Hey jmcbee,
Is there any way any of your shared variables (especially strings, if you have any) are growing beyond the space you have reserved for them?
I had a similar problem -- spurious shared variable errors (number -1950679035). It appeared occassionally when attempting to read a particular shared variable: a simple boolean. What was odd is that there appeared to be no pattern to the errors. My system would run flawlessly for 10-15 minutes with no problems, then it would start. A single shared variable error every 2 or 3 minutes when attempting to read this particular boolean. The value for this particular variable is polled 2-3 times per second, so it was working the vast majority of the time, and fairly only occasionally. My network utilization was somewhere around 2-3%, so network collisions or something related to the network seemed highly unlikely.
The clue came for me while driving home after puzzling on it for a couple of hours. I remembered that I had changed the NUMBER of buffered items for another shared variable (-- a cluster flattened to a string) , but I had neglected to increase the number of character allocated for that particular shared variable. And even though the shared variable that contained the flattened string appeared to be working properly, I wondered if LabVIEW was moving pointers or somehow compensating for the fact that my buffer was no longer large enough to accomodate the larger number of buffered items I had specifed in the Shared Variable set-up dialog.
When I increased the number of characters in the buffer for the flattened string shared variable, lo and behold... the read errors disappeared from the boolean shared variable!
I'm not going to suggest I really understand what is happening under the hood, but based on these observations, it appears that if you don't allocate enough space for shared variables that can change size (specifically strings or buffered shared variables), then error -1950679035 might show up when attempting to read another, totally unrelated shared variable. I suppose the problem could also be related to shared variable arrays -- presuming that LabVIEW must resize the data space as the array size changes. If a resize operation is underway when a read request comes in, and a pointer is temporarily unavailable, it makes sense that an error like this could occur. But again, all of this is based on speculation and observation. What I can say is that care should be taken to make sure you have allocated enough space for buffered shared variables. If you encounter intermittent -1950679035, have a look for potential buffer over-runs and under-allocated buffer spaces.
Hope this is useful.
Regards,
-- D.