Feature Request: To get Scheduler to auto-trigger jobs in external cluster

[UmmulkiramR]

Feature Request: To get Scheduler to auto-trigger jobs in external cluster

[joneubank]

How will the scheduler prioritize which directory to choose? Will it try to completely use up the resources of one directory before moving to the next, or will it try to spread the load evenly? How will it decide between two directories that both have capacity? Should we be able to set a priority order to use up directories or is there some other configurations needed for this strategy?

Also, our workflow requests right now contain information on the cluster and/or directory we want them to run in. After this feature is implemented how will the workflow requests change? Can we still specify the cluster we want it to run in? Do we need to include/exclude specific settings to get the scheduler to decide where to run the workflow?

Lastly, how will the RDPC admin be able to see which cluster a workflow is running in? Is this information visible in the workflow UI and/or workflow API (through the RDPC Gateway)? Do we currently have a view of how many workflows are running in each cluster or is that another task that we should investigate in the future?

[joneubank]

Instead of combining directory path and cluster name into a single string, is there a way to split these into separate lists in the config:

directories:
  - cluster1:
    - "nfs-local/nfs-1"
    - "nfs-local/nfs-2"
  - cluster2:
    - "nfs-external/nfs-1"
    - "nfs-external/nfs-2"