From owner-freebsd-questions  Wed Mar  7 21: 5:42 2001
Delivered-To: freebsd-questions@freebsd.org
Received: from alleyoop.m8.ntu.edu.tw (Alleyoop.m8.ntu.edu.tw [140.112.251.238])
	by hub.freebsd.org (Postfix) with ESMTP id 7464237B718
	for <freebsd-questions@FreeBSD.ORG>; Wed,  7 Mar 2001 21:05:36 -0800 (PST)
	(envelope-from ml@alleyoop.m8.ntu.edu.tw)
Received: (from ml@localhost)
	by alleyoop.m8.ntu.edu.tw (8.11.2/8.11.0) id f2854us48596
	for freebsd-questions@FreeBSD.ORG; Thu, 8 Mar 2001 13:04:56 +0800 (CST)
	(envelope-from ml)
Date: Thu, 8 Mar 2001 13:04:55 +0800
From: AlleyOop <ml@alleyoop.m8.ntu.edu.tw>
To: freebsd-questions@FreeBSD.ORG
Subject: Re: every day kernel panic in "setlock"
Message-ID: <20010308130455.A47551@alleyoop.m8.ntu.edu.tw>
References: <20010307121701.A5577@alleyoop.m8.ntu.edu.tw>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5i
In-Reply-To: <20010307121701.A5577@alleyoop.m8.ntu.edu.tw>; from ml@alleyoop.m8.ntu.edu.tw on Wed, Mar 07, 2001 at 12:17:01PM +0800
Sender: owner-freebsd-questions@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

On Wed, Mar 07, 2001 at 12:17:01PM +0800, AlleyOop wrote:
> Hi all:
>   Excuse me if here is the wrong place to report this kind question.
> 
>     Our machine is dual P3-750 with 2.5 GB RAM, 
>   running bbsd for about 3000-4000 on-line in the same time.
>   Meanwhile , there will be over 4000 process .
> options         NMBCLUSTERS=20480
> options         PMAP_SHPGPERPROC=720
> options         SHMMAXPGS=8192
> options         SHMALL=8192
> options         VM_KMEM_SIZE_MAX="(180*1024*1024)"
After read more about the LINT and some  note in 
/sys/vm/vm_pageout.c
It seems the panic occurs because the heavy use with SHM
thus we tune the MIB vm.v_free* mentioned in LINT. 
and change the value of (vm.v_free_target,vm.v_free_min,vm.v_free_reserved)
to double.
and I also change vm.defer_swapspace_pageouts -> 1 

It seems all ok now , and yesterday night , even though there are 4200
people on-line , our FreeBSD box runs well and stable.

There is still a mmm severe or trivial problem .
If we use sysctl -a to show MIB , FreeBSD crashed immediately.
and the panic message show the "current process" is sysctl.
What is the problem of that ?


Regards,
	Alleyoop.

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message