skonto commented on a change in pull request #24796: [SPARK-27900][CORE] Add
uncaught exception handler to the driver
URL: https://github.com/apache/spark/pull/24796#discussion_r290976095
##########
File path: core/src/main/scala/org/apache/spark/deploy/SparkSubmit.scala
##########
@@ -861,6 +861,16 @@ private[spark] class SparkSubmit extends Logging {
* running cluster deploy mode or python applications.
*/
private def runMain(args: SparkSubmitArguments, uninitLog: Boolean): Unit = {
+ // set the handler as early as possible
+ // we need the driver check because spark-submit is used by mesos to launch
+ // executors (via the SparkLauncher)
+ val isExecutor = args.childArgs.contains("--executor-id")
+ args.deployMode match {
+ case "client" | null if !isExecutor =>
+ Thread.setDefaultUncaughtExceptionHandler(
Review comment:
There should be documentation or something about what is provided (because
Spark on K8s is affected for example) but even if the user adds a handler,
there will be a deadlock here (could be fixed I guess). In general I dont see
people setting one and what happens if the same code is implemented in python
you need a default handler in that case I guess.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
