Fix timeouts related to metalearnings tests

[eddiebergman]

Adding debug statements for now to find the problem on the github action servers as it cleanly finished locally.

Closes #1506

[codecov]

Codecov Report

Merging #1508 (879e2df) into development (e9c0318) will increase coverage by 0.06%.
The diff coverage is 86.76%.

@@               Coverage Diff               @@
##           development    #1508      +/-   ##
===============================================
+ Coverage        83.77%   83.83%   +0.06%     
===============================================
  Files              152      153       +1     
  Lines            11667    11694      +27     
  Branches          2037     2047      +10     
===============================================
+ Hits              9774     9804      +30     
+ Misses            1344     1339       -5     
- Partials           549      551       +2     

[eddiebergman]

I switched to using import tempfile and using gettempdir() instead of explicitly using /tmp. It seemed to not trigger as a result. Maybe the github action servers have some special folder for it and the stochastic nature of the problem was due to node allocation and usage.

I'm going to run the actions a few time and see if it comes up again.

[eddiebergman]

All the timeouts that occur seem to have moved to the next script where there was a hardcoded /tmp again. Renamed it and hopefully it solves it.

[mfeurer]

Looks good to me. Does it solve all issues?

[eddiebergman]

I just set explicit timeouts for each step and bumped the overall timeout per test from 300s to 600s. If one of them actually does hang then at least we'll get a debug message in the future

[eddiebergman]

I removed the comment you left review on, if all the tests pass then I think that can be merged