On Fri, 6 Jul 2018 13:47:56 -0500
Tim Camp <[email protected]> wrote:
> Well we had previously used a cifs connection to the windows box, but the
> mount point being in the user directory would cause the file manager
> sometimes 15 minutes to populate waiting on the windows box. This made it
> rdlogmanger lockup when trying to create a log as it was looking for that
> mount point
> for traffic import. we thought going the other way would stop this problem
> (which it did) but now we have this problem.
>
> As I said it only happens during the day when that mount point is being
> used by the traffic system.
>
> rivendell only needs that share for the short time it takes to import
> traffic, but of course the traffic system needs to see it all the time.
>
> Wondering if I am going to have to go back to the old way of just sending
> the log over via ftp, which was the method years ago.
>
> maybe I can write a short batch file for the traffic girl to run after she
> finishes the log.
Gave it some thought.
If I understand correctly, the Rivendell problem is that the *network* is
saturated, thus preventing the real time transfer of priority stuff, like
audio.
Everything else works.
OK, here's two solutions...
A gigabit card in the Riv. machine, AND a 100BaseTx card in the M$ machine.
Hardware will limit M$ to 10% max of what's available to Riv.
A second card in the Riv. machine, and a super private point-to-point LAN
between the Riv. machine and M$ via a cross-over cable.
Network cards are cheap enough these days, and if you have one in the junk box,
then the only cost is a little time. Probably less time than most any other
solution we could invent.
--
Cowboy
He was a fiddler, and consequently a rogue.
-- Jonathan Swift
_______________________________________________
Rivendell-dev mailing list
[email protected]
http://caspian.paravelsystems.com/mailman/listinfo/rivendell-dev
