[jira] [Created] (HIVE-19883) QTestUtil: initDataset() can be affected by the settings of the previous test

["Jason Dere (JIRA)" <ji...@apache.org>]

Jason Dere created HIVE-19883:
---------------------------------

             Summary: QTestUtil: initDataset() can be affected by the settings of the previous test
                 Key: HIVE-19883
                 URL: https://issues.apache.org/jira/browse/HIVE-19883
             Project: Hive
          Issue Type: Bug
          Components: Testing Infrastructure
            Reporter: Jason Dere


Tried creating a test that set metastore.create.as.acid/hive.create.as.insert.only, and I found that the built-in table default.src was being created as an insert-only transactional table, which will cause errors in other tests that do not set the TxnManager to one that supports transactional tables.

It appears that initDataset() uses the old CliDriver that was used for the previous test, which has any settings used during that test:
{noformat}
java.lang.Exception: Creating src
        at org.apache.hadoop.hive.ql.exec.DDLTask.createTable(DDLTask.java:4926) [hive-exec-4.0.0-SNAPSHOT.jar:4.0.0-SNAPSHOT]
        at org.apache.hadoop.hive.ql.exec.DDLTask.execute(DDLTask.java:428) [hive-exec-4.0.0-SNAPSHOT.jar:4.0.0-SNAPSHOT]
        at org.apache.hadoop.hive.ql.exec.Task.executeTask(Task.java:205) [hive-exec-4.0.0-SNAPSHOT.jar:4.0.0-SNAPSHOT]
        at org.apache.hadoop.hive.ql.exec.TaskRunner.runSequential(TaskRunner.java:97) [hive-exec-4.0.0-SNAPSHOT.jar:4.0.0-SNAPSHOT]
        at org.apache.hadoop.hive.ql.Driver.launchTask(Driver.java:2659) [hive-exec-4.0.0-SNAPSHOT.jar:?]
        at org.apache.hadoop.hive.ql.Driver.execute(Driver.java:2311) [hive-exec-4.0.0-SNAPSHOT.jar:?]
        at org.apache.hadoop.hive.ql.Driver.runInternal(Driver.java:1982) [hive-exec-4.0.0-SNAPSHOT.jar:?]
        at org.apache.hadoop.hive.ql.Driver.run(Driver.java:1683) [hive-exec-4.0.0-SNAPSHOT.jar:?]
        at org.apache.hadoop.hive.ql.Driver.run(Driver.java:1677) [hive-exec-4.0.0-SNAPSHOT.jar:?]
        at org.apache.hadoop.hive.ql.reexec.ReExecDriver.run(ReExecDriver.java:157) [hive-exec-4.0.0-SNAPSHOT.jar:4.0.0-SNAPSHOT]
        at org.apache.hadoop.hive.ql.reexec.ReExecDriver.run(ReExecDriver.java:218) [hive-exec-4.0.0-SNAPSHOT.jar:4.0.0-SNAPSHOT]
        at org.apache.hadoop.hive.cli.CliDriver.processLocalCmd(CliDriver.java:239) [hive-cli-4.0.0-SNAPSHOT.jar:?]
        at org.apache.hadoop.hive.cli.CliDriver.processCmd(CliDriver.java:188) [hive-cli-4.0.0-SNAPSHOT.jar:?]
        at org.apache.hadoop.hive.cli.CliDriver.processLine(CliDriver.java:402) [hive-cli-4.0.0-SNAPSHOT.jar:?]
        at org.apache.hadoop.hive.cli.CliDriver.processLine(CliDriver.java:335) [hive-cli-4.0.0-SNAPSHOT.jar:?]
        at org.apache.hadoop.hive.ql.QTestUtil.initDataset(QTestUtil.java:1277) [classes/:?]
        at org.apache.hadoop.hive.ql.QTestUtil.initDataSetForTest(QTestUtil.java:1259) [classes/:?]
        at org.apache.hadoop.hive.ql.QTestUtil.cliInit(QTestUtil.java:1328) [classes/:?]
        at org.apache.hadoop.hive.cli.control.CoreCliDriver.runTest(CoreCliDriver.java:176) [classes/:?]
        at org.apache.hadoop.hive.cli.control.CliAdapter.runTest(CliAdapter.java:104) [classes/:?]
        at org.apache.hadoop.hive.cli.TestMiniLlapLocalCliDriver.testCliDriver(TestMiniLlapLocalCliDriver.java:59) [test-classes/:?]
{noformat}

A new CliDriver is created for the new test, but only after we've created the dataset tables for the next test (see the line numbers for QTestUtil.cliInit() in both stack traces).
{noformat}
CliSessionState(SessionState).getConf() line: 317	
CliDriver.<init>() line: 110	
QTestUtil.cliInit(File, boolean) line: 1360	
CoreCliDriver.runTest(String, String, String) line: 176	
CoreCliDriver(CliAdapter).runTest(String, File) line: 104	
TestMiniLlapLocalCliDriver.testCliDriver() line: 59	
{noformat}

I think fix is to move the creation of the new CliDriver higher up in QTestUtil.cliInit(), before we call initDataset().



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)