[
https://issues.apache.org/jira/browse/HIVE-13392?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15325068#comment-15325068
]
Alan Gates commented on HIVE-13392:
-----------------------------------
Patch seems fine, though it seems to contains some stuff unrelated to the
stated purpose of the JIRA (e.g. moving ValidCompactorTxnList around).
+1
> disable speculative execution for ACID Compactor
> ------------------------------------------------
>
> Key: HIVE-13392
> URL: https://issues.apache.org/jira/browse/HIVE-13392
> Project: Hive
> Issue Type: Bug
> Components: Transactions
> Affects Versions: 1.0.0
> Reporter: Eugene Koifman
> Assignee: Eugene Koifman
> Attachments: HIVE-13392.2.patch, HIVE-13392.3.patch,
> HIVE-13392.4.patch, HIVE-13392.patch
>
>
> https://developer.yahoo.com/hadoop/tutorial/module4.html
> Speculative execution is enabled by default. You can disable speculative
> execution for the mappers and reducers by setting the
> mapred.map.tasks.speculative.execution and
> mapred.reduce.tasks.speculative.execution JobConf options to false,
> respectively.
> CompactorMR is currently not set up to handle speculative execution and may
> lead to something like
> {code}
> 2016-02-08 22:56:38,256 WARN [main] org.apache.hadoop.mapred.YarnChild:
> Exception running child :
> org.apache.hadoop.ipc.RemoteException(org.apache.hadoop.hdfs.protocol.AlreadyBeingCreatedException):
> Failed to CREATE_FILE
> /apps/hive/warehouse/service_logs_v2/ds=2016-01-20/_tmp_6cf08b9f-c2e2-4182-bc81-e032801b147f/base_13858600/bucket_00004
> for DFSClient_attempt_1454628390210_27756_m_000001_1_131224698_1 on
> 172.18.129.12 because this file lease is currently owned by
> DFSClient_attempt_1454628390210_27756_m_000001_0_-2027182532_1 on
> 172.18.129.18
> at
> org.apache.hadoop.hdfs.server.namenode.FSNamesystem.recoverLeaseInternal(FSNamesystem.java:2937)
> at
> org.apache.hadoop.hdfs.server.namenode.FSNamesystem.startFileInternal(FSNamesystem.java:2562)
> at
> org.apache.hadoop.hdfs.server.namenode.FSNamesystem.startFileInt(FSNamesystem.java:2451)
> at
> org.apache.hadoop.hdfs.server.namenode.FSNamesystem.startFile(FSNamesystem.java:2335)
> at
> org.apache.hadoop.hdfs.server.namenode.NameNodeRpcServer.create(NameNodeRpcServer.java:688)
> at
> org.apache.hadoop.hdfs.protocolPB.ClientNamenodeProtocolServerSideTranslatorPB.create(ClientNamenodeProtocolServerSideTranslatorPB.java:397)
> 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:2151)
> {code}
> Short term: disable speculative execution for this job
> Longer term perhaps make each task write to dir with UUID...
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
