[comp.protocols.appletalk] Running CAP with K-star 8.0

arnold@synopsys.synopsys.com (Arnold de Leon) (07/06/90)

Hello,

  Can I run cap 5.0 with K-star 8.0?  What does it take to run
cap?  Am I asking the right questions?

Background:

  I have 5 Fastpaths on the same ether.  Previously they were used
only to connect Localtalk networks.  Now I want our Suns to do be
able to share printers with the macs, enter cap.

  I've compiled cap and have been trying it in "debug" mode.  Running
"getzones" prints out the zones we have available.  But running "atlook"
prints out nothing.  Also "atistest" fails even when I have the atis
daemon running.

  Any pointers would be most appreciated.  Thanks.

arnold@synopsys.com (Arnold de Leon) (07/07/90)

After more testing I noticed that if I put 

at-rtmp         201/udp                         # udp: rtmp
at-nbp          202/udp                         # udp: nbp
at-echo         204/udp                         # udp: echo
at-zip          206/udp                         # udp: zip

in the YP map more /etc/services atistest would report:

CAP distribution 5.00 using UDP encapsulation, March 1988
Copyright (c) 1986,1987,1988 by The Trustees of Columbia University in the Cityk

abInit: [ddp:   0.42, 11] starting
debugging NBP
Registering "atis test:testing@*"
NBP SndNBP: sending
NBP nbp_timeout: 4 tick timeout on -536872120, 3 remain
NBP SndNBP: sending
NBP nbp_timeout: 4 tick timeout on -536872120, 2 remain
NBP SndNBP: sending
NBP nbp_timeout: 4 tick timeout on -536872120, 1 remain
NBP SndNBP: sending
NBP nbp_timeout: 4 tick timeout on -536872120, 0 remain
NBP SndNBP: sending
NBP nbp_timeout: 4 tick timeout on -536872120, 3 remain
NBP SndNBP: sending
NBP nbp_timeout: 4 tick timeout on -536872120, 2 remain
NBP SndNBP: sending
NBP nbp_timeout: 4 tick timeout on -536872120, 1 remain
NBP SndNBP: sending
NBP nbp_timeout: 4 tick timeout on -536872120, 0 remain
nbp register error because: couldn't register name - is atis running?

If services are removed and atis is restarted I get:

CAP distribution 5.00 using UDP encapsulation, March 1988
Copyright (c) 1986,1987,1988 by The Trustees of Columbia University in the City
of New York

abInit: [ddp:   0.42, 11] starting
debugging NBP
Registering "atis test:testing@*"
NBP SndNBP: sending
NBP nbp_timeout: 4 tick timeout on -536872120, 3 remain
NBP SndNBP: sending
NBP nbp_timeout: 4 tick timeout on -536872120, 2 remain
NBP SndNBP: sending
NBP nbp_timeout: 4 tick timeout on -536872120, 1 remain
NBP SndNBP: sending
NBP nbp_timeout: 4 tick timeout on -536872120, 0 remain
NBP SndNBP: sending
NBP status done: found -536872120
Okay

I somehow think I am missing something obvious.

	arnold

wcc@cup.portal.com (wcc - usa) (07/09/90)

> After more testing I noticed that if I put 
> 
> at-rtmp         201/udp                         # udp: rtmp
> at-nbp          202/udp                         # udp: nbp
> at-echo         204/udp                         # udp: echo
> at-zip          206/udp                         # udp: zip
  ^^^^^^
Where'd you get this from? It's "at-zis", not "at-zip". Grep the CAP 
sources for "zis" and zip" just to make sure.

Note that it doesn't really matter which socket range you use (the "old"
768-series or the "new" 200-series), but the only essential matter is that
ALL devices using these sockets are using the SAME ones. ALL CAP hosts and
ALL encapsulating gateways.

If you want to use the 200-series, you have to reconfigure the gateway
(kbox?) to use the 200 series as well.

Note it is possible (likely - it happens a lot) that the Sun's "portmapper"
has grabbed the 768-series ports. If you start aufs early enough in your
boot sequence (before portmapper) you can beat it to them. Until the next
time someone changes your rc.local, that is.

> in the YP map more /etc/services atistest would report:

> nbp register error because: couldn't register name - is atis running?

> If services are removed and atis is restarted I get:

> NBP status done: found -536872120
> Okay

Meaning your kbox is running the 768-series.

Tom Evans
wcc@cup.portal.com
Webster Computer Corporation
Suite J, 2109 O'Toole Ave. San Jose, California 95131-1303
Ph (408) 954-8054  FAX (408) 954-1832

Head Office
tom@wcc.oz.au, multigate@wcc.oz.au
1270 Ferntree Gully Rd. Scoresby, Victoria, Australia
Ph 61 3 764-1100  FAX 61 3 764-1179

arnold@synopsys.com (Arnold de Leon) (07/10/90)

In article <31552@cup.portal.com> wcc@cup.portal.com (wcc - usa) writes:
>> After more testing I noticed that if I put 
[ snip snip]
>> at-zip          206/udp                         # udp: zip
>  ^^^^^^
>Where'd you get this from? It's "at-zis", not "at-zip". Grep the CAP 
>sources for "zis" and zip" just to make sure.
>

I got it from the CAP documentation (doc/installation.ms)

>Note that it doesn't really matter which socket range you use (the "old"
>768-series or the "new" 200-series), but the only essential matter is that
>ALL devices using these sockets are using the SAME ones. ALL CAP hosts and
>ALL encapsulating gateways.
>
>If you want to use the 200-series, you have to reconfigure the gateway
>(kbox?) to use the 200 series as well.
>

  This turned out to be the key.  There is an option on the
Fastpaths (enable options 11) for using the "new" 200-series
port numbers. (Thanks Tom Evans!)

  After sorting out /etc/atalk.local (those spaces in Zone
names will kill ya!), I get it the basic stuff working.

  I can now print on printers on the Appletalk network.

	arnold

-- 
Arnold de Leon  			arnold@synopsys.com
Synopsys Inc.				(415) 962-5051
1098 Alta Ave.
Mt. View, CA 94043