arrow-up
      1
      ·
      7 days ago

      I’m not actually sure because I haven’t measured it. But I’ve read that while CPU and memory overhead is small, disk IO is much faster without virtualization.

      • MangoPenguin@lemmy.blahaj.zone
        link
        fedilink
        English
        arrow-up
        1
        ·
        7 days ago

        It could be depending on the underlying filesystem and how it’s set up, I’ve run into that before.

        But my self hosted stuff doesn’t need particularly fast disk IO, so I don’t really notice even if there is a bottleneck lol

        • MotoAsh@lemmy.world
          link
          fedilink
          arrow-up
          2
          ·
          edit-2
          7 days ago

          It definitely depends. If you use qcow2 files for your disks, they’re a copy-on-write format, which is slower than most file systems inherently. If the host system is also a copy-on-write file system, then it doubles your writes outright.

          If you’re using raw disks/partitions, there is minimal overhead that usually comes down to usually miniscule driver differences. Although depending on what the VM platform supports, you might outright lose snapshot/restore support.

          • MangoPenguin@lemmy.blahaj.zone
            link
            fedilink
            English
            arrow-up
            1
            ·
            7 days ago

            Yeah I’m running ZFS with raw disks, and most things are in containers anyways, just a few VMs for Windows or stuff that doesn’t like containers.