search

ni / nifpga-python

Python API for LabVIEW FPGA devices
MIT License
74 stars 18 forks source link

Occasionally Error: RpcServerMissing (-63044) when calling 'NiFpgaDll_Open' #35

Closed BrankoIlic closed 3 years ago

BrankoIlic commented 4 years ago

I'm running unit tests, and after about fourty reads from cRio (by fourty Sessions) I start to get this error:

File "C:\Program Files (x86)\Jenkins\workspace\E-TEST\Test\helpers\motor_monitor_helper.py", line 30, in get_current_position with Session(self.file, self.resource) as session: File "C:\Python27\lib\site-packages\nifpga\session.py", line 109, in init self._session) File "C:\Python27\lib\site-packages\nifpga\status.py", line 100, in internal _raise_or_warn_if_nonzero_status(status, function_name, argument_names, args) File "C:\Python27\lib\site-packages\nifpga\status.py", line 66, in _raise_or_warn_if_nonzero_status raise codes_to_exception_classes[status](function_name, argument_names, *args) RpcServerMissingError: Error: RpcServerMissing (-63044) when calling 'NiFpgaDll_Open' with arguments: bitfile path: 'b'C:\Program Files (x86)\Jenkins\workspace\E-TEST\Test\NiFpga_eTest.lvbitx'' signature: 'b'1EEE087A3560B2C0A53EF8014DCE226E'' resource: 'b'rio://10.0.4.89/RIO0'' attribute: 0x0 session: 0x0L

That line 30 of motor_monitor_helper.py is:

29    def get_current_position(self):
30        with Session(self.file, self.resource) as session:
31            Settings.get_instance().load()
32            positionCounts = session.registers['Position (Counts)'].read()
33            print("session.registers['Position (Counts)']: {}".format(positionCounts))
34            return float(positionCounts) / float(Settings.get_instance().encoder_resolution)

And after that I have to reset cRio to make it works again. So, all that 40 reads go fine, and that it stops. I have changed tests that are running, but it depends only of number of tests, not of tests itself.

My python version is 2.7.16 nifpga is Version: 19.0.0

strainmike commented 4 years ago

What cRIO model are you using? Also what version of NI-RIO is on the cRIO and on the Windows machine (you can view this in MAX under software)?

Do you have 40 sessions open at once or are you opening and closing a bunch of sessions?

BrankoIlic commented 4 years ago

Big thanks for quick response :)

I have fixed this by extended time period between every 4th Session open, from about 5s to about 10s. That was easiest to change. But I do not understand why this change fix works...

cRIO is 9025 NI-RIO on the cRIO is 19.1, but there is also line "NI CompactRIO 19.5 - September 2019" NI-RIO on the Windows machine is 19.5.0 Sessions are open and closed sequentially with time distance, never more at once.

strainmike commented 4 years ago

You mention that you need to reset the cRIO to make it work again. Could you try leaving it idle for a few minutes after it gets into that state and seeing if it recovers? You might also check serial out when the problem occurs.

This sounds a lot like socket exhaustion which our older vxworks targets were prone to running into at times. If that ends up being the case, then there might be a workaround.

BrankoIlic commented 4 years ago

As I said, it works fine now. And to answer your question: I can not change it now, but as I recall it has been sometimes idle for few minutes but it would not recover. Thanks