[
https://issues.apache.org/jira/browse/HIVE-17947?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16227164#comment-16227164
]
Daniel Voros commented on HIVE-17947:
-------------------------------------
I think one of the reasons for the exception above is using RemoteIterator. Its
{{hasNext()}} function seems to fail if the file was removed since creating the
iterator.
Another reason why I've decided to go with the
{{FileUtils#listStatusRecursively()}} (that uses
{{FileSystem#listStatus():FileStatus[]}}) is because there's no (public) method
in FileSystem that accepts a PathFilter and returns a RemoteIterator. Without
that we would fail again at hasNext() when trying to iterate the results to
filter out hidden files that were removed in the meantime.
Directories being removed between finding them and listing their contents would
mess with the current solution as well, but since we're able to filter out
hidden files we're not listing the staging directories that are likely to be
removed.
After a quick look it seems to me that we're already using
{{listStatusRecursively()}} to list files under a table
[here|https://github.com/apache/hive/blob/cd08cd6d3103ed70a8da4cce7eaaad251eb2a12f/ql/src/java/org/apache/hadoop/hive/ql/exec/FetchOperator.java#L662],
so we would probably have come across the memory limit if it could be an issue
in this case.
Please let me know what you think!
> Concurrent inserts might fail for ACID table since HIVE-17526 on branch-1
> -------------------------------------------------------------------------
>
> Key: HIVE-17947
> URL: https://issues.apache.org/jira/browse/HIVE-17947
> Project: Hive
> Issue Type: Bug
> Components: Transactions
> Affects Versions: 1.3.0
> Reporter: Daniel Voros
> Assignee: Daniel Voros
> Priority: Blocker
> Attachments: HIVE-17947.1-branch-1.patch
>
>
> HIVE-17526 (only on branch-1) disabled conversion to ACID if there are
> *_copy_N files under the table, but the filesystem checks introduced there
> are running for every insert since the MoveTask in the end of the insert will
> call alterTable eventually.
> The filename checking also recurses into staging directories created by other
> inserts. If those are removed while listing the files, it leads to the
> following exception and failing insert:
> {code}
> java.io.FileNotFoundException: File
> hdfs://mycluster/apps/hive/warehouse/dvoros.db/concurrent_insert/.hive-staging_hive_2017-10-30_13-23-35_056_2844419018556002410-2/-ext-10001
> does not exist.
> at
> org.apache.hadoop.hdfs.DistributedFileSystem$DirListingIterator.<init>(DistributedFileSystem.java:1081)
> ~[hadoop-hdfs-2.7.3.2.6.3.0-235.jar:?]
> at
> org.apache.hadoop.hdfs.DistributedFileSystem$DirListingIterator.<init>(DistributedFileSystem.java:1059)
> ~[hadoop-hdfs-2.7.3.2.6.3.0-235.jar:?]
> at
> org.apache.hadoop.hdfs.DistributedFileSystem$23.doCall(DistributedFileSystem.java:1004)
> ~[hadoop-hdfs-2.7.3.2.6.3.0-235.jar:?]
> at
> org.apache.hadoop.hdfs.DistributedFileSystem$23.doCall(DistributedFileSystem.java:1000)
> ~[hadoop-hdfs-2.7.3.2.6.3.0-235.jar:?]
> at
> org.apache.hadoop.fs.FileSystemLinkResolver.resolve(FileSystemLinkResolver.java:81)
> ~[hadoop-common-2.7.3.2.6.3.0-235.jar:?]
> at
> org.apache.hadoop.hdfs.DistributedFileSystem.listLocatedStatus(DistributedFileSystem.java:1018)
> ~[hadoop-hdfs-2.7.3.2.6.3.0-235.jar:?]
> at
> org.apache.hadoop.fs.FileSystem.listLocatedStatus(FileSystem.java:1735)
> ~[hadoop-common-2.7.3.2.6.3.0-235.jar:?]
> at
> org.apache.hadoop.fs.FileSystem$6.handleFileStat(FileSystem.java:1864)
> ~[hadoop-common-2.7.3.2.6.3.0-235.jar:?]
> at org.apache.hadoop.fs.FileSystem$6.hasNext(FileSystem.java:1841)
> ~[hadoop-common-2.7.3.2.6.3.0-235.jar:?]
> at
> org.apache.hadoop.hive.metastore.TransactionalValidationListener.containsCopyNFiles(TransactionalValidationListener.java:226)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> org.apache.hadoop.hive.metastore.TransactionalValidationListener.handleAlterTableTransactionalProp(TransactionalValidationListener.java:104)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> org.apache.hadoop.hive.metastore.TransactionalValidationListener.handle(TransactionalValidationListener.java:63)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> org.apache.hadoop.hive.metastore.TransactionalValidationListener.onEvent(TransactionalValidationListener.java:55)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> org.apache.hadoop.hive.metastore.HiveMetaStore$HMSHandler.firePreEvent(HiveMetaStore.java:2478)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> org.apache.hadoop.hive.metastore.HiveMetaStore$HMSHandler.alter_table_core(HiveMetaStore.java:4145)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> org.apache.hadoop.hive.metastore.HiveMetaStore$HMSHandler.alter_table_with_environment_context(HiveMetaStore.java:4117)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at sun.reflect.GeneratedMethodAccessor107.invoke(Unknown Source)
> ~[?:?]
> at
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> ~[?:1.8.0_144]
> at java.lang.reflect.Method.invoke(Method.java:498) ~[?:1.8.0_144]
> at
> org.apache.hadoop.hive.metastore.RetryingHMSHandler.invokeInternal(RetryingHMSHandler.java:148)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> org.apache.hadoop.hive.metastore.RetryingHMSHandler.invoke(RetryingHMSHandler.java:107)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> com.sun.proxy.$Proxy32.alter_table_with_environment_context(Unknown Source)
> [?:?]
> at
> org.apache.hadoop.hive.metastore.HiveMetaStoreClient.alter_table_with_environmentContext(HiveMetaStoreClient.java:299)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> org.apache.hadoop.hive.ql.metadata.SessionHiveMetaStoreClient.alter_table_with_environmentContext(SessionHiveMetaStoreClient.java:325)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at sun.reflect.GeneratedMethodAccessor87.invoke(Unknown Source) ~[?:?]
> at
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> ~[?:1.8.0_144]
> at java.lang.reflect.Method.invoke(Method.java:498) ~[?:1.8.0_144]
> at
> org.apache.hadoop.hive.metastore.RetryingMetaStoreClient.invoke(RetryingMetaStoreClient.java:173)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at com.sun.proxy.$Proxy33.alter_table_with_environmentContext(Unknown
> Source) [?:?]
> at sun.reflect.GeneratedMethodAccessor87.invoke(Unknown Source) ~[?:?]
> at
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> ~[?:1.8.0_144]
> at java.lang.reflect.Method.invoke(Method.java:498) ~[?:1.8.0_144]
> at
> org.apache.hadoop.hive.metastore.HiveMetaStoreClient$SynchronizedHandler.invoke(HiveMetaStoreClient.java:2227)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at com.sun.proxy.$Proxy33.alter_table_with_environmentContext(Unknown
> Source) [?:?]
> at org.apache.hadoop.hive.ql.metadata.Hive.alterTable(Hive.java:626)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at org.apache.hadoop.hive.ql.metadata.Hive.alterTable(Hive.java:608)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at org.apache.hadoop.hive.ql.metadata.Hive.loadTable(Hive.java:2020)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at org.apache.hadoop.hive.ql.exec.MoveTask.execute(MoveTask.java:380)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at org.apache.hadoop.hive.ql.exec.Task.executeTask(Task.java:199)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> org.apache.hadoop.hive.ql.exec.TaskRunner.runSequential(TaskRunner.java:100)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at org.apache.hadoop.hive.ql.Driver.launchTask(Driver.java:1987)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at org.apache.hadoop.hive.ql.Driver.execute(Driver.java:1667)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at org.apache.hadoop.hive.ql.Driver.runInternal(Driver.java:1414)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at org.apache.hadoop.hive.ql.Driver.run(Driver.java:1211)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at org.apache.hadoop.hive.ql.Driver.run(Driver.java:1204)
> [hive-exec-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> org.apache.hive.service.cli.operation.SQLOperation.runQuery(SQLOperation.java:242)
> [hive-service-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> org.apache.hive.service.cli.operation.SQLOperation.access$800(SQLOperation.java:91)
> [hive-service-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> org.apache.hive.service.cli.operation.SQLOperation$BackgroundWork$1.run(SQLOperation.java:336)
> [hive-service-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at java.security.AccessController.doPrivileged(Native Method)
> ~[?:1.8.0_144]
> at javax.security.auth.Subject.doAs(Subject.java:422) [?:1.8.0_144]
> at
> org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1866)
> [hadoop-common-2.7.3.2.6.3.0-235.jar:?]
> at
> org.apache.hive.service.cli.operation.SQLOperation$BackgroundWork.run(SQLOperation.java:350)
> [hive-service-2.1.0.2.6.3.0-235.jar:2.1.0.2.6.3.0-235]
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
> [?:1.8.0_144]
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> [?:1.8.0_144]
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
> [?:1.8.0_144]
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> [?:1.8.0_144]
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
> [?:1.8.0_144]
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
> [?:1.8.0_144]
> at java.lang.Thread.run(Thread.java:748) [?:1.8.0_144]
> {code}
> We should:
> - run checks only when converting to ACID and not on every alter (as we do
> on master)
> - skip hidden files to prevent recursing into staging directories
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
