hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Siddharth Seth (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HIVE-11786) Deprecate the use of redundant column in colunm stats related tables
Date Tue, 06 Oct 2015 21:48:27 GMT

    [ https://issues.apache.org/jira/browse/HIVE-11786?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14945892#comment-14945892
] 

Siddharth Seth commented on HIVE-11786:
---------------------------------------

This patch seems to have caused a pretty big performance regression on the client.

With the patch
{code}
2015-10-06T17:25:01,133 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogBegin(117))
- <PERFLOG method=get_aggr_stats_for from=org.apache.hadoop.hive.metastore.RetryingHMSHandler>
2015-10-06T17:25:21,237 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogEnd(157))
- </PERFLOG method=get_aggr_stats_for start=1444166701133 end=1444166721237 duration=20104
from=org.apache.hadoop.hive.metastore.RetryingHMSHandler threadId=0 retryCount=0 error=false>
2015-10-06T17:25:21,248 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogBegin(117))
- <PERFLOG method=get_aggr_stats_for from=org.apache.hadoop.hive.metastore.RetryingHMSHandler>
2015-10-06T17:25:41,074 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogEnd(157))
- </PERFLOG method=get_aggr_stats_for start=1444166721248 end=1444166741074 duration=19826
from=org.apache.hadoop.hive.metastore.RetryingHMSHandler threadId=0 retryCount=0 error=false>
2015-10-06T17:25:41,170 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogBegin(117))
- <PERFLOG method=get_aggr_stats_for from=org.apache.hadoop.hive.metastore.RetryingHMSHandler>
2015-10-06T17:26:59,585 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogEnd(157))
- </PERFLOG method=get_aggr_stats_for start=1444166741170 end=1444166819585 duration=78415
from=org.apache.hadoop.hive.metastore.RetryingHMSHandler threadId=0 retryCount=0 error=false>
2015-10-06T17:27:01,439 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogBegin(117))
- <PERFLOG method=get_aggr_stats_for from=org.apache.hadoop.hive.metastore.RetryingHMSHandler>
2015-10-06T17:27:01,662 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogEnd(157))
- </PERFLOG method=get_aggr_stats_for start=1444166821439 end=1444166821662 duration=223
from=org.apache.hadoop.hive.metastore.RetryingHMSHandler threadId=0 retryCount=0 error=false>
{code}

Without the patch.
{code}
2015-10-06T17:04:51,205 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogBegin(117))
- <PERFLOG method=get_aggr_stats_for from=org.apache.hadoop.hive.metastore.RetryingHMSHandler>
2015-10-06T17:04:51,440 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogEnd(157))
- </PERFLOG method=get_aggr_stats_for start=1444165491205 end=1444165491439 duration=234
from=org.apache.hadoop.hive.metastore.RetryingHMSHandler threadId=0 retryCount=0 error=false>
2015-10-06T17:04:51,450 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogBegin(117))
- <PERFLOG method=get_aggr_stats_for from=org.apache.hadoop.hive.metastore.RetryingHMSHandler>
2015-10-06T17:04:51,655 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogEnd(157))
- </PERFLOG method=get_aggr_stats_for start=1444165491450 end=1444165491655 duration=205
from=org.apache.hadoop.hive.metastore.RetryingHMSHandler threadId=0 retryCount=0 error=false>
2015-10-06T17:04:51,751 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogBegin(117))
- <PERFLOG method=get_aggr_stats_for from=org.apache.hadoop.hive.metastore.RetryingHMSHandler>
2015-10-06T17:04:52,355 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogEnd(157))
- </PERFLOG method=get_aggr_stats_for start=1444165491751 end=1444165492355 duration=604
from=org.apache.hadoop.hive.metastore.RetryingHMSHandler threadId=0 retryCount=0 error=false>
2015-10-06T17:04:54,224 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogBegin(117))
- <PERFLOG method=get_aggr_stats_for from=org.apache.hadoop.hive.metastore.RetryingHMSHandler>
2015-10-06T17:04:54,624 INFO  [main([])]: log.PerfLogger (PerfLogger.java:PerfLogEnd(157))
- </PERFLOG method=get_aggr_stats_for start=1444165494224 end=1444165494624 duration=400
from=org.apache.hadoop.hive.metastore.RetryingHMSHandler threadId=0 retryCount=0 error=false>
{code}

This is running against a MySQL database with a client-embedded metastore.

Are there specific log lines which would help to debug this ? I have logs with debug logging
for org.apache.hadoop.hive.metastore

> Deprecate the use of redundant column in colunm stats related tables
> --------------------------------------------------------------------
>
>                 Key: HIVE-11786
>                 URL: https://issues.apache.org/jira/browse/HIVE-11786
>             Project: Hive
>          Issue Type: Bug
>          Components: Metastore
>            Reporter: Chaoyu Tang
>            Assignee: Chaoyu Tang
>             Fix For: 1.3.0, 2.0.0
>
>         Attachments: HIVE-11786.1.patch, HIVE-11786.1.patch, HIVE-11786.2.patch, HIVE-11786.patch
>
>
> The stats tables such as TAB_COL_STATS, PART_COL_STATS have redundant columns such as
DB_NAME, TABLE_NAME, PARTITION_NAME since these tables already have foreign key like TBL_ID,
or PART_ID referencing to TBLS or PARTITIONS. 
> These redundant columns violate database normalization rules and cause a lot of inconvenience
(sometimes difficult) in column stats related feature implementation. For example, when renaming
a table, we have to update TABLE_NAME column in these tables as well which is unnecessary.
> This JIRA is first to deprecate the use of these columns at HMS code level. A followed
JIRA is to be opened to focus on DB schema change and upgrade.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message