"Embedded-only" Class when running BigBench on Spark

[Anahita Shayesteh]

Hi,

I am running TPCx-BB (downloaded from TPC website) with spark-sql engine. I have downloaded spark-1.6.1-bin-hadoop2.6 binary, since I could not build spark with hive and yarn profiles.

I am running validation test (sale factor 1) and it seems to be running fine, but I get below error for all query runs:

16/06/21 11:05:03 INFO DataNucleus.Datastore: The class "org.apache.hadoop.hive.metastore.model.MFieldSchema" is tagged as "embedded-only" so does not have its own datastore table.

16/06/21 11:05:03 INFO DataNucleus.Datastore: The class "org.apache.hadoop.hive.metastore.model.MOrder" is tagged as "embedded-only" so does not have its own datastore table.

Can I ignore these errors or does it mean there is a problem in my integration of spark with cloudera manager environment?

Thank you in advance for your help.

Anahita 

[Anahita Shayesteh]

There are other warning messages as well. Please see attached for log file of query 7.

Thank you,
Anahita

[Bhaskar Gowda]

Hi Anahita - you can ignore this error if your objective is to only run the test with single process connecting to the Hive metastore. Please setup a  dedicated hivemetastore if you plan to run large scale tests.

refer to apache hive knowledge base for details: https://cwiki.apache.org/confluence/display/Hive/AdminManual+MetastoreAdmin

-Bhaskar

On Tuesday, June 21, 2016 at 11:16:40 AM UTC-7, Anahita Shayesteh wrote:

[Anahita Shayesteh]

Thank you Bhaskar,

How about this warning? Is it related to Hive metastore as well?

16/06/21 12:02:25 WARN metastore.ObjectStore: Failed to get database default, returning NoSuchObjectException

Thank you,

Anahita

[Yan Tang]

hi,  Anahita

You can also ignore this error. 

Best Regards,

Yan

在 2016年6月22日星期三 UTC+8上午5:34:00，Anahita Shayesteh写道：

[Anahita Shayesteh]

Thank you Yan.

Anahita