We are baffled by a problem that just came up at a customer site. They
are running Universe on Win2003 server with SP1. Have had
DigiPortServer's in place there seems like forever. The digi's where
attached to an NT4 server which we decomissioned and went to new
server running Win2003. Everything was just fine with all the dumb
terminals for the first six or more months after migrating to Win2003
server with Universe 10.something.

Now, 'ports', dumb terminals attached to digi seem to 'hang' at random
- screen freezes - any keystrokes have no effect - Ctrl-Q doesn't free
them up (flow control is xon/xoff). But, you can send a message to any
of these hung serial ports via Universe MESSAGE command and it will
appear on the hung terminal.

Anyone have any ideas about how to troubleshoot this problem?

universe support said it isn't a universe problem

digi support has been not much help. they suggested removing sp1 from
the win2003 server, although they have no other reported problems with
win2003 sp1 and their gear.

Any ideas?

Clueless in Cleveland.

A question from a purely "pick-ish" point of view:

Does Universe have a Pib Status for their 'where' command? What does it
indicate about these "frozen" ports. If a message can get through, there's
no obvious hardware problem. If a port stops, there are a limited number of
reasons: sleeping, waiting, comatized, deactivated, dropped off the pib
links, buffer over-run to name a few.

Mark Brown


"iJah" <iJah (AT) sbcglobal (DOT) net> wrote

Dear Clueless...

I'm grasping at straws from a foggy memory bank as it has been a while since
I had to work with dumb terminals but...

Have the installed any new equipment in the area of the server or the
effected dumb terminals? I'd look for such things as big motors or wire for
large loads.

Check the wiring and ensure that any and all required jumpers between CTS,
RTS, and other signals on the serial lines are setup correctly. You might
need to get a 'break-out-box' to see the serial signals and experiment with
different configurations. If you can get away with just the send, receive,
and signal ground that seems to work the best as I recall. (I seem to
recall on a 25 pin connector pins 6,8, and 20 should be jumpered together,
but check the facts before you start playing.)

If you are running shielded cable ensure that the shield is run to machine
ground only at one end of the cable. I have seen sparks fly 'cause some
dummy manage to wire a plug so that the machine ground was hot on one side.
Having the shield going to machine ground between a badly wire machine and a
correctly wired machine cause a short through the cables shield.

HTH

Dale

And remember for every mile of road there is two miles of ditch!


"iJah" <iJah (AT) sbcglobal (DOT) net> wrote

And since it's Windows 2003. Any chance they are downloading security
patches? And maybe something is now out of sync between drivers and
O/S?


"Dale Benedict" <dale_knowspam_benedict (AT) flightcraft (DOT) ca> wrote in
message news:GcYYe.273423$tt5.253872 (AT) edtnps90 (DOT) ..

We have the identical problem - the Digiboard Portserver II locks up
occasionally - it resets itself after about 20 minutes. We are on D3/
Red hat 7.1 In the past year the lockups are becoming more frequent.

We've been running the Portservers for years without any problems. The
wiring hasn't changed in many years (we used an internal digi card
before the Portserver II) - so it's probably not a cabling issue.

You can telnet to the Digiboard with no problem. There are no errors
in the Realport software on the linux box. We even replaced the
Portserver with a brand new one - no effect.


iJah wrote: