SUMMARY: inetd failing

From: Stephen Pendleton (pendleto@math.ukans.edu)
Date: Wed Feb 22 1995 - 00:17:25 CST


My original post was:
 
I am running 4.1.3_U1 on a Sparc 10 and I have on occasion logged this error:
 
Feb 17 12:57:38 titania inetd[195]: 100068/rpc/udp server failing
(looping), service terminated
 
I then lose all connections to the network. I can't telnet to or from or
ping from the machine.
 
 
The corresponding line (I think) in the inetd.conf file is:
100068/2-3 dgram rpc/udp wait root /usr/etc/rpc.cmsd rpc.cmsd
 
What is this daemon used for?
The only change I have made in the inetd.conf file is to add a daemon
called identd to the file. It is part of a package called pidentd. Could
this conflict?
 
Does anyone have any ideas?
 
Thanks,
Steve
 
*******************************************

Dave Curado ( davec@HK.Super.NET ) replied:
 
The problem is sun's inetd can only accept 40 connections per
minute. The patch is available from Sun, Patch-ID# 100178-09...
 
Then you can then start inetd from rc.local with something like
 
        inetd -r 80 60
 
meaning accept 80 connections per 60 seconds...
 
If you have trouble getting the patch, I have a copy here...
 
davec

*******************************

I think he is correct since I applied the patch and have had no problems
since. We have been getting a tremendous number of hits on our WWW site
lately so maybe this could have overloaded the daemon.
As a side note Mathew Keenan (matt@uts.EDU.AU) suggested I take the
100068/2-3 out of inetd.conf since it is a security hole.

Thanks,

Steve



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:17 CDT