DSXchange

In Datastage 8.5, There is Xmeta in Database and the Repository (DS_JOBS / DS_JOBOBJECTS etc) in hashed files also.
Locks can be still cleared using DS.TOOLS. And aswell using Datastage Admin Webconsole.
May I know why is there two different places the metadata is been maintained?
Or what is maitained in Xmeta and what is in OS hashed file repository?
I understand that there is some extra feature such as DS_AUDIT and more reporting feature is avaialble in XMeta.
Anything else?

I think the goal is to eliminate all the hashed files and only use XMETA. I am guessing that took longer than expected or there were performance issues. So they are eliminating slowly and duplicating data because XMETA is supported across multiple product lines.

Only IBM knows for sure why it is this way.

Thanks for the Reply Kim!
May I know the actual purpose of Xmeta in today's world?
Apart from reporting and Auditing.

If we delete a record in Xmeta (Eg: Delete a job), will it relfect in Datastage repository too? Or the Vice versa?

The local repository (which you have called "hashed files" and which was originally built on the UniVerse RDBMS) is the original, and continues to be used primarily for design-time operations so that you don't commit anything to XMETA until you are ready to. For performance reasons you can also specify local job logging.

There are times you can delete in XMETA and it is okay. If you are running logs only in XMETA then you can delete in XMETA. I would not guess on any other table. Do it at your own risk. I doubt if it is supported.

Thanks Ray! When do it get commited to Xmeta? After Compiling or Running a job first time?

Thanks Kim, Looks like there is some script available from IBM to force sync the both Xmeta and the UV.

It's always connected (via services) to XMETA. Your Information Server roles and DataStage credentials are stored in XMETA - you could not log in were they not there.