mridulm commented on code in PR #47197: URL: https://github.com/apache/spark/pull/47197#discussion_r1670374260
########## core/src/main/scala/org/apache/spark/executor/TaskMetrics.scala: ########## @@ -328,19 +354,16 @@ private[spark] object TaskMetrics extends Logging { */ def fromAccumulators(accums: Seq[AccumulatorV2[_, _]]): TaskMetrics = { val tm = new TaskMetrics - val externalAccums = new java.util.ArrayList[AccumulatorV2[Any, Any]]() for (acc <- accums) { val name = acc.name - val tmpAcc = acc.asInstanceOf[AccumulatorV2[Any, Any]] if (name.isDefined && tm.nameToAccums.contains(name.get)) { val tmAcc = tm.nameToAccums(name.get).asInstanceOf[AccumulatorV2[Any, Any]] tmAcc.metadata = acc.metadata - tmAcc.merge(tmpAcc) + tmAcc.merge(acc.asInstanceOf[AccumulatorV2[Any, Any]]) } else { - externalAccums.add(tmpAcc) + tm._externalAccums += acc Review Comment: It has to do with some nits with Java memory model guarantees - the updates to the list need not be visible when queried from the read/write lock since it did not go through the same barrier when updates were made. (I am reviewing only based on diff though (afk); so not sure if there are mitigating reasons why it might not apply) -- 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-unsubscr...@spark.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org