[SPARK-19545][YARN] Fix compile issue for Spark on Yarn when building…

… against Hadoop 2.6.0~2.6.3 ## What changes were proposed in this pull request? Due to the newly added API in Hadoop 2.6.4+, Spark builds against Hadoop 2.6.0~2.6.3 will meet compile error. So here still reverting back to use reflection to handle this issue. ## How was this patch tested? Manual verification. Author: jerryshao <[email protected]> Closes #16884 from jerryshao/SPARK-19545.
apache · Feb 10, 2017 · 8e8afb3 · 8e8afb3
1 parent d5593f7
commit 8e8afb3
Show file tree

Hide file tree

Showing 2 changed files with 23 additions and 8 deletions.
diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md
@@ -445,7 +445,7 @@ To use a custom metrics.properties for the application master and executors, upd
   This will be used with YARN's rolling log aggregation, to enable this feature in YARN side
   <code>yarn.nodemanager.log-aggregation.roll-monitoring-interval-seconds</code> should be
   configured in yarn-site.xml.
-  This feature can only be used with Hadoop 2.6.1+. The Spark log4j appender needs be changed to use
+  This feature can only be used with Hadoop 2.6.4+. The Spark log4j appender needs be changed to use
   FileAppender or another appender that can handle the files being removed while its running. Based
   on the file name configured in the log4j configuration (like spark.log), the user should set the
   regex (spark*) to include all the log files that need to be aggregated.
@@ -524,8 +524,8 @@ pre-packaged distribution.
 1. In the `yarn-site.xml` on each node, add `spark_shuffle` to `yarn.nodemanager.aux-services`,
 then set `yarn.nodemanager.aux-services.spark_shuffle.class` to
 `org.apache.spark.network.yarn.YarnShuffleService`.
-1. Increase `NodeManager's` heap size by setting `YARN_HEAPSIZE` (1000 by default) in `etc/hadoop/yarn-env.sh` 
-to avoid garbage collection issues during shuffle. 
+1. Increase `NodeManager's` heap size by setting `YARN_HEAPSIZE` (1000 by default) in `etc/hadoop/yarn-env.sh`
+to avoid garbage collection issues during shuffle.
 1. Restart all `NodeManager`s in your cluster.
 
 The following extra configuration options are available when the shuffle service is running on YARN:

diff --git a/resource-managers/yarn/src/main/scala/org/apache/spark/deploy/yarn/Client.scala b/resource-managers/yarn/src/main/scala/org/apache/spark/deploy/yarn/Client.scala
@@ -245,12 +245,27 @@ private[spark] class Client(
     }
 
     sparkConf.get(ROLLED_LOG_INCLUDE_PATTERN).foreach { includePattern =>
-      val logAggregationContext = Records.newRecord(classOf[LogAggregationContext])
-      logAggregationContext.setRolledLogsIncludePattern(includePattern)
-      sparkConf.get(ROLLED_LOG_EXCLUDE_PATTERN).foreach { excludePattern =>
-        logAggregationContext.setRolledLogsExcludePattern(excludePattern)
+      try {
+        val logAggregationContext = Records.newRecord(classOf[LogAggregationContext])
+
+        // These two methods were added in Hadoop 2.6.4, so we still need to use reflection to
+        // avoid compile error when building against Hadoop 2.6.0 ~ 2.6.3.
+        val setRolledLogsIncludePatternMethod =
+          logAggregationContext.getClass.getMethod("setRolledLogsIncludePattern", classOf[String])
+        setRolledLogsIncludePatternMethod.invoke(logAggregationContext, includePattern)
+
+        sparkConf.get(ROLLED_LOG_EXCLUDE_PATTERN).foreach { excludePattern =>
+          val setRolledLogsExcludePatternMethod =
+            logAggregationContext.getClass.getMethod("setRolledLogsExcludePattern", classOf[String])
+          setRolledLogsExcludePatternMethod.invoke(logAggregationContext, excludePattern)
+        }
+
+        appContext.setLogAggregationContext(logAggregationContext)
+      } catch {
+        case NonFatal(e) =>
+          logWarning(s"Ignoring ${ROLLED_LOG_INCLUDE_PATTERN.key} because the version of YARN " +
+            "does not support it", e)
       }
-      appContext.setLogAggregationContext(logAggregationContext)
     }
 
     appContext