Skip to content

Upgrades Technical guide

Oracle Data Integrator 12c to 14c: the repository upgrade

TL;DR

Oracle lists ODI 12.2.1.4 as the source for its 14.1.2 upgrade guide. Plan the master and work repository changes together with the installed agent topology, connected data systems and job schedules. Rehearse a consistent restore and compare representative job outputs before cutover. A Forms release number or another Fusion Middleware product’s upgrade path does not establish ODI compatibility.

Scope and evidence

Scope
ODI upgrade source release, topology distinctions and acceptance scope.
Evidence basis
Checked against Oracle’s ODI 14.1.2 upgrade introduction. The checks below guide planning and do not replace the topology-specific procedure or establish tested compatibility.
Sources checked

Confirm the source and the topology

Oracle’s ODI 14.1.2 guide identifies 12.2.1.4 as the source release and provides different upgrade topologies. Use the procedure that matches the installed agents and domain configuration. Do not borrow a starting version from the Forms or WebLogic upgrade documentation.

Inventory each master and work repository, agent deployment, context, data server, Knowledge Module, mapping, scenario, load plan and schedule. Include external schedulers and scripts that invoke ODI. Identify shared WebLogic or database dependencies where present; a shared product family does not mean every installation shares a domain.

Plan recovery beyond the repository backup

The recovery boundary must cover the affected repositories and configuration at a consistent point. Rehearse restoration before production execution and prevent both environments from running the same schedules.

Repository restoration does not reverse changes already written to a source or target system. For each business-critical load, define how to identify partial output, resume or reverse it safely, and reconcile the result. Agree the owner and decision point for stopping the cutover.

Define acceptance before running the upgrade

AreaWhat to establish
TopologyEach context resolves to the intended physical system and schema
Credentials and driversSupported connectivity works without copying production secrets into uncontrolled test environments
AgentsThe selected runtime starts, connects and executes representative work
Mappings and Knowledge ModulesCustom code and generated operations produce the expected result
Scenarios and load plansRequired runs complete, failures are visible, and restart behavior is understood
SchedulingExactly one intended scheduler owns each production job after cutover
Data outputsCounts, control totals, rejected records and business rules reconcile with the agreed baseline

Follow Oracle’s exact-release upgrade and post-upgrade instructions when deciding which assets require regeneration or other changes. Compare behavior on an isolated representative dataset; successful agent startup alone is insufficient.

Coordinate the cutover

Use rehearsal results to estimate elapsed time, including validation and recovery decisions. Repository size alone cannot determine the window. Confirm downstream availability, job dependencies and the treatment of partially completed loads.

ODI 12.2.x lifecycle planning belongs alongside the rest of the estate. The support dates tool provides the policy dates; hosting a release on OCI does not change its support tier.

ECMWorks can help scope repository, agent and integration dependencies and define acceptance evidence. Discuss the ODI estate before choosing a production sequence.

Questions

What is the documented ODI 14.1.2 starting point?

Oracle’s ODI guide identifies 12.2.1.4. For another source release, establish a supported intermediate route using the relevant Oracle documentation.

Does ODI always share a WebLogic domain?

No. Identify the actual agent deployment and follow the corresponding topology in Oracle’s guide. Coordinate shared dependencies where they exist.

What is the recovery plan for repository changes?

Treat changes as requiring restoration rather than assuming a downgrade. Back up all affected repositories and configuration consistently and rehearse recovery, including schedules and external data effects.

Must every scenario be regenerated?

Do not impose a blanket rule. Follow the exact-release upgrade and post-upgrade instructions, assess custom Knowledge Modules and generated code, and test the scenarios and load plans that matter to the business.

Give the upgrade a clear scope.

Tell us the starting release and the change you are considering. We can discuss dependencies, a supported route, rehearsal and the checks your team needs for acceptance.