Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

dummy mode: consider removing #5961

Open
oliver-sanders opened this issue Feb 8, 2024 · 3 comments
Open

dummy mode: consider removing #5961

oliver-sanders opened this issue Feb 8, 2024 · 3 comments
Assignees
Labels
question Flag this as a question for the next Cylc project meeting.
Milestone

Comments

@oliver-sanders
Copy link
Member

Cylc 7 had two "dummy" modes:

dummy: Basically comments out all script* and environment so that you are running "blank" tasks, but still submitting them to the relevant platforms with the appropriate directives. This was apparently used by some as a validation test.

dummy-local: Also comments out the host and directives to make the "blank" tasks run locally.

Cylc 8 only has one mode called "dummy" which matches the behaviour of Cylc 7's "dummy-local".

Considering that "simulation" mode does everything that "dummy" mode does, without the unnecessary side effect of local submissions (which have the potential to flood the system!), simulation mode is a better tool for learning/testing. It also has options for simulating job failure and will play properly with the optional outputs system in due course.

So it's unclear what purpose the "dummy" mode (aka "dummy-local") has in Cylc 8. I raised this Discourse post to see if the community had any use cases and separately polled MO users. No responses so far.

Given changes currently underway to implement "skip" mode, maintaining the "dummy" mode code is about to become a bit burdensome, so we should consider removing it if unused.

@oliver-sanders oliver-sanders added the question Flag this as a question for the next Cylc project meeting. label Feb 8, 2024
@oliver-sanders oliver-sanders added this to the cylc-8.x milestone Feb 8, 2024
@hjoliver
Copy link
Member

hjoliver commented Feb 9, 2024

I have no problem with ditching dummy mode in favour of simulation mode, so long as we can make simulation work fully correctly.

For historical context, dummy mode predates simulation mode because it was easy to implement without breaking anything - all you have to do is ignore the script and environment content in tasks and everything else works exactly the same.

But simulation mode is fundamentally quite different, and more difficult (at least a bit more difficult) to get right.

Anyhow, that said, let's ditch dummy mode.

@hjoliver hjoliver added question Flag this as a question for the next Cylc project meeting. and removed question Flag this as a question for the next Cylc project meeting. labels Feb 9, 2024
@hjoliver
Copy link
Member

Also see #5975 - that seems a more natural fit for dummy mode than sim mode. In principle we could run some real jobs in sim mode, but I'm not sure how easy that would be to implement.

@hjoliver
Copy link
Member

hjoliver commented Dec 4, 2024

Tentatively assigning @wxtim as it seems like we will agree to do this, and Tim's "cup runneth over" with run-mode knowledge ATM.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
question Flag this as a question for the next Cylc project meeting.
Projects
None yet
Development

No branches or pull requests

3 participants