topic Re: TCP wrapper causing telnet hang time in Operating System - HP-UX

TCP wrapper causing telnet hang time

TheJuiceman — Tue, 06 Nov 2007 20:03:04 GMT

I recently enabled tcpd for telnet on three of our boxes....one is 11.23 and two are 11.11.

Now we have a few PC's that are experiencing connection delays (up to 20 second delays) before getting a login prompt, but only on one of the 11.11 and the 11.23 systems. The other 11.11 system is fine. All of the affected PC's are acting the same way, but it is only a FEW PC's that are having this problem, not all. If I edit the inetd.conf file back to OTB, then the hang time goes away. All three boxes are set up the same. Here is the inetd.conf entry in question...

telnet stream tcp nowait root /usr/lbin/tcpd /usr/lbin/telnetd -b /etc/
issue -TCP_DELAY

What could be causing the delay? And why would it only affect a few PC's? And why only two of the servers? Thanks.

Re: TCP wrapper causing telnet hang time

Bill Hassell — Tue, 06 Nov 2007 21:27:18 GMT

May not be the problem but delays getting a login prompt are very often due to DNS failures. Put the address of a failing PC into /etc/hosts with a dummy name and make sure your nsswitch.conf is setup for files then DNS like this:

hosts: files [NOTFOUND=continue UNAVAIL=continue] dns

Now try the failing PC. If it works, then the telnet daemon's request to validate the incoming IP address is being ignored by the DNS server (or the DNS server is getting blocked).

Re: TCP wrapper causing telnet hang time

TheJuiceman — Tue, 06 Nov 2007 22:00:43 GMT

I would think that it would be a DNS issue too. However, if I set up telnet to NOT use the tcpd, then the connection is instant. It is definitely something with tcpd.

Re: TCP wrapper causing telnet hang time

TheJuiceman — Tue, 06 Nov 2007 23:49:07 GMT

I found the problem. It was the /etc/tcpd.conf file. This file has the following entry:

rfc931_timeout 15

Because of this, the authentication was timing out 15 seconds, thus delaying the connection.

The server that WAS working had an error in the /etc/tcpd.conf file. Removing this entry in the file on all of the servers fixed the problem.