Re: Kernel 2.2.14 OOM killer strikes.

From: Mike A. Harris (mharris@meteng.on.ca)
Date: Sat Jul 08 2000 - 02:06:56 EST


On Fri, 7 Jul 2000, Chris Meadors wrote:

>I only have one question about this scheme. What about a program that
>leaks very slowly over a long period of time? (Wasn't there even a kernel
>leak like that in the early 2.2.x series?) It will eventually end up
>exausting all memory and surely there'll be another program asking for
>memory faster than it as you head the OOM situation.

Right.

>I know, I know, this is a rare case and the program doing that should be
>fixed. But it is something to think about. Perhaps there'll be a new
>class of attacks, ones that take weeks or months to pull off.

Actually, I agree with you, but it isn't really all that rare at
all. Netscape is notorious for memleaking from time to
time. Since my OOM the other day, I've come within 4Mb of
OOM'ing again - this time because of Netscape. I closed ALL
windows but one, and the webpage it was left on, was the redhat
manual. No CGI, no java, no cookies, no javascript - all
disabled. Netscape was consuming 100Mb of VM though, with a
single page left open. Granted, I was viewing some large image
files from NASA prior to that, that were in the 2-4Mb range of
size, however I closed those pages. Netscape seems to leak
memory for every graphic that is loaded. Try opening some 4Mb
gif files. If you have trouble finding ones that large, go to
http://www.enterprisemission.com, or to jpl.nasa.gov...

In this case, netscape SHOULD be killed, but how to distinguish
when it should or shouldn't with computer logic? Impossible. I
need more VM!!!!!!!!! ;o)

-- 
Mike A. Harris                                     Linux advocate     
Computer Consultant                                  GNU advocate  
Capslock Consulting                          Open Source advocate

I've overclocked my keyboard interface. It's quite messy dipping my hands into the mineral oil, but *MAN* is my keyboard ever fast now! - Anonymous Coward

- 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/



This archive was generated by hypermail 2b29 : Sat Jul 15 2000 - 21:00:08 EST