Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755702Ab2BFVTa (ORCPT ); Mon, 6 Feb 2012 16:19:30 -0500 Received: from longford.logfs.org ([213.229.74.203]:35652 "EHLO longford.logfs.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752640Ab2BFVT3 (ORCPT ); Mon, 6 Feb 2012 16:19:29 -0500 X-Greylist: delayed 2201 seconds by postgrey-1.27 at vger.kernel.org; Mon, 06 Feb 2012 16:19:28 EST Date: Mon, 6 Feb 2012 14:02:06 -0500 From: =?utf-8?B?SsO2cm4=?= Engel To: linux-kernel@vger.kernel.org Cc: Joel Becker Subject: Announce: cancd 0.2.0 netconsole capture server Message-ID: <20120206190206.GA8068@logfs.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit User-Agent: Mutt/1.5.20 (2009-06-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2732 Lines: 59 Quoting Documentation/networking/netconsole.txt: The remote host can run either 'netcat -u -l -p ', 'nc -l -u ' or syslogd. Now, both of these options have their issues. Netcat works fine, but only reasonably scales to 1. Having to use a seperate port for every test machine, as well as a seperate netcat process that writes to a proper file becomes a bit of a nightmare by the time you approach three digits of machines. Syslogd would be great, if netconsole actually spoke syslog protocol. As it is netconsole speaks a protocol I would describe as "7bit ascii, no nonsense". Syslog protocal is described in rfc5424, a 38-page document. I think it is fair to say the two are incompatibel. In case someone actually wants to see the full trainwreck, a quick search should come up with enough bug reports to scare your children into bed. Given that sad state of things, here is yet another alternative: http://git.kernel.org/?p=linux/kernel/git/joern/cancd.git It is far from perfect. Bases on cancd 0.1.0, written by Joel Becker, it is completely incompatible to the older version. I decided to just bang on the code long enough until it happened to do what I wanted. No thought went into preserving previous functionality that I didn't want to use. No thought went into making functionality I did actually want optional. So unless your taste and requirements closely match mine, you will have to invest some work. The good new is that you have the source and can do it, just like I did it. If you send me patches that keep things working for me, I will be happy to merge them. If someone else wants to become maintainer, great! Less work for me. Also beware that I changed the code, but not the documentation. Oh, Joel also disagrees with a number of the changes I made. A better programmer with more time could have kept 0.1.0 and 0.2.0 compatible. I didn't, sorry. Jörn PS: Since there is no documentation, here is a quick overview. Invocations: cancd -p -l Cancd will take any messages thrown at that port, prepend a timestamp to each line and write them out to a logfile matching the "machine". Initially, the logfile will be the IP address of the source host. If that IP has a matching reverse DNS entry, the logfile will be "foo" for "foo.example.org". If you have two seperate machines foo.a.example.org and foo.b.example.org, they will use the same logfile. Best used in combination with logrotate, for obvious reasons. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/