SUMMARY: What means: bootp/udp server failing (looping), service terminated?

From: Detlev Habicht (habicht@imes.uni-hannover.de)
Date: Sun Jan 30 1994 - 07:50:24 CST


Hi all!

Thanx to all.

This was my problem:

I have this message on our server:

Jan 27 09:07:33 popeye inetd[562]: bootp/udp server failing (looping),
                                   service terminated

Sometimes i see it every 10 minutes. I don't understand this message.
What can i do? Where is the problem?

And here are the hints:

---------------------------------------------------------------------------------
From: davec@cs.ust.hk (Dave Curado)

On Sun's inetd, they have "loop detection" which means
if there are too many connections to a service in a
short amount of time, inetd will assume there is some
sort of evil loop happening, and will terminate the
service. If you don't use xterms or something that
need the bootp stuff, why not just take it out of
your services map.

If you do use bootp to help xterms and/or diskless clients
boot up, maybe one of them is very talkative and should
be silenced. =-)

hope this helps, Dave C.

BTW - there is a patched inetd out there on the net,
that you can start up with some arguments, ie:

inetd -r 60 30

means, allow 60 connections on a service in a 30 second
time span.

I think sun's stock inetd is set to like... uh, 40
connections in a 60 second period.

(the patch it wicked easy, just mv the file in, and
 edit your rc.local to make the inetd call to
   inetd -f ## ##

Bye. Dave

---------------------------------------------------------------------------------
From: barnes@sde.mdso.vf.ge.com (Barnes William)

You have bootp turned on in inetd.conf and you have a device that is trying to boot from you under protocol bootp (like a PC or Concentrator or bridge ...). However, for some reason, the bootp call keeps failing. You can usually find this by checking the information in the /etc/bootptab file (if you have one) or starting bootp with the -d flag for debug (again, not always available, depends on vendor that supplied it to you.)

Bill Barnes
barnes@mdso.vf.ge.com

---------------------------------------------------------------------------------
From: stephen@biology.ucsc.edu

Well in your /etc directory there is a file called inetd.conf which
the 'inetd' (internet daemon) reads, to start services that are being
requested on some particular port on your server. If inetd finds the
service listed in inetd.conf, it trying to start it up.

It looks like it is trying to start a service called 'popeye' that
is (by my guess) a bootp service request. A bootp daemon is started
when another machine makes a request about its IP address and possibilty
other information that then allows the remote machine to 'boot' and start
its network software.

>Sometimes i see it every 10 minutes. I don't understand this message.

>What can i do? Where is the problem?
>
 Well you can remove or comment out the entry in the /etc/inetd.conf
file. Then wait and see if anyone complains. Or try to find out where
this request is coming from.

Stephen Hauskins

---------------------------------------------------------------------------------
From: cecilp@vancouver.cantel.rogers.com (Cecil Pang)

Hi,

This seems that your udp is not running properly. The bootp have to
interface with udp and you udp is not 100% functioning.

This is usually used for diskless client and peripherals.

Check your /etc and YP files ethers, hosts, bootparams and services for
possible error.

Check your if /etc/inetd.conf has something like:-

bootps dgram udp wait root /usr/etc/bootpd bootpd

use ps -aux | grep rpc
and check if daemon is up and running

root 133 0.0 0.0 52 0 ? IW Jan 25 0:00 rpc.bootparamd

---------------------------------------------------------------------------------
>From lrr@Princeton.EDU Thu Jan 27 19:55:14 1994

Too many connections so inetd dropped the service. See patch 100178-latest for
a fix to this.

 ===== ======= ======= Larry Rogers

---------------------------------------------------------------------------------
>From bern@kleopatra.Uni-Trier.DE Thu Jan 27 20:07:55 1994

Aehhh ... Das muss nicht notwendigerweise am angemahnten Service liegen.
Passieren tut's, wenn der inetd "den Eindruck hat", dass ein Internet-
Service 'ne Endlosschleife fabriziert; Dazu testet er, ob die Rate der
fuer einen Service ankommenden Requests "vernuenftig" ist (wie auch
immer festgelegt). Ist sie laengere Zeit zu hoch, Abbruch. Nun ist diese
Rate aber zu Zeiten nicht ganz so schneller Hardware fest vorgegeben
worden ... Irgendjemand hat mir 'mal einen inetd Jumbo Patch fuer dieses
Problem empfohlen, aber ich fand's bis heute nicht noetig, den zu ziehen,
und die Mail find' ich gerade nicht.

Viel Glueck,
                                                                        J. Bern
---------------------------------------------------------------------------------
>From wpmc!mother!cygan@uu5.psi.com Thu Jan 27 23:53:56 1994

We got this message for a while and traced it back to a
Synoptic hub that had its boot prom set to boot off the
net, but we were not set up that way. After it failed, it
then boots off the prom, so we were not having any problems,
only the error. We changed the defaults so that it booted
off the prom automatically, and it the errors went away.

---------------------------------------------------------------------------------
From: tkevans@fallst.es.dupont.com (Tim Evans)

You probably have a blank line in /etc/services or /etc/inetd.conf

-- 
Tim Evans                     |    E.I. du Pont de Nemours & Co.

--------------------------------------------------------------------------------- From: abeckett@fmlrnd.co.uk

This is due to the inetd receiving two many connections on that port in a short space of time. We've hit the same problem with the shell port for remote shells. Mind you - this doesn't tell you what is causing it to receive too many connections.

*********************************************************************** * Andrew Beckett * *

--------------------------------------------------------------------------------- From: Eelco van Asperen <evas@cs.few.eur.nl>

Inetd is the "super-daemon" that starts off other daemons to handle requests from clients. When a daemon is called more than a certain number of times per second, inetd considers that a failure ("hmm, that daemon appears to terminate almost immediately after I started it, that can't be right") and prints the message above.

Bootp is a protocol that is used by hosts to get network configuration information. Perhaps there is some host on your network that puts out a lot of bootp requests or there are a lot of hosts that are booted at the same time. When a bootp server receives a request, it looks in its database to see if it has got information on that host; if it has, it sends a reply. If it does not have info on the host, it terminates. So: perhaps your bootp server is "triggered" by hosts on your net that is does not know about. If your bootp server supports it, try logging all requests and look at that to see which hosts are sending the bootp requests.

Unfortunately, there is no way to adjust the limit that inetd has for the number of requests per second.

-- Eelco van Asperen. | Erasmus University Rotterdam

Thanx to all answers, i will test the hints!

Detlev

-- Detlev Habicht _______________/\ ____________________________________________________ \/ OFFICE: D-30167 Hannover +49 511 7624992 habicht@imes.uni-hannover.de Institut fuer Mikroelektronische Systeme, Universitaet Hannover HOME: D-30419 Hannover +49 511 2712029 detlev@toppoint.de



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:08:55 CDT