I would like to know whether under battle conditions (lots of users trying
to log in simultaneously at the start of a lesson) I can rely on a client's
preexec script completing its output operations before the client gets its
login batch file.
I am a school network mamager with about 1000 teenage boy users of a new (to
all of us) Linux-Samba server.
(SuSE6.4 and Samba 2.0.10)
I have pre- and post-exec working to record logon & logoff for mischief
tracking, custom stats etc.
I have logon scripts driven by filespace usage stats, time-of-day
restrictions and 'rogues list' entries. These advise the user of the
issues
affecting their ability to log on, and enforce a client switch-off if
necessary.
I need to ensure that a user can only have one active login - otherwise boys
denied access will pressurise their peers to log on for them.
I am using a grep-tail-grep of the historical logon-off record to calculate
an appropriate echo to the user status file that determines the logon
action. The scripts work one logon at a time but I'm not sure I can test
it properly in service - can I trust it?.
Thanks for any help
Chris Cowsley