frankh@durin.sparta.COM (Frank Halsema) (02/21/90)
I seem to have a simple setup problem but just can't get it right.
I am trying to run NCSA Telnet with MacTCP. When I execute it, Telnet
brings up the intial logo window for a few seconds and then exits. I
think the problem is the MacTCP start up. Here is my configuration.
The net has two bridges. A Hayes interbridge and a Kinetics KFPS 4. The
KFPS 4 is running K-star vers. 8. I am using atalkad on a Sun. My KFPS has
CONFIGURATION:
Atalk Phase 2
K-Star IP
Atalk Phase 1
LocalTalk EtherTalk 2 Ethertalk 1
Zone * Start 1 Zone *
Net # 0 End 1 Net 0
Node # 220 Net 1 Node 195
Node 195
IP INFO
FP Box 192.0.0.195
(all else 0.0.0.0)
IP CLIENTS
Dynamic 0 Static 0
ADMIN
Host 192.0.0.7
IPTALK
Zone *
Net 0
Node 195
All other fields are 0
Options 4 and 17 are on.
My atalkatab has:
55.0 N0 192.0.0.0 SpartaLH-Ether #Laguna Eng, appletalk-in-IP
57.0 E 192.0.0.195 SpartaLH-ET #Laguna Eng, ethertalk
56.0 KC 192.0.0.195 SpartaLH-4th-AT #Laguna Eng, 4th floor localtalk
I192.0.0.0 I192.0.0.7 #ipbroad ipname
I192.0.0.7 L0 #ipdebug ipfile
L0 L0 L0 L0 S57.0 S0 #ipother atnetet ddprangestart
LX0 S3 S40 #flags ipstatic ipdynamic
S56.0 S55.0 "SpartaLH-Ether"#atneta atnete zonea
I set up MacTCP to be in the zone SpartaLH-4th-AT and gave it an address
of 192.0.0.196. The Mac host file has these entries
durin.sparta.com. A 192.0.0.7
mac.sparta.com. A 192.0.0.196
Lastly my config.tel for NCSA has these changes
zone="SpartaLH-4th-AT"
name=durin; hostip=192.0.0.7; nameserver=1
If anyone has any ideas I Thanks in advance.
--
Frank Halsema UUCP: durin!frankh
SPARTA, Inc. ARPA: frankh@durin.sparta.com
23041 de la Carlota, Suite 400
Laguna Hills Ca, 92653 (714) 768-8161 EXT 339 (714)583-9114 FAXpwrg6481@uxa.cso.uiuc.edu (02/23/90)
The problem you seem to be having is that MacTCP is not properly
installed. Telnet will kick out like that (and did to me when I first
got the MacTCP version) if MacTCP is not there in memory. Are you sure
it is set up correctly?
pr
--
Pete Resnick (...so what is a mojo, and why would one be rising?)
Graduate assistant - Philosophy Department, Gregory Hall, UIUC
System manager - Cognitive Science Group, Beckman Institute, UIUC
Internet/ARPAnet/EDUnet : resnick@kant.cogsci.uiuc.edu or
pwrg6481@uxa.cso.uiuc.edu
BITNET (if no other way) : FREE0285@UIUCVMDloganj@yvax.byu.edu (03/01/90)
We had a similar problem because we didn't understand how the "Dynamic" and "Static" addresses work in the KFPS with K-star. If I remember correctly, at least one of them must be non-zero or nothing works. These numbers specify a range of legal consecutive node numbers immediately following the KFPS node number (in your case 195) that can be used by the Macs. And if your Mac node number falls out of that range then it doesn't work. :( We specified 30 dynamic nodes and 30 static nodes, which means that all of the Macs on the AppleTalk must have node numbers within 60 AFTER the KFPS node number. Seems to be working okay, but I don't know if you can mix static and dynamic node numbers arbitrarily throughout the specified range. Hope that helps. Regards, jim loganj@yvax.byu.edu loganj@byuvax.bitnet
dent@unocss..unl.edu (Local Submission) (03/02/90)
loganj@yvax.byu.edu writes: >We had a similar problem because we didn't understand how the "Dynamic" and >"Static" addresses work in the KFPS with K-star. If I remember correctly, >at least one of them must be non-zero or nothing works. These numbers >specify a range of legal consecutive node numbers immediately following the >KFPS node number (in your case 195) that can be used by the Macs. And if >your Mac node number falls out of that range then it doesn't work. :( I think the way it is set up, 'dynamic' IP node numbers are handed out more- or-less randomly to a mac whenever it needs it, regardless of it's AppleTalk node number, location on the LocalTalk, etc. So, if you were running, say, NCSA Telnet, and told it to get the IP node number dynamically, it would just ask the FastPath, and be assigned one. The 'static' IP node numbers are sort of ignored by the FastPath; all it seems to dp is move the start of the dynamic IP numbers. So, if you have 5 static numbers, 10 dynamic ones, and the FastPath's IP node number ends in .106, the FastPath would dynamically hand out IP node numbers from .112 to .122. I suppose this is in case you have any macs on your LocalTalk net that have specific IP node numbers that you want constant (which can be done by using MacTCP, for example). So in any case, if you want dynamic IP number handling, then you shouldn't be hand-assigning /any/ Macs IP Node numbers. The corrolary to this is that if you are hand-assigning Mac IP node numbers, you don't need /any/ dynamic ones inthe FastPath. One more thing to note: if you don't care for having all of your Mac IP node nums directly following the FastPath one, K-Star 8.0 lets you move the beginning of the range elsewhere (or to have a few right after the FastPath's, and a few somewhere completely elsewhere). As an anecdote... has anyone else had their FastPaths suddenly seize control of the LocalTalk? We had a power glitch (we think!), and weren't running atalkd anywhere yet, so the FastPath wound up scrambling itself in the user room. Then, it took control of the LocalTalk net and wouldn't let anyone on it! We brought up the chooser on one Mac, to see why the printers were acting weird, and noticed that all 4 of our ImageWriters were blinking in and out of the network. I can imagine the following conversation: "Hi! I'm an Apple ImageWriter just coming on to the AppleTalk, so could I please have, say '114' as my node number? Anyone using that one yet?" Fastpath: "SHUT UP! GET OFF MY NETWORK!" "Yikes! Ok, ok!" I guess the FastPath was bullying around all those poor innocent ImageWriters. :-) (We wound up completely reloading the FastPath to fix the problem...) -/ Dave Caplinger /--------------------------------------------------------- Microcomputer Specialist, Campus Computing, Univ. of Nebraska at Omaha dent@zeus.unomaha.edu ...!uunet!unocss!dent DENT@UNOMA1
morgan@JESSICA.STANFORD.EDU (03/02/90)
> We specified 30 dynamic nodes and 30 static nodes, which means that > all of the Macs on the AppleTalk must have node numbers within 60 > AFTER the KFPS node number. Seems to be working okay, but I don't > know if you can mix static and dynamic node numbers arbitrarily > throughout the specified range. The statically-assigned addresses start with the first one above the IP address of the box itself, and run upwards. The dynamically- assigned addresses run upwards from the end of the statically-assigned ones. Personally, I've never understood why anyone would use statically-assigned addresses, since they're harder to set up and have no obvious benefits. Are there applications that require them? - RL "Bob" Morgan Networking Systems Stanford
jln@accuvax.nwu.edu (John Norstad) (03/03/90)
In article <9003011903.AA18844@jessica.Stanford.EDU> morgan@JESSICA.STANFORD.EDU writes: (writing about FastPath boxes and dynamic and static IP addresses) > Personally, I've never understood why anyone would use > statically-assigned addresses, since they're harder to set up and have > no obvious benefits. Are there applications that require them? We use them for two things here at NU: 1) We have a Mail*Link mail gateway from StarNine on our QuickMail server. It needs a static IP address so that other Internet hosts can find it. 2) For security reasons, we plan to assign static IP addresses to all those staff Macs whose owners use the NetNews HyperCard stack to post USENET news articles. We configure our NNTP server to allow postings only from those fixed IP addresses. It's too easy for students on the public access Macs to use NetNews to masquerade as other people when they post. Unfortunate. Our scheme isn't perfect, but it helps. There's probably other reasons for using static addresses - these are just two of them. John Norstad Northwestern University jln@acns.nwu.edu
billkatt@mondo.engin.umich.edu (billkatt) (03/03/90)
In article <2323@unocss..unl.edu> dent@unocss..unl.edu (Local Submission) writes: >So in any case, if you want dynamic IP number handling, then you shouldn't >be hand-assigning /any/ Macs IP Node numbers. The corrolary to this is that >if you are hand-assigning Mac IP node numbers, you don't need /any/ dynamic >ones inthe FastPath. One more thing to note: if you don't care for having all >of your Mac IP node nums directly following the FastPath one, K-Star 8.0 lets >you move the beginning of the range elsewhere (or to have a few right after >the FastPath's, and a few somewhere completely elsewhere). You don't want to use kboxes to do your IP for Macs with ethernet. It slows things down too much and makes your Macs dependent on the kbox. Doubles net traffic for IP, too. > >As an anecdote... has anyone else had their FastPaths suddenly seize control of >the LocalTalk? We had a power glitch (we think!), and weren't running atalkd >anywhere yet, so the FastPath wound up scrambling itself in the user room. >Then, it took control of the LocalTalk net and wouldn't let anyone on it! We >brought up the chooser on one Mac, to see why the printers were acting weird, >and noticed that all 4 of our ImageWriters were blinking in and out of the >network. I can imagine the following conversation: Um... That is a power supply problem, not a problem with not having atalkad yet. Actually, it works like this: Power glitches->Kbox glitches->kbox stack dumps on the LocalTalk. The problem is that the kbox stack dumps so rapidly and often that the net is so full of kbox packets that it can't carry any other traffic. Also, I don't think the kbox listens before talking when it stack dumps. > > "Hi! I'm an Apple ImageWriter just coming on to the AppleTalk, > so could I please have, say '114' as my node number? Anyone > using that one yet?" Imagewriters come in above 127. They also crash when anything broadcasts excessively, like a crashed kbox. > >Fastpath: "SHUT UP! GET OFF MY NETWORK!" > > "Yikes! Ok, ok!" > >I guess the FastPath was bullying around all those poor innocent ImageWriters. >:-) (We wound up completely reloading the FastPath to fix the problem...) The story of my life. -Steve
schmidt@SUMEX-AIM.STANFORD.EDU (Christopher Schmidt) (03/06/90)
In addition to the others mentioned, I like static addresses for the
following reasons.
I have a couple of nets where I converted all the macs to static
addresses because localtalk macs occasionally (using either SUMacIP 3.0 or
MacTCP) would complain "can't get my IP address" for no obvious reason. This
happened often enough that it was worth assigning static addresses to quench
the user complaints. We're running KSTAR 7.0U on Fastpath-4's, but had the
same problem with 7.0 (and with 7.0U on Fastpath-2U's). I think I had the
problem with KIP, too, but it's so long ago now, that I forget.
Another reason I like static addresses is that I can finger a user who
is telnetted into our Sun from one of our many macs and tell by the
name/address which mac he is actually using so I can find him. (A single
Fastpath-4 serves 25 offices on three floors of one of our buildings.)
I can ping my mac from anywhere in the world.
The main reason I like static addresses, however, is that email I send
from my mac is obviously from my mac (SUMEX-Mac-19.Stanford.edu) and not from
the relatively anonymous "kbox-a" or whatever!
--Christopher
-------lemke@radius.UUCP (Steve Lemke) (03/07/90)
In article <CC1427840A7F201C26@BAYLOR.BITNET> CALIFFM@BAYLOR.CCIS.BAYLOR.EDU (Michael Califf) writes: } }Along this line, has anyone else had problems with a KBox running KStar }losing track of addresses which it has handed out after a period of }inactivity? MacTCP acquires its address when opened. If we run }NCSA Telnet, then quit and go on to other activities and then go back }to Telnet some time (hours) later things lock up, often requiring a }reboot. I'll have more specifics after I get a few minutes to watch }the packets. } }Mike Califf }califfm@baylor.edu Hmmm. We might have something here... I'm using NCSA Telnet's own static addressing, without MacTCP (since I never got around to setting up K-Star for dynamic addressing). I have a Mac SE with a EtherSC on it, and I've had lots of periodic crashes with NCSA. Usually it will happen after many hours of running NCSA, or several launches of it. Lately it's settled down, and I _do_ realize that it could be any one of a million other things, but it always seemed to somehow be related to NCSA. I'm using v2.3, and I guess now is the time to at least mention it - perhaps there are others like your- self and myself out there who have had this problem but have ignored it like I have, thinking that it was merely some init conflict or other... Ideas? Questions? Comments? Post or Email - whatever's easier for you. -- ----- Steve Lemke, Engineering Quality Assurance, Radius Inc., San Jose ----- ----- Reply to: radius!lemke@apple.com (Coming soon: radius.com ...) -----