on 01-22-2014 12:03 AM
Hi,
got some fast growing InfoCubes, set up to small.
They get self repartioned and sef distributed by the HANA DBMS System ?
How does the repartitioning / blade distribution work at all ?
...and is controlled by what and which parameter ?
....or do i need to do repartion and redestibute the tables manually by Landscape Redistribution ?
ThanXs
Martin
...and please, please, please don't tell me to use archiving / nearl line storage....got simply too many infocubes,
but much memory...it's a question on managebility ....
Hello Gabor,
this means for SAP BW :
-the table TABLE_PLACEMENT "replaces" -in a formal sense- the former possibility of the
RSA1 / Modelling / InfoProvider Settings/ Zusätze / DB Performance / Mainatin DB Storage Parameters ?
-the table TABLE_PLACEMENT must be managed manually ?
-execptions for small / very large tables must be defined manually ?
-the Landscape Redistibution must be pushed manually from time to time ?
-how to predict the durartion of redistribution (with intense ressource consumption
and unknow number of affected tables) in a productive system ?
-or otherwise spoken: maneging the table TABLE_PLACEMENT is one of the key challenges
in a SAP BW HANA / HANA System to maintain the productivity & performance due to the impact
of reducing cross blade joins and equal distibution of access to the table of the baldes ?
ThankXs
Martin
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Martin,
The DB Performance is still valid, e.g. you can execute the partitioning from there as well (see e.g.:
Partitioning - Modeling - SAP Library ).
The TABLE_PLACEMENT needs to be maintained manually as well as the execution of the redistribution.
Regards,
Gabor
Hi Martin,
When it comes to partioning the infoproviders in BW on HANA the partioning is no more supported for cubes thats why its greyed out. But in case of DSO you can do partioning .
Partitioning: The fact table of a HANA-optimized InfoCube can (and must) no longer be partitioned
semantically, i.e. using a time characteristic to split the data into disjoint partitions. Query
performance is fast even without such partitions on very large fact tables. However, overall 4
different partitions are created automatically for every fact table: Partition 1 for non-compressed
Requests, Partition 2 for compressed Requests, Partition 3 for reference points of the inventory data,
and Partition 4 for so-called historic movements of inventory data. Partitions 3 and 4 are created also
if the InfoCube does not contain any inventory/non-cumulative keyfigure, they are then always empty.
For more details http://www.saphana.com/servlet/JiveServlet/previewBody/1363-102-2-1810/SDN_HANA_opt_InfoCube%20FINAL...
Let me know if you need any further details.
Thanks & Regards
A.Dinesh
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
94 | |
11 | |
11 | |
10 | |
9 | |
8 | |
6 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.