What Trekport is.
Trekport is an Oracle to PostgreSQL migration platform. It is built for database teams that need to ship a cutover, not produce a feasibility study.
Who this is for
Trekport is built for the database team that owns the migration outcome. The senior DBA who runs the cutover. The platform engineer who has to integrate the deployment into CI. The VP of data platform who has to brief the board on timeline and risk.
If you are evaluating Trekport for the first time, start with the conversion walkthrough — it takes a single schema from extraction to a validated PostgreSQL target and is the shortest path to understanding what the platform produces.
What the pipeline produces
Trekport produces deterministic, reviewable PostgreSQL output. Every Oracle catalog object becomes one of three things:
- A fully equivalent PostgreSQL object compiled from a real parse-tree translation.
- An object backed by a battle-tested PostgreSQL extension from the compatibility layer.
- An explicit, documented out-of-scope report when no equivalent exists. See the troubleshooting guide for how the engine classifies these cases.
No black-box rewrites. No silent placeholders. The output is the artifact a reviewer can sign.
How the pipeline runs
The pipeline runs in seven stages. Each stage emits an artifact the next stage consumes. The artifacts are inspectable and the run is reproducible byte-for-byte against the same inputs.
| Stage | What it does | Artifact |
|---|---|---|
| 1. Catalog extraction | Walks the Oracle data dictionary and serializes every object. | ./catalog/ |
| 2. Parse-tree construction | Builds a real syntax tree from every SQL and PL/SQL artifact. | In-memory AST |
| 3. Rule application | Applies named, versioned translation rules against the tree. | Transformed AST |
| 4. Code emission | Renders the transformed tree back to PostgreSQL source. | ./converted/ |
| 5. Dependency graph | Computes inter-object dependencies for deployment ordering. | Phase plan |
| 6. Deployment | Runs the fourteen-phase orchestrator against the target. | Deploy report |
| 7. Validation | Compares row counts, checksums, and samples against the source. | Validation report |
Stages 1 through 4 are covered in the conversion walkthrough. Stage 6 is covered in detail in deployment phases.
What it does not do
Trekport does not redesign your schema. It does not refactor your application code. It does not promise to convert Oracle features that have no PostgreSQL equivalent. It reports what does not convert, with rationale, so the gap is explicit before cutover — not surprising on go-live night.