Github user juanrh commented on the issue: https://github.com/apache/spark/pull/19267 @tgravescs I was finally able to contribute https://github.com/apache/hadoop/pull/289 which solves [YARN-6483](https://issues.apache.org/jira/browse/YARN-6483). With that patch, and the code in this pull request, in `YarnAllocator.allocateResources` we will receive a `NodeReport` entry in `allocateResponse.getUpdatedNodes` for each node moved to `DECOMMISSIONING` state using [Hadoop's graceful decommission](https://hadoop.apache.org/docs/stable/hadoop-yarn/hadoop-yarn-site/GracefulDecommission.html), which would trigger blacklisting for those nodes. But for now YARN-6483 has only been accepted for Hadoop 3.1.0, so I'll work on [SPARK-21737](https://issues.apache.org/jira/browse/SPARK-21737) to have an alternative solution that doesn't rely on the cluster manager
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org