dave@arnold.UUCP (Dave Arnold) (09/03/87)
I'm having serious problems with my UNIXpc communications, and was wondering if anybody can offer some advise. First, a little background: I was told by an AT&T hotliner that I should have my data line on /dev/ph1, so my L.sys file has all uucp connections made on ph1, and all my office administration has ph1 set as a data line, ph0 nothing (I only have one line connected to the back of UNIXpc, and it's on line 2). And here's a summary of my crash problem, HELP!: 1. "uucico -r1" executed by cron (smgr) @ 2:00am 2. Woke up @ 7:00am, went to check mail and news, uucico still running (apparently hung), /usr/spool/uucp/LOGFILE says OK (conversation complete), setgetty also running (apparently hung). NOTE: I did "ls -l /dev/ph1" and the group was set to "mail", /dev/ph0 is set to "sys". 3. kill -9 uucico, and kill -9 setgetty 4. uuxqt never ran, so I did a uuxqt manually to process uucp work. 5. setgetty ph1 1. Went to work, tried to call 7300 from work, and test my uemacs tam library changes, but 7300 would not answer phone. 6. Arrived home, tried to call some systems using "cu -s1200 -l/dev/ph1", But it reported "Error opening modem port, bad address, error = 14". 7. Thought I might be having hardware problems, so I loaded in my diagnostic disk, and ran a full test of modem and dialer. All passed wonderfully. 8. Hmmm (Actually I wasn't thinking that, I can't use that language here!). 9. I tried going into office and tried resetting stuff in the office, and got errors about trying to open ph0? 10. I went into "Telephone setup", and changed ph0 to the data line, and turned off ph1. I then switched the line on the back from line 2 to line 1. BTW, office automatically changed all my L.sys entries from ph1 to ph0. 11. tried "cu -s1200 -l/dev/ph0". It works. 12. uucp works also. 13. setgetty ph0 1. 14. Called home from work again, it works! But /dev/ph1 is broken!!!! What's going on here? Hardware? Software? Is the driver for ph1 corrupted? I check /usr/adm/unix.log and it doesn't show any problems with any drivers. If the driver is corrupted, how do I fix it? Help! If I lose ph0, I will lose all contact with the world, and won't be able to bug you folks on the net anymore! -- Name: Dave Arnold USmail: 26561 Fresno, Mission Viejo, Ca, 92691 USA DDD: Home: +1 714 586 5894, Work: +1 714 999 3482 UUCP: ...!seismo!uunet!arnold!dave
andrew@teletron.UUCP (Andrew Scott) (09/04/87)
In article <150@arnold.UUCP>, dave@arnold.UUCP (Dave Arnold) writes:
I'm having serious problems with my UNIXpc communications, and was wondering
if anybody can offer some advise. First, a little background:
...
And here's a summary of my crash problem, HELP!:
1. "uucico -r1" executed by cron (smgr) @ 2:00am
2. Woke up @ 7:00am, went to check mail and news,
uucico still running (apparently hung), /usr/spool/uucp/LOGFILE
says OK (conversation complete), setgetty also running (apparently
hung).
NOTE: I did "ls -l /dev/ph1" and the group was set to "mail", /dev/ph0
is set to "sys".
3. kill -9 uucico, and kill -9 setgetty
...
I have had this problem *many* times. uucico calls setgetty to restore the
phone line as a dial in line. The problem lies in setgetty. I believe it
hangs when the line is left off hook for some reason. Mind you, i'm just
guessing. I've brought this problem up with our sales rep. several times,
and I've also spoken to people on the hot line. I did receive a fix disk
with a new uucico, but the problem still persists. I get out of the jam by
doing a system shutdown and re-booting. This is not great, but for a lack of
any other info., it's all I can do to restore our uucp link. I too would
appreciate some answers from people in the know...
Thanks,
Andrewdave@arnold.UUCP (Dave Arnold) (09/05/87)
In article <96@teletron.UUCP>, andrew@teletron.UUCP (Andrew Scott) writes: > > I get out of the jam by > doing a system shutdown and re-booting. > I too would appreciate some answers from people in the know... > > Thanks, > Andrew Rebooting did fix it once for me, but not this time. I can no longer use /dev/ph1 at all. It is completely trashed. Any attempt to use it causes a "Bad address" error, like the driver is munged! -- Name: Dave Arnold USmail: 26561 Fresno, Mission Viejo, Ca, 92691 USA DDD: All +1 714 - Home: 586-5894, Work: 999-3482, BBS: 458-6563 UUCP: ...!uunet!arnold!dave
ford@crash.UUCP (09/08/87)
In article <150@arnold.UUCP> dave@arnold.UUCP (Dave Arnold) writes: >I'm having serious problems with my UNIXpc communications, and was wondering >if anybody can offer some advise. First, a little background: > >1. "uucico -r1" executed by cron (smgr) @ 2:00am >2. Woke up @ 7:00am, went to check mail and news, > uucico still running (apparently hung), /usr/spool/uucp/LOGFILE > says OK (conversation complete), setgetty also running (apparently > hung). > NOTE: I did "ls -l /dev/ph1" and the group was set to "mail", /dev/ph0 > is set to "sys". >3. kill -9 uucico, and kill -9 setgetty > [...] > /dev/ph1 is broken!!!! What's going on here? Hardware? Software? First observation: kill -9 is almost ALWAYS a stupid thing to do. Unix provides a signal, SIGTERM, that tells a process to "clean up" and exit. This signal is even the DEFAULT for the kill command. If you use signal 9 (SIGKILL), you are terminating a process without letting it fix any stuff it may have in a half-finished state. When you kill uucico with signal 9, you cause it to terminate leaving several files in a bad state. First, the lock file for ph1 still says that ph1 is in use, so no other uucp-style programs (like cu) will use it. Second, the system-status file for the system being talked to still says that a conversation is in progress, so no uucico processes are allowed to talk to that system. I recommend using 'kill <pid>' with no -9 except in unusual situations (like if a process refuses to exit after a SIGTERM). Second observation: You mentioned errors that sound like the actual /dev/ph1 could not be opened. If this is true, there is something worse going on than what I mentioned above. If it is just the lockfile problem, you will get messages like "Cannot dial out -- line in use". Have you A) Tried openning /dev/ph1 with a simple program? (even /bin/cat?) What errors are returned by open(2)? B) Rebooted? C) Powered off and on? If you remove /usr/spool/uucp/LCK..ph1 and /usr/spool/uucp/STST.* and you have ph1 properly configured in your /usr/lib/uucp/L-devices (The administration menu does go in and mung this file) then "cu -lph1 <phonenumber>" should work. -- Michael "Ford" Ditto -=] Ford [=- P.O. Box 1721 ford@crash.CTS.COM Bonita, CA 92002 ford%oz@prep.mit.ai.edu