Great ways to begin the day:

Tags: , , , , ,

46 Responses:

  1. luddite says:

    init didn't.

  2. Mark says:

    why nanoseconds???

    • Probably the same reason that df, top, ls etc. by default show you the exact amount of bytes a process or file is taking up, with no formatting, making it impossible to read. "If they wanted a human-readable number", thinks the author, "they'd have added the --human-readable flag / compiled the numbers_for_humans kernel extension".

    • Pavel Lishin says:

      Probably for the same reason that I don't tell people that my birthday is March.

    • Boxcat says:

      Looking at the ship date on that kernel (2011!) I'm surprised that we even have nanosecond resolution. gettimeofday on OS X only has microseconds last I checked.

  3. nicfit says:

    It makes me laugh that all the same shit you ranted on linux for still happens. OS failures, weak tools.. it's all the same.

    • jwz says:

      It's a failing 7-year-old motherboard, genius.

      But if your Stockholm Syndrome requires you to take this as evidence that everything is terrible everywhere and so you might as well keep using Linux, well, knock yourself out.

      • Dan says:

        Amazing to think the Mac Mini first debuted that long ago. It sure doesn't seem that long. I really wanted one, but by the time I could justify getting one, I needed a mobile solution so I got a Macbook Air instead. Still, it's a pretty sweet machine.

        I don't think I've ever used the same motherboard on any of my computers for 7 years. What's the oldest machine you still have in active service?

        • "you" means JWZ but this is a public blog so I'll bore you with my answer instead:

          I have a Proliant 1600 server I've been running more or less continuously* for 14 years.

          (*It used to be turned off at nights until it's equivalent of a CMOS battery died and I had trouble with a replacement)

          • nooj says:

            what was your max uptime?

            • I rebooted it the day I posted that reply, I was replacing the batteries in the UPS the server is attached to and although the batteries are hot-swappable I did a shutdown out of caution. The server's shutdowns are either OS upgrades (no longer relevant as the obscure Unix OS vendor is effectively undead) or something to do with the UPS. I had to replace the Compaq branded UPS I bought with the server because I was having problems getting batteries that fitted it from Compaq DEC HP or anyone else. I think the uptime was around 600 days before the most recent shutdown.

        • fop says:

          Mac Mini debuted earlier than 7 years ago. 7 years ago would be close to the Mini refresh where they changed to an Intel CPU though.

          • Dan says:

            The Mac Mini G4 debuted Jan 2005, and the Mac Mini that JWZ has (based on the Sytem Model Name in the screenshot) debuted Feb 2006.

            • jwz says:

              Incidentally if anyone has a full screen shot or log of that kind of panic, I'd love to add it to BSOD. I didn't think of it at the time so I only have the partial picture.

        • Adolf Osborne says:

          The old mail server at work still churns ~24x7 and has run for more than 9 years (it no longer tosses mail, but does make a fine Samba box for the technical crowd) without a new hardware injection. It's a simple scratch-built P4 box with a Tyan motherboard and a good power supply.

          The hard drives should have given up long ago, but they keep trucking along (80GB WD SATA) in RAID 1.

          When the bearings in the CPU fan started to sound a bit dry after a half-decade or so, I carefully laid the box down its side, took the cover off, peeled the sticker from the fan and added a drop of oil without ever seriously considering shutting it down. It didn't mind, and the new oil still works well.

          Peak uptime is a bit over 2.5 years. It never has crashed*, though it has been rudely interrupted a few times.

          *: I did somewhat back over it with a truck once during a move related to a prolonged bandwidth outage, but that buffed out just fine. I don't know if that's really considered a "crash" or not since it wasn't running at the time...

          • Dan says:

            Pfff. I'm just RMAing my 4th Seagate 2TB drive in a year. They keep failing after a few months, and I don't know why. It got so bad that I now have 4 drives in a RAID 10 configuration and a 5th as a global spare, which has saved my data twice now. I suppose newer drives are higher precision and more sensitive, but I miss being able to trust my hard drives.

            • Yariv says:

              I once had a suspiciously high rate of hard disk failure after I moved to a new apartment. It turned out that unstable electricity was the culprit. Do yourself a favor and get a decent UPS. Not the cheap kind, though. You want one that channels the power through the batteries or does some other magic to keep the supply current clean.

              • Dan says:

                Not a problem for me. I have a high end power supply in the computer, and it's hooked to a good high capacity BackUPS XS 1500. My problems started after the Thailand floods, so I think that quality control while recovering from that is at the root of the problem.

                • Adolf Osborne says:

                  A BackUPS XS counts as a cheap one, as its "line interactive" nature does nothing meaningful toward isolating your system from the every-day badness that floats down the AC wire.

                  It pretty much just sits there doing nothing more than a common power strip unless there is a substantial disturbance -- things need to get pretty ugly before it switches transformer taps or kicks to battery.

                  For an example of a not-cheap UPS, look for terms like "online double-conversion." These units convert the AC to DC, and then create a whole new AC waveform to power attached devices -- all the time.

                  It's about as isolated as one can get without investing in big gnarly chunks of copper and iron like ferroresonant transformers and motor-generators.

  4. 47f274a3faaf says:

    Is osx really still worth checking out? I'm a kid who uses Windows and Linux at work and really has no problems with Linux at work or home. As recently as four years ago linux driver support and flash and such on common hardware was a pain, but everything pretty much just works now is my experience. Is it really better on a mac these days?

    I don't even have any complaints against windows 7; that all works ok and there's good free packages to do everything. It's just less convenient to install stuff than with the good linux distros anymore.

    • Owen W. says:

      OSX really does Just Work, like everybody says. No, really. If you're coming from linux there are a couple of surprises:
      * You'll have to pay for little utilities you think should be free, like text editors or interface tweakers. Don't complain, just pay the 10$.
      * "package management" is better than windows, but not as good as linux mostly because 3rd-party software actually exists.
      * With every new version of the OS, something may get changed for the worse, like an uglier UI, low-level file format change. See also, JWZ's complaints about itunes for a representative experience. There's nothing you can do, no bug you can file.

      I have a friend who says that when he demos OSX to windows and linux users, he shows them his laptop and shuts the lid (it suspends). Then he opens and shuts it, quickly, a dozen times. That's usually enough.

      • Cob says:

        I have a friend who says that when he demos OSX to windows and linux users, he shows them his laptop and shuts the lid (it suspends). Then he opens and shuts it, quickly, a dozen times. That's usually enough.

        I do the same thing when demoing my ThinkPad to OS X and Windows users. The Windows response is "wow that's fast!". The Mac response is "oh, just like OS X.". The latter sounds a little more resentful when they've already seen things like the removable battery or the price.

  5. Jeremy Wilson says:

    I continue to be amazed at your ability to break computers.

    I'm also impressed that a 7 year old Mac lasted that long - that's 2006! Youtube was a year old!

    • Tim says:

      I have a Mac Mini of the same era that's still going strong. I use it to play videos on my TV. I'm thinking of replacing it for the sole reason that it's a little slow to play some HD video (and it would be nice to have 802.11n).

    • Owen W. says:

      these two statements contradict each other. You're amazed that he broke an old computer, but you're amazed his old computer wasn't already broken.

      • Jeremy Wilson says:

        My statement is more that, if anyone is going to break a computer, it's jwz. Age generally has nothing to do with it.

        However in this case it happens to be an old one, so it might not be his electronics destroying distortion field to blame for once.

  6. Karl Shea says:

    It is worth checking out. My only complaint is that there isn't any binary package management available. All the cool kids use Homebrew, which is great, except that compiling everything is kind of annoying.

    Otherwise, it's just like Linux except that the fonts are better and shit just tends to work. Also you don't have to use Gimp anymore.

  7. Karl Shea says:

    God dammit. jwz, your OpenID thing is busted.

    This was supposed to be in reply to "47f274a3faaf "

    • Dan Lyke says:

      Experience suggests that "your OpenID thing is busted" is a tautology. Even when you can get a given provider and consumer to agree, there's almost always some other whackiness.

      I miss LID. There were only two implementations, and it worked on both of them, and actually did the thing OpenID was supposed to.

    • Nick Lamb says:

      It's not OpenID-related, the reply function is just half-busted anyway.

      • jwz says:

        Busted how? I'm not doing anything fancy, this is the default WordPress thing.

        • Nick Lamb says:

          Sometimes (but not now of course, not when I'm trying to reproduce this on demand), when you click Reply, the in-place preview DIV (or whatever it is) appears at the bottom instead of where your reply would naturally go in threaded order. Having once made the mistake of going "Oh, I'm sure they'll fix that in post" only to see my now non-sequiteur comment appear three or four pages away from where it was relevant, I learned my lesson and I now retry until the preview is where I actually want my reply to appear before writing an actual comment. I don't think I have ever taken more than a passing interest in anyone else's WordPress blog, including our own, so I wouldn't know if it happens to other instances.
          I am a scarred veteran of the early DOM wars so I won't be surprised if this never happens in Safari (or in San Francisco, or indeed on machines with odd-numbered IP addresses) but it does happen here, and it happens before you get as far as picking how to authenticate so it can't plausibly be caused by the OpenID relying party code.

          • jwz says:

            So, here's one thing that happens that sucks but wouldn't cause this: If you click on the "You can reply here" thing in the reply-notification email, it takes you to "http://...?replytocom=126584#respond" but the reply field stays at the bottom of the page -- it does not relocate it to under the comment you're replying to. Which sucks. But, it still replies to the proper comment.

            If you click on the "Reply" link on the page, the Javascript runs and relocates the reply field. I've never seen it relocate the reply field and yet post the reply at top level instead of to that comment.

            • nooj says:

              It's easy to forget to click the "Reply" button in the first place, if the reply field (at the bottom of the page) is already visible. In Karl Shea's case, 47f274a3faaf's post was only one from the bottom at the time, with no children.

              Additionally, depending on one's screen/font/window size, the Name, Comment, and Preview fields take up most or all of the visible area, making it easier to accidentally forget to confirm which post is actually being replied to.

            • Nick Lamb says:

              "I've never seen it relocate the reply field and yet post the reply at top level instead of to that comment."

              No, I've never seen it do that either. But I swear I have seen it just hop down to the Reply widget where it starts off, at the bottom of the page rather than moving it. For someone who hasn't seen it "work" normally, or who doesn't expect it to matter, this isn't a sufficient clue that anything is wrong. Now that I have a snowy weekend I've spent some time trying to reproduce the problem on a machine with Firebug, and of course I still can't.

  8. phuzz says:

    The question is; is this a machine that you will mourn and raise a glass to, or one that you will take outside and 'do an Office Space' on?
    (Assuming it's really mounted it's last disk and gone to sing in the data center invisible)

    • Patrick says:

      I'm going to have to remember that line the next time some suicidal app team asks me to upgrade their bog-ancient Sun v440 with RAM from Crazy Jim's SPARC Emporium & Bait Shop. Remember your data? How _well_ do you remember your data?

  9. Are you sure this wasn't just xscreensaver/BSOD? I once sat looking at my screen for a couple minutes trying to understand the crash dump before realizing it was for a VAX.

  10. Nik says:

    I hope we can all agree that when this happens on Windows, it's still Bill Gates' personal fault, even if the machine was a decade old and overclocked like crazy.

  11. John Bloom says:

    Are you just using this topic to update your kill file? Cause I'm totally onto you. <_<