[YARN-5594] Handle old RMDelegationToken format when recovering RM

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.6.0
• Fix Version/s: 3.1.0, 2.10.0
• Component/s: resourcemanager
• Labels: oct16-medium
• Target Version/s: 3.1.0, 2.10.0
• Hadoop Flags: Reviewed
• Assignee: Robert Kanter
• Reporter: Tatyana But
• Created: 30/Aug/16 15:14
• Updated: 16/May/18 23:45
• Resolved: 04/Dec/17 21:21

Description

We've got that error after upgrade cluster from v.2.5.1 to 2.7.0.

2016-08-25 17:20:33,293 ERROR
org.apache.hadoop.yarn.server.resourcemanager.ResourceManager: Failed to
load/recover state
com.google.protobuf.InvalidProtocolBufferException: Protocol message contained
an invalid tag (zero).
at com.google.protobuf.InvalidProtocolBufferException.invalidTag(InvalidProtocolBufferException.java:89)
at com.google.protobuf.CodedInputStream.readTag(CodedInputStream.java:108)
at org.apache.hadoop.yarn.proto.YarnServerResourceManagerRecoveryProtos$RMDelegationTokenIdentifierDataProto.<init>(YarnServerResourceManagerRecoveryProtos.java:4680)
at org.apache.hadoop.yarn.proto.YarnServerResourceManagerRecoveryProtos$RMDelegationTokenIdentifierDataProto.<init>(YarnServerResourceManagerRecoveryProtos.java:4644)
at org.apache.hadoop.yarn.proto.YarnServerResourceManagerRecoveryProtos$RMDelegationTokenIdentifierDataProto$1.parsePartialFrom(YarnServerResourceManagerRecoveryProtos.java:4740)
at org.apache.hadoop.yarn.proto.YarnServerResourceManagerRecoveryProtos$RMDelegationTokenIdentifierDataProto$1.parsePartialFrom(YarnServerResourceManagerRecoveryProtos.java:4735)
at org.apache.hadoop.yarn.proto.YarnServerResourceManagerRecoveryProtos$RMDelegationTokenIdentifierDataProto$Builder.mergeFrom(YarnServerResourceManagerRecoveryProtos.java:5075)
at org.apache.hadoop.yarn.proto.YarnServerResourceManagerRecoveryProtos$RMDelegationTokenIdentifierDataProto$Builder.mergeFrom(YarnServerResourceManagerRecoveryProtos.java:4955)
at com.google.protobuf.AbstractMessage$Builder.mergeFrom(AbstractMessage.java:337)
at com.google.protobuf.AbstractMessage$Builder.mergeFrom(AbstractMessage.java:267)
at com.google.protobuf.AbstractMessageLite$Builder.mergeFrom(AbstractMessageLite.java:210)
at com.google.protobuf.AbstractMessage$Builder.mergeFrom(AbstractMessage.java:904)
at org.apache.hadoop.yarn.server.resourcemanager.recovery.records.RMDelegationTokenIdentifierData.readFields(RMDelegationTokenIdentifierData.java:43)
at org.apache.hadoop.yarn.server.resourcemanager.recovery.FileSystemRMStateStore.loadRMDTSecretManagerState(FileSystemRMStateStore.java:355)
at org.apache.hadoop.yarn.server.resourcemanager.recovery.FileSystemRMStateStore.loadState(FileSystemRMStateStore.java:199)
at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager$RMActiveServices.serviceStart(ResourceManager.java:587)
at org.apache.hadoop.service.AbstractService.start(AbstractService.java:193)
at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.startActiveServices(ResourceManager.java:1007)
at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager$1.run(ResourceManager.java:1048)
at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager$1.run(ResourceManager.java:1044

The reason of this problem is that we use different formats of files /var/mapr/cluster/yarn/rm/system/FSRMStateRoot/RMDTSecretManagerRoot/RMDelegationToken* in these hadoop versions.

This fix handle old data format during RM recover if InvalidProtocolBufferException occures.

Issue Links

• is broken by: YARN-2743 Yarn jobs via oozie fail with failed to renew token (secure) or digest mismatch (unsecure) errors when RM is being killed
• relates to: YARN-8310 Handle old NMTokenIdentifier, AMRMTokenIdentifier, and ContainerTokenIdentifier formats