Hi KNIME Team,
I realized that with KNIME 4.0.1, all running workflows will be set to idle (yellow traffic light at all nodes that have not been executed yet) if a node of in a single workflow runs into an error, whereas the type of the failing node is irrelevant.
Did you also realize such issue?
I am happy to provide a log file but as the type of the failing node is irrelevant and the log file does not really show any plausibel information, I did not attach anything here.
no. And not sure I understand. Can you try explaining it a bit more? Include some screenshots or example workflow?
If a node fails it is expected for every subsequent node to go in state configured -yellow traffic light.
I attached a sample workflow that is supposed to call other workflows.
01_LO 20190919 LOOP.knwf (98.5 KB)
The idea is:
- Search for the newest workflow (based on the name convention “ID YYYMMDD”, ID = Identifier of the workflow) in sub-workflow “00_0510_NEWEST_WF”,
- Execute the newest workflow,
- Repeat everything once in an hour.
The first called workflow (“00_0510_NEWEST_WF”) is executed quite often and takes about 30 seconds to 1 minute to return a result. Sometimes, the call (waiting time) for workflow “00_0510_NEWEST_WF” seems to be too long and the node will fail. If this happens all other running workflows and nodes will be set to idle (except the nodes that have already been executed) (I tried to prevent the situation with the try-catch error nodes but it does not prevent the other workflows to be set to idle).
Please feel free to comment if you need further information to analyze the issue.
Thank you for your help in advance.
…most interestingly, by playing around with the workflow, I now receive another error that is being catched by the try-catch-error node:
ERROR Call Workflow (Table Based) 8:1370:1321 Execution failed in Try-Catch block: Cannot read file “knime_container_20190919_6543162619106371745.bin.snappy”
However, the error did not set the other workflows to idle as previously described.