attilapiros opened a new pull request #32790: URL: https://github.com/apache/spark/pull/32790
### What changes were proposed in this pull request? In `BlockManagerMasterEndpoint` for the disk persisted RDDs (when `spark.shuffle.service.fetch.rdd.enable` is enabled) we are keeping track the block status entries by external shuffle service instances (so on YARN we are basically keeping them by nodes). This is the `blockStatusByShuffleService` member val. And when all the RDD blocks are removed for one external shuffle service instance then the key and the empty map can be removed from `blockStatusByShuffleService`. ### Why are the changes needed? It is a small leak and I was asked to take care of it in https://github.com/apache/spark/pull/32114#discussion_r640270377. ### Does this PR introduce _any_ user-facing change? No. ### How was this patch tested? Existing tests. -- 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. For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
