Migrating Oracle packages to PostgreSQL
Packages can combine APIs, state and business logic. Separate those responsibilities before deciding how each part should live in PostgreSQL.
Read the articlePractical articles on packages, PL/SQL, database links, Scheduler, Oracle-specific features, PostgreSQL/Aurora and the work required to turn a converted schema into an operable target.
Packages can combine APIs, state and business logic. Separate those responsibilities before deciding how each part should live in PostgreSQL.
Read the articleJobs are part of the operating model. Capture schedules, dependencies and failure behaviour before choosing the target scheduler.
Read the articleA database link is more than a connection. Classify reads, writes, consistency and failure behaviour before choosing the replacement.
Read the articleTreat storage type, XML queries and business processing as separate concerns: XMLTYPE is more than stored text.
Read the articlePL/SQL file access affects privileges, operating procedures and error handling. Make the ownership change explicit.
Read the articleDefine durability, acknowledgement, ordering and retry semantics before choosing the replacement technology.
Read the articleDo not migrate only the SELECT statement. Refresh behaviour, data age, availability and query patterns all matter.
Read the articleSynonyms can hide schema boundaries and remote targets. Make the actual dependencies explicit before moving them.
Read the articleReview partition keys, boundaries and maintenance processes. The Oracle design is not automatically the best PostgreSQL design.
Read the articleExtensions can close functional gaps, but availability, privileges, maintenance and exit strategy must be assessed first.
Read the articleTreat schemas, roles, interfaces and operations as one target design before local fixes become permanent dependencies.
Read the articleEvaluate availability, throughput and recovery separately. A PostgreSQL target needs its own operating and failure model.
Read the articleThese are three different migration paths with different levels of change, operational responsibility and compatibility impact.
Read the articleSeparate infrastructure relocation from an engine change. That makes sequencing, risks and dependencies much clearer.
Read the articleManaged operations do not remove compatibility work. Business logic, supported features and application behaviour remain decisive.
Read the articleCombine technical inventory with cost drivers. A useful business case depends on documented assumptions, not a single infrastructure price.
Read the articleAfter choosing the partition model, define a repeatable lifecycle for loading, querying, archiving and maintenance.
Read the articleEquivalent results do not imply equivalent runtime. Compare representative workloads and investigate measured bottlenecks on the target.
Read the articleTurn findings into migration units: dependencies, pilot, testing and cutover in an evidence-based sequence.
Read the articleUnderstand the source, choose a meaningful pilot, and validate data and logic before committing to an engine change.
Read the articleNot every database needs the same path. Prioritise applications by changeability, risk and real modernisation value.
Read the articleClarify what problem should be solved before selecting a platform, then define how success will be proven.
Read the articleKeep analysis, conversion and engineering judgement separate. The right tool combination depends on the task.
Read the article