Skip to main content
Version 3.0

Highlights — TDP Datacenter

Platform

VersionHighlights
TDP 3.0JupyterLab added: interactive environment for data analysis, notebooks and visualisations; support for multiple languages and extensions; modern tab-based interface integrating code editing, execution and charts.
OpenMetadata added: data governance and catalogue platform; enables lineage, data discovery and quality definitions; integration with a wide range of ecosystem services (databases, lakes, pipelines).

Table A. TDP Datacenter platform highlights

Components

See the highlights of the updated components:

ServiceVersionCategoryFeature/Highlight
Airflow2.10.5OrchestrationCleanup tasks guaranteed even on failure or manual success — prevents resource leaks.
Multiple bug fixes: mapped trigger rules, ID validation, XCom with "/", event logs and more.
Helm Chart image updates with newer versions for Airflow, PgBouncer and exporters.
Delta Lake3.3.0Optimised table formatIdentity Columns: automatic generation of unique values for new rows (Python and Scala) — generatedAlwaysAs and generatedByDefaultAs modes.
VACUUM LITE: 5–10× faster cleanup of obsolete files using only the transaction log.
Row Tracking Backfill: enables row-level lineage tracking on existing tables, including Row Id and Row Commit Version metadata.
Version Checksums: per-commit checksums reinforcing consistency, performance and debugging.
UniForm ALTER: enables UniForm format on existing tables without rewriting data.
Type Widening: type widening support in Delta Kernel (Java and Rust).
Druid32.0.0Real-time AnalyticsStability and performance focus: 341+ commits from 52 contributors (~30% more than previous releases).
ANSI SQL compliance: legacy null-handling settings removed — ANSI SQL behaviour is now fixed.
New Projection feature: pre-computes aggregations for significantly improved query performance.
Great Expectations1.3.5Data QualityNew: CheckpointFactory.add_or_update and strict_min/strict_max in ExpectTableRowCountToBeBetween.
row_condition support with datetimes in Pandas and Spark; BigQuery cleanup scheduling; strict parameter in Window type; guaranteed run_id in ValidationDefinition.run.
HBase2.5.7NoSQL DatabaseSwitch to avoid region reopening when editing tables — prevents RIT storms.
isolate_regions command in RegionMover — precise, controlled region isolation and relocation.
Hive4.0.0Data Exploration and AnalyticsMajor leap: ~5,000 commits since version 3.1.3.
Apache Iceberg integration: improved support including compaction via OPTIMIZE TABLE.
Enhanced ACID transactions and locking; compaction for ACID and Iceberg tables.
Docker support: official Hive images on Docker Hub.
Execution optimisations: anti-join, branch pruning, column histograms, HPL/SQL, scheduled queries, materialized views, Tez and LLAP.
Improved replication for external and ACID tables; Apache Ozone support as scalable storage.
Advanced features: native GeoSpatial, Iceberg compaction, JWT auth in Metastore, SAML 2 in HiveServer2.
Iceberg1.8.0Table FormatsEnd of support for Spark 3.3 and Hive Runtime.
Deletion Vectors: new spec, APIs and read/write support.
New types: Variant and UnknownType.
Operational improvements: fast append, removal of unused specs, useful procedures in Spark.
Improved AWS/Azure integration; extended compatibility with Spark, Flink and Hive.
Kafka3.4.1Data StreamingZooKeeper-to-KRaft migration (initial release; not recommended for production) — zero downtime.
New generation field in the consumer protocol — helps manage partition claims.
Time-based metadata snapshots (e.g. hourly); rack-aware consumers for local AZ.
Kafka Streams (KIP-770 & KIP-837): config and cache metric updates; broadcast output records to all partitions.
MirrorMaker2: reads all offset syncs at startup; publishes syncs during task commit; translates offsets across replication flows.
Fixed interceptor resource leak.
NiFi1.28.1Data Flow ManagementFixed logging of sensitive parameter values in flow sync log — not exposed even with debug enabled (1.28.1).
XSS protection: parameter descriptions are now correctly neutralised (1.28.0).
End of support for NiFi 1.x (08 Dec 2024) — migration to 2.x series strongly recommended.
Ozone1.4.1Scalable Object StoragePrevents race condition in datanode when creating VERSION — improves operational reliability.
Refined SCM logs — reduces noise and false errors when handling sequence IDs in closed containers.
S3 Gateway security: secrets management endpoint restricted to administrators.
Spark3.5.3Distributed ComputingThird maintenance release of the 3.5 series — security and bug fixes. Recommended for environments already running 3.5.
Superset4.1.2Data VisualisationNew charts: Big Number with time comparisons, Heatmap, Histogram and Sankey.
Dynamic catalogue in connected databases; more intuitive upload UI with validations.
Time filters and Jinja macros for dynamic dashboards; UX improvements in SQL Lab and permissions.
Security fixes in 4.1.2: prevents resource takeover and access control bypass.
Apache Tez0.10.4Execution Engine~28 fixes and improvements — observability, stability and security.
ZooKeeper3.8.4Distributed CoordinationConfigurable maximum connections/clients per server.
zkCli: new option to wait for connection before executing commands — avoids failures in unstable environments.
Startup logs with time units; clearer network address display; improved error messages.

Table B. TDP Datacenter component highlights