Does your black box report any link statistics? In particular is it logging
any CRC errors?
T1 lines are very sensitive to interference… and there is nothing worse
than trying to diagnose a link that sort of works.
We do a lot of QNX LAN bridging over T1s. Recently we had some problems
with one of our redundant (bridged) LANs that appears to be quite similar to
yours e.g. horrible through put. To make a long story short… after a week
of sniffing the network and scratching our heads, we decided to investigate
for the obvious… and… er… check the T1 wiring. Well, as it turned
out, “someone” had installed some frame-relay equipment for one of our
clients and left the power transformer sitting right on top of the T1 cable
for our LAN bridging (with the power cord snaked around it, just for good
measure). We moved the transformer & power cord to a safer location…
Walla, the T1 and netting returned to normal.
We’ve also experienced other problems where bundles of T1 cables have cross
talked. This usually isn’t a big problem on active lines, but lines that
are actually dead, still have enough signaling on them to make the line
interface hardware think that the line is active.
Lessons learned:
Keep T1 cables away from power sources.
Use shielded (and grounded) cable whenever possible.
Have the telecom guys run the tx & rx pairs on separate blocks when/were
shielded cabling isn’t possible.
Hoping your problems are “obvious” 
Rob
“Mario Charest” <goto@nothingness.com> wrote in message
news:a943om$ijg$1@inn.qnx.com…
“Eric Broussard” <> n5jeq@ev1.net> > wrote in message
news:a93r90$cqg$> 1@inn.qnx.com> …
all of the netmap files are the same and netinfo -l shows no errors just
good clean tx/rx
I tried copying a directory across from one node to the problem node
using
the verbose mode
and it would copy files but I keep getting “no route to host” and it
would
finally die
I keep thinking that the Black Box that I am using for the T1s is not
liking
the QNX network protocol
That’s possible, also the difference in speed (QNX thinks you are on
a 10Mbits network) could cause some timing problem. Maybe
you could try specifing the rate (-r1000000) to the ethernet driver?
Or play with various timeout setting (check doc on Net and Netdriver
and netpoll)
\
Rick Broussard
Email > scada@jala.org
“Mario Charest” <> goto@nothingness.com> > wrote in message
news:a8msq5$1kl$> 1@inn.qnx.com> …
“Eric Broussard” <> n5jeq@ev1.net> > wrote in message
news:a8mpsi$st6$> 1@inn.qnx.com> …
nameloc was running but it is now removed
by slow I mean when we first connected it and ran sin net, there was
no
delay
when running the control system software we could call up screens
without
delay
now it takes approx 10 seconds to get past node 7 when running sin
net
when running the control system software it can take up to 30 - 45
seconds
to call up a screen
the screens that are not on the remote node starts taking longer to
call
up
not as long as they do on node 7 but due to the fact it is a control
system
we will generally take down node 7 before it slows the system down
to
much.
Are netmap all the same on every machines. Does netinfo -l report
abnormal number of errors?
\
Rick Broussard
Email > scada@jala.org
“Mario Charest” <> goto@nothingness.com> > wrote in message
news:a8erkd$79v$> 1@inn.qnx.com> …
“Eric Broussard” <> n5jeq@ev1.net> > wrote in message
news:a8eodj$597$> 1@inn.qnx.com> …
I have several nodes on a local network and am trying to add a
new
node
via
a T1 connection. It was running great for about 2 weeks then
started
to
slow
down the entire system. Any suggestions?
Make sure nameloc is not running on node 7.
Can you elaborate on “slow down the entire system”?
The setup is that nodes 1,2,3,4,5,8,9 are on Logical network 1
nodes 1,6,7 are on Logical network 2
node 7 is being connected via 2 Black Box routers on a T1
connection
node one has dual 3com 10/100 Ethernet cards
both networks go through there own 10/100 switch
If any more information is needed let me know.
Rick Broussard
Email > scada@jala.org
\