Post by SamPost by noelPost by Marco MoockPost by SamThe most convenient way of doing this with sysvinit is to drop an
S50mystuff
When I see this I automatically see a redhat or debian user
Maybe one from 10-15 years ago.
Neither Red Hat, nor Debian, works like this, any more.
Post by noelPost by Marco MoockIs it enough if that includes
#/bin/bash /etc/rc.d/rc.nscd start
Or does it need more?
Do I need a kill script or will it be killed automatically at shutdown?
If I need one, which runlevel would fit?
0?
just add it to rc.M and be done with it (will work *all run levels 1-6*)
That approach probably works for a one-off situation.
one off, no, weve used that for nearly 25 years, in production (ISP
environment) on all servers, incl hosting, its the most robust and
effective, a LOT of slackware servers run by others operate the same way.
Post by Sam- updates to new versions of Slackware, at some bright point in the future,
- or installs, or reinstalls, for any one of several different reasons,
oh, you're only a home user, even then, no, its suitable, rc.M stock,
hasn't changed that much, and keeping all the custom services in same
area makes it simple to include, of course now postfix and dovecot as
stock, some of it is redundant, though we still have pure-ftpd, as well
as a verious other things, like imap-proxy, radius, IRC server, jabber
server, news server :) and no doubt I've forgotten one or two others off
top of my head, so it wont take much to add them back if anythig
catastrophoic occurs, we have however had proprietary software (eg:
billing system ) come with service files and they dont work, so not
always a drop in replacement, so why bother.
Post by SamThen one wants to have some kind of a reproducible, and automated
process. That is, one that's initiated by a simple, basic command.
Manually editing rc.M is not going to fit the bill.
scp does.
we have the template for whatever service it is
if its MX-in or smtp we send over that
if its a imap/pop we send over that
web server we send its over, and so on and so on and so on.
...be they repairs (never happened) or adding to the cluster a new.
I'm betting I can commision a new server in same time or better than your
suggested methods.
but hey, OP can run it how he seem fit, thats the best part about
slackware, you tell the OS what you want, not having debian or redhat or
worse - microsoft, dictate what will happen.