[comp.protocols.appletalk] MacTCP and FastPath 4

hpoppe@ncar.ucar.edu (Herb Poppe) (11/14/90)

I have not been successful in getting MacTCP on LocalTalk nodes to work 
with a FastPath 4 (4.1 ROMs) running the built-in "IP sub-sectioning" code.

I called Shiva Tech Support and was told that this was possible, but 
required patching MacTCP with ResEdit. He claimed that they run FastPaths 
that way in house, but could not tell me how to make the patch. He also 
said that there was no one else there that could help me.

If anyone in netland knows how to do this, I would sure like to know how. 
If, in fact, it is not possible to make MacTCP work with "IP 
sub-sectioning", I would like to hear about that, too. Then I'll stop 
wasting time on this and use K-STAR 8.x.

Thanks.

Herb Poppe             hpoppe@ncar.ucar.edu
NCAR                      (303) 497-1296
1850 Table Mesa Dr.
Boulder, CO  80307-3000

moyman@ECN.PURDUE.EDU (James M Moya) (11/15/90)

>
>From: ncar.ucar.edu!hpoppe@handies.ucar.edu  (Herb Poppe)
>Organization: National Center for Atmospheric Research
>Subject: MacTCP and FastPath 4
>
>I have not been successful in getting MacTCP on LocalTalk nodes to work 
>with a FastPath 4 (4.1 ROMs) running the built-in "IP sub-sectioning" code.
>
>I called Shiva Tech Support and was told that this was possible, but 
>required patching MacTCP with ResEdit. He claimed that they run FastPaths 
>that way in house, but could not tell me how to make the patch.

Wonderful, tech support saying you need to modify *MacTCP* to *fix* the
problem.

> He also 
>said that there was no one else there that could help me.

There never is...

>
>If anyone in netland knows how to do this, I would sure like to know how. 
>If, in fact, it is not possible to make MacTCP work with "IP 
>sub-sectioning", I would like to hear about that, too.

I would also like to know this *fix*


I have also been getting intermittant problems with MacTCP with the new
Shiva Boxes.  The Kinetics FP-4 running KSTAR 7.0 had no problems
what-so-ever with MacTCP.  Since booting up several Shiva Boxes upgrades,
with 7.0.1 and 8.0.1 the IP problems have started to pour in.  Nothing like
upgrading FastPath *1* that hasn't given me a lick of trouble in 3 years,
to a Shiva 4 that has been complaining since I booted it up...they didn't
tell me that was include in the upgrade.  The old FP was slow but hey, it
worked flawlessly.  What a deal!

>Then I'll stop 
>wasting time on this and use K-STAR 8.x.
>
You may be wasting your time with KSTAR also...^^^
--moya

--Mike Moya 
--Macintosh Systems and Networking
--Engineering Computer Network, Purdue University
--moyman@ecn.purdue.edu or ..!pur-ee!moyman

bschmidt@bnr.ca (Ben Schmidt) (11/15/90)

In article <9011141828.AA02286@aquarium.ecn.purdue.edu> 
moyman@ECN.PURDUE.EDU (James M Moya) writes:
> >
> >From: ncar.ucar.edu!hpoppe@handies.ucar.edu  (Herb Poppe)
> >Organization: National Center for Atmospheric Research
> >Subject: MacTCP and FastPath 4
> >
> >I have not been successful in getting MacTCP on LocalTalk nodes to work 
> >with a FastPath 4 (4.1 ROMs) running the built-in "IP sub-sectioning"
> >code.

Yup, MacTCP does an NBP lookup for an IPGATEWAY.  But i don't think that FastPaths running the IP subnetting code in the firmware (I'm too old too learn that new-fangled word Shiva invented: sub-sectioning) don't register themselves as IPGATEWAYs.  So, upon failing to find an IPGATEWAY, MacTCP then closes itself!!?  Talk about giving up easily!  Whose fault?  I don't know.  Seems Apple could help out a litte, since they are only now writing (through John Vezaides' IETF group) the RFC for DDP/IP gateways.  



How about a bit more forgiveness for the old boxes?  Ah well, that's progress...   :^)

> >I called Shiva Tech Support and was told that this was possible, but 
> >required patching MacTCP with ResEdit. He claimed that they run 
> >FastPaths that way in house, but could not tell me how to make the patch.
>

sounds like some legal thing.  Apple's fine print probably says you're not 
allowed to modify system software or your hair colour or anything, and we're all legally in contravention of this restriction for ResEditing our Finder Trashcan icons into IBM logos!   :^)  :^)   So I figure maybe Shiva's lawyers  told their staff not to encourage people to do quasi-legal things?  sigh.  but I work for a big company and these types of legal abstractions happen all the time...  :^)  (maybe Momma was right, I shoulda' gone to law school)
 
> Wonderful, tech support saying you need to modify *MacTCP* to *fix* the
> problem.

Well, non-MacTCP telnets seem to have no problems finding FastPaths running IP subnetting code.  If you configure MacTCP to use a static address within the IP subnet administered by a FastPath running the firmware code, it seems plausible that MacTCP could have been written to work in this situation.

> > He also said that there was no one else there that could help me.
> 
> There never is...
> 
> >
> >If anyone in netland knows how to do this, I would sure like to know 
> >how.  If, in fact, it is not possible to make MacTCP work with "IP 
> >sub-sectioning", I would like to hear about that, too.
> 
> I would also like to know this *fix*
> 
> 
> I have also been getting intermittant problems with MacTCP with the new
> Shiva Boxes.  The Kinetics FP-4 running KSTAR 7.0 had no problems
> what-so-ever with MacTCP.  Since booting up several Shiva Boxes upgrades,
> with 7.0.1 and 8.0.1 the IP problems have started to pour in.  Nothing 
> like upgrading FastPath *1* that hasn't given me a lick of trouble in 3 
> years, to a Shiva 4 that has been complaining since I booted it up...they
> didn't tell me that was include in the upgrade.  The old FP was slow but
> hey, it worked flawlessly.  What a deal!
> 

Well you could just get an Ethernet card.  It works fine with MacTCP.  :^) 

> >Then I'll stop wasting time on this and use K-STAR 8.x.

I'm sure all KFPS users are waiting for K-Star 8.1.  Too many 
things seem to have changed with K-Star 7-8.0, what with NBP filtering, SNMP support, and ATP2 support.  I'm sure that there's a lot of lose ends that a 2nd pass, with no additional features added, just code refinement would address.  Hopefully Shiva has figured the code out by now, and we'll get K-Star 8.1 under our Xmas trees?!   :^)

> You may be wasting your time with KSTAR also...^^^

Are we having one of those days, Mike?    :^)

(PS. We've got 200 FPs on our internet, so i hope i'll be forgiven for my 
tongue in cheek remarks above. :^)

Ben Schmidt     Bell-Northern Research, Ltd.   Ph: (613) 763-3906
Information     P.O. Box 3511, Station C       FAX:(613) 763-3283
Technology      Ottawa Canada K1Y 4H7          bschmidt@bnr.ca

rbrink@hubcap.clemson.edu (Rick Brink) (11/16/90)

From article <9161@ncar.ucar.edu>, by hpoppe@ncar.ucar.edu (Herb Poppe):
> I have not been successful in getting MacTCP on LocalTalk nodes to work 
> with a FastPath 4 (4.1 ROMs) running the built-in "IP sub-sectioning" code.
> 
> I called Shiva Tech Support and was told that this was possible, but 
> required patching MacTCP with ResEdit. He claimed that they run FastPaths 
> that way in house, but could not tell me how to make the patch. He also 
> said that there was no one else there that could help me.
> 
Is there a reason you don't use K-Star IP, and dynamic addressing?  It's 
simpler to set up (assuming it works correctly.)  The only reason I know
of is to get a static address for mail purposes.  And you can get static
addresses with K-Star as well.

rbrink@hubcap.clemson.edu (Rick Brink) (11/16/90)

From article <1990Nov14.232423.1183@bwdls61.bnr.ca>, by bschmidt@bnr.ca (Ben Schmidt):
> 
> I'm sure all KFPS users are waiting for K-Star 8.1.  Too many 
> things seem to have changed with K-Star 7-8.0, what with NBP filtering, SNMP support, and ATP2 support.  I'm sure that there's a lot of lose ends that a 2nd pass, with no additional features added, just code refinement would address.  Hopefully Shiva has figured the code out by now, and we'll get K-Star 8.1 under our Xmas trees?!   :^)
> 
K-Star 8.1 is available on Shiva.com which supports anomymous FTP.  I got it 
from there a few weeks ago.  Don't remember what directory it was in...

bschmidt@bnr.ca (Ben Schmidt) (11/16/90)

In article <11640@hubcap.clemson.edu> rbrink@hubcap.clemson.edu (Rick 
Brink) writes:
> From article <1990Nov14.232423.1183@bwdls61.bnr.ca>, by bschmidt@bnr.ca 
(Ben Schmidt):
> > 
> > I'm sure all KFPS users are waiting for K-Star 8.1.  Too many 
> > things seem to have changed with K-Star 7-8.0, what with NBP 
> > filtering, SNMP support,  and ATP2 support.  I'm sure that 
> > there's a lot of lose ends that a 2nd pass, with no additional 
> > features added, just code refinement would address.  Hopefully 
> > Shiva has figured the code out by now, and we'll get K-Star 8.1
> > under our Xmas trees?!   :^)
> > 
> K-Star 8.1 is available on Shiva.com which supports anomymous FTP.  
> I got it from there a few weeks ago.  Don't remember what directory 
> it was in...

Rick, are you sure you're not getting confused with shiva.com's
 
                  /dist/FastPath/K-Star/8.0.1
                                          ^
which is advertised to be merely a patched version of K-Star 8.0 to
ensure that the new FastPaths with Shiva's Ethernet vendor code don't
blow their brains when they don't find Excelan's "080089" code as part
of their firmware Ethernet address?

Ben Schmidt     Bell-Northern Research, Ltd.   Ph: (613) 763-3906
Information     P.O. Box 3511, Station C       FAX:(613) 763-3283
Technology      Ottawa Canada K1Y 4H7          bschmidt@bnr.ca

PETER@YALEVM.BITNET (Peter Furmonavicius) (11/16/90)

>K-Star 8.1 is available on Shiva.com which supports anomymous FTP.  I got it
>from there a few weeks ago.  Don't remember what directory it was in...
 
I presume you mean version 8.0.1.
Its in the: /dist/FastPath/K-Star/8.0.1 directory.
 
BITnet:  Peter @ YaleVM      InterNet:  Peter @ YaleVM.YCC.Yale.edu
 
| Yale University Computer Center |      Peter Furmonavicius
| 175 Whitney Avenue              |      Senior Research Programmer
| P.O. Box 2112                   |            ^ Macintosh  :-)
| New Haven, CT  06520            |      (203) 432-6600

johnc@synopsys.com (John Crossley) (11/17/90)

We have MacTCP running over FastPath 4's with no problems (after we 
figured how to configure it).  On one of my more satisfying calls to 
Shiva, I was told the setup we are using with good results:
1.  Dump IP Subsectioning
2.  Use K-Star 8.0 (or better)
3.  Use AppleTalk phase 1 (for our net)

We run each LocalTalk network as a Class C subnet (That is, the EtherNet 
is 192.9.200.x and one of the local zones is 192.9.203.x)

If this is a help, I can mail the FastPath configuration.

P.S.  We are evaluating the Webster Multigate.  It was doing fine until it 
took up smoking.  You see, we don't allow smoking in the computer room.


John Crossley
(415) 962-7069

mst@mx.csun.edu (Michael Temkin) (11/19/90)

In article <9161@ncar.ucar.edu> hpoppe@ncar.ucar.edu (Herb Poppe) writes:
>I called Shiva Tech Support and was told that this was possible, but 
>required patching MacTCP with ResEdit. He claimed that they run FastPaths 
>that way in house, but could not tell me how to make the patch. He also 
>said that there was no one else there that could help me.
>
>Thanks.
>
>Herb Poppe             hpoppe@ncar.ucar.edu
>NCAR                      (303) 497-1296
>1850 Table Mesa Dr.
>Boulder, CO  80307-3000

I set this up here and it works great.  I am currently at home so I
don't have the setup here, but I will post it when I get to work.  For
those of you who don't know what the advantage is, here goes...

If you set it up without the subsectioning and you allocate, say, 20 static
and 30 dynamic addresses, then these addresses are the 50 addresses after
the FastPath on the ethernet.  Thus you just lost 51 addresses and you
only have 254 to start with (class B network).  If you use subsectioning,
then the 51 addresses are used on a different subnet and only 1 is lost
on the ethernet.  This is especially important if the FP is placed on the
main backbone.

As far as Shiva goes, if they said that, maybe they should sell off
the rights to the FastPath before they kill off all their business.

Mike.

--
Mike Temkin
mst@csun.edu
Cal. State U. Northridge, School of Engineering and Computer Science
Voice phone: (818) 885-3919