diff --git ql/src/java/org/apache/hadoop/hive/ql/exec/tez/DagUtils.java ql/src/java/org/apache/hadoop/hive/ql/exec/tez/DagUtils.java index 54e3f77..1804076 100644 --- ql/src/java/org/apache/hadoop/hive/ql/exec/tez/DagUtils.java +++ ql/src/java/org/apache/hadoop/hive/ql/exec/tez/DagUtils.java @@ -79,7 +79,7 @@ import org.apache.hadoop.yarn.api.records.URL; import org.apache.hadoop.yarn.util.ConverterUtils; import org.apache.hadoop.yarn.util.Records; -import org.apache.tez.client.PreWarmContext; +import org.apache.tez.client.PreWarmVertex; import org.apache.tez.dag.api.DAG; import org.apache.tez.dag.api.DataSinkDescriptor; import org.apache.tez.dag.api.DataSourceDescriptor; @@ -380,6 +380,15 @@ private Resource getContainerResource(Configuration conf) { } /* + * Helper to setup default environment for a task in YARN. + */ + private Map getContainerEnvironment(Configuration conf, boolean isMap) { + Map environment = new HashMap(); + MRHelpers.updateEnvironmentForMRTasks(conf, environment, isMap); + return environment; + } + + /* * Helper to determine what java options to use for the containers * Falls back to Map-reduces map java opts if no tez specific options * are set @@ -476,9 +485,7 @@ private Vertex createVertex(JobConf conf, MapWork mapWork, map = new Vertex(mapWork.getName(), new ProcessorDescriptor(MapTezProcessor.class.getName()). setUserPayload(serializedConf), numTasks, getContainerResource(conf)); - Map environment = new HashMap(); - MRHelpers.updateEnvironmentForMRTasks(conf, environment, true); - map.setTaskEnvironment(environment); + map.setTaskEnvironment(getContainerEnvironment(conf, true)); map.setTaskLaunchCmdOpts(getContainerJavaOpts(conf)); assert mapWork.getAliasToWork().keySet().size() == 1; @@ -550,11 +557,7 @@ private Vertex createVertex(JobConf conf, ReduceWork reduceWork, reduceWork.isAutoReduceParallelism() ? reduceWork.getMaxReduceTasks() : reduceWork .getNumReduceTasks(), getContainerResource(conf)); - Map environment = new HashMap(); - - MRHelpers.updateEnvironmentForMRTasks(conf, environment, false); - reducer.setTaskEnvironment(environment); - + reducer.setTaskEnvironment(getContainerEnvironment(conf, false)); reducer.setTaskLaunchCmdOpts(getContainerJavaOpts(conf)); Map localResources = new HashMap(); @@ -598,17 +601,16 @@ private LocalResource createLocalResource(FileSystem remoteFs, Path file, /** * @param numContainers number of containers to pre-warm * @param localResources additional resources to pre-warm with - * @return prewarm context object + * @return prewarm vertex to run */ - public PreWarmContext createPreWarmContext(TezConfiguration conf, + public PreWarmVertex createPreWarmVertex(TezConfiguration conf, int numContainers, Map localResources) throws IOException, TezException { ProcessorDescriptor prewarmProcDescriptor = new ProcessorDescriptor(HivePreWarmProcessor.class.getName()); prewarmProcDescriptor.setUserPayload(MRHelpers.createUserPayloadFromConf(conf)); - PreWarmContext context = new PreWarmContext(prewarmProcDescriptor, getContainerResource(conf), - numContainers, null); + PreWarmVertex prewarmVertex = new PreWarmVertex("prewarm", prewarmProcDescriptor, numContainers,getContainerResource(conf)); Map combinedResources = new HashMap(); @@ -616,14 +618,10 @@ public PreWarmContext createPreWarmContext(TezConfiguration conf, combinedResources.putAll(localResources); } - context.setLocalResources(combinedResources); - - /* boiler plate task env */ - Map environment = new HashMap(); - MRHelpers.updateEnvironmentForMRTasks(conf, environment, true); - context.setEnvironment(environment); - context.setJavaOpts(getContainerJavaOpts(conf)); - return context; + prewarmVertex.setTaskLocalFiles(localResources); + prewarmVertex.setTaskLaunchCmdOpts(getContainerJavaOpts(conf)); + prewarmVertex.setTaskEnvironment(getContainerEnvironment(conf, false)); + return prewarmVertex; } /** diff --git ql/src/java/org/apache/hadoop/hive/ql/exec/tez/TezSessionState.java ql/src/java/org/apache/hadoop/hive/ql/exec/tez/TezSessionState.java index c9484e8..19c8060 100644 --- ql/src/java/org/apache/hadoop/hive/ql/exec/tez/TezSessionState.java +++ ql/src/java/org/apache/hadoop/hive/ql/exec/tez/TezSessionState.java @@ -50,11 +50,12 @@ import org.apache.hadoop.hive.shims.ShimLoader; import org.apache.hadoop.security.UserGroupInformation; import org.apache.hadoop.yarn.api.records.LocalResource; -import org.apache.tez.client.PreWarmContext; import org.apache.tez.client.TezClient; +import org.apache.tez.client.PreWarmVertex; import org.apache.tez.dag.api.SessionNotRunning; import org.apache.tez.dag.api.TezConfiguration; import org.apache.tez.dag.api.TezException; +import org.apache.tez.dag.api.Vertex; import org.apache.tez.mapreduce.hadoop.MRHelpers; /** @@ -170,6 +171,15 @@ public void open(HiveConf conf, String[] additionalFiles) // generate basic tez config TezConfiguration tezConfig = new TezConfiguration(conf); tezConfig.set(TezConfiguration.TEZ_AM_STAGING_DIR, tezScratchDir.toUri().toString()); + + if (HiveConf.getBoolVar(conf, ConfVars.HIVE_PREWARM_ENABLED)) { + int n = HiveConf.getIntVar(conf, ConfVars.HIVE_PREWARM_NUM_CONTAINERS); + n = Math.max(tezConfig.getInt( + TezConfiguration.TEZ_AM_SESSION_MIN_HELD_CONTAINERS, + TezConfiguration.TEZ_AM_SESSION_MIN_HELD_CONTAINERS_DEFAULT), n); + tezConfig.setInt(TezConfiguration.TEZ_AM_SESSION_MIN_HELD_CONTAINERS, n); + } + session = new TezClient("HIVE-" + sessionId, tezConfig, true, commonLocalResources, null); @@ -182,10 +192,10 @@ public void open(HiveConf conf, String[] additionalFiles) int n = HiveConf.getIntVar(conf, ConfVars.HIVE_PREWARM_NUM_CONTAINERS); LOG.info("Prewarming " + n + " containers (id: " + sessionId + ", scratch dir: " + tezScratchDir + ")"); - PreWarmContext context = utils.createPreWarmContext(tezConfig, n, + PreWarmVertex prewarmVertex = utils.createPreWarmVertex(tezConfig, n, commonLocalResources); try { - session.preWarm(context); + session.preWarm(prewarmVertex); } catch (InterruptedException ie) { if (LOG.isDebugEnabled()) { LOG.debug("Hive Prewarm threw an exception ", ie);