Details
-
Type:
New Feature
-
Status:
Resolved
-
Priority:
Severe
-
Resolution: Fixed
-
Affects Version/s: None
-
Fix Version/s: 5.0.0 GA (5.0.0 GA Suite Release)
-
Component/s: Big Data (Hadoop etc.)
-
QA Validation Status:Not Yet Validated
Description
The idea would be to have a PMR job launched from spoon contain callback information to an embedded Carte server running inside Spoon.
Then we can have the mappers, reducers and combiners phone home and notify Spoon that they started processing.
At that point you could somehow get hold of the transformation running on the node, gather data from the steps and push it back to Spoon.
We could get hold of the transformation again by starting up a copy of Carte for the duration of the transformation, implement RMI ... or we could use some Hadoop specific interface not yet known to me right now.
Then we can have the mappers, reducers and combiners phone home and notify Spoon that they started processing.
At that point you could somehow get hold of the transformation running on the node, gather data from the steps and push it back to Spoon.
We could get hold of the transformation again by starting up a copy of Carte for the duration of the transformation, implement RMI ... or we could use some Hadoop specific interface not yet known to me right now.
We could even push logging information periodically directly into a specific Spoon perspective, ...