From owner-freebsd-rc@freebsd.org Tue Nov 3 23:13:15 2015 Return-Path: Delivered-To: freebsd-rc@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 53A8DA25B03 for ; Tue, 3 Nov 2015 23:13:15 +0000 (UTC) (envelope-from jilles@stack.nl) Received: from mx1.stack.nl (relay04.stack.nl [IPv6:2001:610:1108:5010::107]) (using TLSv1 with cipher DHE-RSA-CAMELLIA256-SHA (256/256 bits)) (Client CN "mailhost.stack.nl", Issuer "CA Cert Signing Authority" (not verified)) by mx1.freebsd.org (Postfix) with ESMTPS id 1A9DD14DD for ; Tue, 3 Nov 2015 23:13:15 +0000 (UTC) (envelope-from jilles@stack.nl) Received: from snail.stack.nl (snail.stack.nl [IPv6:2001:610:1108:5010::131]) by mx1.stack.nl (Postfix) with ESMTP id 0F2F2B8071; Wed, 4 Nov 2015 00:13:11 +0100 (CET) Received: by snail.stack.nl (Postfix, from userid 1677) id EB88F28494; Wed, 4 Nov 2015 00:13:10 +0100 (CET) Date: Wed, 4 Nov 2015 00:13:10 +0100 From: Jilles Tjoelker To: Dirk Engling Cc: freebsd-rc@FreeBSD.org Subject: Re: etc/rc.subr rc_pid for start_cmd Message-ID: <20151103231310.GA70094@stack.nl> References: <563933AB.5090407@erdgeist.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <563933AB.5090407@erdgeist.org> User-Agent: Mutt/1.5.21 (2010-09-15) X-BeenThere: freebsd-rc@freebsd.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: "Discussion related to /etc/rc.d design and implementation." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 03 Nov 2015 23:13:15 -0000 On Tue, Nov 03, 2015 at 11:22:35PM +0100, Dirk Engling wrote: > while fixing startup scripts for qmail I fell for the trap to assume > rc.subr to do the "${name} already running? " check for me. When I then > compared my rc script to several others from various ports/*/*/files/ > directories I could not spot any difference until I noticed they were > also doing it wrong, and would happily spawn more daemons when their > custom start_cmd was invoked. > I then went over a list returned from "grep -ri start_cmd /usr/ports/" > and found that around a third of tools (that install their rc scripts > from the /files/ directory, i.e. they were written by FreeBSD port > maintainers, not some unwashed Linux hippies, who do not know better) do > not really do the check either. While the other 2/3 hilariously > re-implemented the rc.subr check in the most creative ways. > Has there been any historic reason this check was not done somewhere > around /etc/rc.subr:2088 like > if [ $rc_arg = "start" -a -z "$rc_fast" -a -n "$rc_pid" ]; then > if [ -z "$rc_quiet" ]; then > echo 1>&2 "${name} already running? " \ > "(pid=$rc_pid)." > fi > return 1 > fi > maybe with a knob to turn that off for daemons that really know what > they're doing? If not I propose to move the check before custom > ${XXX_cmd} evaluation and will be glad to provide a diff. The original idea was to skip the "already running" checks while booting the system since they are quite slow, and to perform them when a manual service FOO start is done. -- Jilles Tjoelker