[comp.unix.ultrix] _SLOW_ response to login request

hudson@ux1.cso.uiuc.edu (John Hudson) (09/18/90)

I had read in several previous postings of unusually long waits for
logins under Ultrix and had only seen one solution - use a non-DEC
version of bind.

I am running Ultrix 3.1 (UWS 2.1) on a VaxStation II-GPX.  Typical waits
for valid logins are from 90-120 seconds, sometimes longer.  The problem
seems to be independent of machine load, request location (local, remote,
or console) and communication program (rlogin, telnet, etc.).  

Any suggestions would be appreciated!

John Hudson

pemurray@miavx1.acs.muohio.edu (Peter Murray) (10/03/90)

In article <1990Sep17.203920.26835@ux1.cso.uiuc.edu>, hudson@ux1.cso.uiuc.edu (John Hudson) writes:
> I had read in several previous postings of unusually long waits for
> logins under Ultrix and had only seen one solution - use a non-DEC
> version of bind.
> 
> I am running Ultrix 3.1 (UWS 2.1) on a VaxStation II-GPX.  Typical waits
> for valid logins are from 90-120 seconds, sometimes longer.  The problem
> seems to be independent of machine load, request location (local, remote,
> or console) and communication program (rlogin, telnet, etc.).  

Are you running the Gateway version of telnetd (telnetd.gw)?  This version
of the daemon allows users to use the login: prompt to connect to other
machines on decnet and the internet.  When you mentioned that you replaced
bind, my ears perked up.

Running the non-gateway version of telnetd might increase response time.

Peter
-- 
Peter Murray            Neat UNIX Stunts #6:             pemurray@miavx1.bitnet
176 Thompson Hall         csh> ar m God           pmurray@apsvax.aps.muohio.edu
Oxford, OH 45056                       NeXT Mail:  pmurray@next4.acs.muohio.edu

jenst@dxcsjr.cern.ch (Jens Trier Rasmussen) (10/03/90)

pemurray@miavx1.acs.muohio.edu (Peter Murray) writes:

>In article <1990Sep17.203920.26835@ux1.cso.uiuc.edu>, hudson@ux1.cso.uiuc.edu (John Hudson) writes:
>> I had read in several previous postings of unusually long waits for
>> logins under Ultrix and had only seen one solution - use a non-DEC
>> version of bind.
>> 
>> I am running Ultrix 3.1 (UWS 2.1) on a VaxStation II-GPX.  Typical waits
>> for valid logins are from 90-120 seconds, sometimes longer.  The problem
>> seems to be independent of machine load, request location (local, remote,
>> or console) and communication program (rlogin, telnet, etc.).  

I've seen the same problem when I was using BIND and the nameserver was
down. Check if you are using BIND and the sequence in /etc/svc.order
(maybe put local before bind).

regards
	Jens Trier Rasmussen
	CERN-CN/CS
	jenst@cernvax.cern.ch