SLT for non-SAP – Considerations
This document is prepared based on version or add-on: DMIS 2010_1_700 SP06.
The sources supported for SLT replication are SAP and non-SAP.
The main differences between the source SAP and non-SAP:
- A Database connection is required to connect from non-SAP source to SLT while Remote Function Call (RFC) connection is required between SAP source and SLT.
- T-code used for database connection is DBACOCKPIT while for RFC SM59 is used.
- Tables can be just initially loaded but cannot automatically be replicated (like addition of new tables/columns after the initial load) and table must have the primary key or Index with elementary data types only for non-SAP and requires manual process to do replication while for SAP source replication happens automatically (through Replicate button).
- One to many replication is not possible for non-SAP but up to 4 is possible for SAP source.
- Read modules are located in the SAP LT Replication Server for non-SAP but for SAP source they reside within source system.
- User role for non-SAP vary depending on non-SAP source while SAP_IUUC_REPL_REMOTE is used for SAP source.
- Secondary database connection is required for connecting from SLT to target HANA system for both SAP and non-SAP sources.
- DD02L and DD02T contains the metadata information for both SAP and non-SAP sources.
- Source system tracks database changes by using database triggers. It records information about changes in the logging tables. Read modules transfer the data from the source system to the SAP LT Replication Server. The relevant data is read from the application tables.
For non-SAP, the supported databases are:
- Sybase ASE
To perform the replication and initially load on non-SAP source, a specific table from a given schema, the database user must have privileges for the following actions:
• Selecting from the table
• Creating a table in the given schema (for creating the logging table)
• Selecting from the logging table
• Deleting the logging table
• Creating database triggers for the table
• Deleting the triggers
• Creating synonyms for the specific table
• Deleting the synonyms
Depending on the specific external database system, the process of granting privileges to a user can vary.
After the initial load process, in order to reflect any changes in the non-SAP source system to HANA system then we need to create a local report with name Z_CREATE_DD_TABLES in the SLT system. After executing the report, DD02L, DD02T will be updated accordingly. For more details please refer to SAP note 1751323 for source code.
Thank you Tobias.
I need HANA table data in my APO system, any idea how to configure it so that both the systems start talking each other. ENH PK 2 is all together a big concept, I need only 1-2 tables in APO
this sounds like a replication to non-HANA, right. From a technical point of view this could be achieved with SLT - but it is not general available so far.
So i can have HANA as a source system, use SLT to replicate the data to Oracle as well if required?
I never tried with HANA as source. Lets see what Tobias says.
Do you really have this requirement or for info?
Yes we have a requirement and depending on Tobias reply we may need to architect our solution.
HANA as a source is working.
Replication to non-ABAP targets
is from a technical point of view possible. But to be honest, the replication to non-ABAP can become pretty complex. So we decided to bring this not as a function to the customer, to avoid any setup issues. Usually we work with our customer in a small project to ensure the setup and educate the customer - afterwards the customer can work alone. If you have the requirement - just drop me a mail and I can share further information.
We have a requirement to pull data out of HANA and into Cassandra. We've been looking at both SLT Replication and Smart Data Access out of HANA. We are currently on Rev 74 of HANA.
Do you know of any supported way to do this at this point in time? If so, what method? If not, do you know of any future plans for this support?
puh, have you checked DataServices?
We have a requirement to provision data from ORACLE to HANA and also from ORACLE to MSSQL using the same SAP SLT system.
Could you please throw some light on this, as how this could be done.
Thanks & Regards,
Oracle to HANA is covered in the product. Replication from Oracle to MS SQL is not part of the product, but we can evaluate with you the feasability within a small PoC. Technically all prerequisites are available but some pieces has to be adjust manually. This is the reason because we cannot brand it "officially released". If you have a strong interest pls reach out to your SAP conctact. Your contact can reach out to me for further details.
Is there any plan to remove this pre-requisite
"table must have the primary key or Index with elementary data types for non-SAP".
We are planning to replicate from smart Ops that has tables with no keys.
Thanks for writing some excellent blogs! keep goin HANA Champ..
Hi - Great overview. Though I have some questions.
Can you detail out more the differences between using an ABAP and Non-Abap source?
I have configured replication using an Oracle e-business application successfully to 7 different schemas, but we are now testing various DBA maintenance activities and have concerns about performance of the triggers on the E-business application. We have been told it is "different": on Abap vs non-Abap...performance and where change data is stored, but need more details. Some people have told us expect 25% performance hit, others said 3-5%, depending on the volume of transactions to the tables replicated.
i have also seen this comment
"As a difference, the read modules are created in the SAP LT Replication Server. The connection the non-SAP source system is established as a database connection"
But not sure how this impacts the source system. Other than pushing more of the processing into the SLT server.
We are very concerned about the impact of upgrades, table reorgs and other changes on the source when the vendor is not at all aware of SLT.
Can you elaborate for me please?
BTW - I have read the article on ASUG 2013 by Poornima Ramachandra on the Maindenform example which was very helpful. Just trying to go a little deeper into the differences. Also need clarification of the comment:
Tables that are truncated and reloaded every night in the source system require a different approach and needs a script to initiate the job. What do you mean about a script to initiate the job?
Hello, I have a doubt:
Is DB/4 supported as NON-ABAP source database?
It seems it is not supported but I'm asking to be sure I'm not missunderstanding anything.
Did you check the PAM (Product Availability Matrix) ?
Thanks & Regards,
Which version of ORACLE is supported if I have to do SLT from ORACLE to HANA?
Source NON-SAP /MSSQL2008 to SLT2.0/DMIS 2011_1_731 (SP11) (SLT running on Linux/Sybase) - Target HANA 1.0 SPS11
Is this OSS note 1751323 still valid for the above environment ??
I have read :
"""After the initial load process, in order to reflect any changes in the non-SAP source system to HANA system then we need to create a local report with name Z_CREATE_DD_TABLES in the SLT system. After executing the report, DD02L, DD02T will be updated accordingly. For more details please refer to SAP note 1751323 for source code.""
Is there any documentation that specifically mentions what steps to done to enable the mapping to be in place in SLT when extracting from a nonSAP data source?
I managed to have the connection right and the job for extraction also runs, but there is some missing piece on the data mapping that I am trying to identify.
Hello, I’ve to replicate data from an MII 15 single stack Java on MSS 2012(11.0.5058) to HANA via SLT. How can I achieve this? With an ABAP stack I've found lots of documents, but with a java stack not. I'm on SLT 2.0 SP12
Have I to do a "Non SAP to HANA" way?