rjd@attibr.UUCP (Robert J. Diamond x7650) (05/06/89)
We have a problem. We need to time out inactive users on our 3b2's, running SVR3.0, SVR3.1, SVR3.1.1, and SVR3.2.1 However, we cannot use the usual mean of checking tty idle time because there are people that sit in applications that do things like update an on-screen clock every minute. I need something that in essence will check for keystrokes as the basis for the timeout. I thought of using a streams driver, but I don't want to re-invent the wheel. Any ideas/solutions? -- Robert Jon Diamond | attibr!rjd AT&T | rjd@attibr.att.com International Operations Support | attmail: rdiamond 1300 Mt. Kemble Avenue, Basking Ridge, NJ | (201) 953-7650
gwyn@smoke.BRL.MIL (Doug Gwyn) (05/08/89)
In article <170@attibr.UUCP> rjd@attibr.UUCP (Robert J. Diamond x7650) writes: >We have a problem. We need to time out inactive users on our 3b2's, >... Any ideas/solutions? I suggest you reconsider. Why is it so important that users press a key every so often? Surely there must be applications where that would be a silly requirement. What is your real requirement? A "need to time out users" is not an end in itself.
ccs013@castor.ucdavis.edu (Jason) (05/08/89)
In article <10226@smoke.BRL.MIL> gwyn@brl.arpa (Doug Gwyn) writes: >In article <170@attibr.UUCP> rjd@attibr.UUCP (Robert J. Diamond x7650) writes: >>We have a problem. We need to time out inactive users on our 3b2's, >>... Any ideas/solutions? > >I suggest you reconsider. Surely there must be applications where that >would be a silly requirement. Exactly ... perhaps a combination of idle and cpu usage to check for an anppropriate "time-out" might be needed. ___ ___ __ ___ )___ _____________________________ ( | '__| (__ / / / / | II Corinthians 10:17 | \ | (__)\ __) /__/ / / +---------------------------+ \_| Let the engineers apply; I *Internet: jygabler@ucdavis * | simply want to know why! *BITNET: jygabler@ucdavis * | PURE SCIENCE ------------> *UUCP: ucdavis!jygabler *
rjd@attibr.UUCP (Robert J. Diamond x7650) (05/08/89)
In article <10226@smoke.BRL.MIL>, gwyn@smoke.BRL.MIL (Doug Gwyn) writes: | In article <170@attibr.UUCP> rjd@attibr.UUCP (Robert J. Diamond x7650) writes: | >We have a problem. We need to time out inactive users on our 3b2's, | >... Any ideas/solutions? | | I suggest you reconsider. Why is it so important that users press | a key every so often? Surely there must be applications where that | would be a silly requirement. | | What is your real requirement? A "need to time out users" is not | an end in itself. There IS a need for us to time out users. When you have machines with 16 ports each, and times when 20 users want to get on, and half of the users that ARE logged in are at lunch or a meeting or just ignoring their terminals, it would be nice to get them off. -- Robert Jon Diamond | attibr!rjd AT&T | rjd@attibr.att.com International Operations Support | attmail: rdiamond 1300 Mt. Kemble Avenue, Basking Ridge, NJ | (201) 953-7650
gwyn@smoke.BRL.MIL (Doug Gwyn) (05/08/89)
In article <4241@ucdavis.ucdavis.edu> ccs013@castor.ucdavis.edu (Jason) writes: >Exactly ... perhaps a combination of idle and cpu usage to check for an >anppropriate "time-out" might be needed. Virtually every day I have connections that must remain open but on which no activity, CPU or I/O, occurs for perhaps several hours.
gwyn@smoke.BRL.MIL (Doug Gwyn) (05/08/89)
In article <173@attibr.UUCP> rjd@attibr.UUCP (Robert J. Diamond x7650) writes: -In article <10226@smoke.BRL.MIL>, gwyn@smoke.BRL.MIL (Doug Gwyn) writes: -| What is your real requirement? A "need to time out users" is not -| an end in itself. -There IS a need for us to time out users. When you have machines with -16 ports each, and times when 20 users want to get on, and ... See, you could install 4 more ports (for example). With a bit of creativity I'm sure you can find other solutions.
ado@elsie.UUCP (Arthur David Olson) (05/09/89)
> Virtually every day I have connections that must remain open but on which > no activity, CPU or I/O, occurs for perhaps several hours. I (and any other readers as unimaginative as myself) would appreciate details regarding such situations. -- Space: Canada, 0 tries ever. Arthur David Olson ado@ncifcrf.gov ADO is a trademark of Ampex.
gwyn@smoke.BRL.MIL (Doug Gwyn) (05/09/89)
In article <9103@elsie.UUCP> ado@elsie.UUCP (Arthur David Olson) writes: >> Virtually every day I have connections that must remain open but on which >> no activity, CPU or I/O, occurs for perhaps several hours. >I (and any other readers as unimaginative as myself) would appreciate >details regarding such situations. It's easy to come up with examples. Here is a simple one: a layer attached to my printer process, to which I direct output when I want it printed on the printer attached to my terminal
les@chinet.chi.il.us (Leslie Mikesell) (05/11/89)
In article <10226@smoke.BRL.MIL> gwyn@brl.arpa (Doug Gwyn) writes: >In article <170@attibr.UUCP> rjd@attibr.UUCP (Robert J. Diamond x7650) writes: >>We have a problem. We need to time out inactive users on our 3b2's, >>... Any ideas/solutions? > >I suggest you reconsider. Why is it so important that users press >a key every so often? Surely there must be applications where that >would be a silly requirement. > >What is your real requirement? A "need to time out users" is not >an end in itself. How about this situation: The ports are dial-up, and mostly inbound wats lines, the users are calling in to pick up frequently updated files. There are >1000 login id's and about 8 inbound lines. My solution was to use a login shell that just accepts their requests and hands out the files unless they ask for a menu of other choices. A separate process is forked off which periodically stat's the tty line checking for a maximum idle time and total on-line time mentioned in the comment field of their passwd file entry. They mostly don't have permission to write files or a home directory, either. If either time expires, the other process forces the line to drop. Les Mikesell
mcdonald@uxe.cso.uiuc.edu (05/15/89)
>> Virtually every day I have connections that must remain open but on which >> no activity, CPU or I/O, occurs for perhaps several hours. >I (and any other readers as unimaginative as myself) would appreciate >details regarding such situations. Certainly: Each such port is attached to a sensor which detects missles being fired at you. When such a message is received, the computer instructs a star wars laser to fire at the missle. Other scenarios perhaps exist, such as automated teller machines.