save_cache when node is 0

With the new units based pricing model, we are finding that we can fan-out faster and while doing some repetitive work (and spending more build minutes), we probably shouldn't be doing things like updating a cache from 20 concurrent containers.

  • Guest
  • May 14 2018
  • Duplicate
  • Attach files
  • Guest commented
    May 14, 2018 16:26

    It would seem this is unnecessary - I see this in the log for extra nodes:


    Skipping cache generation, cache already exists for key