About EricL

EricL · ‎10-26-2017

What's the issue in impala? Impala don't see the new columns? Did you update table to include the new columns through Beeline? Have you validated metadata in Impala?

EricL · ‎10-26-2017

You will need to check both HS2 log and Spark application log to get the real error message. "Failed to create spark client" is too generic and it can be anything.

EricL · ‎10-26-2017

You might want to try map parquet tables by index, rather than column name: SET parquet.column.index.access=true;

EricL · ‎10-25-2017

Hmm, For JDBC and ODBC drivers, there is a setting called RowsFetchedPerBlock, the default value is 10,000. For 9.2M rows, this back and forth might cause the delay. Can you try to increase this value to see if it can help? See Doc below: http://www.cloudera.com/documentation/other/connectors/impala-jdbc/latest/Cloudera-JDBC-Driver-for-Impala-Install-Guide.pdf and search for RowsFetchedPerBlock.

EricL · ‎10-25-2017

Can you please provide cloudera.impalaodbc.ini file for review?

EricL · ‎10-25-2017

How big is the file? And if you scp the result file within the cluster, how long does it take? This is to confirm if there are speed differences between within the cluster and across to the outside of cluster. If there is, how much different.

EricL · ‎10-25-2017

If you check the hue.ini file, which is the configuration for Hue, you can find below impala section: [impala] server_host={impala-daemon-host-url} server_port=21050 impala_principal=impala/{impala-daemon-host-url} impersonation_enabled=True server_conn_timeout=120 You can see that to update Hue to connect to a different impala, you can update value for server_host.

EricL · ‎10-25-2017

Do your users also use Hue to run Impala queries? Hue will hold the query handler open so that it can do paginations, and it will only kill the handler after user navigates away from the impala page. If user stays on the page, the handler will be kept open and the query is considered as in flight. Please confirm Hue usage first. Thanks

EricL · ‎10-19-2017

Glad that we identified the issue!

EricL · ‎10-18-2017

If you are using CDH, you should already have parquet-tools command line available. If not, you can get it from here: https://github.com/apache/parquet-mr/tree/master/parquet-tools You just need to run: parquet-tools cat /path/to/parquet-file.parq Please note that the path is on local file system, not in HDFS. And then you can grep the output to search for string "error_message".

Online	Offline
Last Visited	‎08-12-2020 03:17 AM

Member Since	‎03-23-2015 01:24 PM
Last Visited	‎08-12-2020 03:17 AM
Posts	1,288
Kudos received	113

Cloudera Community

Re: max() function generating an error in sqoop

Re: Add a dynamic variable to a Hive view

Re: Hive Server 2 failing to start CDP ,Cloudera M...

Re: Sqoop export from hive to teradata - > issue ...

Re: Cloudera Hadoop internal workings

Re: schema incompatible issue in impala

Re: Hive on Spark CDH 5.7 - Failed to create spark...

Re: Hive select when column name different in parq...

Re: Impala ODBC/JDBC bad performance - rows fetch ...

Re: Impala ODBC connectivity error

Re: Impala ODBC/JDBC bad performance - rows fetch ...

Re: How does Hue decide Impala Node as Coordinator

Re: Impala Queries Executing long time

Re: error_message : You have exceeded your daily r...

Re: error_message : You have exceeded your daily r...