X-Junk-Score: 0 [] X-Cloudmark-Score: 0 [] Return-Path: Received: from [217.112.85.135] (HELO 2kcafe.co.uk) by mail.stalker.com (CommuniGate Pro SMTP 5.2.7) with ESMTP id 47525569 for CGatePro@mail.stalker.com; Fri, 05 Sep 2008 09:09:58 -0700 Received-SPF: none receiver=mail.stalker.com; client-ip=217.112.85.135; envelope-from=murray@liquidthinking.co.uk Received: from [78.86.188.127] (account murray@liquidthinking.co.uk HELO wks003.ggp.local) by 2kcafe.co.uk (CommuniGate Pro SMTP 5.0.13) with ESMTPSA id 10991735 for CGatePro@mail.stalker.com; Fri, 05 Sep 2008 17:09:21 +0100 Date: Fri, 05 Sep 2008 17:09:27 +0100 To: "CommuniGate Pro Discussions" Subject: Re: Server Crashing From: "Murray Crane" Organization: 2K Cafe/LiquidThinking Content-Type: text/plain; format=flowed; delsp=yes; charset=utf-8 MIME-Version: 1.0 References: Content-Transfer-Encoding: 7bit Message-ID: In-Reply-To: User-Agent: Opera Mail/9.52 (Linux) On Wed, 27 Aug 2008 10:03:26 +0100, David Brookfield wrote: > We've got CGP 5.0.13 server running on FreeBSD it's run like a dream for > about 3 years, but in the last couple of days it just stops. Tiny bit more info for the list, and a request for help from any FreeBSD gurus. We've tweaked the max shared memory sysctl value and CGP is staying up a lot better now. When it was being particularly bad a few days ago, I just happened to be watching in 'top' and noticed that free RAM would approach 0 and CGServer (using 500MB+ RAM by that time) would die. I'm guessing that CGS was asking for more than the server was prepared to hand out and so it was being killed. So I've been (secondarily) trying to find a way of keeping the CGServer process up and running via a cronjob, but Google hasn't been my friend so far and my bash/sh scripting is stale to say the least. I can check if CGS is running with 'pgrep' (exit status 1 if it's not found) and the existing rc script will let me start it again, but can someone give me a hint how to join those two concepts together? Murray Crane LiquidThinking