RE: Kernel doesn't free Cached Memory

["Al Boldi" <a1426z@gawab.com>]

Bill Davidsen wrote: {
Al Boldi wrote:
> Dick Johnson wrote: {
> 
>>On Fri, 2005-07-22 at 08:27 -0300, Vinicius wrote:
>>[...]
>>
>>>   I have a server with 2 Pentium 4 HT processors and 32 GB of RAM, 
>>>this server runs lots of applications that consume lots of memory to.
>>>When I stop this applications, the kernel doesn't free memory (the 
>>>memory still in use) and the server cache lots of memory (~27GB).
>>>When I start this applications, the kernel sends  "Out of Memory" 
>>>messages and kill some random applications.
> 
> 
> ...you might even need to turn memory over-commit off:
>   echo "0" > /proc/sys/vm/overcommit_memory
> }
> 
> That's in 2.4. In 2.6 it's:
>   echo "2" > /proc/sys/vm/overcommit_memory

RHEL3 *is* a 2.4 kernel.
> 
> But the kernel doesn't honor no-overcommit in either version, i.e. it
still
> overcommits/pages-out loaded/running procs, thus invoking OOM!
> 
> Is there a way to make the kernel strictly honor the no-overcommit
request?
> 

Don't have swap?
}

Turn off swap and things get worse!

Paolo Ornati wrote:{
Bill Davidsen <davidsen@tmr.com> wrote:
> And IMHO Linux is *way* too willing to evicy clean pages of my 
> programs to use as disk buffer, so that when system memory is full I 
> pay  the overhead of TWO disk i/o's, one to finally write the data to 
> the  disk and one to read my program back in. If free software is 
> about  choice, I wish there was more in the area of how memory is 
> used.

isn't this tuned enough by "/proc/sys/vm/swappiness" ?
}

Swappiness tunes but does not inhibit overcommit!

So the question remains:
	Why Is there no way to make the kernel _strictly_ honor the
no-overcommit request?

--
Al

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