[comp.sys.ibm.pc.rt] RT's and TCP. HELP!!!

barron@eniac.seas.upenn.edu (Daniel P. Barron) (03/20/89)

Hi Folks!

AAAARRRGGHH!!  I have an RT running AIX 2.2.1 and TCP/IP 2.2.1.  Well, sort
of running it.  Some questions, for any help with which I shall be eternally
grateful.

1) When /etc/rc.tcpip runs at startup I get error messages from the named
   daemon such as:

   Mar 19 21:52:02 rt108 named[92]: zoneref: Masters for secondary zone wharton.upenn.edu unreachable

   rt108 is my hostname and wharton.etc is the network I'm on 128.91.11.xx

   It seems I've simply got the wrong stuff in my /etc/named.* files, but I
   can't for the life of me figure out what *should* be in there.  I've been
   at this for only a few months, but still, I don't think it should be
   *this* difficult to figure out.
   An explanation would be nice, but even nicer would be a copy of somebody's
   working named.* files, just to look at.  It's hard as hell figuring out
   what IBM wants you to put in these files.  I'm trying to hook my RT up
   to the Internet directly.  Funny thing is, it sort of works.  I can telnet
   ftp and (sorta) sendmail.  Still, these errors are frustrating.  Which 
   brings me to the second problem:

2) With TCP/IP v. 2.2 I could do use dotted decimal format, i.e.
   telnet 128.91.11.1, but not nameserving, i.e. telnet wharton.  Now with
   this "updated" version, names work and (get this)  numbers don't!  What
   I would like to do is use remote nameserver on our site, only resorting
   to a local file if there is no entry upstream.  Can I do this?

3) Finally, even when things work and I can telnet to remote hosts, full
   screen stuff (like editors, etc.) doesn't work.  Is this due to not
   running X-windows?  I mean, the stuff works, but the keys are all
   screwed up and don't behave properly even though the thing is supposed 
   to be emulating a VT100.  Is keyboard remapping the answer?  Can I do
   that?

Any help, pointers, *anything* would be appreciated.  IBM doc stinks and 
even though my doc/software version numbers match it's obvious the info
in the doc is *wrong* (i.e. it names files that no longer exist, etc.)

I just don't think it should be *this* hard.

HELP!

adTHANKSvance,

db

p.s. When we got this working at first (ver. 2.1) it took me and *two* 
IBM tech refs *and* a very knowledgeable friend over three days to get
it working.  And this update is a nightmare.  I say again, it shouldn't 
be *this* hard!!!
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
| Daniel P. Barron            | E-Mail to:                            |
| Wharton Computing           |           barron@wharton.upenn.edu    |
| University of Pennsylvania  |           barron@eniac.seas.upenn.edu |
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
= "Millions long for immortality who don't know what to do with       =
=  themselves on a rainy Sunday afternoon"                            =
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

melanie@uxe.cso.uiuc.edu (03/22/89)

um, in reference to your first problem, you do have the correct
netmask entered in /etc/net? if youre a class b network but youre
subnetting, your netmask should be 255.255.255.0 or 0xffffff00

>2) With TCP/IP v. 2.2 I could do use dotted decimal format, i.e.
>   telnet 128.91.11.1, but not nameserving, i.e. telnet wharton.  Now with
>   this "updated" version, names work and (get this)  numbers don't!  What
>   I would like to do is use remote nameserver on our site, only resorting
>   to a local file if there is no entry upstream.  Can I do this?

dotted decmial works on my rt, but i am not a slave nameserver, just 
a lowly client. all i have is a resolv.conf file that lists a
couple of nameservers.

>3) Finally, even when things work and I can telnet to remote hosts, full
>   screen stuff (like editors, etc.) doesn't work.  Is this due to not
>   running X-windows?  I mean, the stuff works, but the keys are all
>   screwed up and don't behave properly even though the thing is supposed 
>   to be emulating a VT100.  Is keyboard remapping the answer?  Can I do
>   that?

no, the vt100 emulator in ibm aix telnet is hopelessly wrong. everyone i know 
running aix has this same problem. my local ibm software person even
admitted it was a basically bad implementation. your best bet is to find 
someone who has typed in a termcap entry for the 5151. (dont look at me!)
the good news is, their 3270 emulator is great.

>p.s. When we got this working at first (ver. 2.1) it took me and *two* 
>IBM tech refs *and* a very knowledgeable friend over three days to get
>it working.  And this update is a nightmare.  I say again, it shouldn't 
>be *this* hard!!!

bunda@cs.utexas.edu (John Bunda) (03/22/89)

In article <8949@netnews.upenn.edu>, barron@eniac.seas.upenn.edu (Daniel P. Barron) writes:

>    It seems I've simply got the wrong stuff in my /etc/named.* files, but I
>    can't for the life of me figure out what *should* be in there.  
...
>    An explanation would be nice, but even nicer would be a copy of somebody's
>    working named.* files, just to look at.

Look in /usr/lpp/tcpip/samples, where you'll find examples of what these
name database files should look like.  There is also (or there was in 2.2)
a shell script to convert an /etc/hosts format file to a name database.
You'll also need to define a resolv.conf file.  See the TCP/IP book
for setting up "domain name service".

> 2) With TCP/IP v. 2.2 I could do use dotted decimal format, i.e.
>    telnet 128.91.11.1, but not nameserving, i.e. telnet wharton.  Now with
>    this "updated" version, names work and (get this)  numbers don't!  What
>    I would like to do is use remote nameserver on our site, only resorting
>    to a local file if there is no entry upstream.  Can I do this?

Numbers should work fine.  I suspect other problems are causing your
hostname resolver to not work properly.

> 3) Finally, even when things work and I can telnet to remote hosts, full
>    screen stuff (like editors, etc.) doesn't work.  Is this due to not
>    running X-windows?  I mean, the stuff works, but the keys are all
>    screwed up and don't behave properly even though the thing is supposed 
>    to be emulating a VT100.  Is keyboard remapping the answer?  Can I do
>    that?

I suspect you are running into problems with telnet eating control chars.
There are several ways around this - one is to use rlogin, the other
is to use stty to disable the normal xon/xoff handshaking but I forget
the exact details of how to do this.

-John






-- 
...................................
John Bunda            UT CS Dept.      
bunda@cs.utexas.edu   Austin, Texas    

eap@bu-cs.BU.EDU (Eric Pearce) (03/23/89)

In article <221600003@uxe.cso.uiuc.edu> melanie@uxe.cso.uiuc.edu says:
>no, the vt100 emulator in ibm aix telnet is hopelessly wrong. everyone i know 
>running aix has this same problem. my local ibm software person even
>admitted it was a basically bad implementation. your best bet is to find 
>someone who has typed in a termcap entry for the 5151. (dont look at me!)
>the good news is, their 3270 emulator is great.

 I could probably supply a termcap, since we have the 4.3 RT's here too.
 I took the IBM specific entries and added them to our Suns.  The 8514
 one needed a little modification.  I was able to add terminfo entries
 on the AIX machine with tic source produced by "infocmp" on a Sun4.

 What "3270 emulator" are you talking about?  The only one we have is
 "telnet" in "em 3270" mode, which locks up your terminal and doesn't
 respond to any keystrokes other than telnet escape character.  Has
 anybody got that "xant" thing I've hear about?

-- 
-------------------------------------------------------------------------------
 Eric Pearce                                   ARPANET eap@bu-it.bu.edu
 Boston University Information Technology      CSNET   eap%bu-it@bu-cs
 111 Cummington Street                         JNET    jnet%"ep@buenga" 
 Boston MA 02215                               UUCP    !harvard!bu-cs!bu-it!eap 
 617-353-2780 voice  617-353-6260 fax          BITNET  ep@buenga

njs@scifi.UUCP (Nicholas J. Simicich) (03/24/89)

In article <5068@cs.utexas.edu> bunda@cs.utexas.edu (John Bunda) writes:
-In article <8949@netnews.upenn.edu>, barron@eniac.seas.upenn.edu (Daniel P. Barron) writes:
 (Q&A about named elided.)

-> 2) With TCP/IP v. 2.2 I could do use dotted decimal format, i.e.
->    telnet 128.91.11.1, but not nameserving, i.e. telnet wharton.  Now with
->    this "updated" version, names work and (get this)  numbers don't!  What

-Numbers should work fine.  I suspect other problems are causing your
-hostname resolver to not work properly.

We've reported this bug, unless someone else has beaten us to it.
Call support, and get the fix if there is one, or get them to send you
the fix when they build one.

-> 3) Finally, even when things work and I can telnet to remote hosts, full
->    screen stuff (like editors, etc.) doesn't work.  Is this due to not
->    running X-windows?  I mean, the stuff works, but the keys are all
->    screwed up and don't behave properly even though the thing is supposed 
->    to be emulating a VT100.  Is keyboard remapping the answer?  Can I do
->    that?

-I suspect you are running into problems with telnet eating control chars.
-there are several ways around this - one is to use rlogin, the other
-is to use stty to disable the normal xon/xoff handshaking but I forget
-the exact details of how to do this.

If this is your problem, you can issue stty -ixon before doing your
telnet.  As far as what is transmitted when you press a key, yes,
keyboard remapping may be your answer.  I presume from your
descriptions that your screens are being displayed correctly, but when
you type pfk's and cursor keys that odd things are happening.  

-- 
Nick Simicich --- uunet!bywater!scifi!njs --- njs@ibm.com (Internet)