maestrowf
maestrowf copied to clipboard
LLNL
A tool to easily orchestrate general computational workflows both locally and on supercomputers
As specified in #212 when maestro is run from outside a virtual environment it cannot find conductor. Maestro can run from within an activated virtualenv: ``` $ venv/bin/activate (venv) $...
The celery tasks in the merlin package require blocking parallel launches to be substituted for $(LAUNCHER) in the step scripts. This is due to the way celery handles task completions....
Currently depends waits for either all inbound edges or one inbound edge. Using depends: [parentnode] or depends:[parentnode_*] I would like to wait for a subset of inbound edges chosen based...
It's the case where the conductor fails in execution due to unexpected issues (system crash or file permission errors due to unexpected file system issues). When the conductor crashes, the...
When user creates a generator, currently it is not able to access data from the original yaml file. Also the "variable" from env are parsed first which means we cannot...
I have a workflow that tries to run a serial code multiple times interactively on a machine with multiple CPU's. What I hoped would happen was that the serial code...
Maestro currently logs the run command it issues for each step to ``DEBUG``; however, this would be useful information for Maestro to always log in the output for debugging and...
For dedicated application time (DAT) runs, we need a "#SLURM --qos=exempt" in the batchfile. Can this be added to maestrowf/maestrowf/interfaces/script/slurmscriptadapter.py, perhaps as a "qos: exempt" option? Gratefully, -Chris
It would be useful to have the ability to run a given step or a range of steps in the workflow. This would be really helpful for debugging a workflow.
When using the `FluxAdapter` (and by proxy, probably the `FluxSpectrumAdapter`) the wall time for a scheduled step cannot exceed 24 hours. This bug is probably related to the use of...