Hi All,
I was in a discussion to highlight the difference between Connector stages and Enterprise stages. With the extent that I worked in Oracle Enterprise stage, I replied saying the Enterprise stages are Sequential. Then there was a followup question as how to make Oracle Enterprise stage to run in Parallel. I wondered how can we change any default execution mode for any stage. Please let me know your opinion as you had come across changing the Oracle Enterprise stage to run in parallel mode or is it possible in some way at all.
Thanks
Oracle Enterprise stage in parallel mode?
Moderators: chulett, rschirm, roy
-
priyadarshikunal
- Premium Member
- Posts: 1735
- Joined: Thu Mar 01, 2007 5:44 am
- Location: Troy, MI
Generalized statement that the oracle enterprise stage runs in sequential mode is incorrect. The main difference between Connectors and enterprise stage is the library and drivers being used and the options provided by the stage.
If you check the execution of enterprise, Loads are generally parallel unless you set APT_ORACLE_LOAD_OPTIONS to PARALLEL=FALSE, In this case you must set the stage to run in sequential mode on the Advanced tab of the Stage page.
In case of reads its sequential unless you specify Partition Table property, which causes parallel execution on the processing nodes containing a
partition derived from the named table
If you check the execution of enterprise, Loads are generally parallel unless you set APT_ORACLE_LOAD_OPTIONS to PARALLEL=FALSE, In this case you must set the stage to run in sequential mode on the Advanced tab of the Stage page.
In case of reads its sequential unless you specify Partition Table property, which causes parallel execution on the processing nodes containing a
partition derived from the named table
Priyadarshi Kunal
Genius may have its limitations, but stupidity is not thus handicapped.
Genius may have its limitations, but stupidity is not thus handicapped.
i second that..
As a general rule, Datastage Enterprise Edition automatically combines pipelining, partitioning and parallel processing. The concept is hidden from a Datastage programmer. The job developer only chooses a method of data partitioning and the Datastage EE engine will execute the partitioned and parallelized processes.
Enterprise Edition (EE) introduces Parallel jobs. Parallel jobs implement the scalable and parallel data processing mechanisms.
As a general rule, Datastage Enterprise Edition automatically combines pipelining, partitioning and parallel processing. The concept is hidden from a Datastage programmer. The job developer only chooses a method of data partitioning and the Datastage EE engine will execute the partitioned and parallelized processes.
Enterprise Edition (EE) introduces Parallel jobs. Parallel jobs implement the scalable and parallel data processing mechanisms.
Thanks for the couple of responses.
1. I am not still clear on the DEFAULT execution mode of Oracle Enterprise stage or any Enterprise as per the response. Are you implying that it is SEQUENTIAL for read and PARALLEL for write?
2. Does the default mode vary for any stage whether it is read or write based on your response? Long time ago, when I was checking Oracle Enterprise on read side for some checking purpose, I saw the execution mode was Sequential and it was greyed out too so I couldnt try changing. So, I got the understanding as Enterprise stages used on source/target side are Sequential by default.
Note: I am not refering to any load feature in which case only, I guess APT_ORACLE_LOAD_OPTIONS is used.
1. I am not still clear on the DEFAULT execution mode of Oracle Enterprise stage or any Enterprise as per the response. Are you implying that it is SEQUENTIAL for read and PARALLEL for write?
2. Does the default mode vary for any stage whether it is read or write based on your response? Long time ago, when I was checking Oracle Enterprise on read side for some checking purpose, I saw the execution mode was Sequential and it was greyed out too so I couldnt try changing. So, I got the understanding as Enterprise stages used on source/target side are Sequential by default.
Note: I am not refering to any load feature in which case only, I guess APT_ORACLE_LOAD_OPTIONS is used.
Here are the answers.
1. Yes. By default READ method is sequential unless your table is partitioned and you are enforcing stage to read in parallel from partitions. Write always in parallel bydefault.
2. Bydefault stage will try to read in sequential while extraction and in parallel while loading. If you want to write in sequential, then you need to change the execution mode on writing stage to sequential.
1. Yes. By default READ method is sequential unless your table is partitioned and you are enforcing stage to read in parallel from partitions. Write always in parallel bydefault.
2. Bydefault stage will try to read in sequential while extraction and in parallel while loading. If you want to write in sequential, then you need to change the execution mode on writing stage to sequential.
--
Swathi Ch
Swathi Ch