[
https://issues.apache.org/jira/browse/HIVE-14660?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15449205#comment-15449205
]
Benjamin BONNET commented on HIVE-14660:
----------------------------------------
Here are the properties we set before creating the table and before
inserting/deleting rows:
{code}
set hive.support.concurrency=true;
set hive.enforce.bucketing=true;
set hive.exec.dynamic.partition.mode=nonstrict;
set hive.txn.manager=org.apache.hadoop.hive.ql.lockmgr.DbTxnManager;
set hive.compactor.initiator.on=true;
set hive.compactor.worker.threads=1;
{code}
Apart from those properties, settings are "standard" (we use a HDP 2.3
cluster).
mapred.reduce.tasks is not set
We have (by default):
mapreduce.job.reduces=-1
mapreduce.reduce.speculative=true
Table has 36 columns, is clustered into 4 buckets (on a single column), ORC
formatted, transactional and partitioned by year/month. It has about half
billion rows.
The genuine query that fails is a kind of :
{code}DELETE FROM table WHERE string_operations_on_some_columns IN (
select_from_another_table );
{code}
Concerning the mapred.reduce.tasks setting, I talk about a work-around (not a
solution) since reading the FileSinkOperator, one sees it has be designed to
operate on multiple buckets. In my opinion, the only mistake was the use of an
array instead of a map, or a circular array (if you are guaranteed the way
buckets are dealtwith is circular for the latter, which I could not assume
when I wrote the patch).
> ArrayIndexOutOfBoundsException on delete
> ----------------------------------------
>
> Key: HIVE-14660
> URL: https://issues.apache.org/jira/browse/HIVE-14660
> Project: Hive
> Issue Type: Bug
> Components: Query Processor, Transactions
> Affects Versions: 1.2.1
> Reporter: Benjamin BONNET
> Assignee: Benjamin BONNET
> Attachments: HIVE-14660.1-banch-1.2.patch
>
>
> Hi,
> DELETE on an ACID table may fail on an ArrayIndexOutOfBoundsException.
> That bug occurs at Reduce phase when there are less reducers than the number
> of the table buckets.
> In order to reproduce, create a simple ACID table :
> {code:sql}
> CREATE TABLE test (`cle` bigint,`valeur` string)
> PARTITIONED BY (`annee` string)
> CLUSTERED BY (cle) INTO 5 BUCKETS
> TBLPROPERTIES ('transactional'='true');
> {code}
> Populate it with lines distributed among all buckets, with random values and
> a few partitions.
> Force the Reducers to be less than the buckets :
> {code:sql}
> set mapred.reduce.tasks=1;
> {code}
> Then execute a delete that will remove many lines from all the buckets.
> {code:sql}
> DELETE FROM test WHERE valeur<'some_value';
> {code}
> Then you will get an ArrayIndexOutOfBoundsException :
> {code}
> 2016-08-22 21:21:02,500 [FATAL] [TezChild] |tez.ReduceRecordSource|:
> org.apache.hadoop.hive.ql.metadata.HiveException: Hive Runtime Error while
> processing row (tag=0)
> {"key":{"reducesinkkey0":{"transactionid":119,"bucketid":0,"rowid":0}},"value":{"_col0":"4"}}
> at
> org.apache.hadoop.hive.ql.exec.tez.ReduceRecordSource$GroupIterator.next(ReduceRecordSource.java:352)
> at
> org.apache.hadoop.hive.ql.exec.tez.ReduceRecordSource.pushRecord(ReduceRecordSource.java:274)
> at
> org.apache.hadoop.hive.ql.exec.tez.ReduceRecordProcessor.run(ReduceRecordProcessor.java:252)
> at
> org.apache.hadoop.hive.ql.exec.tez.TezProcessor.initializeAndRunProcessor(TezProcessor.java:150)
> at
> org.apache.hadoop.hive.ql.exec.tez.TezProcessor.run(TezProcessor.java:139)
> at
> org.apache.tez.runtime.LogicalIOProcessorRuntimeTask.run(LogicalIOProcessorRuntimeTask.java:344)
> at
> org.apache.tez.runtime.task.TezTaskRunner$TaskRunnerCallable$1.run(TezTaskRunner.java:181)
> at
> org.apache.tez.runtime.task.TezTaskRunner$TaskRunnerCallable$1.run(TezTaskRunner.java:172)
> at java.security.AccessController.doPrivileged(Native Method)
> at javax.security.auth.Subject.doAs(Subject.java:415)
> at
> org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1657)
> at
> org.apache.tez.runtime.task.TezTaskRunner$TaskRunnerCallable.callInternal(TezTaskRunner.java:172)
> at
> org.apache.tez.runtime.task.TezTaskRunner$TaskRunnerCallable.callInternal(TezTaskRunner.java:168)
> at org.apache.tez.common.CallableWithNdc.call(CallableWithNdc.java:36)
> at java.util.concurrent.FutureTask.run(FutureTask.java:262)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
> at java.lang.Thread.run(Thread.java:745)
> Caused by: java.lang.ArrayIndexOutOfBoundsException: 5
> at
> org.apache.hadoop.hive.ql.exec.FileSinkOperator.process(FileSinkOperator.java:769)
> at org.apache.hadoop.hive.ql.exec.Operator.forward(Operator.java:838)
> at
> org.apache.hadoop.hive.ql.exec.SelectOperator.process(SelectOperator.java:88)
> at
> org.apache.hadoop.hive.ql.exec.tez.ReduceRecordSource$GroupIterator.next(ReduceRecordSource.java:343)
> ... 17 more
> {code}
> Adding logs into FileSinkOperator, one sees the operator deals with buckets
> 0, 1, 2, 3, 4, then 0 again and it fails at line 769 : actually each time you
> switch bucket, you move forwards in a 5 (number of buckets) elements array.
> So when you get bucket 0 for the second time, you get out of the array...
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
