Apache
/
hadoop
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

YARN-2019. Retrospect on decision of making RM crashed if any exception throw in ZKRMStateStore. Contributed by Jian He. 

(cherry picked from commit ee98d6354b)
This commit is contained in:
Junping Du 2015-07-22 17:52:35 -07:00
parent be2334ba3a
commit 6772c3f4dd
4 changed files with 37 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
hadoop-yarn-project/CHANGES.txt
View File

@ -89,6 +89,9 @@ Release 2.8.0 - UNRELEASED
YARN-2003. Support for Application priority : Changes in RM and Capacity
Scheduler. (Sunil G via wangda)
YARN-2019. Retrospect on decision of making RM crashed if any exception throw
in ZKRMStateStore. (Jian He via junping_du)
IMPROVEMENTS
YARN-644. Basic null check is not performed on passed in arguments before

11
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/YarnConfiguration.java
View File

@ -401,6 +401,11 @@ public class YarnConfiguration extends Configuration {
public static final String RECOVERY_ENABLED = RM_PREFIX + "recovery.enabled";
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 String RM_FAIL_FAST = RM_PREFIX + "fail-fast";
@Private
public static final String RM_WORK_PRESERVING_RECOVERY_ENABLED = RM_PREFIX
+ "work-preserving-recovery.enabled";
@ -2018,6 +2023,12 @@ public class YarnConfiguration extends Configuration {
YARN_HTTP_POLICY_DEFAULT));
}
public static boolean shouldRMFailFast(Configuration conf) {
return conf.getBoolean(YarnConfiguration.RM_FAIL_FAST,
conf.getBoolean(YarnConfiguration.YARN_FAIL_FAST,
YarnConfiguration.DEFAULT_YARN_FAIL_FAST));
}
@Private
public static String getClusterId(Configuration conf) {
String clusterId = conf.get(YarnConfiguration.RM_CLUSTER_ID);

16
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/resources/yarn-default.xml
View File

@ -323,6 +323,22 @@
<value>false</value>
</property>
<property>
<description>Should RM fail fast if it encounters any errors. By defalt, it
points to ${yarn.fail-fast}. Errors include:
1) exceptions when state-store write/read operations fails.
</description>
<name>yarn.resourcemanager.fail-fast</name>
<value>${yarn.fail-fast}</value>
</property>
<property>
<description>Should YARN fail fast if it encounters any errors.
</description>
<name>yarn.fail-fast</name>
<value>true</value>
</property>
<property>
<description>Enable RM work preserving recovery. This configuration is private
to YARN for experimenting the feature.

7
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/RMStateStore.java
View File

@ -44,6 +44,7 @@ import org.apache.hadoop.yarn.api.records.ApplicationAttemptId;
import org.apache.hadoop.yarn.api.records.ApplicationId;
import org.apache.hadoop.yarn.api.records.ApplicationSubmissionContext;
import org.apache.hadoop.yarn.api.records.impl.pb.ApplicationSubmissionContextPBImpl;
import org.apache.hadoop.yarn.conf.YarnConfiguration;
import org.apache.hadoop.yarn.event.AsyncDispatcher;
import org.apache.hadoop.yarn.event.Dispatcher;
import org.apache.hadoop.yarn.event.EventHandler;
@ -855,6 +856,7 @@ public abstract class RMStateStore extends AbstractService {
* @param failureCause the exception due to which the operation failed
*/
protected void notifyStoreOperationFailed(Exception failureCause) {
LOG.error("State store operation failed ", failureCause);
if (failureCause instanceof StoreFencedException) {
updateFencedState();
Thread standByTransitionThread =
@ -862,8 +864,11 @@ public abstract class RMStateStore extends AbstractService {
standByTransitionThread.setName("StandByTransitionThread Handler");
standByTransitionThread.start();
} else {
if (YarnConfiguration.shouldRMFailFast(getConfig())) {
rmDispatcher.getEventHandler().handle(
new RMFatalEvent(RMFatalEventType.STATE_STORE_OP_FAILED, failureCause));
new RMFatalEvent(RMFatalEventType.STATE_STORE_OP_FAILED,
failureCause));
}
}
}