From owner-freebsd-questions Mon Apr 2 13:51:53 2001 Delivered-To: freebsd-questions@freebsd.org Received: from neko.cts.com (neko.cts.com [209.68.192.150]) by hub.freebsd.org (Postfix) with ESMTP id 10DB237B71F for ; Mon, 2 Apr 2001 13:51:45 -0700 (PDT) (envelope-from ctsmhn@cts.com) Received: from CARTMAN (cartman.cts.com [205.163.23.192]) by neko.cts.com (8.9.3/8.9.3) with SMTP id NAA03329; Mon, 2 Apr 2001 13:51:42 -0700 (PDT) From: "Matthew H. North" To: "Yavuz Maslak" Cc: Subject: RE: squid settings Date: Mon, 2 Apr 2001 13:55:38 -0700 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0) X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6700 In-Reply-To: <20010329093807.A18904@myhakas.matti.ee> Importance: Normal Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG I'm not sure 'limits' will work in this case. I think (someone with more knowledge please say otherwise if I'm wrong) that the kernel actually has a hard limit of 512MB ram, particularly given your MAXDSIZ setting of 512*1024*1024, which is 512MB. Increasing the limit configuration in login.conf won't overcome this. Instead, recompile your kernel with MAXDSIZ set to something larger (somewhere shy of 2048*1024*1024 unless you want to run into physical limit problems). I ran into this exact problem with Squid: the squid binary would start, begin loading the resident part of the massive cache into memory, reach about 512MB resident, crash and start over (with the error message you describe previously). We first looked at limits configuration in login.conf and found that squid was already running in a class that had full privilege. Further, running a simple C program, *as root*, that allocated chunks of RAM at 2MB a pop, died at the same point: about 512MB resident. So we concluded that there must be a hard limit coded into the kernel. We found MAXDSIZ, upped it to 700*1024*1024 (768MB physical ram in the system), and the problem was fixed. - Matt | -----Original Message----- | From: owner-freebsd-questions@FreeBSD.ORG | [mailto:owner-freebsd-questions@FreeBSD.ORG]On Behalf Of Vallo Kallaste | Sent: Wednesday, March 28, 2001 11:38 PM | To: Yavuz Maslak | Cc: questions@FreeBSD.ORG | Subject: Re: squid settings | | | On Sun, Nov 12, 2000 at 10:34:14PM +0200, Yavuz Maslak | wrote: | | > I use squid2.3 on Freebsd4.2 stable and it works as transparent | > | > I have set max users =256 in my custom kernel | > | > Because I have a problem I look at /var/log/messages and I see | once xcalloc: Unable to allocate 4096 blocks of 1 bytes!" | > After that my squid restart by itself However mysquid hadn't used swap | > What shall I do ? Where can I findout about that ? | | Check the limits(1). Create appropriate section in login.conf and | modify squid startup script as necessary. | -- | | Vallo Kallaste | vallo@matti.ee | | To Unsubscribe: send mail to majordomo@FreeBSD.org | with "unsubscribe freebsd-questions" in the body of the message | To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message