Comment 8 for bug 809646

On 07/13/2011 10:05 PM, gholt wrote:
> Ah, the existence check and the log*msg stuff make sense, just wish we
> (well, you guys) didn't have to duplicate the command list every time
> but I don't really see a way around that.
>
> As far as 'shutdown' I meant that the command keyword should be added to
> the list, not replacing 'stop' with it. Both commands exist.

I don't think that "shutdown" is a standard parameter for a init.d
script. Why do you want to add it? Does it do anything different than
"stop"? I see: "shutdown: allow current requests to finish on supporting
servers". But then, how do we implement this in a "normal" script that
doesn't use swift-init?

> Similarly with 'restart' and 'reload'.

I believe that all init.d scripts I wrote have the following:

restart|force-reload|reload)
        $0 stop
        sleep 1
        $0 start
;;

so they do have restart and reload. Is there anything I missed?

> I'm not sure either on the daemonizing thing. Let us know if it's
> broken; we haven't had problems in our use of it so it's probably
> something a bit more subtle to us.

To what I tested, some daemon just wont create PID, or wont go in the
background, and all sorts of "jokes". I cared having the package working
and uploaded in SID, but truth is, Soren is right, I should have
reported it. I just thought there was some weird reasons behind it.

Now, I think we should get rid of swift-init on all startup scripts
completely if possible.

Thomas