10-16-2010 02:12 PM
Hi,
We're getting communication errors while connecting to our FP-2010 module from our Windows XP workstation intermittenly. Our C++ application cannot connect to it via ethernet. At the time of disconnection, we cannot ping to our FP-2010 module.
- The IP address of our module is unique in our network.
- Operating temperature of the module is always under control.
- The only way to get the module pingable is hard reset.
- We have searched the www.ni.com for possible answers but nothing was helpful.
We're thinking a firmware update to our FP-2010 ethernet module. Is it possible and how can we obtain it?
PS: Do we need to specify serial number of our FP-2010 module?
Best regards,
Arda Ulas UNLU
Integration Engineer
HAVELSAN A.S.
+90 312 292 74 67
+90 532 684 61 32
10-17-2010 06:19 AM
Hi Arda. Are you running LabVIEW RT on the -2010? If you are, is your program on the module still running?
Is there other network traffic that the module is exposed to? We had one site where they were using a network
monitoring tool called 'What's Up' where the monitoring packet itself would cause the TCP stack in the module
to lock up as you have described. If it is on a shared network, can you test on a direct, dedicated link?
Is this a new problem, i.e. it worked fine for years and now doesn't?
How frequently does the crash happen? Any particular time of day?
We have also seen a few instances, maybe 6-10 in several hundred controller-years, where the communication would
just lock up as you have described without ever being able to determine why or have it repeat.
Matt
10-17-2010 10:57 PM
Hi,
Yes, LabView RT. I mean, after the disconnection starts, program is not running. Because we can easily understand it from the others ubsystem.
We're also using What's Up. It sees the module on its list. And we've disabled the SNMP component of it. SO we think it's not running any unnecessary traffic. But we will check again.
We will also double-check IP addresses. Very small possibility that we might miss out something.
No chance to test on dedicated link.
This problem has been observed for 2 years, on both two same environments. Happens once in a 3 or 4 weeks.
Will observe again...
10-18-2010 08:37 AM - edited 10-18-2010 08:38 AM
Hi Arda. Take a look here, especially the first message:
Crashing-cFP-with-http-port-scan
This is from when we were troubleshooting the Whats's Up crashes.
Just a plain old HTTP request was enough to hang the cFP module,
I expect the FP would be similar.
Are you able to completely remove the FP module from the What's Up
routine for a while?
Matt