Re: keyspace timeout?

Scott Fendley (scott@nemesis.uark.edu)
Tue, 20 May 1997 00:54:41 -0500 (CDT)


On Mon, 19 May 1997, Andrew James Alan Welty wrote:

> The timeout is a minimum of two hours. Depending on when it is the
> timeout can be 2 hours to 26 hours. Though this may have changed by now.
> What happens is the server can reissue the key after two hours, but only
> does so at a certain time. Rocke was thinking of changing this from once
> every 24 hours to once every six hours though.

I don't see a problem at this point in time with the timeout rate... I
know I have around 20 Mac IIs(and 386/486s) that are putting in about 5
blocks of 2^22 pairs at night...they are taking several hours to complete
each time. If Rocke manages to get it to loop until it can contact the
server to tell it that the key was or wasn't found, then at times it might
be wise to keep it somewhere between 6-10 hours...anything less will throw
out some loaded down fast machines and some slower machines also. Maybe
24 hours is appropriate for bad network problems....but 2 hours is a bad
idea. As a student I would still like to help my school make a good
showing in the statistics anyway that we can. My point: there isn't a
problem at this moment in time about the timeout...but as we get closer
to the end of it all...then it might.

Scott Fendley
Department of Computer Science
Unix Systems Administrator