I have not yet scheduled etc. and thus far only us...
# ask-metaflowh
happy-wolf-7852
10/26/2023, 7:03 AMI have not yet scheduled etc. and thus far only used the below to kickoff a process in the cloud, which runs for a couple of hours. Yesterday, I had to leave work before it finished. Just curious, does this matter (i.e. my local machine was “disconnected”)? I think, as I used “batch”, it should carry on running “in the cloud” anyway and does not require the local “client” to be connected/to complete? Sorry for this stupid question.
python3 SomeGreatBatchProcess.py --environment=conda --package-suffixes .sql run \--parameter1 1.00 \--parameter2 1.2 \--parameter3 xyz --with batchf
flaky-plumber-70709
10/26/2023, 2:00 PMI think what’s going on when using metaflow in this way is that your local client is acting as the orchestrator even though it’s not actually running the steps. For truly remote execution you should create a step function out of the flow and use the trigger cli command to kick off the run.
flaky-plumber-70709
10/26/2023, 2:01 PMI tend to do this early even when just experimenting since I can trigger a lot of parameterized executions and then analyze the results later using the metaflow client
h
happy-wolf-7852
10/26/2023, 2:01 PMthanks - yeah that’s what I thought/think too …