hdinsight-migration
Kiểm tra cập nhật — MỘT LẦN MỖI PHIÊN (bắt buộc) Lần đầu tiên kỹ năng này được sử dụng trong một phiên, hãy chạy kỹ năng kiểm tra cập nhật trước khi tiếp tục.
npx skills add https://github.com/microsoft/skills-for-fabric --skill hdinsight-migrationUpdate Check — ONCE PER SESSION (mandatory) The first time this skill is used in a session, run the check-updates skill before proceeding.
- GitHub Copilot CLI / VS Code: invoke the
check-updatesskill.- Claude Code / Cowork / Cursor / Windsurf / Codex: compare local vs remote package.json version.
- Skip if the check was already performed earlier in this session.
CRITICAL NOTES
- To find workspace details (including its ID) from a workspace name: list all workspaces, then use JMESPath filtering
- To find item details (including its ID) from workspace ID, item type, and item name: list all items of that type in that workspace, then use JMESPath filtering
- HDInsight has no
mssparkutilsordbutilsequivalent —notebookutilsis net-new capability being introducedHiveContextandSQLContextare legacy Spark 1.x/2.x APIs — Fabric uses Spark 3.xSparkSessionexclusivelywasb://paths are deprecated and require a Storage Account key or SAS — replace with OneLake shortcuts
HDInsight → Microsoft Fabric Migration
Prerequisite Knowledge
Read these companion documents before executing migration tasks:
- COMMON-CORE.md — Fabric REST API patterns, authentication, token audiences, item discovery
- COMMON-CLI.md —
az rest,az login, token acquisition, Fabric REST via CLI - SPARK-AUTHORING-CORE.md — Notebook deployment, lakehouse creation, Spark job execution
For notebook and Lakehouse creation, see spark-authoring-cli. For Fabric Warehouse DDL/DML authoring, see sqldw-authoring-cli.
Table of Contents
| Topic | Reference |
|---|---|
| Migration Workload Map | § Migration Workload Map |
| SparkSession & Context API Changes | § SparkSession API Changes |
| WASB / ABFS → OneLake Path Migration | path-migration.md |
| Hive DDL → Delta Lake / Lakehouse Schemas | hive-to-delta.md |
| Oozie → Fabric Pipelines | § Oozie → Fabric Pipelines |
Introducing notebookutils | § Introducing notebookutils |
| Before/After Code Patterns | code-patterns.md |
| Spark Configuration Differences | § Spark Configuration Differences |
| Must / Prefer / Avoid | § Must / Prefer / Avoid |
| Authentication & Token Acquisition | COMMON-CORE.md § Authentication |
| Lakehouse Management | SPARK-AUTHORING-CORE.md § Lakehouse Management |
Migration Workload Map
| HDInsight Component | Fabric Target | Notes |
|---|---|---|
| Spark cluster (notebooks, scripts) | Fabric Spark (Lakehouse / Notebooks / SJD) | No persistent cluster — Starter Pool or Custom Pool provides on-demand Spark |
| Hive / HiveServer2 | Lakehouse SQL Endpoint + Lakehouse schemas | Delta Lake replaces Hive metastore; schemas provide namespace equivalent |
| HBase | Fabric Warehouse or Azure Cosmos DB (separate from Fabric) | HBase has no direct Fabric equivalent — assess workload access patterns |
| Oozie workflows | Fabric Data Pipelines | Map Oozie actions to Fabric activities; see § Oozie → Fabric Pipelines |
| YARN Resource Manager | Fabric Spark monitoring (Spark UI, Monitoring Hub) | No YARN — Fabric manages compute automatically |
| Ambari | Fabric Monitoring Hub + Admin Portal | Cluster health, capacity, and job monitoring |
| WASB / ABFS storage | OneLake Shortcuts → abfss://[email protected]/ | See path-migration.md |
| Ranger policies | Fabric workspace roles + OneLake data access roles | Map Ranger row/column filters to Lakehouse row-level security |
| Livy REST server | Fabric Livy API | Compatible endpoint — see SPARK-AUTHORING-CORE.md |
SparkSession & Context API Changes
HDInsight Spark clusters often use legacy Spark 1.x / 2.x API styles. Replace all of these with the unified SparkSession:
| Legacy HDInsight Pattern | Fabric Spark 3.x Replacement |
|---|---|
from pyspark import SparkContext; sc = SparkContext() | Not needed — sc = spark.sparkContext (pre-instantiated) |
from pyspark.sql import HiveContext; hc = HiveContext(sc) | Not needed — spark session has Hive-compatible SQL support via Delta schemas |
from pyspark.sql import SQLContext; sqlc = SQLContext(sc) | Not needed — use spark.sql(...) directly |
SparkSession.builder.enableHiveSupport().getOrCreate() | Not needed in Fabric — spark is pre-built and available |
sc.textFile("wasb://[email protected]/path") | spark.read.text("abfss://[email protected]/lh.Lakehouse/Files/path") |
sqlContext.sql("CREATE TABLE ... STORED AS ORC") | See hive-to-delta.md for Delta DDL equivalent |
In Fabric notebooks,
spark(SparkSession) andsc(SparkContext) are pre-instantiated — do not callSparkContext()orSparkSession.builder...getOrCreate()at the top of migrated notebooks.
Oozie → Fabric Pipelines
Map Oozie workflow actions to Fabric Data Pipeline activities:
| Oozie Action Type | Fabric Pipeline Activity | Notes |
|---|---|---|
<spark> action | Notebook activity or Spark Job Definition activity | Pass parameters via notebook cell parameters or SJD arguments |
<hive> action | Script activity (SQL) against Lakehouse SQL Endpoint | Convert HiveQL to Spark SQL or Delta SQL |
<shell> action | Azure Function activity or Web activity | Shell scripts must be refactored; no direct shell execution in Fabric Pipelines |
<java> action | Azure Batch activity (external) or refactor to PySpark | Java MapReduce jobs must be rewritten |
<sqoop> action | Copy Data activity (Fabric Data Factory connector) | Sqoop import/export maps to Fabric Copy Data with JDBC source/sink |
<coordinator> (time-based schedule) | Pipeline schedule trigger | Set recurrence in pipeline trigger; supports cron-like expressions |
<coordinator> (data-triggered) | Storage Event trigger | Trigger on OneLake file arrival |
Delegate to
spark-authoring-clifor notebook and SJD creation after mapping pipeline activities.
Introducing notebookutils
HDInsight Spark had no built-in utility framework equivalent to mssparkutils or dbutils. When migrating to Fabric, introduce notebookutils for common operations:
| Operation | Old HDInsight Approach | notebookutils Equivalent |
|---|---|---|
| List files | dbutils (N/A) / HDFS CLI | notebookutils.fs.ls("abfss://...") |
| Copy file | HDFS API / shutil | notebookutils.fs.cp(src, dest) |
| Read secret | Azure Key Vault REST call | notebookutils.credentials.getSecret(keyVaultUrl, secretName) |
| Get notebook context | Not available | notebookutils.runtime.context — returns workspace ID, notebook ID, etc. |
| Run child notebook | Not available | notebookutils.notebook.run("notebook_name", timeout, {"param": "value"}) |
| Exit notebook with value | sys.exit() | notebookutils.notebook.exit("value") |
| Mount storage | WASB config in spark-defaults.conf | OneLake Shortcut (no runtime mount needed) |
Spark Configuration Differences
| HDInsight Concept | Fabric Spark Equivalent | Migration Action |
|---|---|---|
spark-defaults.conf (cluster-wide) | Fabric Spark Workspace Settings + Environment item | Move config properties to Environment or use %%configure in notebooks |
%%configure magic | %%configure magic — identical | No change needed |
| YARN queue / resource allocation | Fabric Spark pool node size and autoscale settings | Map queue SLAs to Custom Pool configuration |
| Ambari service configs (HDFS, YARN tuning) | Not applicable — Fabric manages infrastructure | Remove; focus on application-level Spark configs |
| HDI Spark version (e.g., Spark 2.4) | Fabric Runtime 1.3 = Spark 3.5 (latest) | Test for deprecated API removals (e.g., HiveContext, RDD-style ML) |
Conda environment / bootstrap.sh | Fabric Environment item with custom libraries | Recreate conda/pip dependencies in a Fabric Environment |
hive-site.xml (metastore connection) | Not needed — Delta Lake IS the metastore in Fabric | Remove metastore config; use Lakehouse schemas for namespace organization |
Must / Prefer / Avoid
MUST DO
- Replace all
wasb:///wasbs://paths with OneLakeabfss://paths or OneLake Shortcuts —wasb://requires storage account keys which are not the Fabric-preferred auth model - Replace
HiveContext,SQLContext, and standaloneSparkContext()— use the pre-instantiatedsparksession in Fabric notebooks - Migrate Hive DDL (
STORED AS ORC,LOCATION,TBLPROPERTIES) to Delta Lake DDL — see hive-to-delta.md - Introduce
notebookutilsfor file system operations, secret retrieval, and child notebook orchestration where HDInsight used custom scripts or direct API calls - Replace Oozie XML workflows with Fabric Data Pipelines — see § Oozie → Fabric Pipelines
- Align library management to Fabric Environments — remove
bootstrap.sh, conda envs, and runtime%pip installpatterns for production workloads
PREFER
- OneLake Shortcuts over copying data — mount existing ADLS Gen2 containers as shortcuts to avoid re-ingestion during migration
- Delta Lake for all tables migrated from Hive ORC/Parquet — ACID guarantees, time travel, and schema enforcement improve data quality
- Fabric Starter Pool for initial migration validation — no pool configuration overhead, fast session startup
- Lakehouse schemas (database namespaces) for organizing migrated Hive databases — one schema per Hive database within a single Lakehouse
- Medallion architecture for restructuring migrated data layers during migration — align Bronze/Silver/Gold with raw Hive → validated Delta → serving Gold patterns
AVOID
- Do not use
SparkContext()orHiveContext()constructors in Fabric notebooks — they conflict with the pre-instantiatedsparksession and will raise errors - Do not use
hive-site.xmlor external Hive metastore configuration — Fabric's Delta Lake-backed Lakehouse IS the metastore - Do not assume YARN queue mappings translate to Fabric pools — re-design resource allocation based on Fabric Spark pool SLAs
- Do not attempt to run Oozie shell actions or Java MapReduce jobs directly in Fabric — these must be refactored (see § Oozie → Fabric Pipelines)
- Do not use
%shmagic for file system operations in production notebooks — usenotebookutils.fs.*for portability and OneLake token-based auth
Examples
See code-patterns.md for full before/after examples. Key quick references:
Legacy context → Fabric pre-instantiated session
# HDInsight (remove entirely)
from pyspark.sql import HiveContext
hc = HiveContext(sc)
# Fabric — use pre-instantiated spark directly
df = spark.sql("SELECT * FROM sales.fact_orders")
WASB path → OneLake path (after shortcut creation)
# HDInsight
df = spark.read.parquet("wasb://[email protected]/orders/")
# Fabric
df = spark.read.parquet("Files/raw/orders/")
Hive DDL → Delta DDL
-- HDInsight
CREATE TABLE sales_db.fact_orders (...) STORED AS ORC LOCATION 'wasb://...';
-- Fabric
CREATE SCHEMA IF NOT EXISTS sales_db;
CREATE TABLE sales_db.fact_orders (...) USING DELTA;