Hi All,
Hopefully this is the right place to ask about how development and production is best managed on KNIME (SERVER).
Background
We started to use KNIME to set-up a ETL pipeline and now plan to deploy V1.
Where I got a potential issue is the manner how the “components” seems organized.
Or if there is alternative structure I should use to get the following issue fixed I am happy to hear this
My example:
Assume this is my map structure
I developed the following workflow, which should be called multiple times as a seperate workflow ( or maybe component) depending on the client/instance it is run for.
In the workflow I have a component (in practise I have over 30 which makes it a issue) that is available on the server (the example is local). Using this I can make a single adjustment to this component to make sure all the workflows that are called are updated.
This works well in development mode. However, as I would like to deploy this I got into the situation that the path is linked to the 0_DEVELOPMENT calculation path. This implies that I should first copy the example into production and then unlink all the components and then create a new link to the 1_PRODUCTION folder to not mix my development and production.
the manual steps takes alot of time and is sensitive to errors. Therefore:
Is there s simpele alternative in which I can work on a development version and directly deploy all the work of the delpotment version at once on the Knime server? Taking into account that i work with components that are called by multiple workflows of the same instance as outline above.
Looking forward to a reply,
Best regards / vriendelijke groeten,
Tom