Impala Queries with Where Clauses/Other Conditions Perform Type Conversions on "null" Data Causing Timeouts, Excessive Run Time and Unmanageable Server errors on the BDA (Doc ID 2054680.1)

Last updated on SEPTEMBER 11, 2015

Applies to:

Big Data Appliance Integrated Software - Version 4.2.0 and later
Linux x86-64

Symptoms

On BDA V4.2/CDH 5.4.0 running an impala query with a where clause or other condition like:

select * from <table> where id=<value>;

has the following symptoms:

1. In the Impala Query Editor in Hue, "Unmanageable server error occurred: undefined" may be raised.

2. In the Impala Query Editor in Hue, the query may timeout.

3. Running the query directly in the Impala shell may appear to take forever to complete.

4. On Node 4 the node where the Impala daemon resides, /var/log/impalad/impalad.INFO contains hundreds of lines like below:

Error converting column: <X> to DOUBLE (Data is: null)

It is also observed that:

1. Running the same query without a where clause or other condition is successful.

2. Only queries on certain tables have these results.  On other tables the same query is successful.

3. All such failing queries run successfully in beeline.

4. The table is imported weekly from an Oracle DB via Sqoop.


Cause

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms