I have two WFs (one caller and one callee) in the same KNIME community hub team’s space. The space is shared with at least one more member of the team with edit access.
The main (caller) WF is deployed as schedule job. The execution fails each time with error since it can’t locate the callee WF. However, when I try to run the adhoc job manually job it just works fine.
I suspect the reason could be that the original deployment was setup with “team” execution scope. I have changed it to “user” scope now and will see if it runs successfully next week.
My question is if the callee WF resides in a space shared with team (though not with all the members), does the deployment need to have user execution scope?
If both the caller and callee are in the same team space, the scheduler should be able to find the callee normally, user scope is needed if the workflow requires personal secrets/Kerberos (stored under user account). Otherwise team scope is fine.
To help diagnose this, could you show how the caller invokes the callee?
In the Call Workflow node, did you select the callee via the Hub picker (Workflow on Hub), or are you using a relative path?
Is the node pinned to a specific revision?
Was the callee moved or renamed after deployment?
A quick screenshot of that node’s configuration would be perfect hiding any confidential data.
In your dialog the path ../deduplicate_cleanup means go to the parent folder of the caller, then open the sibling workflow named deduplicate_cleanup. If the callee isn’t exactly there, the scheduler won’t find it.
For scheduled runs on Hub, it’s more suitable to reference the callee via the Hub se:
In Call Workflow Service, set Workflow location to Hub space (instead of Relative to → Current workflow).
Browse to your team space and select the callee workflow.
If the caller and callee are in team-owned spaces that the team executor can read, selecting the callee via the Hub space option will be fine under team scope.
