[GitHub] [spark] cloud-fan commented on a change in pull request #34092: [SPARK-36827][CORE] Improve the perf and memory usage of cleaning up stage UI data

[GitBox <gi...@apache.org>]

cloud-fan commented on a change in pull request #34092:
URL: https://github.com/apache/spark/pull/34092#discussion_r715428573



##########
File path: core/src/main/scala/org/apache/spark/status/AppStatusListener.scala
##########
@@ -1253,12 +1254,47 @@ private[spark] class AppStatusListener(
     toDelete.foreach { j => kvstore.delete(j.getClass(), j.info.jobId) }
   }
 
-  private def cleanupStages(count: Long): Unit = {
-    val countToDelete = calculateNumberToRemove(count, conf.get(MAX_RETAINED_STAGES))
-    if (countToDelete <= 0L) {
-      return
+  private case class StageCompletionTime(
+      stageId: Int,
+      attemptId: Int,
+      completionTime: Long)
+
+  private def cleanupStagesWithInMemoryStore(countToDelete: Long): Seq[Array[Int]] = {
+    val stageArray = new ArrayBuffer[StageCompletionTime]()
+    val stageDataCount = new mutable.HashMap[Int, Int]()
+    kvstore.view(classOf[StageDataWrapper]).forEach { s =>

Review comment:
       I think `KVUtils.viewToSeq(view, countToDelete.toInt)` is better and cleaner if the in-memory store implements it with top-n algorithm. We can do it in followups.




-- 
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: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org