Subject: Re: /etc/rc.d source of boottime breakage?
To: None <current-users@netbsd.org>
From: Laine Stump <lainestump@rcn.com>
List: current-users
Date: 03/15/2000 17:21:43
At 06:35 AM 3/15/00 -0800, Jason R Thorpe wrote:
>On Wed, Mar 15, 2000 at 01:48:18AM -0500, Laine Stump wrote:
> > 2) Any ideas why the @reboot cronjobs of a non-root user are no longer
> >    running? Do other people have this problem, or is it something weird
> >    with my setup? (Note that these point to a script that is on another
> >    disk, but that disk is in the list of critical_filesystems in
> >    rc.conf).
>
>Yes.  cron is shut down by rc.shutdown now.  It probably gets a different
>signal now than it used to get from init.

Actually this one turned out to be my own problem - when reconstructing the
disk I had neglected to put in a symlink from /homes to the real home
directories. I didn't notice this problem in normal use because /etc/passwd
lists all home directories as "/usr/homes/..." Chalk this one up to a bad
sysadmin. (BTW, @reboot commands are executed while the system is coming
*up*, not while it is going down, which seems a bit counterintuitive, since
the reboot command shuts the machine down. But that's another discussion...)

However, I see that rc.wscons is no longer called properly -
/etc/rc.d/wscons runs it with:

	. /etc/rc.wscons

which means that rc.wscons thinks that its argv is:

	/etc/rc.d/wscons start

and this of course throws it for a loop (since it's normally expecting to
have no arguments). I fixed my copy by changing the "." to "sh", but don't
know if that's the ideal solution. After all, "wscons stop" still isn't
supported...

Oh, and also the "usage" error in rc.wscons is only echoed to the console;
it doesn't go into the syslog or into dmesg (and it's nonsensical anyway).

Shall I file a PR, or is someone already working through these?