on 04-03-2016 9:59 AM
Hi,
I have perhaps over 36 million records in my planning area (sounds a lot but it is not really given the size of the business). All good so far but when it comes to back up (extraction and loading to a cube) this is now a 6 to 9 hours affair, impacting of course productive time of users.
Beyond this basic extraction to DS > DTP to ODS > DTP to Info-cube, is there anything else that people do to manage such data backups? The purpose of this back up is.. yes backup in case of issues in planning area like data loss. So at all points in time i need a snapshot of planning area to quickly rebuild the lost data if any.
How do I accelerate this backup process?
- Will multiple info-packages by some filters work faster ?
- Multiple DTPs?
- Multiple ODS or cubes ?
-Serial ?
-Parallel?
-PP profiles?
What else?
Can the planning area be dumped into some directory? as a csv fil in the background ?
Any latest BW animals in latest versions to handle this ?
Then there is no 'Delta' extraction concept in planning area (even though I am willing to compromise a bit). I have to extract in full everyday and then save the snapshot everyday.. Quite a nuisance if this if this will take like say 24 hours an year later... it might.
Thanks
Lok
Hi Lok
First of all I am not sure why data back would impact productive users.
Once you have through with default macros which you would like to execute before extraction, you should be fine. As far as I know, users can access the data when you are extracting the data from Planning area data source into PSA but still for data consistency it´s good to finish your data loads upfront.
Yes surely you can use multiple info packages loading the data into same target i.e. PSA. The same can be done with DTPs and compress them once you loaded all the requests. Based on available workprocess within the system, you can use maximum parallel workprocess which could still make your data load faster.
Not sure about your scenario but you could also have multiple infocubes based on very concrete selections. Surely SAP HANA could definitely bring huge performance boost to data loads with in SAP BW.
Please do let me know the battle neck process, I might provide you some more hints.
Thanks
Amol
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
12 | |
4 | |
3 | |
2 | |
1 | |
1 | |
1 | |
1 | |
1 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.