Note: for some inputs on the common issues, please refer to this article.
The code of the project is split between a common code-base and specific code in the Dataiku recipes, as highlighted in the project architecture.
The actions to take are different depending where the issue is located.
Option 1: Issue related to the code base (no change on the Dataiku project itself)
- Create a hotfix branch from the code base repository master branch.
- If you need it to test your development, create a branch from the Dataiku project in design following these guidelines and switch the library to your new code base hotfix branch.
- Make sure the hotfix does not break anything related to other GBU projects (you can create a branch based on your new code base in the CS or SPP project and run a "Full reload" scenario to make sure it is still working).
- Make sure the Gitlab pipeline with the unit tests passes successfully.
- Merge your hotfix branch for the code-base on master (you can still refer to this page).
- Update the Dataiku master project library for the right GBU by selecting the "Reset from remote HEAD" option.
- Launch a "Full reload" scenario on dev to make sure it runs properly
- Merge your hotfix branch for the code-base on dev (you can still refer to this page) to make sure it has the same status as master.
- Create a new bundle of the project and deploy it in automation.
Option 2: Issue related to the Dataiku project Flow (no change on the code base)
- Create a hotfix branch by duplicating the SPOT_SPP_MASTER Dataiku project (refer to How to : start a new development for the step but be careful that you are not starting from dev but master)
- Apply the hotfix and merge back to master, run a "Full reload" scenario to make sure everything works (do not rebase in master !)
- Rebase the dev to apply the hotfix changes.
Option 3: changes on both the code base and the Dataiku project
- Create a hotfix branch by duplicating the SPOT_SPP_MASTER Dataiku project (refer to How to : start a new development for the step but be careful that you are not starting from dev but master)
- Create a hotfix branch from the code base repository master branch.
- Apply the changes to fix the issue
- Make sure the hotfix does not break anything related to other GBU projects (you can create a branch based on your new code base in the CS or SPP project and run a "Full reload" scenario to make sure it is still working).
- Make sure the Gitlab pipeline with the unit tests passes successfully.
- Merge your hotfix branch for the code-base on master (you can still refer to this page).
- Merge back the Dataiku branch to master, run a "Full reload" scenario to make sure everything works (do not rebase in master !)
- Update the Dataiku master project library for the right GBU by selecting the "Reset from remote HEAD" option.
- Launch a "Full reload" scenario on dev to make sure it runs properly.
- Merge your hotfix branch for the code-base on dev (you can still refer to this page) to make sure it has the same status as master.
- Rebase the hotfix Dataiku branch on dev to apply the hotfix changes.
- Create a new bundle of the project and deploy it in automation.
