root@mjbtn.JOBSOFT.COM (Mark J. Bailey) (03/10/91)
Hi, I am having a problem with Xenix TCP/IP that is about to drive me crazy! I have SCO Xenix 2.3.3 running TCP/IP 1.0.1h. I have a 386 DOS system that is ethernetted to the Xenix system using WD8003e Ethercard Plus's. The card is set on Interrupt 2, I/O Address 0x240, Mem Addres 0xD000. I am trying to use NCSA Telnet 2.2 and 2.3b14, WNQVTNET with Clarkson drivers, and CMU PCIP with Clarkson drivers. I did have an old copy of Locus TCP/IP for MSDOS that I tried too. What is happening is this. I also have an SCO Unix System (ODT 1.0 which is Unix 3.2.1) that contains TCP/IP 1.1 (SCO/Lachman). If I go from the dos system to the ODT/Unix system, everything is *PERFECT*!!!!! BUT, when i try to go from the DOS system to the Xenix system (using telnet), I get the "header" from the Xenix system: System V.2.3.3 XENIX (jcbmdpc.JOBSOFT.COM) but the login never shows up! As a matter of fact, *ALL* of the packages above except the Locus package just drop back out to the DOS prompt with *NO* error messages. That is normal behavior when you logout from a telnet session with the dos packages. It is almost like the Xenix system is dropping out before the login prompt comes up. As I said the Locus package stays in and you can login. Another weird thing is that the NCSA/WNQVTNET programs will work the *FIRST* time you telnet the xenix system after the xenix system is newly booted. If you try and open subseqeunt sessions with the one successful one, they drop out before the login prompt as before. Then if you logout the one session, and try to connect back, it fails from then on. Now, going from Unix-->Xenix and Xenix-->Unix works fine all the time. What urks me the most is that I can't seem to get any type of output from NCSA to tell me what is transpiring to the point xenix steps out. The fact that I get the remote system login header (above) tells me that it is likely not refusing telnet. The netstat -A on xenix shows sessions trying to start from the dos machine, but they have gone TIME WAIT. FTP from NCSA to Xenix works flawlessly as does REXEC. It is only the TELNET. What gives? What is in Unix (or that Unix does) that isn't in Xenix TCP/IP? Why would the Locus TCP/IP stay on to login to xenix? You might ask why not use Locus, well, terminal emulation sucks. It claims ANSI X3.64, but is nowhere near entries in termcap. As I said, it is a very old version. What is it about the DOS programs that might kick them out? If they didn't give up, might they still be able to login? For what it is worth, my Xenix IP address is 192.9.150.100 and the DOS uses 192.9.150.200. I have been told before that Xenix TCP/IP is very buggy - I agree! I have also been told by sales people at SCO the TCP/IP 1.2 for Unix/Xenix will be out 2nd or 3rd quarter of 1991 (maybe around June/July). I feel certain that that will solve my problems. However, in the meantime, what can I do about Xenix TCP/IP 1.0.1h? Is there some parameter tweaking on Xenix and/or the DOS packages I can do to get it to work? How can I get NCSA to be verbose and tell me what it is doing up until it kicks back to DOS? Questions, questions, questions! :-) I apologize for the long posting. I have been struggling with this for a couple of weeks now. I am beginning to get a twitch! :-) Any help, advise, suggestions, and/or consolations would be very much appreciated! Thanks in advance! Mark. -- Mark J. Bailey, N4XHX _______/====X11====\_______ USMAIL: 511 Memorial Blvd., Murfreesboro, TN 37129 | JobSoft | VOICE: +1 615 893 0098 | Design & Development Co.| UUCP: ...!uunet!mjbtn!mjb, ...!raider!mjbtn!mjb | Murfreesboro, TN USA | DOMAIN: mjb@mjbtn.JOBSOFT.COM CIS: 76314,160 --------------------------- <KA9Q-UNIX-USERS Mailing List-Subscribe: ka9q-unix-requests@mjbtn.jobsoft.com>
root@crash.cts.com (Bill Blue) (03/11/91)
In <3868@mjbtn.JOBSOFT.COM> root@mjbtn.JOBSOFT.COM (Mark J. Bailey) writes: }I am having a problem with Xenix TCP/IP that is about to drive me crazy! }I have SCO Xenix 2.3.3 running TCP/IP 1.0.1h. I have a 386 DOS system that }is ethernetted to the Xenix system using WD8003e Ethercard Plus's. The }card is set on Interrupt 2, I/O Address 0x240, Mem Addres 0xD000. I am }trying to use NCSA Telnet 2.2 and 2.3b14, WNQVTNET with Clarkson drivers, }and CMU PCIP with Clarkson drivers. I did have an old copy of Locus TCP/IP }for MSDOS that I tried too. } }What is happening is this. I also have an SCO Unix System (ODT 1.0 which }is Unix 3.2.1) that contains TCP/IP 1.1 (SCO/Lachman). If I go from the }dos system to the ODT/Unix system, everything is *PERFECT*!!!!! BUT, when }i try to go from the DOS system to the Xenix system (using telnet), I get }the "header" from the Xenix system: } }System V.2.3.3 XENIX (jcbmdpc.JOBSOFT.COM) } }but the login never shows up! As a matter of fact, *ALL* of the packages }above except the Locus package just drop back out to the DOS prompt with }*NO* error messages. That is normal behavior when you logout from a telnet }session with the dos packages. It is almost like the Xenix system is dropping }out before the login prompt comes up. As I said the Locus package stays in }and you can login. Another weird thing is that the NCSA/WNQVTNET programs }will work the *FIRST* time you telnet the xenix system after the xenix system }is newly booted. If you try and open subseqeunt sessions with the one }successful one, they drop out before the login prompt as before. Then if you }logout the one session, and try to connect back, it fails from then on. }Now, going from Unix-->Xenix and Xenix-->Unix works fine all the time. Interestingly, I went through a similar experience just this week, with Xenix tcp/ip and the ncsa telnet package. Same cards, same situation. You'll probably find that if you do a netstat -i or -r it will just hang. The clue is that you can't 'ping localhost' but you can 'ping 127.0.0.1'. The lack of name lookup, the inability to rlogin, telnet or ftp TO the Xenix tcp/ip, and the hanging netstat is due to an entry in resolve.conf which is read by routed when it starts up. If you have a 'nameserver' line in there, but the defined nameserver is not active at the moment, all these things will happen. Just delete or rename the file, and it'll all work. Of course, deleting or renaming the file is just to get you going. I haven't figured out how all the various config files and daemons interract, and the docs don't really tell much of the story, so I can't tell you much more than I have. If anyone listening would care to provide an informative tutorial, I'm sure a lot of people including myself would be greatly appreciative. --Bill
shevett@mccc.edu (Dave Shevett) (03/14/91)
In article <3868@mjbtn.JOBSOFT.COM> root@mjbtn.JOBSOFT.COM (Mark J. Bailey) writes: =Hi, = =I am having a problem with Xenix TCP/IP that is about to drive me crazy! =I have SCO Xenix 2.3.3 running TCP/IP 1.0.1h. I have a 386 DOS system that =is ethernetted to the Xenix system using WD8003e Ethercard Plus's. I have the identical setup, sans 2.3.3 (I have 2.3.1). = =System V.2.3.3 XENIX (jcbmdpc.JOBSOFT.COM) = ditto. =but the login never shows up! As a matter of fact, *ALL* of the packages =above except the Locus package just drop back out to the DOS prompt with =*NO* error messages. I have telnetted from my PC to the Xenix box, but you have to try it around 20 times, and then it connects, but is a little flakey. This is an acknowledged bug in TCP/IP 1.0.1h from SCO. I posted about it about a month and a half ago, and got 2 or 3 replies. The seemingly correct one is to go grab my 1.0.0 controlled release of TCP/IP, and snarf the telnetd off it and install that. Unfortunately, I'm down an ESDI controller at the moment, and have not had a chance to try this. NOTE - it only happens when gonig from a PC. WIN/QVT, FTP Software, NCSA, the whole shebang (I have a Unix-pC with Wollongong TCP/IP on it, and that can log in fine). = =What gives? What is in Unix (or that Unix does) that isn't in Xenix TCP/IP? A real implementation of this system. (pseudo real, it's still not BSD) =I have =also been told by sales people at SCO the TCP/IP 1.2 for Unix/Xenix will =be out 2nd or 3rd quarter of 1991 (maybe around June/July). I feel certain =that that will solve my problems. Me too. I have projects that have to be done, and working along a 19200 line just isn't the same. I WANT MY MULTISCREENS! =Mark J. Bailey, N4XHX _______/====X11====\_______ =USMAIL: 511 Memorial Blvd., Murfreesboro, TN 37129 | JobSoft | =VOICE: +1 615 893 0098 | Design & Development Co.| =UUCP: ...!uunet!mjbtn!mjb, ...!raider!mjbtn!mjb | Murfreesboro, TN USA | =DOMAIN: mjb@mjbtn.JOBSOFT.COM CIS: 76314,160 --------------------------- =<KA9Q-UNIX-USERS Mailing List-Subscribe: ka9q-unix-requests@mjbtn.jobsoft.com> ,___, .--------------------------------------------------------------. \o.o/ | Dave Shevett | Unix-PC BOF | The shortest distance between | ( \_/ ) | Lawrenceville,NJ | TCF '91 | two puns is a Straight Line | |_|_| | shevett@mccc.EDU | Apr 20-21st | -- Doc Webster | '--------------------------------------------------------------'
aris@tabbs.UUCP (Aris Stathakis) (03/14/91)
In <3868@mjbtn.JOBSOFT.COM> root@mjbtn.JOBSOFT.COM (Mark J. Bailey) writes: >Hi, >I am having a problem with Xenix TCP/IP that is about to drive me crazy! >I have SCO Xenix 2.3.3 running TCP/IP 1.0.1h. I have a 386 DOS system that >is ethernetted to the Xenix system using WD8003e Ethercard Plus's. The >card is set on Interrupt 2, I/O Address 0x240, Mem Addres 0xD000. I am >trying to use NCSA Telnet 2.2 and 2.3b14, WNQVTNET with Clarkson drivers, >and CMU PCIP with Clarkson drivers. I did have an old copy of Locus TCP/IP >for MSDOS that I tried too. Apparently this problem is specifically (TCP/IP with NCSA telnet) is fixed in SCO TCP/IP 1.2. Aris -- Aris Stathakis | Bang: ..!uunet!ddsw1!olsa99!tabbs!aris or aris@tabbs.UUCP - - - You cannot achieve the impossible without attempting the absurd -
root@crash.cts.com (Bill Blue) (03/18/91)
In <1991Mar13.191407.23820@mccc.edu> shevett@mccc.edu (Dave Shevett) writes: }In article <3868@mjbtn.JOBSOFT.COM> root@mjbtn.JOBSOFT.COM (Mark J. Bailey) writes: }=I am having a problem with Xenix TCP/IP that is about to drive me crazy! }=I have SCO Xenix 2.3.3 running TCP/IP 1.0.1h. I have a 386 DOS system that }=is ethernetted to the Xenix system using WD8003e Ethercard Plus's. ... }=but the login never shows up! As a matter of fact, *ALL* of the packages }=above except the Locus package just drop back out to the DOS prompt with }=*NO* error messages. } }I have telnetted from my PC to the Xenix box, but you have to try it around }20 times, and then it connects, but is a little flakey. } }This is an acknowledged bug in TCP/IP 1.0.1h from SCO. I posted about it }about a month and a half ago, and got 2 or 3 replies. The seemingly }correct one is to go grab my 1.0.0 controlled release of TCP/IP, and snarf }the telnetd off it and install that. Unfortunately, I'm down an ESDI }controller at the moment, and have not had a chance to try this. } }NOTE - it only happens when gonig from a PC. WIN/QVT, FTP Software, NCSA, }the whole shebang (I have a Unix-pC with Wollongong TCP/IP on it, and that }can log in fine). I find that 2.2D NCSA Telnet works perfectly with SCO Xenix TCP/IP 1.0.1h on version 2.3.3 Xenix. However, the beta version of 2.3 Telnet behaves as described above. I have been suspicious that it is due to the particular version of packet drivers (ne1000) that I am using. I'd really like to upgrade to 2.3 because of the Kermit-like ability to redefine the keyboard (I really hope they carry it further than they have so far, though). Anyway, anyone know anything more about the login problem on 2.3? --Bill