cancel
Showing results for 
Search instead for 
Did you mean: 

Distubution of BDLS run between work process

Former Member
0 Kudos

Hi all,

We have the R/3 4.6c with 4.2 TB and BW 7.0 with 7.4 TB , during the refresh from prod to quality the BDLS steps takes nealry 24 to 40 hrs . We have break down the tables and kicking off mutliple BDLS jobs like Z * tables in BDLS run and other set of tables in another run so that these jobs will run parallel . During this BDLS conversion only one background process

is running for one BDLS run is there any way to make the load distubution among available background process to so that the conversion will complete sooner.

For example : In R/3 i have trigger one BDLS run for CATSDB table in background this will occupy only one background process, how can we make to use of available all background process to run for CATSDB so that conversion will be completed soon .

Please suggest..

Thanks,

Subhash.G

Accepted Solutions (0)

Answers (2)

Answers (2)

Former Member
0 Kudos

Hi,

First of all how many back ground processed you have in your system.

Are you releaseing the jobs in parallel.

I shall suggest you not to run parallel Jobs for BDLS as you can have error as you dont have proper knowledgr that the table are interlinked or not.

So please run One BDLS program to convert all the tables because if you miss some you shall be in problem.

Thanks

Rishi Abrol

Former Member
0 Kudos

Thanks

Former Member
0 Kudos

There is nothing wrong is running BDLS jobs in parallel. We were facing a similar situation in a project a few years back and then I had devised a solution after taking inputs from various sources and usage of parallel processing was one of steps of achieving BDLS faster. Parallel processing is not an issue at all. However parallel processing needs to be done smartly. Bigger tables ought to be excluded and for them dedicated processes can be used. For the rest you can tables for bracket them like A* to D* ,E* to F* and so on making sure you are exclusing the larger tables. However usage of temporary indexes for big tables where not all the fields need to be filled in can be helpful. The index needs to include the field for logical system .No point in creating temporary indexes for tables where the logical system field is always full since anyways a complete table scan would be performed. Create indexes for 6-7 largest tablest of this kind. Of course increase thnumber of background processes as well temporarily.

Regards.

Ruchit.

Former Member
0 Kudos

Hi Subhash

Check this [Link|http://www.erpgenie.com/sap-technical/basis/bdls-taking-a-long-time-to-run] may help you

Regards

Uday