lkml.org 
[lkml]   [1998]   [Oct]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectOOM and policy

PS: I wasn't criticizing the OOM algorithm or the idea of having a
system-wide ENOMEM handler that takes over for processes that don't handle
it themselves. It's just that if a process running on a server has to be
killed because all of ram and swap are used up, I don't want the machine
to choose which one to terminate, *I* want to choose which processes get
dumped out of memory first. The dynamic swap file created on demand idea
was just a suggestion for deferring the actual termination of programs
until an operator has a chance to review the OOM algorithms decisions.
How practical this is in a networked environment or in general (process
migration heuristic) is of course a different question, a non-policy
question.

Since this is policy, how can I tell the kernel at bootup which processes
I want it to kill first, regardless of resource usage heuristics?

("Don't send sigterm to the dbms manager unless you've already killed off
all of these other trivial pursuits that might be running. Don't ever
wonder about it.")

Regards, Clayton Weaver cgweav@eskimo.com (Seattle)






-
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:0.023 / U:0.056 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site