From owner-freebsd-hackers Fri Aug 31 1:17:55 2001 Delivered-To: freebsd-hackers@freebsd.org Received: from tepid.osl.fast.no (tepid.osl.fast.no [213.188.9.130]) by hub.freebsd.org (Postfix) with ESMTP id 7A2F337B40A for ; Fri, 31 Aug 2001 01:17:51 -0700 (PDT) (envelope-from raw@fast.no) Received: from raw.grenland.fast.no (fw-oslo.fast.no [213.188.9.129]) by tepid.osl.fast.no (8.9.3/8.9.1) with ESMTP id IAA96332; Fri, 31 Aug 2001 08:55:50 +0200 (CEST) (envelope-from raw@fast.no) Received: (from raw@localhost) by raw.grenland.fast.no (8.11.3/8.11.3) id f7V6tmB90506; Fri, 31 Aug 2001 08:55:48 +0200 (CEST) (envelope-from raw) From: Raymond Wiker MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <15247.13555.979008.237761@raw.grenland.fast.no> Date: Fri, 31 Aug 2001 08:55:47 +0200 To: Subject: Re: memory + apache In-Reply-To: <20010830113325.E6569-100000@gandalf.bravenet.com> References: <200108301719.f7UHJtR51738@web.cs.ndsu.nodak.edu> <20010830113325.E6569-100000@gandalf.bravenet.com> X-Mailer: VM 6.92 under 21.1 (patch 12) "Channel Islands" XEmacs Lucid Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Dan writes: > > I will give it a try. > touch /var/account/acct > accton > > how long does it take for anything to get written to that file? > > As far as fork storms, I did noticed 1. I had the junior admin write > a script to restart apache if LA got to high ....doing a truss on the pid > i did noticed mad processes and his perl script hitting 10% cpu at times. > Looking at it it was just a basic infinite while loop checking `uptime`. > I have taken that script off but I don;t think that is what is causing > this swap issue. Depending on the exact way that uptime was called, this is quite likely to breed zombie processes (you may need to explicitly call wait or waitpid to reap the processes.) The zombies may or may not be contributing to the swap problems, but they would definitely fill up your process table. Disclaimer: Perl may have changed in this regard during the last few years, but I was bitten by exactly this behaviour some years back. Oh, the embarassment of writing a system sanity check that brings the machine down ;-) > Checking the hardware as well I have confirmed that the manufacter > for these machines only included a heat sink and this tiny fan not > big enough for these boxes....so cooling of the chip may be an > issue as well.....gonna have to check over a bunch of things next > couple days. //Raymond. -- Raymond Wiker Raymond.Wiker@fast.no To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message