By Ayush Ganeriwal-Oracle on Nov 12, 2015
This is the second article in the series of three articles covering Oracle Data Integrator (ODI) 12.2.1 lifecycle management features. In the previous article, Managing Versions in Apache Subversion, I discussed ODI Studio features for creating, viewing, comparing and restoring versions in Apache Subversion. In this article I will cover the Branch and Tag management capabilities that are required for parallel development.
Parallel Development Using Subversion Branches
First, let’s take a look at the high level setup of the Subversion branches in ODI for any parallel development. As shown in the below diagram, you can configure a Subversion branch/trunk with one master and work repository combination. In this example, the Subversion trunk is used by user set 1 as the main code line and contains all ODI objects. Branch 01 is configured for user set 2, who are working on a subset of ODI objects. User set 3 is working on a release branch and fixing bugs on the release code line. Each of these user set work in parallel and leverages Subversion branches as the channel to merge their changes across different branches.
A Tag is a text name applied to the current versions of a set of objects to take a snapshot of their current state. It freezes the current state of objects and can be referred back with the Tag name at some future point in time. You can create a Tag directly from the ODI Studio and you can create it for all the repository objects or for a selected set of objects.
ODI takes following actions on the objects present in the Tag to ensure consistency.
- Add objects to Subversion if they are not yet versioned.
- Create versions for the objects if their latest changes are not yet versioned.
- Automatically include all the dependencies of the selected objects so that all the relevant objects can be found in the Tag when you refer back to it.
Thus, a Tag creation is a good way of ensuring all the relevant objects are being version controlled and it is recommended to create Tags at different logical points during your development cycle.
There are two type of Tags created in ODI
- Full Tag – It is created with all the objects in the repository. It will first sync all the repository object state to Subversion versions before applying the Tag.
- Partial Tag – It is created for a selected set of objects and user drags and drops an object into the Partial Tag creation wizard. Al the dependencies for the selected object are automatically added to the Tag.
In the example, a mapping was added to the Partial Tag and all the required model and topology objects are automatically added by ODI Studio.
Creating a Subversion Branch
You can create a Subversion branch from a Tag. The branch creation screen lists all the relevant Subversion Tags and their respective comments to easily locate a Tag for branch creation. The objects versions of the selected Tag becomes the branch point and may not necessarily be the latest version of the objects.
Populating ODI Repository from a Branch or Trunk Objects
Once you configure your ODI repository with a Subversion branch/trunk, you can populate it with the branch/trunks objects through a couple of ODI Studio menu options.
- Populating repository – It is used when you want to populate a fresh repository with the branch content. For example when you create a project branch and want to setup an ODI repository to the branch.
- Populating restored repository – This option is used when your current repository gets corrupted and you restore an older repository backup. It syncs the objects present in the restored repository with the latest contents from the Subversion branch/trunk. It will pull all the objects from the Subversion branch/trunk and imports them in a special Import Mode that only inserts or updates the objects presented in the repository while leaves behind any extra objects in the repository.
Merging changes from one development branch to another is an essential need for any parallel development. ODI Studio allows you to merge changes from a Branch, Trunk or Tag to the repository it is connected.
The merge process automatically merges the object changes that do not have any conflicts. The Merge summary details all the affected objects in the merge. It also lists the objects that could not be automatically merged and their conflicts should me manually resolved. The summary report can be saved in the file system in multiple formats so that it can be shared with the team.
The merge result window assists you in resolving conflicts from a branch merge. There are a couple of tabs available in this window.
Merge Object Selection Tab
It lists the objects affected in the merge and provides different search and filtering options. It allows following operations on the conflicted objects
- Mark the conflict as resolved for the selected object.
- Open Merge Conflict Resolution window for the selected object.
- Assign ownership of the conflict resolution to a particular user responsible for conflict resolution of that object. By default the merge process assigns conflict ownership to the user who has last modified it.
Merge Conflict Resolution Tab
It assists you in manually resolving an object’s conflict by presenting the object properties side by side.
A number of branch management features introduced in ODI 12.2.1 are vital for organizing objects across branches and allow parallel development between functionally and geographically distributed teams.
Stay tuned for the third and final article on lifecycle management, which will cover the release management capabilities using newly introduced Deployment Archives.