Data model cleanup

[thijslemmens]

Dear all

A big client of ours is using Alfresco in an archiving use case. They are putting in documents and custom metadata for over 5 years now, and have been extending the model on a regular basis. The repository contains about 8 million documents. The document model has a lot of unused fields and also duplicated properties with a slightly different name. Now we want to cleanup the document models. This is what I have in mind right now:

1) Design the new document models
2) Setup a new Alfresco with the new models
3) Do a migration from the old Alfresco to the new one, mapping the old properties to the new ones. (We have our own migration product, move2alf, supporting CMIS as a source.

Since this is a big Alfresco, this process will not be easy. Does anyone have experience in doing something similar, or advice to take a different approach?

kind regards

Thijs

[mitpatoliya]

This is very tricky job even though you have all the tools.
I must say unless it is really necessary don't go for this.
If you have properties which are not used anywhere and you want to remove them you can directly change the model and do full re-indexing. That way all the references of old property will be removed from the contents.

[thijslemmens]

Thanks for the comment. I agree that it is tricky, but it is not only unused properties. Some properties are duplicated with a slightly different name and should be merged. Some inheritance structures should be changed. It would need a long path of compatible document model changes and scripts. At least, we will test it thoroughly first before concluding it is possible.