Announcement

Collapse
No announcement yet.

Linux 4.14-rc6 Released: Linux 4.14 Kernel Final In 2~3 Weeks

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

  • Linux 4.14-rc6 Released: Linux 4.14 Kernel Final In 2~3 Weeks

    Phoronix: Linux 4.14-rc6 Released: Linux 4.14 Kernel Final In 2~3 Weeks

    Linus Torvalds released the Linux 4.14-rc6 kernel this Monday morning rather than on his usual Sunday cadence due to Internet connection woes. RC6 is also bigger than he would prefer...

    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
    Does anyone know the status of bcachefs? The bcachefs site seems out of date.

    Still hoping btrfs can become stable before final but my home partition still keeps remounting ro with 4.14-rc5 so I don't have a lot of hope of that. It ran out of space a few weeks ago and still gives out of space errors despite its own df tool claiming there's free space on data, metadata, reserve, etc. No amount of scrubbing, balancing and checking seems to be able to fix it.

    Comment


    • #3
      Originally posted by ResponseWriter View Post
      Does anyone know the status of bcachefs? The bcachefs site seems out of date.

      Still hoping btrfs can become stable before final but my home partition still keeps remounting ro with 4.14-rc5 so I don't have a lot of hope of that. It ran out of space a few weeks ago and still gives out of space errors despite its own df tool claiming there's free space on data, metadata, reserve, etc. No amount of scrubbing, balancing and checking seems to be able to fix it.
      *lol* Sounds like a real mess you've created there. Might be a perfect opportunity to learn from the failures, wipe the disks and to start over. I hope it's not an important system.

      Comment


      • #4
        Originally posted by sdack View Post
        *lol* Sounds like a real mess you've created there. Might be a perfect opportunity to learn from the failures, wipe the disks and to start over. I hope it's not an important system.
        Nothing on it I can't afford to lose. *All* my btrfs partitions report errors after what I'd consider normal usage so I'm inclined not to trust it again. In this case it's a laptop so gets even lighter usage than my main desktop system (web browsing, saving files, compiling, playing audio).

        I want the compression features as I have a lot of source code files but not a lot of real space (small SSD, shared with Windows) and no other in-kernel filesystem provides that yet.

        Comment


        • #5
          ... or maybe you just get a bigger SSD for Linux and use the old one for Windows. :'P

          Comment


          • #6
            Originally posted by ResponseWriter View Post
            Still hoping btrfs can become stable before final but my home partition still keeps remounting ro with 4.14-rc5 so I don't have a lot of hope of that. It ran out of space a few weeks ago and still gives out of space errors despite its own df tool claiming there's free space on data, metadata, reserve, etc. No amount of scrubbing, balancing and checking seems to be able to fix it.
            Might be a good idea to report it in the mailing list, for the very least.

            There there are long-time users that can also help out with issues, not just developers.

            I'm suspecting that your issue is about defragging, btrfs allocated a ton of mostly-empty blocks and can't go on until you consolidate that back with a defrag and reclaim actual unused space. Did you use "autodefrag" mount option?

            Did you use the "nossd" mount option (even for SSDs)? see this discussion https://www.mail-archive.com/linux-b.../msg63056.html

            If you can't get it to go rw, add another drive to the filesystem (you will remove this when you're done).

            In the mailing list this script https://github.com/jtek/ceph-utils has been already been recommended to be used to keep defragmentation in check on production systems that don't like the performance hit of a full defrag, but for a laptop it's probably perfectly fine to run a full btrfs defrag at night every now and then.

            Comment


            • #7
              Originally posted by sdack View Post
              *lol* Sounds like a real mess you've created there. Might be a perfect opportunity to learn from the failures, wipe the disks and to start over. I hope it's not an important system.
              Man, the sprinkle of medieval "repent sinners" statement from a moral judge above that you add to this thread is great and very useful, keep it up.

              Deus Vult!

              Comment


              • #8
                Originally posted by starshipeleven View Post
                Might be a good idea to report it in the mailing list, for the very least.

                There there are long-time users that can also help out with issues, not just developers.

                I'm suspecting that your issue is about defragging, btrfs allocated a ton of mostly-empty blocks and can't go on until you consolidate that back with a defrag and reclaim actual unused space. Did you use "autodefrag" mount option?

                Did you use the "nossd" mount option (even for SSDs)? see this discussion https://www.mail-archive.com/linux-b.../msg63056.html

                If you can't get it to go rw, add another drive to the filesystem (you will remove this when you're done).

                In the mailing list this script https://github.com/jtek/ceph-utils has been already been recommended to be used to keep defragmentation in check on production systems that don't like the performance hit of a full defrag, but for a laptop it's probably perfectly fine to run a full btrfs defrag at night every now and then.
                Thanks for taking the time to offer useful suggestions! From that thread, nossd option might be the one to fix it because it happens after some hours of use, which may indicate something is happening in the background. I'm sure a defrag can't hurt either.

                Originally posted by cRaZy-bisCuiT View Post
                ... or maybe you just get a bigger SSD for Linux and use the old one for Windows. :'P
                That's the plan eventually, but it'll get the SSD from my desktop when I can afford to replace that. However, I have more important things I need to spend money on before that. Such is the price of growing up...

                Comment


                • #9
                  Originally posted by starshipeleven View Post
                  Man, the sprinkle of medieval "repent sinners" statement from a moral judge above that you add to this thread is great and very useful, keep it up.

                  Deus Vult!
                  Thanks, will do. Seeing how it got you out of under your bridge and into the light *lol* ...
                  Last edited by sdack; 24 October 2017, 10:30 AM.

                  Comment


                  • #10
                    ResponseWriter the patch for that issue has been sent and landed in mainline (so it should not be necessary to mount as "nossd" in future kernels), and has a pretty damn long commit description that is worth a read. https://git.kernel.org/pub/scm/linux...b4e43710293875

                    Comment

                    Working...
                    X