I got quite a few responses to this question. Thanks,
everyone!
Most people thought it probably was a blank line in
the master file for the services database. I had
looked at the file, but not specifically for blank
lines, so, lo and behold, there one was at the end
of the file! I removed it, of course, but I have to
wait for machines to reboot to see if it really does
fix it. If enough machines reboot without the error,
I'll be convinced. (The thunder storms predicted for
today should do the trick)...
Several others are having the same problem. Sun has advised
some to put a sleep before and after the inetd startup, so
that it has time to finish reading the services database
before it tries to start anything else. Others suggested
a blank line in /etc/inetd.conf, but that would have to
be everywhere, and it's not an NIS database, so I don't
think that's the case.
There is a patch which fixes inetd breaks on fast machines.
The ID is 100178-04, for 4.1 and 4.1.1. I'm going to order
the patch, just in case it's not the blank line in services.
I got many similar responses, so I won't include them all here.
If you would like to see them, send me an e-mail request. That
includes the patch README.
Thanks to the following respondents:
Bill Shorter bill@aloft.att.com
Brett Lymn blymn@baobab.awadi.com.AU
Brent A. Wiese brent@bcm.tmc.edu
Jeff Beadles jeff@onion.rain.com
Birger birger@vest.sdata.no
Gordon gorpong@sbcoc.com
Dave Plummer dlp@mph.sm.ucl.ac.uk
Bertil Roslund bertilr@dit.lth.se
Geert Jan geertj@ica.philips.nl
Mike Sullivan mbs@bae.bellcore.com
Julann Santos Julann.Santos@East.Sun.COM
John Kilheffer amp19263@garfield.amp.com
Tim White tim@otis.hssc.scarolina.edu
mikem@ll.mit.edu
Jeff Pack jfp@inel.gov
michael pearlman canuck@rice.edu
dan@breeze.bellcore.com
Bala Vasireddi bala@Synopsys.COM
Josef jgaehwil@itr.ch
miker@sbcoc.com
stephen dowdy stephen@urel.medtronic.com
Angela Thomas Hoynowski angela@tandem.com
Ralph Howard rahjr@bae.bellcore.com
mikey@ccs.carleton.ca
The original (though edited) request follows:
> Oh, great sun-managers!
>
> We have been having a problem with many of our SPARCstations
> here in the last few weeks. When a machine boots, it comes
> up normally, then, just after the login prompt appears, the
> following messages come up:
>
> lpd[122]: printer/tcp: unknown service
> inetd[119]: login/tcp: unknown service
>
> or:
>
> inetd[114]: shell/tcp: unknown service
> inetd[114]: comsat/udp: unknown service
> lpd[117]: printer/tcp: unknown service
>
> In every case, doing a kill -HUP on inetd would clear the problem.
> For some reason, inetd couldn't deal with life at boot time, though.
>
> It looks like something is wrong with the services database, but
> what?
>
> What am I missing here??
>
Leslie Dreyer
AT&T Bell Laboratories
Allentown, PA
lbd@alux2.att.com
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:43 CDT