From 9b78e6e33d8c117c1e909df414f20d9db56efe4b Mon Sep 17 00:00:00 2001 From: Xuan Date: Mon, 7 Sep 2015 17:45:47 -0700 Subject: [PATCH] YARN-4087. Followup fixes after YARN-2019 regarding RM behavior when state-store error occurs. Contributed by Jian He --- hadoop-yarn-project/CHANGES.txt | 3 +++ .../hadoop/yarn/conf/YarnConfiguration.java | 2 +- .../src/main/resources/yarn-default.xml | 5 ++++- .../resourcemanager/recovery/RMStateStore.java | 15 +++++++++------ 4 files changed, 17 insertions(+), 8 deletions(-) diff --git a/hadoop-yarn-project/CHANGES.txt b/hadoop-yarn-project/CHANGES.txt index db96dd69e73..a16207064a0 100644 --- a/hadoop-yarn-project/CHANGES.txt +++ b/hadoop-yarn-project/CHANGES.txt @@ -889,6 +889,9 @@ Release 2.7.2 - UNRELEASED YARN-4105. Capacity Scheduler headroom for DRF is wrong (Chang Li via jlowe) + YARN-4087. Followup fixes after YARN-2019 regarding RM behavior when + state-store error occurs. (Jian He via xgong) + Release 2.7.1 - 2015-07-06 INCOMPATIBLE CHANGES diff --git a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/YarnConfiguration.java b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/YarnConfiguration.java index 5e1bab281f5..4879ca1d926 100644 --- a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/YarnConfiguration.java +++ b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/YarnConfiguration.java @@ -402,7 +402,7 @@ public class YarnConfiguration extends Configuration { public static final boolean DEFAULT_RM_RECOVERY_ENABLED = false; public static final String YARN_FAIL_FAST = YARN_PREFIX + "fail-fast"; - public static final boolean DEFAULT_YARN_FAIL_FAST = true; + public static final boolean DEFAULT_YARN_FAIL_FAST = false; public static final String RM_FAIL_FAST = RM_PREFIX + "fail-fast"; diff --git a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/resources/yarn-default.xml b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/resources/yarn-default.xml index 436bfb04e63..59bfb569865 100644 --- a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/resources/yarn-default.xml +++ b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/resources/yarn-default.xml @@ -349,9 +349,12 @@ Should YARN fail fast if it encounters any errors. + This is a global config for all other components including RM,NM etc. + If no value is set for component-specific config (e.g yarn.resourcemanager.fail-fast), + this value will be the default. yarn.fail-fast - true + false diff --git a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/RMStateStore.java b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/RMStateStore.java index b7f1e6c72f1..aa5caf96e9a 100644 --- a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/RMStateStore.java +++ b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/RMStateStore.java @@ -45,6 +45,7 @@ import org.apache.hadoop.yarn.api.records.ApplicationId; import org.apache.hadoop.yarn.api.records.ApplicationSubmissionContext; import org.apache.hadoop.yarn.api.records.ReservationId; import org.apache.hadoop.yarn.api.records.impl.pb.ApplicationSubmissionContextPBImpl; +import org.apache.hadoop.yarn.conf.HAUtil; import org.apache.hadoop.yarn.conf.YarnConfiguration; import org.apache.hadoop.yarn.event.AsyncDispatcher; import org.apache.hadoop.yarn.event.Dispatcher; @@ -1013,18 +1014,20 @@ public abstract class RMStateStore extends AbstractService { */ protected void notifyStoreOperationFailed(Exception failureCause) { LOG.error("State store operation failed ", failureCause); - if (failureCause instanceof StoreFencedException) { + if (HAUtil.isHAEnabled(getConfig())) { + LOG.warn("State-store fenced ! Transitioning RM to standby"); updateFencedState(); Thread standByTransitionThread = new Thread(new StandByTransitionThread()); standByTransitionThread.setName("StandByTransitionThread Handler"); standByTransitionThread.start(); + } else if (YarnConfiguration.shouldRMFailFast(getConfig())) { + LOG.fatal("Fail RM now due to state-store error!"); + rmDispatcher.getEventHandler().handle( + new RMFatalEvent(RMFatalEventType.STATE_STORE_OP_FAILED, + failureCause)); } else { - if (YarnConfiguration.shouldRMFailFast(getConfig())) { - rmDispatcher.getEventHandler().handle( - new RMFatalEvent(RMFatalEventType.STATE_STORE_OP_FAILED, - failureCause)); - } + LOG.warn("Skip the state-store error."); } }