We are trying to implement BLUE-GREEN deployments for our hybris nodes. For any classic application using database, blue-green deployments poses problems with 2 aspects
1) Schema changes - SAP's rolling cluster update solution addresses this aspect by having 2 typesystems simultaneously on the DB (https://help.sap.com/viewer/d0224eca81e249cb821f2cdf45a82ace/1811/en-US/8c455268866910149b25f7b53d1af3e1.html)
2) Data changes - The data changes (impexes present as part of projectdata/essentialdata) that are imported are common to both old and new code. How are these suggested to be addressed ?
Every release, we will have overlapping data changes on content/product catalog which is heavily dependent on code. Couple of trivial examples are shared below. But we do have a lot more scenarios where business logic is driven by data.
Examples :
So in all such cases,
Though we can build a custom solution (#1: having catalog versions as blue and green and seggregating the release changes) addressing the above use cases I am more keen to understand,