lkml.org 
[lkml]   [1999]   [Jul]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Memory hogs
Bernd Paysan wrote:
>
> Hi!
>
> I've myself seen some lethal memory hoggers in 2.2.x, and a friend also
> reported one to me. The symptom is simple: the system starts swapping like
> mad, and with the current hard disk speeds and not-too-large swap partitions,
> the swap space is full after a few seconds - while it's swapping that hard,
> you don't have a chance to recover, anyway. After the attack, not only the
> memory hogging app is killed, but a lot of other apps (often the X server
> itself) are kicked out, too. The X server is especially nasty, since it leaves
> the video card in an "unknown" state (for the OS), and with VC switch
> blocked.
>
> IMHO the behaviour to react on out of memory condition should be to kill
> the application that allocated most pages recently, and not randomly shoot at
> well-behaving apps. And second, the system should send a signal first,
> before the situation becomes so awful that the kernel has to take immediate
> actions. This would allow the X server to quit gracefully (the X server can be a
> memory hog, too: just try to create a pixmap several thousands pixel width
> and height).

I find syslogd to be one of the early victims of OOM conditions. Almost
always when we run out of memory on something syslogd died (trying to
allocate memory to report the problem, one supposes), so we almost never
find out what ran away.

D

-
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 13:53    [W:1.072 / U:0.032 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site