Re: tlug: Time

["Frank Bennett (=?iso-2022-jp?B?GyRCJVUlaSVzJS8kWSVNJUMlSBsoQg==?= )" <bennett@example.com>]

Concerning the "time/tcp" entries in /etc/inetd.conf, which were commented
out.  I had edited these to invoke in.timed, which was wrong, wrong,
wrong.  When I restored the original lines and uncommented them, I still
couldn't get inetd to respond to rdate requests after a kill -HUP <pid>.

We return to our story as Frank responds to Stephen's post...

On Mon, Sep 06, 1999 at 06:16:58PM +0900, Stephen J. Turnbull wrote:

Thank you for a clear and careful response again, Stephen.  I had in fact
tampered with the "time" line to invoke in.timed on port 37, and that was
obviously a wrong thing to do.

> >>>>> "FB" == "Frank Bennett" <bennett@example.com> writes:
...
> According to the rdate manpage, RFC 868 is "usually implemented" in
> inetd, not a separate server.  And in fact grepping /etc/inetd.conf
> and /etc/services gives
> 
> daytime		stream	tcp	nowait	root	internal
> #daytime	dgram	udp	wait	root	internal
> time		stream	tcp	nowait	root	internal
> #time		dgram	udp	wait	root	internal
> 
> daytime		13/tcp
> daytime		13/udp
> time		37/tcp		timserver
> time		37/udp		timserver
> timed		525/udp		timeserver
> 
> Are you sure you've got timed plugged in to the right service?  It
> sure looks to me like it is a different protocol from rdate, and
> probably what's happening is that the naive rdate protocol is trying
> to interpret a magic number from timed as a date.

I set things back to the status quo ante, and started getting "Connection
refused" again at the rdate end.  I began composing a reply to you, but
kept fiddling during the writing, because your explanation jives perfectly
with the documentation, and no one else seems to be having any trouble
with this.

The trick turned out to be to fix the contents of the inetd.conf file,
AND to kill inetd and restart it --- kill -HUP apparently wasn't enough.
Makes sense now in retrospect: the appropriate lines were commented out in
the beginning, I added the "in.timed" line, and kill -HUP got inetd to
reread its config tables and invoke the external utility.  I eliminated
the value, and inetd reverted to its native self.  Killing and restarting
caused its native self to change to the settings I wanted.

I know how to do this now, but from the admin's point of view, if kill
-HUP is meant to cause inetd to reread its config file, it seems to me
that the entire file should take effect, to keep the interface clean. 
Failing that, if some entries require different incantations to give
effect to them, it should be documented in the config file itself, to
quash the gotchas.

Not a whinge, really, just a comment.  I'm copying this to the
maintainers listed in inetd.conf.

Cheers,
-- 
-x80
Frank G Bennett, Jr         @@
Faculty of Law, Nagoya Univ () email: bennett@example.com
Tel: +81[(0)52]789-2239     () WWW:   http://rumple.soas.ac.uk/~bennett/

-------------------------------------------------------------------
Next Nomikai: September 17 (Fri), 19:30 Tengu TokyoEkiMae 03-3275-3691
Next Technical Meeting: October 9 (Sat), 13:00     place: Temple Univ.
-------------------------------------------------------------------
more info: http://www.tlug.gr.jp        Sponsor: Global Online Japan