From owner-freebsd-hackers  Fri Jul 14 16:11:13 1995
Return-Path: hackers-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id QAA04436
          for hackers-outgoing; Fri, 14 Jul 1995 16:11:13 -0700
Received: from mailhost1.cac.washington.edu (mailhost1.cac.washington.edu [140.142.32.2])
          by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id QAA04430
          for <freebsd-hackers@freebsd.org>; Fri, 14 Jul 1995 16:11:10 -0700
Received: from hope96.apl.washington.edu by mailhost1.cac.washington.edu
	(5.65+UW95.05/UW-NDC Revision: 2.33 ) id AA13020;
	Fri, 14 Jul 95 16:11:09 -0700
Received: from mac-pal.apl.washington.edu by apl.washington.edu (4.1/SMI-4.1)
	id AA18557; Fri, 14 Jul 95 16:11:07 PDT
Message-Id: <9507142311.AA18557@apl.washington.edu>
Date: Fri, 14 Jul 1995 16:09:43 -0800
From: kargl@apl.washington.edu (steve)
To: freebsd-hackers@freebsd.org
Subject: SOLVED: rsh/rlogin problems in 2.0.5-RELEASE
Sender: hackers-owner@freebsd.org
Precedence: bulk

This was posted in the newsgroup.   I thouhgt a core team member
might want to investigate.

> From: jdp@polstra.com (John Polstra)
> Newsgroups: comp.unix.bsd.freebsd.misc
> Subject: SOLVED: rsh/rlogin problems in 2.0.5-RELEASE
> Date: 14 Jul 1995 10:37:58 -0700
> Organization: Polstra & Co., Seattle, WA
> Lines: 53
> Message-ID: <3u6a1m$q30@seattle.polstra.com>
> NNTP-Posting-Host: seattle.polstra.com
> Summary: Caused by #comments in /etc/hosts.equiv
> 
> Many people have posted accounts of long delays or hangs when trying to
> use rsh or rlogin to a system running FreeBSD 2.0.5-RELEASE.  Today I
> discovered the cause of the problem, at least on my system.  Once you
> understand the cause, the fix is trivial.
> 
> On attempts to rsh from a SVR4 machine into the FreeBSD-2.0.5 machine, I
> was observing delays of almost exactly 2.5 minutes.  Luckily, there is
> also a third machine on the same ethernet, running SunOS.  That machine
> has a sniffer program called "etherfind".  I ran it to observe what was
> going on during the execution of rsh.  Immediately, I saw that the long
> delays were related to unanswered DNS queries that were being sent by the
> FreeBSD machine to the SVR4 machine (which is the name server for the
> local net).
> 
> On the name server machine, I turned on debugging in the "in.named"
> program.  Here are the strange queries that were coming from the FreeBSD
> machine:
> 
>     req: nlookup(#localhost) id 1280 type=1
>     req: nlookup(#my_very_good_friend.domain) id 1536 type=1
> 
> These queries are of course bogus, and the name server (rightly or
> wrongly) did not respond to them.  On the FreeBSD side, the resolver
> timed out and tried again, several times, adding up to (surprise) 2.5
> minutes of delay.
> 
> A quick "grep my_very_good_friend /etc/*" took me straight to the
> "/etc/hosts.equiv" file, which looked like this:
> 
>     #localhost
>     #my_very_good_friend.domain
> 
> I had never touched this file, so it must be the default one that gets
> installed when one installs FreeBSD 2.0.5-RELEASE.
> 
> THE SOLUTION:  Get rid of (or repair) that bogus "/etc/hosts.equiv" file.
> That fixes the problem instantly.
> 
> THE MORAL:  Don't try to put comments into "/etc/hosts.equiv".  (A quick
> look at "/usr/src/lib/libc/net/rcmd.c" confirms that there is no comment
> processing for this file.)
> 
> A STRANGE OBSERVATION:  Strangely, I had no problems with rsh or rlogin
> until today.  And I have not changed any of the relevant files, as far
> as I know.  So I suspect that the failure mode for this particular
> problem depends in some way on the vagaries of the particular name
> server that is being used.  Even if you're not currently having
> problems, you should repair your "/etc/hosts.equiv" file.
> 
> -- 
>    John Polstra                                       jdp@polstra.com
>    John D. Polstra & Co., Inc.                Seattle, Washington USA
>    "Self-knowledge is always bad news."                 -- John Barth