Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?
Open Source Operating Systems Upgrades Linux

Linux 3.2 Has Been Released 271

diegocg writes "Linux 3.2 has been released. New features include support for Ext4 block size bigger than 4KB and up to 1MB, btrfs has added faster scrubbing, automatic backup of critical metadata and tools for manual inspection; the process scheduler has added support to set upper limits of CPU time; the desktop responsiveness in presence of heavy writes has been improved, TCP has been updated to include an algorithm which speeds up the recovery of connection after lost packets; the profiling tool 'perf top' has added support for live inspection of tasks and libraries. The Device Mapper has added support for 'thin provisioning' of storage, and a support for a new architecture has been added: Hexagon DSP processor from Qualcomm. New drivers and small improvements and fixes are also available in this release. Here's the full list of changes."
This discussion has been archived. No new comments can be posted.

Linux 3.2 Has Been Released

Comments Filter:
  • Btrfs (Score:5, Interesting)

    by davegaramond ( 632107 ) on Wednesday January 04, 2012 @10:09PM (#38591738)
    So does this mean I can start using btrfs, at least for personal workstations? I've got a new box at the office waiting to be setup, with a 120GB Corsair SSD as the main system disk, normal 2TB harddisk as backup/media storage. Will be using Debian. Should I use btrfs?
    • Re:Btrfs (Score:5, Interesting)

      by Anonymous Coward on Wednesday January 04, 2012 @10:15PM (#38591786)

      Short answer: no.

      Long answer: Please! The more people exercising the code, the more bugs will be revealed, and the more confident the developers can be. But you will have to be ready for some performance regressions and data loss danger. For the brave.

      • Re:Btrfs (Score:5, Interesting)

        by Anonymous Coward on Wednesday January 04, 2012 @10:32PM (#38591924)

        Performance is still pretty bad, especially when deleting many small files. It can take minutes with BTRFS, while EXT4 does it almost instantly in comparison.

        • by RemyBR ( 1158435 )


          Don't know about this new version, but I tried btrfs on ubuntu a few months ago, on a box I use mostly with photo and video editing software. It was slow to the point of being completely unusable. Specially for non destructive photo editing, where the software creates and modifies (replaces) small files with metadata for each action you perform on the images. Have being using ext4 since then, and everything is ok.

      • Re:Btrfs (Score:5, Funny)

        by Sloppy ( 14984 ) on Thursday January 05, 2012 @01:50AM (#38592982) Homepage Journal

        Short answer: no.

        Long answer: Please!

        In other words, "better you than me."

    • Re:Btrfs (Score:5, Informative)

      by francium de neobie ( 590783 ) on Wednesday January 04, 2012 @10:17PM (#38591794)
      There's no fsck [kernel.org].. So unless you're 100% sure your Linux machine never crashes and your power supply is never interrupted - don't.
      • Re: (Score:3, Interesting)

        by Anonymous Coward

        Even without any crashes or power fails, I've managed to corrupt BTRFS in testing, with just a defrag (as recently as 3.2-RC6). I wouldn't look at BTRFS for a while, at least until it's no longer marked experimental. By all means test away, but don't assume you'll be able to get to any data you put in it.

      • Well, technically there is [ubuntu.com], but since it doesn't repair errors, it's not much good.
    • by nzac ( 1822298 )

      I have herd a number of recommend that for a personal desktop you could use it for non user files (ext for home and boot and btrfs for ever thing else).

      The idea being that if something goes wrong you can just reinstall (30 min for me if I cache my updates).

      You just have weigh up the risk and minimize the damage. Unless you are using it with something like snapper (fs snapshot and roll back) or a stopwatch you will never know the difference unless it gets corrupted.

    • Re:Btrfs (Score:4, Interesting)

      by mattbee ( 17533 ) <matthew@bytemark.co.uk> on Wednesday January 04, 2012 @11:11PM (#38592176) Homepage

      btrfs is tanatlizing for VMs because of the copy-on-write file behaviour (i.e. "cp --reflink a b" creates b instantly regardless of the size of a), but http://lists.fedoraproject.org/pipermail/devel/2011-July/154251.html [fedoraproject.org] is still an issue, as far as I'm aware. So storing VMs, where you access them with O_SYNC, just gets slower over time until it's unusable. I'm not quite brave enough to suggest that any of our customers use it, at least until there's a working fsck.

  • by inflex ( 123318 ) on Wednesday January 04, 2012 @10:11PM (#38591754) Homepage Journal

    Waiting to see the usual fanatical wars over filesystems... people calling for the death of the EXT3/4 system.

    Personally the whole fanatical thing seems a bit silly - who'd have ever thought that people would lynch each other over having different options for different purposes/tasks, the very core of the whole idea of what we do and strive for. I'm fine with ext4, thanks :)

    • Re: (Score:3, Interesting)

      I still say we should lynch EXT3/4(even though I use it) due to it's complete /inability/ to undelete files.
      Because, as we all know, people /never/ manage to accidentally delete files and /always/ have recent backups handy.

      • by inflex ( 123318 )

        People generally only make that mistake once or twice before they become a bit more clued up and invest in a backup option, even on OS's that provide undelete (Windows). Agreeably it doesn't save you when you create and then lose a file between the backup times.

        It might be a nice option to have, so long as it doesn't inhibit/hinder the existing system. I think an entirely different filesystem would be a better option, something with inbuilt versioning/history.

      • by DarkOx ( 621550 ) on Wednesday January 04, 2012 @10:33PM (#38591928) Journal

        Then what you should do is change you shell so rm is a functIon that moves stuff to the "trash" rather than compromise the on disk format of the file system so an operation "unlink" which is supposed to be destructive can be undone. Solve the problem in the correct place.

        • Agreed. I alias rm to 'rm -i' so that I have to at least type -f if I really mean it :)

          • rm -i (Score:5, Insightful)

            by steveha ( 103154 ) on Thursday January 05, 2012 @01:54AM (#38592998) Homepage

            I used to alias "rm" as "rm -i".

            Then, one day, I was using someone else's computer. I used "rm" with the expectation that it would prompt me, but this person never bothered to set it up that way, and I had the fearful experience of worrying whether it was deleting too much. I hadn't been too careless that time, but it got me thinking. It's dangerous to use "rm" when I really mean "rm -i"; habits are strong things.

            So I made a change that I still use. I now alias "r" as "rm -i". "r" by itself does not have default behavior on most computers. Now if I absent-mindedly type "r *.txt" on someone else's computer, I get "r: command not found" and I edit the command to say "rm -i".

            I suppose I should have used "rmi" or something like that, just in case I am a guest somewhere that "r" was aliased to something crazy. In practice, it hasn't been a problem. I use more aliases than most people seem to; they seem to be content with the defaults. I seem to be the only one I know who likes one-letter aliases.

            Hmm, I guess I might accidentally run the R statistics package someday?


            • by huge ( 52607 )

              I now alias "r" as "rm -i". "r" by itself does not have default behavior on most computers.

              I have a friend that used to alias "r" for "rm" and "e" for "emacs". Once he had to restore his thesis from day or two old backup he stopped doing that :)

            • Re:rm -i (Score:5, Funny)

              by dotancohen ( 1015143 ) on Thursday January 05, 2012 @05:24AM (#38593934) Homepage

              It's dangerous to use "rm" when I really mean "rm -i"; habits are strong things.

              Especially muscle-memory habits.:wq

          • by Urkki ( 668283 )

            Agreed. I alias rm to 'rm -i' so that I have to at least type -f if I really mean it :)

            ...and then, you get into habit of using -f when you're sure you want to delete, and soon you get the habit of using -f with rm. Which, needless to say, is pretty dangerous habit to have.

            • LOL, yes I'm aware of that danger. Unless I'm wiping out an entire folder recursively or something, I don't bother with the -f flag. FreeBSD has a -I now (capital i) which only prompts you if more than 3 files are going to be deleted. This isn't available on Mac, but I think Linux has it as well.

      • If thats a problem then just replace the rm command with something less final.

      • Re: (Score:2, Informative)

        by Anonymous Coward

        One workaround we did on our server over a decade ago was to hardlink basically everything to /shadow

        Entries in /shadow went through periodic timed deletes, but that avoided irreparable stupidity.

      • That's not a bug, that's a feature!
    • by martin-boundary ( 547041 ) on Wednesday January 04, 2012 @10:24PM (#38591860)

      Personally the whole fanatical thing seems a bit silly

      That's not silly! There are two reasons for silliness. Surprise and fear. Fear and surprise... and ruthless efficiency. There are *three* reasons for silliness, these being fear, surprise, and ruthless efficiency... and an almost fanatical devotion to the Pope. *Amongst* the reasons for silliness are such elements as fear, surprise, ruthless efficiency and ... Ok, you're right, fanatical is silly after all.

    • I tried btrfs when I recently installed ubuntu on a new netbook. It was taking 15 minutes to fsck the disk on startup. This file system seems a long way from being used by default.

      • by msk ( 6205 )

        Where'd you get fsck for btrfs?

        • Thats a good question. I thought it was running fsck because it took forever to mount on startup. But maybe the mount was slow for other reasons.

    • I bet you use emacs

    • Ext4's only major flaw is that it does not have snapshots. That can be fixed, actually.

    • Why? The last time I had a look at BTRFs it was at least in a VM significantly slower than EXT4.
      This was noticable at every corner of the OS.
      I would not use it for production, at least not in a scenario where I/O performance matters, but the features are neat. I love
      the snapshotting. But my personal guess is it is at least a year away from being a viable EXT4 replacement.

    • by Hatta ( 162192 )

      Do you remember when Ext4 came out? A bunch of people had data loss after system crashes. It was explained that this was intentional behavior, as specified by the POSIX standard, and the safer behavior of Ext3 was unintentional. Of course, no one cares when they're looking at a tree full of zero byte files, so the file system fanatics rose hell and got them to patch Ext4. So file system fanatics have their uses.

  • Good stuff! (Score:5, Funny)

    by ickleberry ( 864871 ) <web@pineapple.vg> on Wednesday January 04, 2012 @10:16PM (#38591792) Homepage
    I never did like the number "3.1" for some reason
  • by wbr1 ( 2538558 ) on Wednesday January 04, 2012 @10:26PM (#38591880)
    The first kernel I compiled was 1.2.10, I know there are people who have here who have been it longer than I, so this is not an ego-trip. I just feel old. I need doctor Carol Marcus to make me .... "Feel young, as when the earth was new."
  • Wake me when we get to 7.1. At this rate it ought to be sometime this fall.

    • by wbr1 ( 2538558 )
      And FireFox will be upto 321.6.98
  • This looks to be a really strong, likely to be long-supported, kernel. Providing that the Googleification of the TCP stack doesn't hurt local 1-10Gbps performance, that is. Have a care if you do your own kernel compiles... the whole Ethernet driver subsystem has been merged together.

  • Engineers kept using ext3 (or reiserfs...) for a while because ext4 was "too new".
    Now that it's stable and used, is it safe to extend it with such a powerful "block" option, and risk a potential regression?
    • IIRC weren't they not using ext4 because it was missing some crucial features like fsck support?

  • Very good news: standard Linux box will read my NAS' drives without using specific package and user space FS tools.
    The NAS has a Debian Sparc system with ext3 16k blocks. Recovering the data when the NAS is dead (or has problems) is always a concern. Knowing I will be able to start a PC with Linux Live CD and plug those disks to recover my data is a relief.

    I hope this NAS will also accept USB drives with ext3 16k FS made by x86 Linux (it doesn' read ext3 4k FS). I've prepared some with thorough blocks rw

You will never amount to much. -- Munich Schoolmaster, to Albert Einstein, age 10