[HIVE-13513] cleardanglingscratchdir does not work in some version of HDFS

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.3.0, 2.1.0
• Target Version/s: 1.3.0, 2.1.0
• Hadoop Flags: Reviewed
• Assignee: Daniel Dai
• Reporter: Daniel Dai
• Created: 14/Apr/16 05:58
• Updated: 27/Feb/24 22:24
• Resolved: 25/May/16 22:26

Description

On some Hadoop version, we keep getting "lease recovery" message at the time we check for scratchdir by opening for appending:

Failed to APPEND_FILE xxx for DFSClient_NONMAPREDUCE_785768631_1 on 10.0.0.18 because lease recovery is in progress. Try again later.
	at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.recoverLeaseInternal(FSNamesystem.java:2917)
	at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.appendFileInternal(FSNamesystem.java:2677)
	at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.appendFileInt(FSNamesystem.java:2984)
	at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.appendFile(FSNamesystem.java:2953)
	at org.apache.hadoop.hdfs.server.namenode.NameNodeRpcServer.append(NameNodeRpcServer.java:655)
	at org.apache.hadoop.hdfs.protocolPB.ClientNamenodeProtocolServerSideTranslatorPB.append(ClientNamenodeProtocolServerSideTranslatorPB.java:421)
	at org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos$ClientNamenodeProtocol$2.callBlockingMethod(ClientNamenodeProtocolProtos.java)
	at org.apache.hadoop.ipc.ProtobufRpcEngine$Server$ProtoBufRpcInvoker.call(ProtobufRpcEngine.java:616)
	at org.apache.hadoop.ipc.RPC$Server.call(RPC.java:969)
	at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:2137)
	at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:2133)
	at java.security.AccessController.doPrivileged(Native Method)
	at javax.security.auth.Subject.doAs(Subject.java:415)
	at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1657)
	at org.apache.hadoop.ipc.Server$Handler.run(Server.java:2131)

and

16/04/14 04:51:56 ERROR hdfs.DFSClient: Failed to close inode 18963
java.io.IOException: Failed to replace a bad datanode on the existing pipeline due to no more good datanodes being available to try. (Nodes: current=[DatanodeInfoWithStorage[10.0.0.12:30010,DS-b355ac2a-a23a-418a-af9b-4c1b4e26afe8,DISK]], original=[DatanodeInfoWithStorage[10.0.0.12:30010,DS-b355ac2a-a23a-418a-af9b-4c1b4e26afe8,DISK]]). The current failed datanode replacement policy is DEFAULT, and a client may configure this via 'dfs.client.block.write.replace-datanode-on-failure.policy' in its configuration.
	at org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.findNewDatanode(DFSOutputStream.java:951)
	at org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.addDatanode2ExistingPipeline(DFSOutputStream.java:1017)
	at org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.setupPipelineForAppendOrRecovery(DFSOutputStream.java:1165)
	at org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.run(DFSOutputStream.java:470)

The reason is not clear. However, if we remove hsync from SessionState, everything works as expected. Attach patch to remove hsync call for now.

Issue Links

• is related to: HIVE-13429 Tool to remove dangling scratch dir