xev@hstbme.mit.edu (Xev Gittler) (02/18/88)
When a person dials in over a modem line, and then hangs up without logging off, the computer should know enough to kill those processes, right? (It should get a hangup signal?). Could someone suggest why on my 4.3 system, peoples jobs would be just staying around, and the next person that dials in to the same modem will come in in the middle of a session? Xev Gittler xev@hstbme.mit.edu, or xev@athena.mit.edu
mikel@codas.att.com (Mikel Manitius) (02/18/88)
In article <2995@bloom-beacon.MIT.EDU>, xev@hstbme.mit.edu (Xev Gittler) writes: > [ ... ] Could someone suggest why on > my 4.3 system, peoples jobs would be just staying around, and the next > person that dials in to the same modem will come in in the middle of a > session? This could be caused by a number of reasons, the most common that come to mind are: a. You've got very old hardware that doesn't understand hangups (ie: VAX dh boards, I beleive). b. For some reason, the processes are ingoring SIGHUP. c. A front end of some sort (ie: a port selector) that isn't smart enough to tell the host the hangup occured. d. Badly configured, or cheap modems that don't drop DTR when they lose carrier. Also note that "init" won't clean up after the login, until the child that it spwaned for that tty (presumably the login shell) exists. So if it ignores hangups, he user won't logout. -- Mikel Manitius mikel@codas.att.com
bzs@bu-cs.BU.EDU (Barry Shein) (02/18/88)
Posting-Front-End: GNU Emacs 18.41.4 of Mon Mar 23 1987 on bucsd (berkeley-unix) >When a person dials in over a modem line, and then hangs up without >logging off, the computer should know enough to kill those processes, >right? (It should get a hangup signal?). Could someone suggest why on >my 4.3 system, peoples jobs would be just staying around, and the next >person that dials in to the same modem will come in in the middle of a >session? > > Xev Gittler (have you ever tried anonymous login to hstbme? it's, um, interesting, in a minimalist kind of way.) One of two things, probably the former. Either you don't have the bits CLEAR in the tty mux line in the config file (flags 0x00, see the man page) or you're not wired up correctly and the mux never sees/hears/feels the hangup, or both. -Barry Shein, Boston University
bzs@bu-cs.BU.EDU (Barry Shein) (02/19/88)
Posting-Front-End: GNU Emacs 18.41.4 of Mon Mar 23 1987 on bucsd (berkeley-unix) > a. You've got very old hardware that doesn't understand > hangups (ie: VAX dh boards, I beleive). Actually the dh supported an astounding amount of modem control not to mention other configurations (20mA and some sort of "teletype" interface etc.) Unfortunately DEC never supported it on the VAX, mostly they were used on "real" systems on DEC10's and 20's (BSD Unix supported dh's on Vaxen, you could plug dh's in and they worked much better than supported options due to modem control and real on-board DMA buffering tho you usually got flak from field circus if they broke, "we can't fix those, we only took your money cause we needed it".) The mux's they *did* support on the Vax were pretty weak on modem control (DZ's were acceptable but had no DMA at all, DMF's came later with DMA but 6 out of the 8 ports lacking any modem control, anything after that wouldn't qualify as "older".) I don't know that anyone ever ran DL's on a Vax, only the "E" had modem control (at the cost of using an entire board for one serial line) as I remember, but they were PDP-11 interfaces (I guess they could be run on a Vax Unibus, maybe not enough address bits in the control registers to make them practical.) Anyhow, the answer to the question was the flag bits in his config file (we spoke off-line), tho the other suggestions are always worth checking. -Barry Shein, Boston University