Microsoft's SQL Server 2025 (17.x) Preview release candidate 1 (RC1) now includes preview support for Red Hat Enterprise Linux 10 (RHEL). This milestone expands SQL Server's footprint on Linux, giving developers and IT professionals more choices for modern, high-performance deployments.
SQL Server 2025 Preview is packed with AI and developer-friendly features, making RHEL 10 a great platform for experimentation. With RHEL 10 support, you can start validating SQL Server 2025 Preview in your development and test environments ahead of general availability.
What’s new in SQL Server 2025 Preview RC1
In addition to RHEL 10 support, SQL Server 2025 Preview introduces several exciting updates:
New AI and vector data
- Vector data type with support for half-precision floats (2-byte) for up to 3,996 dimensions in a single vector.
- Vector indexes for similarity search.
- External AI model management that integrates with REST inference endpoints and embedding tasks.
Improved developer experience
- Regular expressions: Use regex directly in SQL queries.
- JSON data type: Store, query, and transform JSON natively.
- External REST endpoint invocation: Call REST/GraphQL services directly from SQL Server.
Analytics and data connectivity
- PolyBase on Linux now supports ODBC data sources.
- Native support for Parquet, Delta, and CSV without a requirement for PolyBase services.
- Managed identity for Azure Blob Storage and Data Lake.
Continued reliability
- Fixes for in-place upgrades (replication issues) and PolyBase encryption.
- TDS 8.0 support for PolyBase with ODBC.
RHEL 10 support in SQL Server 2025 RC1
You can now deploy SQL Server 2025 Preview on RHEL 10 for your development and test environments using the Enterprise Evaluation Edition, which is valid for 180 days. For production workloads, we currently recommend using SQL Server 2022 on RHEL 9.
Getting started
To get started with SQL Server 2025 public preview container images, use Podman:
podman pull mcr.microsoft.com/mssql/rhel/server:2025-latestStart the instance:
podman run -e 'ACCEPT_EULA=Y' -p 1433:1433 \
-e 'SA_PASSWORD=yourStrong(!)Password' \
-d mcr.microsoft.com/mssql/rhel/server:2025-latestWith the instance running, you can connect to your SQL Server using the sqlcmd tool inside the container. Run this command on the host:
podman exec -it <container_id or container_name> \
/opt/mssql-tools/bin/sqlcmd \
-S localhost -U sa -P <your_password>You can also use the tools in an entrypoint.sh script to do things like creating databases or logins, attaching databases, importing data, and other setup tasks. For information and examples, read using an entrypoint.sh and setup.sql script to create a database.
There are many ways to connect to your container, including GUI tools on the host or a remote computer. See the connect and query topic in the SQL Server on Linux documentation for more information.
Next steps
With SQL Server 2025 Preview now on RHEL 10, you get more flexibility, modern workloads, and AI-native capabilities, paired with mission-critical performance. Find out what's new in SQL Server 2025, and then follow the quickstart guide and deploy SQL Server 2025 RC1 on RHEL 10 in your test environment.
Red Hat Enterprise Linux 10 is a powerful,and high-performance platform, built with enhanced security capabilities in mind, that’s ready to run SQL Server 2025. Its built-in AI management tools (Red Hat Lightspeed) simplify everyday tasks, freeing IT teams to focus on innovation rather than routine maintenance. With advanced security features like post-quantum cryptography and integrated hardware security support, RHEL 10 delivers reliability and compliance you can trust, making it the ideal choice for both development and production SQL Server workloads. Try it out today on the Red Hat Ecosystem Catalog.
关于作者
Amit is a Principal Product Manager at Microsoft with over 15 years of experience. He has been pivotal in the development and evolution of SQL Server on Linux, significantly contributing to Microsoft's cross-platform solutions. Currently, he oversees SQL Server on Linux and Containers. With over a decade of database experience, he has designed SQL Server-based data platforms for Tier 1 customers across diverse business segments.
