Github user pwendell commented on a diff in the pull request:
https://github.com/apache/spark/pull/204#discussion_r11469656
--- Diff:
core/src/main/scala/org/apache/spark/scheduler/EventLoggingListener.scala ---
@@ -90,9 +97,115 @@ private[spark] class EventLoggingListener(appName:
String, conf: SparkConf)
logEvent(event, flushLogger = true)
override def onUnpersistRDD(event: SparkListenerUnpersistRDD) =
logEvent(event, flushLogger = true)
+ override def onApplicationStart(event: SparkListenerApplicationStart) =
+ logEvent(event, flushLogger = true)
+ override def onApplicationEnd(event: SparkListenerApplicationEnd) =
+ logEvent(event, flushLogger = true)
+
+ /**
+ * Stop logging events.
+ * In addition, create an empty special file to indicate application
completion.
+ */
+ def stop() = {
+ logger.newFile(APPLICATION_COMPLETE)
+ logger.stop()
+ }
+}
+
+private[spark] object EventLoggingListener extends Logging {
+ val LOG_PREFIX = "EVENT_LOG_"
+ val SPARK_VERSION_PREFIX = "SPARK_VERSION_"
+ val COMPRESSION_CODEC_PREFIX = "COMPRESSION_CODEC_"
+ val APPLICATION_COMPLETE = "APPLICATION_COMPLETE"
--- End diff --
Would it make sense to have two cases here? APPLICATION_COMPLETE_SUCCESS
and APPLICATION_COMPLETE_FAILURE? or maybe APPLICATION_COMPLETE and
APPLICATION_FAILURE? The former would be more consistent with some other parts
of Spark where "complete" does not necessarily mean success or failure, just
that it's finished.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
