|
Idleout in SQUEEZE
Good morning fellow,
I am in process of migrating to Unix, SCO Debian Squeeze ( only console, NO GRAPHIC gdm ) and I can not solve the problem of expired connections. In SCO idelout use the instruction set in / etc / default / idleout idle values = seconds. In Linux I do not see where to set this situción. TMOUT there is, but it is not valid, TMOUT only "kill" inactive sessions are in "prompt" or bash shell. I have in production an "ERP" users to "login" with the same name from "putty" ssh, but very often closed "putty" without logging off, so it is eternally busy / dev / PTSX. ****************************************** Also see the configuration of / etc / ssh / sshd_config adding: keepalive yes ClientAliveInterval 3600 (1 hour) but has no effect NONE. ********************************************** Also i was looking at the packages autolog "aptitude install autolog", last updated in 2002, but performed the searches of inactive username and serves me, all my users enter the same login to "ERP" ********************************************** Also see the package "idleout" "http://sourceforge.net/projects/idleout/" contains the sources but I can not understand, is imperative to have installed "finger" for installation and the "daemon" in / etc/rc2.d must be performed manually. its performance is quite good, but inexplicably the 'daemon' stops working after 2 hours of boot from / etc / init.d / idleout start. I will appreciate if anyone knows any package or script that succeed logout users of a specified period of idle time. I find it really surprising that Linux does not incorporate the function "idleout" as the basis of its installation. 1salu2 and thanks in advance for your attention. |
IDLEOUT , TMOUT --> Logout users of a specified period of idle time.
Good morning fellow,
I am in process of migrating to Unix SCO --> Debian Squeeze (single console, not graphics) and I can not solve the problem of expired connections. In SCO idelout use the instruction set in /etc/default/idleout idle values = seconds. In Linux I do not see where to set this situción. TMOUT there is, but it is not valid, TMOUT only "kill" inactive sessions are in "prompt" or bash shell. I have in production an "ERP" users to "login" with the same name/login from "putty" ssh, but very often closed "putty" without logging off, so it is eternally busy /dev/ptsX. ****************************************** Also see the configuration of /etc/ssh/sshd_config adding: keepalive yes ClientAliveInterval 3600 (1 hour) but has no effect NONE. ********************************************** Also i was looking at the packages autolog "aptitude install autolog", last updated in 2002, but performed the searches of inactive username and serves me, all my users enter the same login to "ERP" ********************************************** Also see the package "idleout" contains the sources but I can not understand. ** is imperative to have installed "finger" for installation and the "daemon" in /etc/rc2.d must be performed manually. its performance is quite good, but inexplicably the 'daemon' stops working after 2 hours (aprox) of boot from /etc/init.d/idleout start. I will appreciate if anyone knows any package or script to perform properly the "kill" process "idle". I find it really surprising that Linux does not incorporate the function "idleout" as the basis of its installation. 1salu2 and thanks in advance for your attention. |
|
Thanks for your interest and references "bigrigdriver"
but unfortunately that TMOUT comment and is not valid for my intention, TMOUT only works if the user was prompt, but if you are working within a process, cobol, java, dialog, Oracle, Informix, etc ... and closes his notebook without 'logout', does not perform TMOUT expulsion "kill PID" being active user / dev / pts |
I merged your two *cough* closely related threads, and politely ask
that you don't double-post in the future. Thanks! Cheers, Tink |
Quote:
I may do a double click by mistake and created two thread |
I think that the easiest solution is to modify the program that you are using throught putty to actually have a timeout (so it will return to the prompt and get flushed after some time). I think also that this has to be the desidered behavior, let me explain:
Il something kills the program after a timeout then there is the risk of data loss, if the program closes itself then there is no more risk. Now if this is feasible or not that depends on the program that you are using. |
Thanks "Celyr" for your opinion,
to people who connect to the server SQUEEZE, I try to continually educate them to close properly but the connection is not always possible. Took weeks looking for a package "idleout" similar to that in SCO-Unix but can not find anything really useful and I think if Debian is really intended for data server for "ERP" with Informix Software, Oracle, Cobol, Sculptor, etc .... or is only for web Serbs and / or starting to think to create competition with "windows" and forgetting their desks console functions only, text only. I'm not a espert in-shell script, but I started to gather information and I think the situation can be resolved: 1) Initially locating PID exceeding 60 minutes inactive. Code:
who-Tu | awk '$ 6> "01:00" {print $ 1, $ 3, $ 7}'Code:
/ bin / kill -13-Tu `who | awk '$ 6>" 01:00 "{print $ 7}'`For now the point (2) does not work :cry: If someone wants to help me do it I'll be happy. |
Can you please use code-tags rather than quotes?
And what is kill -13-Tu supposed to do? Also I don't understand how you (or SCO's idleout, for that matter) can possibly tell the difference between a session that's being actively used and one that's just sitting there if it's NOT bash. Killing everything arbitrarily after one hour seems like a recipe for disaster to; but of course just knowing that you're using some sort of ERP doesn't tell us much about use-cases, or usage patterns. Cheers, Tink |
| All times are GMT -5. The time now is 03:08 AM. |