I am intermittently getting 'Warning: Connection to target (NI-cRIO9024) has been lost.' error. The cRIO is running long term tests, and can crash with this error every day at the moment, and then successfully run for a few days, it isn't consistent.
-There are no timed loops I thought it was a 'greedy' loop, but when the error occurs, I can restart it remotely on NI Max, indicating the cRIO isn't wound up in a loop unable to communicate. -I have stripped back the tests, and only a very basic setup is currently running, with little complexity. -I use a shared variable library, sharing Network published shared variables with the host PC for displaying on the GUI. Only some essential NPSV's are buffered. All are hosted on the RT target, not host PC. -The host PC and cRIO are connected on a LAN, via 1 x ethernet switch within the building. -I analysed the network with wireshark in early 2020 and found no network issues, despite the problem occurring then. -For logging to a TDMS file, I set the NIMinBufferSize to 10,000, and clear the buffer every 15 mins. I suspected a buffering problem, perhaps with too much data in the RAM. Error still occurring. -The CPU usage can be seen to increase sometimes before the connection is lost. (see log of CPU usage, 3rd may ~10am). Other high CPU usages in that log were at program startups. The high CPU usage pre-crash does not consistently occur. There have been instances with no CPU increase. -If the GUI is running on the host PC when the cRIO connection is lost, the error Error -1950679023 is seen, as it can not access the NPSVs.
If you need any additional info about the setup please ask.