sebwrede commented on pull request #1390: URL: https://github.com/apache/systemds/pull/1390#issuecomment-937626683
> When I resolved the conflicts with the master branch, I saw that Spark tests were added to the FederatedRowAggregate tests. Since Spark uses (Map)MM instead of the AggregateBinary instruction, I had to make the very same changes to the (Map)MM instruction as I did to the AggregateBinary instruction. There is now a StackOverflowError in FederatedMSVMTest.federatedMSVM1CP. I think it is because of MatrixObject.java:438, where an `acquireReadAndRelease()` call is made when the matrix object is federated. This does not work when it has to read from HDFS. I am not entirely sure how your changes causes this, but we need to change it so that it does not call `acquireReadAndRelease()` whenever the data is federated. It should instead make some check to see if it needs to call `DataConverter.readMatrixFromHDFS` even if the object itself is federated. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
