pjh@mccc.UUCP (Peter J. Holsberg) (02/17/88)
I have a 3b2 direct-connected to a UNIX PC and to a second 3b2. The
UNIX PC's name is 'pc0'; the second 3b2's is 'mc3'. pc0 is on tty14;
mc3, on tty81. Here are the relevant lines from BNU files:
Devices:
mc3 tty81 - 9600 direct
pc0 tty14 - 9600 direct
Direct tty81 - 9600 direct \D
Direct tty14 - 9600 direct \D
Systems:
mc3 Any mc3 9600 - "" \r\d\r\d\r\d\r in:--in: nuucp
pc0 Any pc0 9600 - "" \r\d\r\d\r\d\r in:-- in: nuucp
/etc/inittab
81:234:respawn:/usr/lib/uucp/uugetty -r tty81 9600
14:234:respawn:/usr/lib/uucp/uugetty -r tty14 9600
cu -ltty81 gets me a connected message, while cu -ltty14 gets "CAN'T
ACCESS DEVICE". ???
I switched the wires between tty81 and 14. Now cu -ltty[either] gets
the "CAN'T ACCESS DEVICE". ???
(Oh, I rebooted after making the changes to /etc/inittab.)
Can anyone shed some light here? Thanks.
--
Peter Holsberg UUCP: {rutgers!}princeton!mccc!pjh
Technology Division CompuServe: 70240,334
Mercer College GEnie: PJHOLSBERG
Trenton, NJ 08690 Voice: 1-609-586-4800scl@virginia.acc.virginia.edu (Steve Losen) (02/17/88)
In article <218@mccc.UUCP> pjh@mccc.UUCP (Peter J. Holsberg) writes: > >I have a 3b2 direct-connected to a UNIX PC and to a second 3b2. The >UNIX PC's name is 'pc0'; the second 3b2's is 'mc3'. pc0 is on tty14; >mc3, on tty81. Here are the relevant lines from BNU files: > ... >cu -ltty81 gets me a connected message, while cu -ltty14 gets "CAN'T >ACCESS DEVICE". ??? > >I switched the wires between tty81 and 14. Now cu -ltty[either] gets >the "CAN'T ACCESS DEVICE". ??? ... >-- >Peter Holsberg UUCP: {rutgers!}princeton!mccc!pjh >Technology Division CompuServe: 70240,334 >Mercer College GEnie: PJHOLSBERG >Trenton, NJ 08690 Voice: 1-609-586-4800 The two most likely problems here are 1) lack of permissions or 2) lack of carrier detect (CD). (Or both, yuck!) If the "CAN'T ACCESS DEVICE" error prints immediately, you probably have a permission problem. See who owns the /dev/tty files (it should be uucp) and be sure that cu and uucico are setuid to uucp. If the error message appears after a delay of several seconds, the open is timing out due to lack of carrier detect (CD). When the 3b2 opens a tty port it raises DTR and if CD is not up, the 3b2 waits for CD to come up. Normally the 3b2 will wait FOREVER, but cu (and uucico) set an alarm before attempting to open the line. If the open doesn't return before the alarm goes off, cu terminates with the "CAN'T ACCESS DEVICE" message. Since you have a direct connection between your 3b2s you need to construct a cable that connects DTR at one end to CD at the other. You also need to cross Send Data and Receive Data (The pin that one 3b2 sends data on must obviously be connected to the pin that the other 3b2 reads data from.) This is tricky with those crazy phone jack cables that AT&T uses. You can get the desired cable by connecting two cables together. Use the adaptors that convert the phone jack to the familiar 25 pin RS232 connector. Some of these adaptors are marked "acu modem" and some are marked "terminal". Try this: (3b2) (cable) (terminal adaptor) (modem adaptor) (cable) (3b2) Hopefully you can find a modem and terminal adaptor that are different sexes or put a sex-change adaptor in between them. I've never seen a Unix PC. Presumably it also has phone jacks so the above advise might also work. To ensure that signals are being raised on the "slave" 3b2 and Unix PC, enable logins on the appropriate ports on those machines. You might want to disable logins on the outgoing port on the "master" 3b2 to avoid the "babbling getty syndrome" (Two machines simultaneously thinking that the other machine is trying to login). -- Steve Losen scl@virginia.edu University of Virginia Academic Computing Center
itkin@cup.portal.com (02/19/88)
>cu -ltty81 gets me a connected message, while cu -ltty14 gets "CAN'T >ACCESS DEVICE". ??? > >I switched the wires between tty81 and 14. Now cu -ltty[either] gets >the "CAN'T ACCESS DEVICE". ??? > >(Oh, I rebooted after making the changes to /etc/inittab.) > >Can anyone shed some light here? Thanks. > >-- >Peter Holsberg UUCP: {rutgers!}princeton!mccc!pjh First, since you're running uugetty, are you also running uugetty on the other systems? My understanding of uugetty is that they must be paired. Thus, if the UNIX PC is running getty, you may get into a feedback loop and the port may shutdown. Clearly, there is something wrong with the connection to pc0. Second thought is have you verified that the cable to pc0 is a good cable. The cabling on the 3B2 is a little different, and a cable that works for a 3b2<->3b2 connection probably wouldn't work directly with a UNIX PC. If you switch the cables BEFORE booting, does the connection to mc3 (what the system thinks is pc0) work? Good Luck.
pjh@mccc.UUCP (Peter J. Holsberg) (02/22/88)
In article <3315@cup.portal.com> itkin@cup.portal.com writes: |First, since you're running uugetty, are you also running uugetty on the |other systems? Uugetty on the 3b2, but getty on the 3b1. No HDB is available for my 3b1s. |My understanding of uugetty is that they must be paired. I don't think that needed. |Thus, if the UNIX PC is running getty, you may get into a feedback loop |and the port may shutdown. Clearly, there is something wrong with the |connection to pc0. Clearly! | |Second thought is have you verified that the cable to pc0 is a good |cable. The cabling on the 3B2 is a little different, and a cable that |works for a 3b2<->3b2 connection probably wouldn't work directly with a |UNIX PC. What's a "good cable"? It works fine when I use the 3b1 as a dumb terminal to the 3b2. -- Peter Holsberg UUCP: {rutgers!}princeton!mccc!pjh Technology Division CompuServe: 70240,334 Mercer College GEnie: PJHOLSBERG Trenton, NJ 08690 Voice: 1-609-586-4800
sullivan@vsi.UUCP (Michael T Sullivan) (02/24/88)
In article <231@mccc.UUCP>, pjh@mccc.UUCP (Peter J. Holsberg) writes: > In article <3315@cup.portal.com> itkin@cup.portal.com writes: > |First, since you're running uugetty, are you also running uugetty on the > |other systems? > Uugetty on the 3b2, but getty on the 3b1. No HDB is available for my 3b1s. > > |My understanding of uugetty is that they must be paired. > I don't think that needed. Why would you be running a uugetty to a getty? The uugetty can call the getty, but the getty can't call the uugetty, so the uugetty is only an outgoing line (sorry for all those gettys :-). So why run the uugetty at all? -- Michael Sullivan {uunet | attmail}!vsi!sullivan {ucbvax | ihnp4 | sun}!amdcad!uport!vsi!sullivan