Subject: Re: rc.d and rcorder
To: Malcolm Herbert <firstname.lastname@example.org>
From: Andrew Reilly <email@example.com>
Date: 04/24/2007 12:12:08
On Mon, Apr 23, 2007 at 11:14:35AM +1000, Malcolm Herbert wrote:
> Having recently been exposed to the Sun Solaris smf subsystem, are there
> any thoughts on extending NetBSD's rc system to provide monitoring of
> the started services? ie, who/what is responsible for monitoring the
> health of a particular service provided by the rc.d script?
Most of the daemons look after themselves in various ways, for
historical reasons. Consequently, I don't think that there's a
*huge* amount of pressure to scratch that itch: it's just not
Personally, I really like the daemon management that I get from
DJB's daemontools suite, which I've used to wrap a few other
services besides SMTP. The configuration and set-up, while
logical and minimal in a mathematical sense, is decidedly on the
funky and fiddly side, compared to the traditional Unix flat
text file strategy. DJB seems to have a completely different
approach to dependencies, which is interesting: go ahead and
start everything in parallel. If anything doesn't see an
environment conducive to constructive running, it should exit,
and the management framework will try again a few seconds
later. This has the nice characteristic that it behaves well
when services/devices/routes appear or go away after startup.
I'm not sure how well this strategy would extend to some of the
fundamental system initialization things that get taken care of
in the rc.d script scheme, though.
I also like OSX's launchd, in use, but I haven't looked too
closely at it's innards or how it is set up.