Bulk Operation

YARN-128 RMProxy should retry EOFException

YARN-128 Add leveldb-based implementation for RMStateStore

YARN-128 Move RM recovery related proto to yarn_server_resourcemanager_recovery.proto

YARN-128 Remove ApplicationAttemptState and ApplicationState class in RMStateStore class

YARN-128 RM should honor NM heartbeat expiry after RM restart

YARN-128 Better reporting of finished containers to AMs

YARN-128 NPE on registerNodeManager if the request has containers for UnmanagedAMs

YARN-128 Succeeded application remains in accepted after RM restart

YARN-128 Work preserving recovery of Unmanged AMs

YARN-128 Job stays in PREP state for long time after RM Restarts

YARN-128 Execessive logging for app and attempts on RM recovery

YARN-128 Revisit RMApp transitions from NEW on RECOVER

YARN-128 Fix invalid RMApp transition from NEW to FINAL_SAVING

YARN-128 Apps should be saved after it's accepted by the scheduler

YARN-128 Change killing application to wait until state store is done

YARN-128 Check time cost for recovering max-app-limit applications

YARN-128 RM hangs on shutdown if calling system.exit in serviceInit or serviceStart

YARN-128 Implement a RMStateStore cleaner for deleting application/attempt info

YARN-128 Batching optimization for ZKRMStateStore

YARN-128 Rethink znode structure for RM HA

YARN-128 Save version information in the state store

YARN-128 Register ClientToken MasterKey in SecretManager after it is saved

YARN-128 During RM restart, RM should start a new attempt only when previous attempt exits for real

YARN-128 AM fails to register if RM restarts within 5s of job submission

YARN-128 RM may relaunch already KILLED / FAILED jobs after RM restarts

YARN-128 FileSystemRMStateStore can leave partial files that prevent subsequent recovery

YARN-128 RMStateStore should flush all pending store events before closing

YARN-128 Populate AMRMTokens back to AMRMTokenSecretManager after RM restarts

YARN-128 Recovery issues on RM Restart with FileSystemRMStateStore

YARN-128 Handle app recovery differently for AM failures and RM restart

YARN-128 Document RM Restart feature

YARN-128 Change FileSystemRMStateStore to use directories

YARN-128 List of applications at NM web UI is inconsistent with applications at RM UI after RM restart

YARN-128 Apps Completed metrics on web UI is not correct after RM restart

YARN-128 "Active users" field in Resourcemanager scheduler UI gives negative values

YARN-128 RM crashes if it restarts while the state-store is down

YARN-128 Store completed application information in RM state store

YARN-128 verify that new jobs submitted with old RM delegation tokens after RM restart are accepted

YARN-128 Slow or failing DelegationToken renewals on submission itself make RM unavailable

YARN-128 RMStateStore's removeApplication APIs should just take an applicationId

YARN-128 Restore RMDelegationTokens after RM Restart

YARN-128 Restore clientToken for app attempt after RM restart

YARN-128 Restore appToken and clientToken for app attempt after RM restart

YARN-128 Test and verify that app delegation tokens are added to tokenRenewer after RM restart

YARN-128 NM should reject containers allocated by previous RM

YARN-128 Race condition causing RM to potentially relaunch already unregistered AMs on RM restart

YARN-128 AM max attempts is not checked when RM restart and try to recover attempts

YARN-128 Delayed store operations should not result in RM unavailability for app submission

YARN-128 Create common proxy client for communicating with RM

YARN-128 Add HDFS based store for RM which manages the store using directories

YARN-128 Add Zookeeper-based store implementation for RMStateStore

YARN-128 Security related work for RM restart

YARN-128 Add FileSystem based store for RM

YARN-128 Add FS-based persistent store implementation for RMStateStore

YARN-128 Make changes for RM restart phase 1

YARN-128 Remove old code for restart