Bulk Operation

SPARK-41550 Log `exitCode` when `SparkContext.stop` starts

SPARK-41550 Ignore `IOException` in `handleBlockRemovalFailure` if SparkContext is stopped

SPARK-41550 Support `local[*]` in `spark-submit` in K8s environment

SPARK-41550 Add `TOTAL_SHUFFLE_WRITE` executor roll policy

SPARK-41550 Add `DISK_USED` executor roll policy

SPARK-41550 Remove `Dynamic Resource Allocation` from K8s Future Work

SPARK-41550 Keep removed executor info in decommission state

SPARK-41550 Fix wrong remained shuffles log in BlockManagerDecommissioner

SPARK-41550 Populate ExecutorDecommission with more informative messages

SPARK-41550 Ignore stage fetch failure caused by decommissioned executor

SPARK-41550 Propagate decommission executor loss reason during onDisconnect in K8s

SPARK-41550 Add decomTestTag to K8s Integration Test

SPARK-41550 Randomize the orders of peer in BlockManagerDecommissioner

SPARK-41550 Enable spark.storage.decommission.(rdd|shuffle)Blocks.enabled by default

SPARK-41550 Add numberDecommissioningExecutors metric

SPARK-41550 Support PEAK_JVM_(ON|OFF)HEAP_MEMORY executor rolling policy

SPARK-41550 Enable spark.dynamicAllocation.shuffleTracking.enabled by default

SPARK-41550 Reuse PVCs by default

SPARK-41550 Log ExecutorDecommission as INFO level in TaskSchedulerImpl

SPARK-41550 Remove `Experimental` from `spark.dynamicAllocation.shuffleTracking.enabled`

SPARK-41550 Add PySpark cleanShuffleDependencies API

SPARK-41550 Remove `Experimental` from `RDD.cleanShuffleDependencies` API

SPARK-41550 FallbackStorage shouldn't attempt to resolve arbitrary "remote" hostname

SPARK-41550 ExecutorMonitor.onExecutorRemoved should handle ExecutorDecommission as finished

SPARK-41550 ExecutorMonitor.timedOutExecutors should be deterministic

SPARK-41550 Enable spark.storage.replication.proactive by default