Re: init.d

[prev] [thread] [next] [lurker] [Date index for 2006/06/23]

From: A. Pagaltzis
Subject: Re: init.d
Date: 16:43 on 23 Jun 2006
* Paul Orrock <paulo@xxxxxxxxxxxxxxxx.xxx> [2006-06-23 16:35]:
> You're using the init (boot start up) system to try and start
> it as though it were being booted via init.d and it's rightly
> telling you "You've told me not to start on boot so I won't"

* imacat <imacat@xxxx.xxxxxx.xxx.xx> [2006-06-23 17:20]:
>     1. Sys-V init scripts are for the boot process to start and
>        stop.
>     2. Whoever run init scripts is acting as the boot process.
>     3. You told the boot process to skip it.
>   => It is skipped.

Err, have you two actually looked at the script in question?

Whether or not to start a service at boot time is controlled by
the regular hateful SysV symlink forest in `/etc/rc*.d`, but that
is not what's going on here.

The startup script in the Apache2 package of Debian includes an
*additional* layer in the form of that `/etc/defaults/apache2`
config option which determines whether the script will refuse to
bring up Apache *regardless* of whether you have configured it to
be brought up at boot time.

I couldn't think of any reason why one would want to do something
like that, so I dug into the package metadata. It appears that
this is a lever used by the post-install script uses to castrate
the boot script if port 80 is found to be occupied at package
installation time.

The mind boggles.

This isn't hateful, it's insane.

Regards,
-- 
Aristotle Pagaltzis // <http://plasmasturm.org/>;

Generated at 04:01 on 28 Jun 2006 by mariachi 0.52