You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hive.apache.org by "Kevin Wilfong (JIRA)" <ji...@apache.org> on 2012/07/26 02:49:33 UTC
[jira] [Created] (HIVE-3302) Race condition in query plan for
merging at the end of a query
Kevin Wilfong created HIVE-3302:
-----------------------------------
Summary: Race condition in query plan for merging at the end of a query
Key: HIVE-3302
URL: https://issues.apache.org/jira/browse/HIVE-3302
Project: Hive
Issue Type: Bug
Components: Query Processor
Affects Versions: 0.10.0
Reporter: Kevin Wilfong
Assignee: Kevin Wilfong
Priority: Critical
In the query plan that's used to merge files at the end of a query, the dependency tree looks something like:
MoveTask(1)
/ \
...ConditionalTask MoveTask(2)...
\ /
MergeTask
Here MoveTask(1) moves the partition data to a temporary location, and MoveTask(2) moves it to the final location.
However if there are dynamic partitions generated and some of these partitions are merged and others are moved, the dependency tree is changed at runtime to:
...ConditionalTask MoveTask(2)...
\ /
MergeTask
\
MoveTask(1)
This produces a race condition between the two MoveTasks where if MoveTask(2) runs before MoveTask(1) the partitions moved by MoveTask(1) will get moved to an intermediate location and never moved to the final location. In this case those partitions are quietly lost.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (HIVE-3302) Race condition in query plan for
merging at the end of a query
Posted by "Kevin Wilfong (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HIVE-3302?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Kevin Wilfong updated HIVE-3302:
--------------------------------
Attachment: HIVE-3302.1.patch.txt
> Race condition in query plan for merging at the end of a query
> --------------------------------------------------------------
>
> Key: HIVE-3302
> URL: https://issues.apache.org/jira/browse/HIVE-3302
> Project: Hive
> Issue Type: Bug
> Components: Query Processor
> Affects Versions: 0.10.0
> Reporter: Kevin Wilfong
> Assignee: Kevin Wilfong
> Priority: Critical
> Attachments: HIVE-3302.1.patch.txt
>
>
> In the query plan that's used to merge files at the end of a query, the dependency tree looks something like:
> MoveTask(1)
> / \
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> Here MoveTask(1) moves the partition data to a temporary location, and MoveTask(2) moves it to the final location.
> However if there are dynamic partitions generated and some of these partitions are merged and others are moved, the dependency tree is changed at runtime to:
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> \
> MoveTask(1)
> This produces a race condition between the two MoveTasks where if MoveTask(2) runs before MoveTask(1) the partitions moved by MoveTask(1) will get moved to an intermediate location and never moved to the final location. In this case those partitions are quietly lost.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HIVE-3302) Race condition in query plan for
merging at the end of a query
Posted by "Kevin Wilfong (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HIVE-3302?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13428236#comment-13428236 ]
Kevin Wilfong commented on HIVE-3302:
-------------------------------------
Updated diff and patch, tests all pass now.
> Race condition in query plan for merging at the end of a query
> --------------------------------------------------------------
>
> Key: HIVE-3302
> URL: https://issues.apache.org/jira/browse/HIVE-3302
> Project: Hive
> Issue Type: Bug
> Components: Query Processor
> Affects Versions: 0.10.0
> Reporter: Kevin Wilfong
> Assignee: Kevin Wilfong
> Priority: Critical
> Attachments: HIVE-3302.1.patch.txt, HIVE-3302.2.patch.txt
>
>
> In the query plan that's used to merge files at the end of a query, the dependency tree looks something like:
> MoveTask(1)
> / \
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> Here MoveTask(1) moves the partition data to a temporary location, and MoveTask(2) moves it to the final location.
> However if there are dynamic partitions generated and some of these partitions are merged and others are moved, the dependency tree is changed at runtime to:
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> \
> MoveTask(1)
> This produces a race condition between the two MoveTasks where if MoveTask(2) runs before MoveTask(1) the partitions moved by MoveTask(1) will get moved to an intermediate location and never moved to the final location. In this case those partitions are quietly lost.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (HIVE-3302) Race condition in query plan for
merging at the end of a query
Posted by "Kevin Wilfong (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HIVE-3302?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Kevin Wilfong updated HIVE-3302:
--------------------------------
Status: Patch Available (was: Open)
> Race condition in query plan for merging at the end of a query
> --------------------------------------------------------------
>
> Key: HIVE-3302
> URL: https://issues.apache.org/jira/browse/HIVE-3302
> Project: Hive
> Issue Type: Bug
> Components: Query Processor
> Affects Versions: 0.10.0
> Reporter: Kevin Wilfong
> Assignee: Kevin Wilfong
> Priority: Critical
> Attachments: HIVE-3302.1.patch.txt, HIVE-3302.2.patch.txt
>
>
> In the query plan that's used to merge files at the end of a query, the dependency tree looks something like:
> MoveTask(1)
> / \
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> Here MoveTask(1) moves the partition data to a temporary location, and MoveTask(2) moves it to the final location.
> However if there are dynamic partitions generated and some of these partitions are merged and others are moved, the dependency tree is changed at runtime to:
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> \
> MoveTask(1)
> This produces a race condition between the two MoveTasks where if MoveTask(2) runs before MoveTask(1) the partitions moved by MoveTask(1) will get moved to an intermediate location and never moved to the final location. In this case those partitions are quietly lost.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (HIVE-3302) Race condition in query plan for
merging at the end of a query
Posted by "Kevin Wilfong (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HIVE-3302?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Kevin Wilfong updated HIVE-3302:
--------------------------------
Status: Open (was: Patch Available)
Missed some testcases, will update.
> Race condition in query plan for merging at the end of a query
> --------------------------------------------------------------
>
> Key: HIVE-3302
> URL: https://issues.apache.org/jira/browse/HIVE-3302
> Project: Hive
> Issue Type: Bug
> Components: Query Processor
> Affects Versions: 0.10.0
> Reporter: Kevin Wilfong
> Assignee: Kevin Wilfong
> Priority: Critical
> Attachments: HIVE-3302.1.patch.txt
>
>
> In the query plan that's used to merge files at the end of a query, the dependency tree looks something like:
> MoveTask(1)
> / \
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> Here MoveTask(1) moves the partition data to a temporary location, and MoveTask(2) moves it to the final location.
> However if there are dynamic partitions generated and some of these partitions are merged and others are moved, the dependency tree is changed at runtime to:
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> \
> MoveTask(1)
> This produces a race condition between the two MoveTasks where if MoveTask(2) runs before MoveTask(1) the partitions moved by MoveTask(1) will get moved to an intermediate location and never moved to the final location. In this case those partitions are quietly lost.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (HIVE-3302) Race condition in query plan for
merging at the end of a query
Posted by "Namit Jain (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HIVE-3302?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Namit Jain updated HIVE-3302:
-----------------------------
Resolution: Fixed
Hadoop Flags: Reviewed
Status: Resolved (was: Patch Available)
Committed. Thanks Kevin
> Race condition in query plan for merging at the end of a query
> --------------------------------------------------------------
>
> Key: HIVE-3302
> URL: https://issues.apache.org/jira/browse/HIVE-3302
> Project: Hive
> Issue Type: Bug
> Components: Query Processor
> Affects Versions: 0.10.0
> Reporter: Kevin Wilfong
> Assignee: Kevin Wilfong
> Priority: Critical
> Attachments: HIVE-3302.1.patch.txt, HIVE-3302.2.patch.txt
>
>
> In the query plan that's used to merge files at the end of a query, the dependency tree looks something like:
> MoveTask(1)
> / \
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> Here MoveTask(1) moves the partition data to a temporary location, and MoveTask(2) moves it to the final location.
> However if there are dynamic partitions generated and some of these partitions are merged and others are moved, the dependency tree is changed at runtime to:
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> \
> MoveTask(1)
> This produces a race condition between the two MoveTasks where if MoveTask(2) runs before MoveTask(1) the partitions moved by MoveTask(1) will get moved to an intermediate location and never moved to the final location. In this case those partitions are quietly lost.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (HIVE-3302) Race condition in query plan for
merging at the end of a query
Posted by "Kevin Wilfong (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HIVE-3302?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Kevin Wilfong updated HIVE-3302:
--------------------------------
Attachment: HIVE-3302.2.patch.txt
> Race condition in query plan for merging at the end of a query
> --------------------------------------------------------------
>
> Key: HIVE-3302
> URL: https://issues.apache.org/jira/browse/HIVE-3302
> Project: Hive
> Issue Type: Bug
> Components: Query Processor
> Affects Versions: 0.10.0
> Reporter: Kevin Wilfong
> Assignee: Kevin Wilfong
> Priority: Critical
> Attachments: HIVE-3302.1.patch.txt, HIVE-3302.2.patch.txt
>
>
> In the query plan that's used to merge files at the end of a query, the dependency tree looks something like:
> MoveTask(1)
> / \
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> Here MoveTask(1) moves the partition data to a temporary location, and MoveTask(2) moves it to the final location.
> However if there are dynamic partitions generated and some of these partitions are merged and others are moved, the dependency tree is changed at runtime to:
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> \
> MoveTask(1)
> This produces a race condition between the two MoveTasks where if MoveTask(2) runs before MoveTask(1) the partitions moved by MoveTask(1) will get moved to an intermediate location and never moved to the final location. In this case those partitions are quietly lost.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HIVE-3302) Race condition in query plan for
merging at the end of a query
Posted by "Kevin Wilfong (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HIVE-3302?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13423488#comment-13423488 ]
Kevin Wilfong commented on HIVE-3302:
-------------------------------------
Uploaded a diff here https://reviews.facebook.net/D4395
> Race condition in query plan for merging at the end of a query
> --------------------------------------------------------------
>
> Key: HIVE-3302
> URL: https://issues.apache.org/jira/browse/HIVE-3302
> Project: Hive
> Issue Type: Bug
> Components: Query Processor
> Affects Versions: 0.10.0
> Reporter: Kevin Wilfong
> Assignee: Kevin Wilfong
> Priority: Critical
>
> In the query plan that's used to merge files at the end of a query, the dependency tree looks something like:
> MoveTask(1)
> / \
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> Here MoveTask(1) moves the partition data to a temporary location, and MoveTask(2) moves it to the final location.
> However if there are dynamic partitions generated and some of these partitions are merged and others are moved, the dependency tree is changed at runtime to:
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> \
> MoveTask(1)
> This produces a race condition between the two MoveTasks where if MoveTask(2) runs before MoveTask(1) the partitions moved by MoveTask(1) will get moved to an intermediate location and never moved to the final location. In this case those partitions are quietly lost.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HIVE-3302) Race condition in query plan for
merging at the end of a query
Posted by "Kevin Wilfong (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HIVE-3302?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13422782#comment-13422782 ]
Kevin Wilfong commented on HIVE-3302:
-------------------------------------
JIRA screwed up the formatting of my diagrams, the first is supposed to show MoveTask(1) and MergeTask stemming from ConditionalTask and MoveTask(2) dependent upon MoveTask(1) and MergeTask.
The second is supposed to show MergeTask stemming from ConditionalTask, MoveTask(1) and MoveTask(2) dependent on MergeTask, and most importantly, no dependency relationship between MoveTask(1) and MoveTask(2).
> Race condition in query plan for merging at the end of a query
> --------------------------------------------------------------
>
> Key: HIVE-3302
> URL: https://issues.apache.org/jira/browse/HIVE-3302
> Project: Hive
> Issue Type: Bug
> Components: Query Processor
> Affects Versions: 0.10.0
> Reporter: Kevin Wilfong
> Assignee: Kevin Wilfong
> Priority: Critical
>
> In the query plan that's used to merge files at the end of a query, the dependency tree looks something like:
> MoveTask(1)
> / \
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> Here MoveTask(1) moves the partition data to a temporary location, and MoveTask(2) moves it to the final location.
> However if there are dynamic partitions generated and some of these partitions are merged and others are moved, the dependency tree is changed at runtime to:
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> \
> MoveTask(1)
> This produces a race condition between the two MoveTasks where if MoveTask(2) runs before MoveTask(1) the partitions moved by MoveTask(1) will get moved to an intermediate location and never moved to the final location. In this case those partitions are quietly lost.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (HIVE-3302) Race condition in query plan for
merging at the end of a query
Posted by "Kevin Wilfong (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HIVE-3302?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Kevin Wilfong updated HIVE-3302:
--------------------------------
Status: Patch Available (was: Open)
> Race condition in query plan for merging at the end of a query
> --------------------------------------------------------------
>
> Key: HIVE-3302
> URL: https://issues.apache.org/jira/browse/HIVE-3302
> Project: Hive
> Issue Type: Bug
> Components: Query Processor
> Affects Versions: 0.10.0
> Reporter: Kevin Wilfong
> Assignee: Kevin Wilfong
> Priority: Critical
> Attachments: HIVE-3302.1.patch.txt
>
>
> In the query plan that's used to merge files at the end of a query, the dependency tree looks something like:
> MoveTask(1)
> / \
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> Here MoveTask(1) moves the partition data to a temporary location, and MoveTask(2) moves it to the final location.
> However if there are dynamic partitions generated and some of these partitions are merged and others are moved, the dependency tree is changed at runtime to:
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> \
> MoveTask(1)
> This produces a race condition between the two MoveTasks where if MoveTask(2) runs before MoveTask(1) the partitions moved by MoveTask(1) will get moved to an intermediate location and never moved to the final location. In this case those partitions are quietly lost.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HIVE-3302) Race condition in query plan for
merging at the end of a query
Posted by "Hudson (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HIVE-3302?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13428782#comment-13428782 ]
Hudson commented on HIVE-3302:
------------------------------
Integrated in Hive-trunk-h0.21 #1592 (See [https://builds.apache.org/job/Hive-trunk-h0.21/1592/])
HIVE-3302 Race condition in query plan for merging at the end of a query
(Kevin Wilfong via namit) (Revision 1369375)
Result = SUCCESS
namit : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1369375
Files :
* /hive/trunk/contrib/src/test/results/clientpositive/serde_typedbytes.q.out
* /hive/trunk/contrib/src/test/results/clientpositive/serde_typedbytes2.q.out
* /hive/trunk/contrib/src/test/results/clientpositive/serde_typedbytes3.q.out
* /hive/trunk/contrib/src/test/results/clientpositive/serde_typedbytes5.q.out
* /hive/trunk/ql/src/java/org/apache/hadoop/hive/ql/optimizer/GenMRFileSink1.java
* /hive/trunk/ql/src/java/org/apache/hadoop/hive/ql/plan/ConditionalResolverMergeFiles.java
* /hive/trunk/ql/src/test/queries/clientpositive/merge_dynamic_partition5.q
* /hive/trunk/ql/src/test/results/clientpositive/binary_output_format.q.out
* /hive/trunk/ql/src/test/results/clientpositive/bucketmapjoin1.q.out
* /hive/trunk/ql/src/test/results/clientpositive/bucketmapjoin2.q.out
* /hive/trunk/ql/src/test/results/clientpositive/bucketmapjoin3.q.out
* /hive/trunk/ql/src/test/results/clientpositive/bucketmapjoin4.q.out
* /hive/trunk/ql/src/test/results/clientpositive/bucketmapjoin5.q.out
* /hive/trunk/ql/src/test/results/clientpositive/bucketmapjoin_negative.q.out
* /hive/trunk/ql/src/test/results/clientpositive/bucketmapjoin_negative2.q.out
* /hive/trunk/ql/src/test/results/clientpositive/case_sensitivity.q.out
* /hive/trunk/ql/src/test/results/clientpositive/cast1.q.out
* /hive/trunk/ql/src/test/results/clientpositive/index_auto.q.out
* /hive/trunk/ql/src/test/results/clientpositive/index_auto_file_format.q.out
* /hive/trunk/ql/src/test/results/clientpositive/index_auto_mult_tables_compact.q.out
* /hive/trunk/ql/src/test/results/clientpositive/index_auto_multiple.q.out
* /hive/trunk/ql/src/test/results/clientpositive/index_auto_partitioned.q.out
* /hive/trunk/ql/src/test/results/clientpositive/index_auto_update.q.out
* /hive/trunk/ql/src/test/results/clientpositive/index_compression.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input11.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input12.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input13.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input34.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input35.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input36.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input38.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input6.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input7.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input8.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input9.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input_dynamicserde.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input_part1.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input_part2.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input_part5.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input_testsequencefile.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input_testxpath.q.out
* /hive/trunk/ql/src/test/results/clientpositive/input_testxpath2.q.out
* /hive/trunk/ql/src/test/results/clientpositive/insert1.q.out
* /hive/trunk/ql/src/test/results/clientpositive/insert_into4.q.out
* /hive/trunk/ql/src/test/results/clientpositive/insert_into5.q.out
* /hive/trunk/ql/src/test/results/clientpositive/insert_into6.q.out
* /hive/trunk/ql/src/test/results/clientpositive/join25.q.out
* /hive/trunk/ql/src/test/results/clientpositive/join26.q.out
* /hive/trunk/ql/src/test/results/clientpositive/join27.q.out
* /hive/trunk/ql/src/test/results/clientpositive/join28.q.out
* /hive/trunk/ql/src/test/results/clientpositive/join29.q.out
* /hive/trunk/ql/src/test/results/clientpositive/join32.q.out
* /hive/trunk/ql/src/test/results/clientpositive/join34.q.out
* /hive/trunk/ql/src/test/results/clientpositive/join35.q.out
* /hive/trunk/ql/src/test/results/clientpositive/join36.q.out
* /hive/trunk/ql/src/test/results/clientpositive/join37.q.out
* /hive/trunk/ql/src/test/results/clientpositive/join39.q.out
* /hive/trunk/ql/src/test/results/clientpositive/join_map_ppr.q.out
* /hive/trunk/ql/src/test/results/clientpositive/lineage1.q.out
* /hive/trunk/ql/src/test/results/clientpositive/load_dyn_part1.q.out
* /hive/trunk/ql/src/test/results/clientpositive/load_dyn_part14.q.out
* /hive/trunk/ql/src/test/results/clientpositive/merge1.q.out
* /hive/trunk/ql/src/test/results/clientpositive/merge2.q.out
* /hive/trunk/ql/src/test/results/clientpositive/merge3.q.out
* /hive/trunk/ql/src/test/results/clientpositive/merge4.q.out
* /hive/trunk/ql/src/test/results/clientpositive/merge_dynamic_partition.q.out
* /hive/trunk/ql/src/test/results/clientpositive/merge_dynamic_partition2.q.out
* /hive/trunk/ql/src/test/results/clientpositive/merge_dynamic_partition3.q.out
* /hive/trunk/ql/src/test/results/clientpositive/merge_dynamic_partition4.q.out
* /hive/trunk/ql/src/test/results/clientpositive/merge_dynamic_partition5.q.out
* /hive/trunk/ql/src/test/results/clientpositive/multi_insert.q.out
* /hive/trunk/ql/src/test/results/clientpositive/multi_insert_move_tasks_share_dependencies.q.out
* /hive/trunk/ql/src/test/results/clientpositive/pcr.q.out
* /hive/trunk/ql/src/test/results/clientpositive/ppd_constant_expr.q.out
* /hive/trunk/ql/src/test/results/clientpositive/quote1.q.out
* /hive/trunk/ql/src/test/results/clientpositive/rand_partitionpruner2.q.out
* /hive/trunk/ql/src/test/results/clientpositive/rcfile_createas1.q.out
* /hive/trunk/ql/src/test/results/clientpositive/rcfile_merge1.q.out
* /hive/trunk/ql/src/test/results/clientpositive/rcfile_merge2.q.out
* /hive/trunk/ql/src/test/results/clientpositive/rcfile_merge3.q.out
* /hive/trunk/ql/src/test/results/clientpositive/rcfile_merge4.q.out
* /hive/trunk/ql/src/test/results/clientpositive/sample1.q.out
* /hive/trunk/ql/src/test/results/clientpositive/sample2.q.out
* /hive/trunk/ql/src/test/results/clientpositive/sample4.q.out
* /hive/trunk/ql/src/test/results/clientpositive/sample5.q.out
* /hive/trunk/ql/src/test/results/clientpositive/sample6.q.out
* /hive/trunk/ql/src/test/results/clientpositive/sample7.q.out
* /hive/trunk/ql/src/test/results/clientpositive/stats0.q.out
* /hive/trunk/ql/src/test/results/clientpositive/stats11.q.out
* /hive/trunk/ql/src/test/results/clientpositive/stats4.q.out
* /hive/trunk/ql/src/test/results/clientpositive/subq.q.out
* /hive/trunk/ql/src/test/results/clientpositive/udf1.q.out
* /hive/trunk/ql/src/test/results/clientpositive/udf_10_trims.q.out
* /hive/trunk/ql/src/test/results/clientpositive/udf_length.q.out
* /hive/trunk/ql/src/test/results/clientpositive/udf_reverse.q.out
* /hive/trunk/ql/src/test/results/clientpositive/union.q.out
* /hive/trunk/ql/src/test/results/clientpositive/union10.q.out
* /hive/trunk/ql/src/test/results/clientpositive/union12.q.out
* /hive/trunk/ql/src/test/results/clientpositive/union18.q.out
* /hive/trunk/ql/src/test/results/clientpositive/union28.q.out
* /hive/trunk/ql/src/test/results/clientpositive/union29.q.out
* /hive/trunk/ql/src/test/results/clientpositive/union30.q.out
* /hive/trunk/ql/src/test/results/clientpositive/union4.q.out
* /hive/trunk/ql/src/test/results/clientpositive/union6.q.out
* /hive/trunk/ql/src/test/results/compiler/plan/case_sensitivity.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/input1.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/input2.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/input3.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/input6.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/input7.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/input9.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/input_testsequencefile.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/sample2.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/sample3.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/sample4.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/sample5.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/sample6.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/sample7.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/subq.q.xml
* /hive/trunk/ql/src/test/results/compiler/plan/union.q.xml
> Race condition in query plan for merging at the end of a query
> --------------------------------------------------------------
>
> Key: HIVE-3302
> URL: https://issues.apache.org/jira/browse/HIVE-3302
> Project: Hive
> Issue Type: Bug
> Components: Query Processor
> Affects Versions: 0.10.0
> Reporter: Kevin Wilfong
> Assignee: Kevin Wilfong
> Priority: Critical
> Attachments: HIVE-3302.1.patch.txt, HIVE-3302.2.patch.txt
>
>
> In the query plan that's used to merge files at the end of a query, the dependency tree looks something like:
> MoveTask(1)
> / \
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> Here MoveTask(1) moves the partition data to a temporary location, and MoveTask(2) moves it to the final location.
> However if there are dynamic partitions generated and some of these partitions are merged and others are moved, the dependency tree is changed at runtime to:
> ...ConditionalTask MoveTask(2)...
> \ /
> MergeTask
> \
> MoveTask(1)
> This produces a race condition between the two MoveTasks where if MoveTask(2) runs before MoveTask(1) the partitions moved by MoveTask(1) will get moved to an intermediate location and never moved to the final location. In this case those partitions are quietly lost.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira