Editéieren

Overview of migrating Azure Synapse Spark to Fabric

Use this article as the starting point for migrating Azure Synapse Spark workloads to Microsoft Fabric. It helps you decide which guidance to use, what can be migrated directly, and where manual refactoring or validation is still required.

Fabric Data Engineering supports lakehouse, notebook, environment, Spark job definition, and pipeline items. Most Synapse Spark migrations involve some combination of item migration, data access changes, metadata migration, code refactoring, and post-migration validation.

Before you migrate

Before you begin, confirm that Fabric Data Engineering is the right destination for your workload. Review the Spark runtime, security model, pool model, environment model, and data access patterns that your current Synapse implementation depends on.

Start with these articles:

If you're migrating an existing Synapse workspace, plan to create or use an existing Fabric workspace as the migration target. This article doesn't cover full workspace provisioning or non-Spark workload migration.

What can you migrate?

Synapse-to-Fabric migration usually spans several workstreams.

Migration area Typical scope Primary guidance
Planning and assessment Inventory Spark pools, notebooks, Spark Job Definitions, lake databases, linked services, and blockers Phase 1: Migration strategy and planning
Items, code refactoring, pools, configs, and libraries Notebooks, Spark Job Definitions, Spark pools, lake database mappings, mssparkutils, linked services, file paths, catalog APIs, connector auth, environments, custom pools, Spark properties, library compatibility Phase 2: Spark workload migration
Hive Metastore and lake metadata Databases, tables, partitions, managed vs. external tables Phase 3: Hive Metastore and data migration
Data access and pipelines OneLake shortcuts, ADLS Gen2 access, copy activities, pipeline migration Migrate data and pipelines
Security, validation, and cutover Roles, connections, governance, verification, cutover planning Phase 4: Security and governance migration

Choose your migration path

Use the path that matches your goal.

  • You need an end-to-end migration plan. Start with the 4-phase best practices series. This is the best entry point for most production migrations.
  • You want to move supported Spark items quickly. Start with the Spark Migration Assistant and then use the refactoring and validation articles to close the gaps.
  • You only need help with one area. Use the task-specific articles for notebooks, Spark Job Definitions, pools, libraries, Hive Metastore metadata, or data/pipeline migration.

For most teams, the fastest way to approach a Synapse Spark migration is:

  1. Review Compare Fabric and Azure Synapse Spark: Key Differences.
  2. Read Phase 1: Migration strategy and planning.
  3. Run the Spark Synapse to Fabric Spark Migration Assistant where applicable.
  4. Refactor notebooks, Spark jobs, pools, and libraries using Phase 2: Spark workload migration.
  5. Validate data access, metadata, security, and cutover readiness using the remaining best-practices articles.

Screenshot showing the migration scenarios.

Migration from Synapse Spark to Fabric is usually a copy-and-adapt process rather than a direct in-place move. You can migrate many assets quickly, but you should still expect to validate runtime behavior, replace Synapse-specific integrations, and align security, metadata, and operational patterns with Fabric.

Best practices series

Use the best practices series for a structured, end-to-end migration path:

Task-specific migration articles

If you need targeted guidance for a specific migration task, use these articles:

Related content

  • Last updated on