SAP HANA Extended SQL Executor (ESX) FAQs
1) Define SAP HANA Extended SQL Executor (ESX)?
SAP HANA Extended SQL Executor (ESX) is a front-end execution engine which replaced the row engine. it recovers the database request from session level and assign it to lower-level engines such HEX engine, join engine or calculation engine:
2) What are some important benefits of ESX?
The following are the benefits and components are provided by ESX:
- It minimizes the runtimes and memory requirements by pipelining
- It uses itabs as common format, therefore, making communication with other engines becomes easy.
- It is light-weight and separate library which reduces dependencies from other components
3) How to control the usage of ESX?
Following parameter can be used on a global level in order to activate or deactivate ESX
- indexserver.ini -> [sql] -> esx_level
The user can also USE_ESX_PLAN and NO_USE_ESX_PLAN hints on statement level in order to activate or deactivate ESX
4) How to trace ESX?
Please, following parameters exist to trace ESX:
|indexserver.ini -> [trace] -> esx_dump_plan||debug||Detailed ESX plan information including node operation tree and input itab type information (available with SAP HANA 2.0 >= SPS 03)|
5) What are the various indications that ESX is used for processing database request?
The modules related ESX in the call stack of threads indicate the use of ESX such as:
6) Mention some known issues related to ESX?
The following known issues listed below in the table exist in relation to ESX:
|Termination with unable to update attribute store||2.00.020|
|Crash in ESX::ParallelItabBuildJob::runInternal||2.00.020|
|Termination with NOT NULL constraint violation||<= 2.00.022|
|Crash in ESX::Window::prepareFetchInfo||<= 2.00.022|
|Termination with unknown attribute error||<= 2.00.023|
|Memory Leak in Pool/ESX When Using PlanViz||<= 2.00.023|
|Crash in UnifiedTable::Alphanum::set||<= 2.00.024.01
7) How can we configure ESX?
Please use these following SAP HANA parameters in order adjust the ESX behaviour:
indexserver.ini -> [sql] -> esx_level
Activation (1) or deactivation (0) of ESX
indexserver.ini -> [sql] -> use_esx_project
If set to true, ESX can write search results directly into a stream without using an itab