on 03-24-2011 7:25 PM
Hi All,
Looking at the BDLS run logs on our SAP system, looks like only a small number of tables are updated among the tables it checked. For example:
ECC: 90 out of 1537 tables were updated
CRM: 85 out of 1683 tables were updated
BI: (for ECC), 48 out of 830 tables were updated
(for CRM), 86 out of 830 tables were updated
I was thinking if we can specify the tables to be converted, then the BDLS run will be much faster. But then wonu2019t know for sure which tables to specify.
So please suggest is there any way to figure out the table name which will be Updated by BDLS run so that we can mention those table names only in BDLS execution and reduce its execution time.
Regards,
Shivam Mittal
I would run it for the whole range and not selected tables. But you may be able to get the list from test run. Please check performance section of following link.
[http://help.sap.com/saphelp_sm32/helpdata/en/33/c823dbaea911d6b29500508b6b8a93/content.htm]
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Well to find the records that need to be adjusted is one of the main big parts of BDLS. Depending on your data the tables involved are different, and we cannot give you a list.
But it is possible to speed up BDLS, especially if you do it regularly. Check this blog for details: [Execute conversion of logical system names (BDLS) in short time and in parallel - Intermediate|http://www.sdn.sap.com/irj/scn/weblogs?blog=/pub/wlg/4796] [original link is broken] [original link is broken] [original link is broken];
Cheers Michael
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
Having Indexes on the big tables helps a lot.
if you have the resources, create them in parallel with
create index .... compute statistics online nologging parallel 12;
This way the stats are sampled during creation and there is no need for additional stats calculation.
Do sql monitoring while BDLS runs to detect the expensive statements that need adjustment
I have a system down to 35minutes from 16 hours.
I need 45 Minutes to create all indexes beforehand (most of them take 1-3 minutes with parallel 12) with sqlplus.
So it is a 14+ hours gain.
Having noarchive mode during bdls helps, if you have really many record to convert like
on BW systems or with a big workflow usage.
Volker
User | Count |
---|---|
88 | |
23 | |
11 | |
9 | |
8 | |
5 | |
5 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.