mdietze
commented
10 years ago When generating the new settings details for the local run, if any files are missing (available in the inputs table but not on the machine of interest) list the missing inputs and available dbfiles records showing which machines have that file and where it's located. Also, if files are missing don't start runs and instead return an error code that stops the rest of the benchmark workflow from running (including returning the above list of inputs and dbfiles for reference, rather than only printing them to screen)
github-actions[bot]
Proposed new function to be used with the Benchmarking workflow
Inputs would be the benchmark_reference_run.id,, and machine.id
Internally it would query required info from the database, generate the full settings from the reference run, start the run (or runs for ensemble-based runs), and insert the new run and benchmark records in the database
Output would be full setting object (R list) for the run, which would internally include things like the the workflow.id, ensemble.id, benchmark_ensemble.id, etc.