Subject: Re: Configuring memory parameters for large network-based services
To: David Brownlee <abs@anim.dreamworks.com>
From: Jason Thorpe <thorpej@nas.nasa.gov>
List: tech-kern
Date: 03/22/1999 12:07:38
On Mon, 22 Mar 1999 03:10:39 -0800 (PST) 
 David Brownlee <abs@anim.dreamworks.com> wrote:

 > > (1)  What controls how much memory the pool system can use?
 > > 
 > 	It looks like its still dependent on NMBCLUSTERS
 > 	arch/XXX/XXX/machdep.c defined the mb_map size in terms of
 > 	VM_MBUF_SIZE, which is defined as a multiple of NMBCLUSTERS.

Correct.

 > > (2)  How do I make that space larger?
 > > 
 > 	options         NMBCLUSTERS=4096 or some suitable large number.
 > 
 > > (3)  Will the pages marked as idle ever be reclaimed by the pool
 > >      system, or released back into the system via kernel free()?
 > > 
 > 	Uh, pass.

Idle pool pages are reclaimed by the pagedaemon when the pagedaemon needs
more pages.

 > > (4)  What knobs should I fiddle with to make this machine into a large
 > >      network-based server, cable of handling a few thousand clients?
 > > 
 > 	Iff mb_map is the only barrier you're hitting, then NMBCLUSTERS
 > 	_should_ be enough.

Someone annoying that the message that mb_map is full is no longer
displayed, tho... I'll take a look.

        -- Jason R. Thorpe <thorpej@nas.nasa.gov>