February 7, 2018 at 7:12 am
Is there anyone out there who has a slick way of managing MDS Models & is prepared to share their experience?
What I am interested in specifically:
* Efficient use of source control to keep a history of MDS models (and, optionally, data in the models)
* Automated deployment of MDS models from source control to QA, Prod etc
Thanks.
The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin
July 5, 2018 at 11:20 am
Bumping this, just in case anyone with any answers missed it the first time.
The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin
October 9, 2018 at 6:11 am
Bit of info from my experience...
1) MDS is a source control system. you do not need to use anything else. Production MDS is your versioned source control and changes should be made there by business users. This method works best but the MDS FE can be very slow and is a bit unwieldy.
2) If you are looking at some kind of waterfall release process through environments, MDS is definitely not the best tool for this. We automated the export of packages to apply to other environments and then you end up progressing the data through to Prod. It is clunky and causes lots of headaches.
We have used both ways and now only keep the first one.
If you have master data that is changed by IT (and not business) and the changes are released through environments then it is far simpler to just chuck the data in a table with a soft-delete/history process (or something similar).
October 9, 2018 at 6:39 am
Tom OShea - Tuesday, October 9, 2018 6:11 AMBit of info from my experience...1) MDS is a source control system. you do not need to use anything else. Production MDS is your versioned source control and changes should be made there by business users. This method works best but the MDS FE can be very slow and is a bit unwieldy.
2) If you are looking at some kind of waterfall release process through environments, MDS is definitely not the best tool for this. We automated the export of packages to apply to other environments and then you end up progressing the data through to Prod. It is clunky and causes lots of headaches.
We have used both ways and now only keep the first one.
If you have master data that is changed by IT (and not business) and the changes are released through environments then it is far simpler to just chuck the data in a table with a soft-delete/history process (or something similar).
Tom, thanks for your input. I am not interested in source-controlling MDS data. As you suggest, that's what MDS does by itself.
It is the models themselves which are of interest. The entities, subscription views, business rules and permissions, to be more specific. How did you automate their deployment?
The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply