On Tue, 4 Mar 1997, Alan L. Wendt wrote:
> Can't one address the race conditions by waiting five minutes
> and checking to see if the one user is still logged into the
> same two ports? Or better yet, wait five minutes and then
> telnet into the PM's and show sessions; that will eliminate
> the lost-logoff problem. Since duplicates will be rare,
> the extra overhead of the telnet should be small (once per hour or so).
I have found that sometimes the PM2 thinks a user is still connected
but that is not so. Before kicking someone off I always ping the
connection to check that it is up.
That is why bpmtools only sends warning mail to duplicate users.
BTW, bpmtools uses the "secret" portmaster protocol to check who is
on. It seems much faster, and more reliable than running a psuedo
telnet session.
RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 2680 Bayshore Pkwy, #202 Mountain View, CA 94043-1009
gem@rellim.com Tel:+1(415)964-1186 Fax:+1(415)964-1176