[question]: What's the correct approach for updating the pipeline?

[xdlin]

Hi, first of all, many thanks to the team for creating such a great engine!

When I tried to update an existing pipeline (to add another sink table or just make some minor change), with my local service, the only field I can change is the Parallelism count (even if I stop it first), does that meet the initial design?

If that’s the case, what’s the correct approach for updating the SQL query? Am I supposed to copy it and create a new version? If so, how can I make use of the existing checkpoint?

[indiedev02]

Hey @xdlin — currently I’m using the API that Arroyo provides to make these kinds of changes.
The steps I follow are:

1. Stop the currently running pipeline
2. Delete the pipeline
3. Re-create the pipeline

I don’t think Arroyo supports mutating an existing query yet, and this is the only solution I could come up with using the current API. From the UI, it might be a bit difficult to follow these steps though.

[xdlin]

hi @indiedev02, thanks for the responding, when you re-create the pipeline, is it possible to reuse the existing checkpoint?

[indiedev02]

@xdlin Not really sure about that tho