Impala Performance Issue with ODBC Driver – CDH 6.2 / Impala 3.2.0 / SAS 9.4 M7

[#1]

Hello everyone,

I’m currently facing a performance issue with Cloudera Impala when using it through the ODBC driver.

Here is my environment:

• CDH: 6.2

• Impala: 3.2.0 (Impala 2.7.1 ODBC Driver)

• Integration: SAS 9.4 M7 (SAS Access to Impala)

• Authentication: Active Directory (user/mdp)

• Security: SSL enabled

I have configured my /etc/odbc.ini as follows:

[Impala_KRB14]
Description=ODBC Cloudera Impala SSL Node 14
Driver=/opt/cloudera/impalaodbc/lib/64/libclouderaimpalaodbc64.so
HOST=""""""""" ##hidden
PORT=21050
Database=""""""""""" ##hidden
AuthMech=3
SSL=1
TrustedCerts=/etc/pki/ca-trust/extracted/pem/tls-ca-bundle.pem
LogLevel=2
LogPath=/tmp/impala_odbc.log
SSP_BATCH_SIZE=60536
RowsFetchedPerBlock=60536
TSaslTransportBufSize=4000

I also tried adding several ODBC options/tuning as seen in the odbc.ini parameters, but unfortunately I did not see any performance improvement.

The issue is mainly that:

• Opening a table through SAS takes a long time

• Even simple queries like SELECT ... WHERE ... are very slow

I would like to understand:

1. How can I properly tune the Impala ODBC driver to improve query performance?

2. Are there any known issues or recommended configurations for improving performance with SAS 9.4 M7 and Impala 3.2 on CDH 6.2?

3. What logs or debug options can I enable to provide more details if needed?

4. Changing into Impala ODBC Connector 2.8.4 does it really help ?

Any help or guidance would be highly appreciated.

[#2]

@Karma12, Welcome to our community! To help you get the best possible answer, I have tagged in our Impala experts @kbanik @SVB, who may be able to assist you further.

Please feel free to provide any additional information or details about your query. We hope that you will find a satisfactory solution to your question.

[#3]

Thank you very much for trying to help. I’ve been stuck on this subject for the past two weeks, trying to improve queries on Impala.

I’m currently on RHEL 9 and have unixODBC installed. I’ve configured my environment via /etc/profile.d/oracle.sh as follows:

# Oracle Instant Client configuration
export TNS_ADMIN=/opt/oracle/instantclient_12_2/network/admin
export PATH=/opt/oracle/instantclient_12_2:$PATH
export LD_LIBRARY_PATH=/opt/oracle/instantclient_12_2:/usr/lib64:$LD_LIBRARY_PATH

# ODBC configuration
export ODBCINI=/etc/odbc.ini
export ODBCSYSINI=/etc

# Cloudera Impala ODBC configuration
export CLOUDERAIMPALAINI=/opt/cloudera/impalaodbc/lib/64/cloudera.impalaodbc.ini

I’ve tested with: `isql Impala_KRB14 user password` and it connects successfully, but queries are taking too long. I’m looking for guidance on how to optimize these queries effectively.

Here are the version of installed packages:
unixODBC 2.3.9
DRIVERS............: /etc/odbcinst.ini
SYSTEM DATA SOURCES: /etc/odbc.ini
FILE DATA SOURCES..: /etc/ODBCDataSources
USER DATA SOURCES..: /etc/odbc.ini
SQLULEN Size.......: 8
SQLLEN Size........: 8
SQLSETPOSIROW Size.: 8
 
rpm -qa | grep cyrus-sasl
cyrus-sasl-lib-2.1.27-21.el9.x86_64
cyrus-sasl-gssapi-2.1.27-21.el9.x86_64
cyrus-sasl-plain-2.1.27-21.el9.x86_64
cyrus-sasl-2.1.27-21.el9.x86_64