Some diagnostic scripts do not work with the distributed scheduler. Typically this happens because they insert large numpy arrays into the task graph, which subsequently leads to the scheduler killing the worker because they exceed their maximum allowed amount of memory. This adds an option to use the default scheduler for some diagnostic scripts by specifying no_distributed: true in the diagnostic script settings.
The diagnostic script in recipe_eady_growth_rate.yml does not work with the distributed scheduler. This pull request disables using the distributed scheduler for that recipe.
It is the responsibility of the author to make sure the pull request is ready to review. The icons indicate whether the item will be subject to the 🛠 Technical or 🧪 Scientific review.
Description
Some diagnostic scripts do not work with the distributed scheduler. Typically this happens because they insert large numpy arrays into the task graph, which subsequently leads to the scheduler killing the worker because they exceed their maximum allowed amount of memory. This adds an option to use the default scheduler for some diagnostic scripts by specifying
no_distributed: truein the diagnostic script settings.The diagnostic script in
recipe_eady_growth_rate.ymldoes not work with the distributed scheduler. This pull request disables using the distributed scheduler for that recipe.Link to documentation:
Before you get started
Checklist
It is the responsibility of the author to make sure the pull request is ready to review. The icons indicate whether the item will be subject to the 🛠 Technical or 🧪 Scientific review.
To help with the number of pull requests: