Github user mengxr commented on a diff in the pull request:
https://github.com/apache/spark/pull/22001#discussion_r209277357
--- Diff: core/src/main/scala/org/apache/spark/scheduler/DAGScheduler.scala
---
@@ -929,6 +955,28 @@ class DAGScheduler(
// HadoopRDD whose underlying HDFS files have been deleted.
finalStage = createResultStage(finalRDD, func, partitions, jobId,
callSite)
} catch {
+ case e: Exception if e.getMessage ==
+
DAGScheduler.ERROR_MESSAGE_BARRIER_REQUIRE_MORE_SLOTS_THAN_CURRENT_TOTAL_NUMBER
=>
+ logWarning("The job requires to run a barrier stage that requires
more slots than the " +
+ "total number of slots in the cluster currently.")
+ jobIdToNumTasksCheckFailures.putIfAbsent(jobId, 0)
+ val numCheckFailures = jobIdToNumTasksCheckFailures.get(jobId) + 1
+ if (numCheckFailures <
DAGScheduler.DEFAULT_MAX_CONSECUTIVE_NUM_TASKS_CHECK_FAILURES) {
+ jobIdToNumTasksCheckFailures.put(jobId, numCheckFailures)
+ messageScheduler.schedule(
+ new Runnable {
+ override def run(): Unit =
eventProcessLoop.post(JobSubmitted(jobId, finalRDD, func,
+ partitions, callSite, listener, properties))
+ },
+ timeIntervalNumTasksCheck * 1000,
+ TimeUnit.MILLISECONDS
+ )
+ return
+ } else {
+ listener.jobFailed(e)
--- End diff --
do you expect the same job submitted again? if not, we should remove the
key from the hashmap.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
