Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 14 Jun 1996 12:29:40 -0700 (PDT)
From:      Dave Babler <dbabler@Rigel.orionsys.com>
To:        questions@FreeBSD.org
Subject:   Rlogin delay
Message-ID:  <Pine.BSF.3.91.960614105035.8253A-100000@Rigel.orionsys.com>

next in thread | raw e-mail | index | archive | help
I've observed something concerning rlogin that is a bit odd to me, and 
increasingly annoying for my users. When a user logs off and then tries 
to re-log within a reasonable delay from another machine on the local 
network, the login will be ingored for a very long time (10-60 seconds 
seems typical). I checked the /etc/inetd.conf to make sure it was okay:

---- /etc/inet.conf
#login	stream	tcp	nowait	root	/usr/libexec/rlogind	rlogind
login	stream	tcp	nowait	root	/usr/local/libexec/tcpd	rlogind
----

I've recently added the tc_wrappers port, hence the tcpd entry (it works 
great, BTW). Contrary to the setup, it seems to behaving as though it had 
been setup as 'wait' instead of 'nowait'. As a test, I logged in from the 
other machine and then logged off and periodically ran netstat. This is 
what I see:

---- connection active, test is from port 1022 ----
Active Internet connections
Proto Recv-Q Send-Q  Local Address          Foreign Address        (state)
tcp        0      0  rigel.login            orionsys.com.1022      ESTABLISHED
---- 1022 just logged off ----
tcp        0      0  localhost.hermes       localhost.domain       TIME_WAIT
tcp        0      0  rigel.login            orionsys.com.1022      TIME_WAIT
tcp        0      0  rigel.1235             204.162.115.56.rwhois  FIN_WAIT_2
---- logged off several seconds ago ----
tcp        0      0  localhost.1249         localhost.domain       TIME_WAIT
tcp        0      0  rigel.login            orionsys.com.1022      TIME_WAIT
tcp        0      0  rigel.1235             204.162.115.56.rwhois  FIN_WAIT_2
---- logoff +10 seconds ----
tcp        0      0  localhost.1250         localhost.domain       TIME_WAIT
tcp        0      0  rigel.login            orionsys.com.1022      TIME_WAIT
---- logoff +30 seconds ----
tcp        0      0  localhost.1253         localhost.domain       TIME_WAIT
tcp        0      0  localhost.1252         localhost.domain       TIME_WAIT
tcp        0      0  rigel.login            orionsys.com.1022      TIME_WAIT
---- logoff +45 seconds ----
tcp        0      0  localhost.1256         localhost.domain       TIME_WAIT
tcp        0      0  localhost.1255         localhost.domain       TIME_WAIT
tcp        0      0  rigel.login            orionsys.com.1022      TIME_WAIT
---- finally gone! over 60 seconds later ----
udp        0      0  localhost.domain       *.*                   
udp        0      0  rigel.domain           *.*                   

This is somewhat trimmed down.. it was running with 2 other users using 
TinTin. I don't know if it has any relevance to the re-log problem, but I've 
also had users report significant delays (approaching 60 *seconds*) on 
their connections to the MUD server (Mortal Realms) using TinTin and I've 
been putting that down to overloaded MCI routers in the SF bay area.

Thanks!

-Dave Babler




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.91.960614105035.8253A-100000>