Re: Precommit result failures snowballing ?

[Thejas Nair]

Some recent failures -

org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver_auto_join_stats2
org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver_index_auto_partitioned
org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver_stats_filemetadata
org.apache.hadoop.hive.cli.TestEncryptedHDFSCliDriver.testCliDriver_encryption_insert_partition_dynamic
org.apache.hadoop.hive.cli.TestMiniLlapCliDriver.testCliDriver_stats_filemetadata
org.apache.hadoop.hive.cli.TestNegativeCliDriver.testNegativeCliDriver_authorization_uri_import
org.apache.hadoop.hive.cli.TestNegativeCliDriver.testNegativeCliDriver_columnstats_partlvl_multiple_part_clause
org.apache.hadoop.hive.ql.exec.spark.session.TestSparkSessionManagerImpl.testMultiSessionMultipleUse
org.apache.hadoop.hive.ql.exec.spark.session.TestSparkSessionManagerImpl.testSingleSessionMultipleUse
org.apache.hadoop.hive.ql.security.authorization.plugin.TestHiveOperationType.checkHiveOperationTypeMatch
org.apache.hive.jdbc.TestSSL.testSSLVersion
org.apache.hive.spark.client.TestSparkClient.testAddJarsAndFiles
org.apache.hive.spark.client.TestSparkClient.testCounters
org.apache.hive.spark.client.TestSparkClient.testErrorJob
org.apache.hive.spark.client.TestSparkClient.testJobSubmission
org.apache.hive.spark.client.TestSparkClient.testMetricsCollection
org.apache.hive.spark.client.TestSparkClient.testRemoteClient
org.apache.hive.spark.client.TestSparkClient.testSimpleSparkJob
org.apache.hive.spark.client.TestSparkClient.testSyncRpc


On Mon, Dec 21, 2015 at 10:44 AM, Thejas Nair  wrote:
> The number of tests that fail in precommit seems to be on the rise.
> The larger the number of seemingly unrelated failures to a patch, the
> higher are the chances of missing out on actual failures caused by a
> the patch, resulting in further increase in failures.
>
> I think we need to give this some attention soon. Should we create an
> umbrella jira to track the current set ?
>
> Are there any bugs already open to track them ? I didn't see any when
> i searched by component test/tests.
>
> Thanks,
> Thejas

Precommit result failures snowballing ?

[Thejas Nair]

The number of tests that fail in precommit seems to be on the rise.
The larger the number of seemingly unrelated failures to a patch, the
higher are the chances of missing out on actual failures caused by a
the patch, resulting in further increase in failures.

I think we need to give this some attention soon. Should we create an
umbrella jira to track the current set ?

Are there any bugs already open to track them ? I didn't see any when
i searched by component test/tests.

Thanks,
Thejas