[
https://issues.apache.org/jira/browse/HIVE-5664?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14213120#comment-14213120
]
Venki Korukanti commented on HIVE-5664:
---------------------------------------
bq. Proposed in current form, a db containing 1000 tables, we will make 1000
trips to metastore from client while doing cascade drop. That will be
expensive, it will be good to avoid if we can.
In current Hive (without the fix attached to this jira), HiveMetastoreClient is
getting all table names first and for each table name it is getting the
{{Table}} object as part of dropTable() method. So with this fix, we are not
making it any more expensive, but I agree that it is very expensive to make n+1
metastore calls.
bq. I wonder if instead of getAllTables() , if we can use getTables() or
listTableNamesByFilter() to only retrieve base tables and no index tables.
Since index tables follow specific naming pattern, we can construct and pass
that pattern to one of the above methods to get only base tables. This will
avoid multiple round trips to metastore from client.
Index tables can have custom names like in {{CREATE INDEX temp_tbl3_idx ON
TABLE temp_tbl3(id) AS 'COMPACT' with DEFERRED REBUILD IN TABLE
temp_tbl3_idx_tbl;}}
Couple of other alternatives:
1. Failure is because there is no {{Table}} object is found in {{dropTable}}
method for index tables. Simple fix is to ignore {{NoSuchObjectExceptions}}.
2. Add a method {{getTables(TableType)}} to MetaStore interface to retrieve
{{Table}} objects based on table type. We may need to fetch the tables in
batches to avoid memory issues.
> Drop cascade database fails when the db has any tables with indexes
> -------------------------------------------------------------------
>
> Key: HIVE-5664
> URL: https://issues.apache.org/jira/browse/HIVE-5664
> Project: Hive
> Issue Type: Bug
> Components: Indexing, Metastore
> Affects Versions: 0.10.0, 0.11.0, 0.12.0, 0.13.0, 0.14.0
> Reporter: Venki Korukanti
> Assignee: Venki Korukanti
> Attachments: HIVE-5664.1.patch.txt, HIVE-5664.2.patch.txt
>
>
> {code}
> CREATE DATABASE db2;
> USE db2;
> CREATE TABLE tab1 (id int, name string);
> CREATE INDEX idx1 ON TABLE tab1(id) as 'COMPACT' with DEFERRED REBUILD IN
> TABLE tab1_indx;
> DROP DATABASE db2 CASCADE;
> {code}
> Last DDL fails with the following error:
> {code}
> FAILED: Execution Error, return code 1 from
> org.apache.hadoop.hive.ql.exec.DDLTask. Database does not exist: db2
> Hive.log has following exception
> 2013-10-27 20:46:16,629 ERROR exec.DDLTask (DDLTask.java:execute(434)) -
> org.apache.hadoop.hive.ql.metadata.HiveException: Database does not exist: db2
> at
> org.apache.hadoop.hive.ql.exec.DDLTask.dropDatabase(DDLTask.java:3473)
> at org.apache.hadoop.hive.ql.exec.DDLTask.execute(DDLTask.java:231)
> at org.apache.hadoop.hive.ql.exec.Task.executeTask(Task.java:151)
> at
> org.apache.hadoop.hive.ql.exec.TaskRunner.runSequential(TaskRunner.java:65)
> at org.apache.hadoop.hive.ql.Driver.launchTask(Driver.java:1441)
> at org.apache.hadoop.hive.ql.Driver.execute(Driver.java:1219)
> at org.apache.hadoop.hive.ql.Driver.runInternal(Driver.java:1047)
> at org.apache.hadoop.hive.ql.Driver.run(Driver.java:915)
> at
> org.apache.hadoop.hive.cli.CliDriver.processLocalCmd(CliDriver.java:268)
> at org.apache.hadoop.hive.cli.CliDriver.processCmd(CliDriver.java:220)
> at
> org.apache.hadoop.hive.cli.CliDriver.processLine(CliDriver.java:422)
> at
> org.apache.hadoop.hive.cli.CliDriver.executeDriver(CliDriver.java:790)
> at org.apache.hadoop.hive.cli.CliDriver.run(CliDriver.java:684)
> at org.apache.hadoop.hive.cli.CliDriver.main(CliDriver.java:623)
> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
> at
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> at java.lang.reflect.Method.invoke(Method.java:606)
> at org.apache.hadoop.util.RunJar.main(RunJar.java:160)
> Caused by: NoSuchObjectException(message:db2.tab1_indx table not found)
> at
> org.apache.hadoop.hive.metastore.HiveMetaStore$HMSHandler.get_table(HiveMetaStore.java:1376)
> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
> at
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> at java.lang.reflect.Method.invoke(Method.java:606)
> at
> org.apache.hadoop.hive.metastore.RetryingHMSHandler.invoke(RetryingHMSHandler.java:103)
> at com.sun.proxy.$Proxy7.get_table(Unknown Source)
> at
> org.apache.hadoop.hive.metastore.HiveMetaStoreClient.getTable(HiveMetaStoreClient.java:890)
> at
> org.apache.hadoop.hive.metastore.HiveMetaStoreClient.dropTable(HiveMetaStoreClient.java:660)
> at
> org.apache.hadoop.hive.metastore.HiveMetaStoreClient.dropTable(HiveMetaStoreClient.java:652)
> at
> org.apache.hadoop.hive.metastore.HiveMetaStoreClient.dropDatabase(HiveMetaStoreClient.java:546)
> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
> at
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> at java.lang.reflect.Method.invoke(Method.java:606)
> at
> org.apache.hadoop.hive.metastore.RetryingMetaStoreClient.invoke(RetryingMetaStoreClient.java:89)
> at com.sun.proxy.$Proxy8.dropDatabase(Unknown Source)
> at org.apache.hadoop.hive.ql.metadata.Hive.dropDatabase(Hive.java:284)
> at
> org.apache.hadoop.hive.ql.exec.DDLTask.dropDatabase(DDLTask.java:3470)
> ... 18 more
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
