My original post was:
> Has anyone had trouble with user idle time being set back
> to zero when a user's terminal RECEIVES output from commands
> like wall, talk, new mail notification, etc?
>
> I've got DU v3.2d-1 running on an AlphaServer 2100 model 4/200.
> All our users us telnet connections. But anyone logged in at
> the console (DECwindows) has the same problem. Any terminal
> output to a user sets that user's idle time back to zero, just
> as if the user had entered keyboard input. I've tested several
> idle terminal watchers, and of course none of them ever logs
> out idle users because they first send a warning message to the
> user's terminal, which resets the idle time.
>
> Digital technical support is currently stumped but still
> investigating the problem. Thank you in advance for any
> information that sheds light on this problem.
Apparently many of you running Digital UNIX already know about
this problem and just live with it. I should have mentioned that
the idle time I was referring to is that which is displayed by
the finger and w commands. I believe these commands actually
read the idle time from the /var/adm/utmp file.
The Digital support person I was dealing with said that she
has submitted this to engineering (as a result of investigating
the problem for me). However, I find it difficult to believe
this is the first time Digital has heard of the bug. Mike
Crider specifically mentions the bug on his idled home page:
www.cs.hope.edu/~crider/idled
My thanks to the following for good feedback and information:
Jon Trulson trulsonj_at_mscd.edu
jo fxjwk_at_aurora.alaska.edu
Randy Hayman haymanr_at_icefog.alaska.edu
Also Randy Styka says he has a product (Logmon) for sale which
would avoid the bug altogether by using cpu time only as a
determination of user idle time. Details at info_at_computron.com.
Paul Youngblood
University of Alabama in Huntsville
youngbloodp_at_email.uah.edu
Received on Fri Apr 26 1996 - 20:43:51 NZST