• Hasseman (unregistered)

    The frist guy putting fingers in dykes was Hans Brinkner

  • Hasseman (unregistered)

    I have always been proposing before you're allowed to call yourself a programmer you should have worked a year as an operator and then a year or two as maintenance programmer. Then you can start real programming.

  • Alex (unregistered)

    From my sysadmin experience in the mid-90ies: Yeah we did such things shortly after Perl 5 was released, because, well, we knew how to shell script and perl was just an addition to the shell toolbox (together with grep, awk, sed, ...).

    Aww. And we were so smug about it.

  • TK (unregistered)

    The shell command they were looking for is: uptime -s

  • Herr Otto Flick (unregistered) in reply to Hasseman

    The article is saying the complete opposite though - Ops, despite all their heroics and excellent work, have never/not for a long time been software developers. They see code as a tool to get what they want, and the resultant code looks like it too. It's like when scientists need to program, they know what they have, they know what they want, but it works best when they have a developer to explain that to, in order to design the system that produces that answer.

  • Hasseman (unregistered) in reply to Herr Otto Flick

    My thoughts are that as an operator or system administrator you get all crappy software to get working for end users. As maintenance programmer you get all crappy code to maintain so you know how it works down the chain so when you are a programmer you know the problems of maintenance and operations

  • ZB (unregistered)

    WTF is a "medior"?

  • No fun (unregistered) in reply to ZB

    It's when an asdaroyd enters the admasfear

  • SottoVoce (unregistered)

    In my neck of the woods, the Ops folks know the language they use daily (bash) quite well. It's the developers who copy/paste the 1990s-era code they find via search engines because they don't know the language.

  • (nodebb) in reply to ZB

    Geez, ZB, clearly someone less than "SENior" but above the lowest rung of slaves.

  • (nodebb)

    TRWTF is why would anyone care if the recorded uptime has an accuracy of +/- 1 second? What would you do with that information?

  • Pjrz (unregistered) in reply to cellocgw

    In my experience, that extremely-important-company-will-die-without-it information would go into an Excel spreadsheet and then used to produce at least one line chart and one pie chart that will be produced at the next company meeting (but which nobody will actually pay attention to).

  • Mark Ferguson (unregistered) in reply to ZB

    Neither junior nor senior :)

  • Sikhandtake (unregistered) in reply to ZB

    More to the point, why does "Medior" end up pointing to AOL?

    https://en.wikipedia.org/w/index.php?title=Medior&redirect=no

  • (nodebb)

    With respect, it's "dikes" that hold back the rising tide, where we can plug small leaks with fingers. "Dykes" are a different category entirely. 'nuff said.

  • Little Bobby Tables (unregistered) in reply to OllieJones

    With respect, dike is the American mis-spelling of the perfectly good English word dyke which loosely means "earthwork for holding back a body of water".

    The other point to note is that if there were a hole in a dyke leaking water, you could not block it by inserting your digit. A dyke is a raised rampart with a width of anything up to the order of kilometres, on the top of which are often built major roads. So if you had a leak in a dyke it would mean the entire infrastructure of the dyke would be unsalvageable without a major engineering undertaking, and if it was a leak small enough for someone to block by sticking a finger in it, there would be a number of other leaks around the area which I'm afraid you would not be able to stop.

    So: file under fiction, along with men walking on water, lemmings throwing themselves off cliffs in masses, and ostriches burying their heads in the sand.

  • (nodebb) in reply to Herr Otto Flick

    It's like when scientists need to program

    No. Scientists are often worse, as their “programs” will often consist of a bunch of snippets in a text file that they cut-n-paste into an interactive prompt in some order not related to the order in the file in order to get something that they like. I've seen that several times, and it's pretty damn horrifying.

  • RLB (unregistered) in reply to Little Bobby Tables

    Oh, it's worse than that.

    I'm Dutch. Mary Maples Dodge was a merkin, and knew jack squat about water management.

    If you're in a position to stick your finger into a leak in a dyke and not have the pressure blow it straight back out again and break your finger in the process... you're on the side where the higher pressure is, and therefore, drowning.

  • That Other Guy (unregistered)

    "Developers sometimes fail to appreciate how difficult a job Operations really is."

    THAAAAAAANK YOU

  • ZZartin (unregistered)

    How often is this system being rebooted that it needs to have even second precision in restart time?

  • aaaaaa123456789 (github)

    While anyone can see that a one-second inaccuracy doesn't matter, the issue here is probably that the last boot time changed at all — that would normally indicate a reboot, but all that happened was that the calculation was inaccurate. Moreover, some of those changes would push the last boot time backwards in time, which can cause unfunny (okay, probably funny) errors in lots of broken software.

    Of course, the real solution is to round the time to the nearest minute and solve 98.33% of the errors that way :)

  • SuperVirus (unregistered)

    How about "uptime --since | sed -e 's/ /T/'"...

  • SmiddyP (unregistered) in reply to cellocgw

    Possibly it's part of an inventory collection process that submits the information, with changes historically recorded, and the original dev never noticed that every time it differed out by a second, it was creating a new history item with a before and after...

  • DCL (unregistered) in reply to Hasseman

    I second your comment. I started my IT career as a tape and printer monkey before moving to programming. I found my experience as an operator helped write programs and job suites in a way that simplified things for the operators, such as not unloading a tape if the same tape was needed by the following program.

  • Nobody (unregistered)

    This isn't the brightest solution to the problem, but if that's as vague as "how long ago the machine was booted" then this at least isn't wrong, and do you know off the top of your head the precise incantation to get the real number of wallclock seconds since boot? In perl?

    As a timestamp to be stored somewhere precision may be warranted but for "X seconds ago", not. Myself I'd parse it on whitespace as a float and round it properly and either feed it into a date library or make a note about the precision if I remembered, but this could just be a brainfart or the result of a junior with something more important to deal with.

Leave a comment on “Shell Out”

Log In or post as a guest

Replying to comment #:

« Return to Article