object history

[SVanSchoonlandt]

When history of objects is saved, it is for every object in the repository. When a lot of history is available, it's slowing down performance in the tool. We are currently looking for solutions in order to keep history and optimal performance. We keep history for one year in order to find back who changed something to a model.

I know there is this option on metamodel level to say exclude from history. Not all objects it's necessary for us to keep history, so would it be a good option to set for the metaclasses for which we don't need history to exclude it? What are possible known drawbacks of this option (besides not having the history)

eg for library objects we don't require the history.

Are there any other possible options we could have a look at regarding history?

thanks

stijn

[jhorber]

Hello Stijn

This is not an easy subject

A a stored procedure (SP_CONSOLIDATE_MEGA_DATABASE) enables to reduce the size  of the repository log by consolidating certain commands. It is recommended to schedule its execution.

An administration feature enables to purge elements of the repository log for RDBMS storage

It is run manually from Windows Administration Console (Administration.exe). It enables to target specific MetaClass

See online documentation
HOPEX Administration : Administrator Guide : Managing Repositories : Optimizing Repository Access Performance : Managing Log Size

For the functional impacts, see this KB

http://community.mega.com/t5/custom/page/page-id/mega-kb-solution?sid=501D0000000LtINIA0