Implementation and Change Management
Q: What kind of structural changes are permitted if a cube/DSO has sent data to NLS already?
Please refer to http://service.sap.com/sap/support/notes/1005040.
Q: How do BW and NLS tables sync up again after an InfoProvider has been extended?
Adding a new field to an InfoProvider deactivates an associated Data Archiving Process (DAP). Then reactivating the DAP adds the new field to the NLS table. Please note that the added field is empty. If the added field should be filled you have to reload the data, update the new field and archive the data again. DAP activation can be included as part of the transport to adjust QA and Production systems.
Q: What happens when a data archiving process gets deleted in BW?
When a DAP is deleted in BW, the corresponding tables and all archiving requests are deleted from the Sybase IQ database.
Q: What happens when an InfoProvider with an associated Data Archiving Process (DAP) gets deleted?
If archiving has not been executed before, the DAP and corresponding tables from IQ will be deleted.
On the other hand, if archiving has been executed previously, the DAP as well as previously archived data on Sybase IQ remain in place for future usage.
Q: What to consider when transporting Data Archive Processes and associated NLS connections?
In general you should use the same logical name for your NLS connection in all systems throughout your transport landscape pointing to the particular NLS target database (such as Sybase IQ) per system. Then you can transport the DAPs and generate the archiving requests per system. There is no automatic conversion of NLS connection names in BW transport management (like for BW source systems).
Q: Can I archive data from a DataStore Object that has no time characteristic included?
It is not possible to archive data in an InfoProvider that has no time characteristics / date field at all. It is currently not possible to select the flag “Free selection for archiving” in the Data Archiving Process.
Reporting
Q: Is it possible to use SAP BusinessObjects Explorer on BW generated views, including NLS data?
In order to achieve this, we recommend to use the HANA Model Generation for BW InfoProviders provided with BW 7.40 SP8. These generated HANA Views on top of BW InfoProviders can also access NLS archived data.
For more information on HANA Model generation with BW 7.40, see http://scn.sap.com/docs/DOC-52790 and the respective documentation: http://help.sap.com/saphelp_nw74/helpdata/en/66/33d851345c4770bd4e523701b9f5b0/content.htm?frameset=/en/66/33d851345c4770bd4e523701b9f5b0/frameset.htm (the documentation is on the information level of the latest SP released)
Before BW 7.40 SP5 (and since BW 7.30 SP8), you could use the BW model import wizard in the HANA Studio to generate a HANA view based on the BW InfoProvider metadata directly on the BW tables. These generated HANA view are not able to access NLS archived data. In that case a workaround is possible that involves a HANA Calculation View (union) on top of that generated HANA View and a HANA View based on virtual table on the NLS table in Sybase IQ (using HANA Smart Data Access). As of the complexity of that workaround we strongly recommend the usage of the HANA Model Generation with BW 7.40 SP5.
Q: What's new in SAP BW 7.4 SP8/SP9 with regards to BW Query Execution on archived data?
As of BW 7.40 SP8 and HANA SP08, BW Queries can use the so-called so called "HANA-API" to read from archived data. This means that more query features can be pushed down to HANA to reduce post-processing in the ABAP server. This can have a big effect on the query-runtime especially when restricted keyfigures are used. Technically speaking, as of SP8 the features of query-mode 2 can be used to read from archived data if the query mode 2,3 or 6 is used in the query setting. The query modes are explained in this blog: https://blogs.saphana.com/2013/06/10/bw-on-hana-and-the-query-execution-mode/
As of BW 7.40 SP8, InfoCubes with non-cumulative key figures can be archived, too. HANA Smart Data Access isrequired for query processing. More information can be found at SAP Help:
Note 2165650 - FAQ: BW Near-Line Storage with HANA Smart Data Access
Data Loading
Q: What happens if an NLS job is running and a load DTP also kicks off?
It is not possible to load data during archiving, especially during the delete step. It is a best practice to build a process step within the process chain for data achieving, so that data loads will not conflict with the archiving process.
Please also refer to the documentation provided here: http://help.sap.com/saphelp_nw73/helpdata/EN/4a/40bf4197ea1d0fe10000000a42189c/frameset.htm
Q: Will there be duplicate records between objects stored in the NLS and BW Objects at any time?
Sybase IQ uses snapshot versioning: a write operation cannot block a read operation or vice versa. Readers will see the data valid at the beginning of their extract, even if more recent changes have been applied and committed in the meanwhile. There is a time window around the archiving process during which data exists in two identical instances (once in the primary BW DB, once in NLS). BW is be able to correctly distribute the search criteria so they are restricted that only one instance is respected.
Q. Is it possible to parallelize the DData Loadingata Archiving Process (DAP) request in SAP NetWeaver BW?
Currently the parallel processing option for DAP requests is not supported at batch process level (transaction code: RSBATCH). However, there is a work around based on data modeling level: By splitting the archiving request into several data archiving processes and scheduled them in parallel in a Process Chain, it is possible to achieve parallelism. See the workaround description
Monitoring
Q: How can Nearline Storage Tables and Partitions be accessed and managed from BW?
In order to access and manage partitions on Sybase IQ from BW, SAP provides the program “RSDA_SYB_PARTITION_MONITOR” which is the “Partition Monitor for Nearline Table in Sybase IQ”. This report provides a detailed overview of all Nearline Tables in Sybase IQ and the corresponding partitions for each table.
Since the number of partitions for a data table on Sybase IQ is limited to 1024, the program allows users to define threshold values in order to create alerts and monitor all critical tables which exceed the upper threshold value. In this case the report can be also used to drop or/and merge the individual partitions. See detailed example.
Q: How can I check performance for NLS query access as well as how many records have been read from NLS during reporting?
You can check it with the query monitor in BW (transaction RSRT). Please enter your respective BW Query --> “Execute + Debug” --> Others > check “Display Statistics Data” (and “Do Not Use Cache” – optional). Then, you can check how many records were read from NLS and how long it took for reading form NLS.
Suffix $X reflects the active data and $N reflects Near-line Storage (Sybase IQ).
System Operations and Set Up
Q: Can one NLS system connect to multiple BW instances?
Yes. One Sybase IQ server can talk to multiple BW systems. You should create separate databases in IQ for the different BW systems.