lkml.org 
[lkml]   [1998]   [Aug]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: A true story of a crash.

On Sun, 16 Aug 1998 linker@z.ml.org wrote:

> Actually it doesn't sound bad, until you get to the manual intervention
> part.. I admin MANY systems, and I dont log into all of them weekly much
> less daily.. Sure, I could set it up to page me or something but I doubt
> many processes will like getting zero cpu time for the peroid of hours it
> will take me to fix the problem.. Furthermore if there is an attacker he
> could just relog and attack again..
>
> at 95% (sysctl tunable) your flag goes on and processes goto sleep, any
> requests to swaped ram for nonsleeping processes get stasitfied by
> swapping out chuncks of sleeping tasks into the place you are loading
> from.. Things like X and daemons wont be elegiable for sleeping.
>
> Ok, by putting things to sleep your've stopped your system from swapping
> to death.. Now a reaper decides what to kill, since the system isn't
> already up in flames it's got a little time to decide..
>
> Once memory is freed things go back to normal.

Exactly. Nobody's saying you have to be at the console in the flesh.
Perhaps you have a oneliner daemon that waits for this condition and then
does a telinit. The point is that at this point, you are in a situation
where you can handle things _OUTSIDE THE KERNEL_ with arbitrary
complexity. If you want to kill everything, or reboot, fine. Each admin
can do whatever he thinks is best and the kernel doesn't need to take a
heavy-handed one-size-fits-all aproach.

-Matt



-
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.altern.org/andrebalsa/doc/lkml-faq.html

\
 
 \ /
  Last update: 2005-03-22 13:44    [W:0.090 / U:0.320 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site