Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 9 Jul 2003 17:28:08 -0500 (CDT)
From:      Mike Silbersack <silby@silby.com>
To:        Thomas Chaffic Mowad <tm@owlnet.rice.edu>
Cc:        tm@rice.edu
Subject:   Re: kern.ipc.maxsockets limits with 4GB of memory
Message-ID:  <20030709172625.M11890@odysseus.silby.com>
In-Reply-To: <Pine.GSO.4.33.0307091558090.29385-100000@short-eared.owlnet.rice.edu>
References:  <Pine.GSO.4.33.0307091558090.29385-100000@short-eared.owlnet.rice.edu>

next in thread | previous in thread | raw e-mail | index | archive | help

On Wed, 9 Jul 2003, Thomas Chaffic Mowad wrote:

> Someone thought that the kernel might need more memory so we tried
> increasing VM_KMEM_MAXSIZE too 500 megs and then that caused a panic when
> we booted up.
>
> So we're kinda stuck here and any help would be greatly appreciated.
>
> Thanks,
> --Tom Mowad

Have you tried increasing KVA_PAGES?

#
# Change the size of the kernel virtual address space.  Due to
# constraints in loader(8) on i386, this must be a multiple of 4.
# 256 = 1 GB of kernel address space.  Increasing this also causes
# a reduction of the address space in user processes.  512 splits
# the 4GB cpu address space in half (2GB user, 2GB kernel).
#
options         KVA_PAGES=260

If that doesn't help, you may also consider trying 4.8-RELEASE or -STABLE;
I believe that some additional changes to fix handling of 4GB machines
were committed after 4.7-RELEASE.  (You'll have to search the mailinglist
archives to confirm this, I can't remember exactly when they occured.)

Also, how are you exhausting 128K sockets to do a test of 2000
simultaneous connections?  Are the rest all in the TIME_WAIT state?

Mike "Silby" Silbersack



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030709172625.M11890>