Additional debugging when adding monte carlo nodes

[apshoemaker]

I am in the midst of setting up monte carlo simulations within a containerized kubernetes environment and had to work through a few different network-based gotchas. It became helpful for me to fork the repo and add this simple additional debug value to the socket initialization error in order to track down a connection issue (which helped quite a bit!). I thought it might be advantageous to add something like this to the upstream code base, but most likely perhaps a much more though-through design (perhaps with some sort of explanation of error given the return code, etc).

I mostly wanted to bring this to the attention of the community as a discussion point. If there is interest and thought around what we would like to see here, I might be able to help contribute.

[sharmeye]

Thanks for the input. We'll run this through our testing, but the change is quite reasonable. If you'd like to discuss more robust and thorough error reporting for Monte Carlo that you think might benefit users who are using a similar setup to yours, you can open some issues or contact us directly to discuss.

[coveralls]

coverage: 62.079% (-0.01%) from 62.091% when pulling d43827697b2f82bea94e9b91153e0ba96905b3db on apshoemaker:master into 682f2183919bdf700106c6991d2a344db834f39a on nasa:master.

[apshoemaker]

I am going to close this pull request and create a branched PR so that I can keep it clean from other debug code that you may not want to pull in.