Inconsistent MultiRead errors about pylogix HOT 8 CLOSED

Can you email me the exact script you are using and possibly a wireshark of the issue?
Can you give me some context? Like number of tags you are reading, data types, etc. Does this happen every read, or every 10th read or every 100th read, etc. Maybe mention the firmware of your L75 and the model and firmware of the Ethernet module in your CLX chassis.

I don't have a 1768 series but I do have a L73 at work, I'll have access to it Monday.

edit: also, if you reduce the number of tags you are attempting to read each time, does it make a difference?

from pylogix.

Sure, what address should I send it to? Thanks! I'll check on the reduced tags.

from pylogix.

[email protected]. Thanks

from pylogix.

I've sent you a .pcap file. I have not yet had a chance to try again with fewer tags, but I'll let you know when I do. The pcap file is from running the example I posted above in the interactive console. It happens every MultiRead. I have a workaround in my program running right now which is that for those two PLCs, I loop through the tag list with single Reads and it's working fine for now. The data types are ints, floats, and a bool. Thanks for your help!

from pylogix.

Looks like your analysis from our email conversation was correct. So far, I've split all my MultiReads into two sets and the errors haven't recurred.
For others who might be reading, there is a 500 byte limit on tag names in a request, so if you have too many long tag names in one MultiRead, it'll error out. Thanks!

from pylogix.

IIRC, if you either use unconnected messaging or your PLC only supports ForwardOpen, but not ForwarOpenEx, you only get a payload size of ~500 bytes. It might be a few more, but you can only express 511 in the ForwardOpen request.

from pylogix.

Thanks for the update mainsteam.

Just to add a little info on why this appeared to only be an issue on CLX and the L43. Although the UDT was the same structure in all the PLC's, the base tag name was longer in the CLX and L43, which made the packet longer. I'll look at a way to handle this more gracefully.

Kyle, thanks for the input. I'll do some experimenting with unconnected sends.

from pylogix.

I have started experimenting with ForwardOpenEx. At least in some cases you can use packets of about 4000 bytes. I have not tried to see if I can pack a multi-tag request into that.

from pylogix.