August 23rd Meeting

Present: Eli Manolo Ippei Matthew Vineet Hanming Eric

Presentation by Manolo

Manolo has created an entry point script for python to get the system information out of a system. Validation was done in bin/bash so its usable for whatever is on the machine. Runs docker inspect to get the documentation. The script gives information about base images, as well as actual images. Lists software specs, distro, python version, list of python packages from requirements.txt and what's installed. Keep the validation as language agnostic as we can, if we start validating multiple languages. This script builds in layers base -> build -> validation, this makes the construction easier if you need to build on a container. You can have a couple of base images then build off of those validation.

Updates

EM: I haven't had a lot of time to work on this but I had some questions. Was anyone able to see their orgs system validation documents? Are we missing anything? Eric: I haven't but thinking about this process, I don't know how they would react to this because the validation document has to be static. We can't use anything other than a static pdf and rasterized images. We're trying to draw a line with our department, system is your responsibility and the packages are our responsibility. We don't validate our hardware anymore, if its on the cloud you can't validate that hardware. I think the knitted document will actually work, just run this and the output is your proof. MT: I'm guessing at Jazz our experience is closer to Erics.