lkml.org 
[lkml]   [1998]   [Sep]   [27]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    From
    SubjectRe: 2.0.3x I found something unique...
    Date
    Well, we need to either hunt it down and kill it or come up with
    some sort of work around. The side effect, (besides a little wasted
    memory and a slot in the process table) is that each of these is counted
    in the run queue. Thus system load = (real system load) + (number of
    processes stuck in run queue.)

    Most large companies are metric freaks. The load factor on this server
    increments 1 2 3 with each stuck process, about one every couple of
    weeks, as I mentioned before. You see my point. Management starts to
    freak out because the server is overloaded..... It takes a reboot
    (outage) to clear it. Also another negative metric.

    Is it a technical problem? Not such a bad one no. Is it a political
    problem? You bet. Maybe that is why NT doesn't even give you an
    'uptime' command. You can't abuse what you don't have. :-)

    Still, I don't want to see kernel bloat, and I'd of course like to find
    the cause and fix it. In the meantime, I'm wondering about a module
    that could be inserted, maybe with a /proc interface to trigger it
    manually to 'clean' this stuff up? Doesn't effect people who don't need
    it, and doesn't even have to be resident for those who do, except when
    this happens.

    I used to do some OS level programming on IBM big iron many years ago,
    and I've recently purchased the 'Linux Device Drivers' book, so maybe I
    will be able to move this along myself.

    Anybody see a problem with my approach?

    > > Linus, maybe it's an idea to include this paranoia code for
    > > 2.2 as well. I know we should be solving the real bugs, but
    > > on a production machine you'll also need fix-up code for when
    > > you run into actual bugs...
    >
    > Please don't.
    >
    > This shows serious bug. Bug is visible enough (unkillable processes),
    > but it is not fatal - they are almost harmless. You do not want to
    > workaround this one (it is possible to live with few 'r'unning
    > processes.)


    --
    +--------------------------------------------+ Rick Bressler
    |Mushrooms and other fungi have several | G-4810 (425)342-1554
    |important roles in nature. They help things| Pager 1-800-800-8596
    |grow, they are a source of food, they | bressler@mushroom.ca.boeing.com
    |decompose organic matter and they |
    |infect, debilitate and kill organisms. | Linux: Because a PC is a
    +--------------------------------------------+ terrible thing to waste.

    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.rutgers.edu
    Please read the FAQ at http://www.tux.org/lkml/

    \
     
     \ /
      Last update: 2005-03-22 13:44    [W:3.713 / U:0.004 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site