Merge pull request apache#478 from sryza/sandy-spark-1033

pwendell · pwendell · commit 576c4a4c502c · 2014-01-22T14:10:07.000-08:00
SPARK-1033. Ask for cores in Yarn container requests

Tested on a pseudo-distributed cluster against the Fair Scheduler and observed a worker taking more than a single core.
diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md
@@ -133,7 +133,7 @@ See [Building Spark with Maven](building-with-maven.html) for instructions on ho
 
 # Important Notes
 
-- We do not requesting container resources based on the number of cores. Thus the numbers of cores given via command line arguments cannot be guaranteed.
+- Before Hadoop 2.2, YARN does not support cores in container resource requests. Thus, when running against an earlier version, the numbers of cores given via command line arguments cannot be passed to YARN.  Whether core requests are honored in scheduling decisions depends on which scheduler is in use and how it is configured.
 - The local directories used for spark will be the local directories configured for YARN (Hadoop Yarn config yarn.nodemanager.local-dirs). If the user specifies spark.local.dir, it will be ignored.
 - The --files and --archives options support specifying file names with the # similar to Hadoop. For example you can specify: --files localtest.txt#appSees.txt and this will upload the file you have locally named localtest.txt into HDFS but this will be linked to by the name appSees.txt and your application should use the name as appSees.txt to reference it when running on YARN.
 - The --addJars option allows the SparkContext.addJar function to work if you are using it with local files. It does not need to be used if you are using it with HDFS, HTTP, HTTPS, or FTP files.
diff --git a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala
@@ -532,15 +532,15 @@ private[yarn] class YarnAllocationHandler(
       priority: Int
     ): ArrayBuffer[ContainerRequest] = {
 
-    val memoryResource = Records.newRecord(classOf[Resource])
-    memoryResource.setMemory(workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD)
+    val memoryRequest = workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD
+    val resource = Resource.newInstance(memoryRequest, workerCores)
 
     val prioritySetting = Records.newRecord(classOf[Priority])
     prioritySetting.setPriority(priority)
 
     val requests = new ArrayBuffer[ContainerRequest]()
     for (i <- 0 until numWorkers) {
-      requests += new ContainerRequest(memoryResource, hosts, racks, prioritySetting)
+      requests += new ContainerRequest(resource, hosts, racks, prioritySetting)
     }
     requests
   }
