In some cases, yarn does not automatically restart the container

pdeyhim · May 30, 2014 · e00b656 · e00b656
1 parent 79fa8fd
commit e00b656
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 0 deletions.
diff --git a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala
@@ -261,6 +261,10 @@ class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration,
           finishApplicationMaster(FinalApplicationStatus.FAILED,
             "max number of executor failures reached")
         }
+        val numExecutorsFailed = yarnAllocator.getNumExecutorsFailed
+        if (numExecutorsFailed > 0) {
+          yarnAllocator.addResourceRequests(numExecutorsFailed)
+        }
         yarnAllocator.allocateResources()
         ApplicationMaster.incrementAllocatorLoop(1)
         Thread.sleep(100)

diff --git a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/ExecutorLauncher.scala b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/ExecutorLauncher.scala
@@ -206,6 +206,10 @@ class ExecutorLauncher(args: ApplicationMasterArguments, conf: Configuration, sp
     yarnAllocator.addResourceRequests(args.numExecutors)
     while ((yarnAllocator.getNumExecutorsRunning < args.numExecutors) && (!driverClosed)) {
       yarnAllocator.allocateResources()
+      val numExecutorsFailed = yarnAllocator.getNumExecutorsFailed
+      if (numExecutorsFailed > 0) {
+        yarnAllocator.addResourceRequests(numExecutorsFailed)
+      }
       Thread.sleep(100)
     }