How Package Splitter Work?

[former_member182967]

Hello all,

Can somebody detailed explain how the package splitter working in migration and what is the benefit.

And also, does the splitter parameters affect the package splitter?

Thanks a lot for your knowledge sharing.

Best Regards,

Tong Ning

[Former Member · Accepted solution]

Hello Tong,

Standard systm copy process, all tables of the SAP system are grouped into packages,whereby all tables vd the same data class belong to the same package. The processing unit of unload/load process is a packge.For the system copy, more and smaller R3load packages need to be created during the export, than would generally be the case. To do this, you need to split the STR and EXT files. By not splitting the STR files you end up with the 16 default SAP exports which can be quite large. Each export file is loaded by an individual r3load process. Splitting the STR files allows you to break out the largest tables into their own r3load process.

You can find information about that in the System Copy Guide. An example of the improvement of performing a split can be seen in the figure below. The idea demonstrated in the figure applies for both package and table splitting.

Regards

S.BSIS

[former_member182967 · Accepted solution]

Hello S Basis,

Firstly, thanks for your quick response!

As far as I know, the package splitter can be started manually (See from Package Splitter Users' Guide). I am some confusing about the "Splitting Tool Parameters", does the parameter 'Split packages with size more than' just automatic call the Package Splitter tools?

Wait for your reply, Thanks!

Best Regards,

Tong Ning

[Former Member · Accepted solution]

Hello Tang,

The tool SPLITSTR.PL is used to split the STR files. It is unpacked into the installation directory as of Release 4.6B.
It can be called with the following options:
[-top <count> | -limit <size> | -package <size>] [-p <path>] [<file(s)>
| -d <path>]

-top <count>

transfers the <count> largest tables into separate STR files. <count> specifies the number of the largest tables that are to be unloaded in separate processes. This is the option that is used most frequently.
-top 5 is the default value, that is, if nothing else is specified, the 5 largest tables are transferred into separate STR files.

-limit <size>

All tables that are larger than the <limit> are transferred to separate STR files. You can use k, M, G or T as a unit. With -limit 4G the system transfers for example all tables whose size (which is read from the EXT files) exceeds 4 GB into separate STR files.

-package <size>

With this option you can split the existing STR files into equal-sized packages, that is, new STR files whose size does not exceed <size>.

-p <path>

Path to the target directory, that is, the directory in which the newly created files are stored.

<file(s)>

List of the STR or EXT files that are to be used for splitting.

-d <path>

  Path to the directory, in which you can find the STR and EXT files. The path must end with '/' or '\\'.

Yes,..parameter 'Split packages with size more than' just automatic call the Package Splitter tools? Not require manual intervention. Just, put optimize value in GUI i.e. JAVA splitting tool.

Regards

S.BASIS

[former_member182967 · Accepted solution]

Hello S Basis,

Thanks for your clear explanation.

Best Regards,

Tong Ning