Date: Mon, 18 Nov 2002 04:16:34 +0100 From: Thomas Seck <tmseck-lists@netcologne.de> To: freebsd-stable@freebsd.org Subject: Re: -STABLE was stable for long time Message-ID: <20021118031634.GA755@laurel.tmseck.homedns.org> In-Reply-To: <20021117222804.I23359-100000@hub.org> References: <20021118022259.062A1286A5@mail.tmseck.homedns.org> <20021117222804.I23359-100000@hub.org>
next in thread | previous in thread | raw e-mail | index | archive | help
* Marc G. Fournier (scrappy@hub.org):
> On Mon, 18 Nov 2002, Thomas Seck wrote:
>
> > Always happy to serve. But please read my mail Marc had replied to -
> > especially the part not quoted in his reply to get my meaning. If I were
> > in Marc's position I would run -RELEASE and generate patches for new
> > features and bugfixes myself from -STABLE after review and test. Yes,
> > this means definitely a lot more work than the occasional CVSup - make
> > world - cycle. But after all, that's what system administration is all
> > about.
>
> 'K, and in my position, how would you test? These are bugs that are
> triggered by "real live, high load" scenarios ... how long do you test for
> before you feel it is stable? One of my servers ran 20 days straight,
> with no problems, before it crashed ... if I had tested for 19 days and
> decided it was stable, I would have missed that 'last day' ...
Test programs that bevave like your users - spawning lots of processes
and the like? I am not a developer - are there some "load simulating"
suites available? If so, I could give them a try regularly on various
i386 machines at work. If not, could we talk some of the core developers
into developing a test suite?
> Plus, you are advocating a proactive approach of preventing the crashes
> before they happen, which means that someone has to find the crashes in
> the first place ... I'm advocating that some of us are willing to provide
> "real world, high load" servers to generate those crashes so that they can
> be fixed, and are looking for some support from those that know enough to
> fix it ...
I am talking with my system administrator hat on. My production servers
did not run into troubles with -RELEASE so far. My -STABLE machines at
home did not, too. Maybe I am just lucky... But I agree with you that
having some dedicated testing boxes would be a "good thing"(tm).
--Thomas
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20021118031634.GA755>