Skip to content

Latest commit

 

History

History
281 lines (198 loc) · 12.1 KB

sql-server-linux-availability-group-cross-platform.md

File metadata and controls

281 lines (198 loc) · 12.1 KB
title description author ms.author ms.reviewer ms.date ms.service ms.subservice ms.topic ms.custom monikerRange
Configure SQL Server Always on Availability Group on Windows and Linux
Learn how to create a SQL Server Always On Availability Group (AG) with one replica on a Windows server and the other replica on a Linux server.
rwestMSFT
randolphwest
vanto
11/18/2024
sql
linux
how-to
linux-related-content
>=sql-server-2017

Configure SQL Server Always On availability group on Windows and Linux (cross-platform)

[!INCLUDE tsql-appliesto-sslinux-only]

This article explains the steps to create an Always On availability group (AG) with one replica on a Windows server and the other replica on a Linux server.

Important

[!INCLUDE ssnoversion-md] cross-platform availability groups, which include heterogeneous replicas with complete high-availability and disaster recovery support, is available with DH2i DxEnterprise. For more information, see SQL Server Availability Groups with Mixed Operating Systems.

View the following video to find out about cross-platform availability groups with DH2i.

[!VIDEO https://learn-video.azurefd.net/vod/player?show=data-exposed&ep=get-started-with-sql-server-ags-across-windows-linux-and-container-replicas]

This configuration is cross-platform because the replicas are on different operating systems. Use this configuration for migration from one platform to the other or disaster recovery (DR). This configuration doesn't support high availability.

:::image type="content" source="media/sql-server-linux-availability-group-cross-platform/cluster-type-none.png" alt-text="Diagram of Availability group with cluster type of None.":::

Before proceeding, you should be familiar with installation and configuration for SQL Server instances on Windows and Linux.

Scenario

In this scenario, two servers are on different operating systems. A Windows Server 2022 named WinSQLInstance hosts the primary replica. A Linux server named LinuxSQLInstance host the secondary replica.

Configure the AG

The steps to create the AG are the same as the steps to create an AG for read-scale workloads. The AG cluster type is NONE, because there's no cluster manager.

For the scripts in this article, angle brackets < and > identify values that you must replace for your environment. The angle brackets themselves aren't required for the scripts.

  1. Install [!INCLUDE sssql22-md] on Windows Server 2022, enable Always On Availability Groups from SQL Server Configuration Manager, and set mixed mode authentication.

    [!TIP]
    If you're validating this solution in Azure, place both servers in the same availability set to ensure they are separated in the data center.

    Enable Availability Groups

    For instructions, see Enable or disable Always On availability group feature.

    :::image type="content" source="media/sql-server-linux-availability-group-cross-platform/configuration-manager.png" alt-text="Screenshot showing how to enable Availability Groups." lightbox="media/sql-server-linux-availability-group-cross-platform/configuration-manager.png":::

    SQL Server Configuration Manager notes that the computer isn't a node in a failover cluster.

    After you enable Availability Groups, restart SQL Server.

    Set mixed mode authentication

    For instructions, see Change server authentication mode.

  2. Install [!INCLUDE sssql22-md] on Linux. For instructions, see Installation guidance for SQL Server on Linux. Enable hadr with mssql-conf.

    To enable hadr via mssql-conf from a shell prompt, issue the following command:

    sudo /opt/mssql/bin/mssql-conf set hadr.hadrenabled 1

    After you enable hadr, restart the SQL Server instance:

    sudo systemctl restart mssql-server.service

  3. Configure the hosts file on both servers, or register the server names with DNS.

  4. Open up firewall ports for TCP 1433 and 5022 on both Windows and Linux.

  5. On the primary replica, create a database login and password.

    CREATE LOGIN dbm_login
    WITH PASSWORD = '<password>';

CREATE USER dbm_user FOR LOGIN dbm_login;
GO

    [!CAUTION]
    [!INCLUDE password-complexity]

  6. On the primary replica, create a master key and certificate, then back up the certificate with a private key.

    CREATE MASTER KEY ENCRYPTION BY PASSWORD = '<master-key-password>';

CREATE CERTIFICATE dbm_certificate
    WITH SUBJECT = 'dbm';

BACKUP CERTIFICATE dbm_certificate TO FILE = 'C:\Program Files\Microsoft SQL Server\MSSQL16.MSSQLSERVER\MSSQL\DATA\dbm_certificate.cer'
    WITH PRIVATE KEY (
         FILE = 'C:\Program Files\Microsoft SQL Server\MSSQL16.MSSQLSERVER\MSSQL\DATA\dbm_certificate.pvk',
         ENCRYPTION BY PASSWORD = '<private-key-password>'
);
GO

    [!CAUTION]
    [!INCLUDE password-complexity]

  7. Copy the certificate and private key to the Linux server (secondary replica) at /var/opt/mssql/data. You can use pscp to copy the files to the Linux server.

  8. Set the group and ownership of the private key and the certificate to mssql:mssql.

    The following script sets the group and ownership of the files.

    sudo chown mssql:mssql /var/opt/mssql/data/dbm_certificate.pvk
sudo chown mssql:mssql /var/opt/mssql/data/dbm_certificate.cer

    In the following diagram, ownership and group are set correctly for the certificate and key.

    :::image type="content" source="media/sql-server-linux-availability-group-cross-platform/certificate-key-owner-group.png" alt-text="Screenshot of a Git Bash window showing the .cer and the .pvk in the /var/opt/mssql/data folder.":::

  9. On the secondary replica, create a database login and password and create a master key.

    CREATE LOGIN dbm_login
    WITH PASSWORD = '<password>';

CREATE USER dbm_user FOR LOGIN dbm_login;
GO

CREATE MASTER KEY ENCRYPTION BY PASSWORD = '<master-key-password>';
GO

    [!CAUTION]
    [!INCLUDE password-complexity]

  10. On the secondary replica, restore the certificate you copied to /var/opt/mssql/data.

    CREATE CERTIFICATE dbm_certificate
    AUTHORIZATION dbm_user
    FROM FILE = '/var/opt/mssql/data/dbm_certificate.cer'
    WITH PRIVATE KEY (
        FILE = '/var/opt/mssql/data/dbm_certificate.pvk',
        DECRYPTION BY PASSWORD = '<private-key-password>'
);
GO

    In the previous example, replace <private-key-password> with the same password you used when creating the certificate on the primary replica.

  11. On the primary replica, create an endpoint.

    CREATE ENDPOINT [Hadr_endpoint]
    AS TCP
(
            LISTENER_IP = (0.0.0.0),
            LISTENER_PORT = 5022
)
    FOR DATABASE_MIRRORING
(
            ROLE = ALL,
            AUTHENTICATION = CERTIFICATE dbm_certificate,
            ENCRYPTION = REQUIRED ALGORITHM AES
);

ALTER ENDPOINT [Hadr_endpoint]
    STATE = STARTED;

GRANT CONNECT
    ON ENDPOINT::[Hadr_endpoint] TO [dbm_login];
GO

    [!IMPORTANT]
    The firewall must be open for the listener TCP port. In the preceding script, the port is 5022. Use any available TCP port.

  12. On the secondary replica, create the endpoint. Repeat the preceding script on the secondary replica to create the endpoint.

  13. On the primary replica, create the AG with CLUSTER_TYPE = NONE. The example script uses SEEDING_MODE = AUTOMATIC to create the AG.

    [!NOTE]
    When the Windows instance of SQL Server uses different paths for data and log files, automatic seeding fails to the Linux instance of SQL Server, because these paths don't exist on the secondary replica. To use the following script for a cross-platform AG, the database requires the same path for the data and log files on the Windows server. Alternatively you can update the script to set SEEDING_MODE = MANUAL and then back up and restore the database with NORECOVERY to seed the database.

    This behavior applies to Azure Marketplace images.

    For more information about automatic seeding, see Automatic Seeding - Disk Layout.

    Before you run the script, update the values for your AGs.

    • Replace <WinSQLInstance> with the server name of the primary replica SQL Server instance.

    • Replace <LinuxSQLInstance> with the server name of the secondary replica SQL Server instance.

    To create the AG, update the values and run the script on the primary replica.

    CREATE AVAILABILITY
GROUP [ag1]
WITH (CLUSTER_TYPE = NONE)
FOR REPLICA
    ON N'<WinSQLInstance>'
WITH (
    ENDPOINT_URL = N'tcp://<WinSQLInstance>:5022',
    AVAILABILITY_MODE = ASYNCHRONOUS_COMMIT,
    SEEDING_MODE = AUTOMATIC,
    FAILOVER_MODE = MANUAL,
    SECONDARY_ROLE(ALLOW_CONNECTIONS = ALL)
    ),
    N'<LinuxSQLInstance>'
WITH (
    ENDPOINT_URL = N'tcp://<LinuxSQLInstance>:5022',
    AVAILABILITY_MODE = ASYNCHRONOUS_COMMIT,
    SEEDING_MODE = AUTOMATIC,
    FAILOVER_MODE = MANUAL,
    SECONDARY_ROLE(ALLOW_CONNECTIONS = ALL);
    )
GO

    For more information, see CREATE AVAILABILITY GROUP.

  14. On the secondary replica, join the AG.

    ALTER AVAILABILITY GROUP [ag1] JOIN WITH (CLUSTER_TYPE = NONE);
ALTER AVAILABILITY GROUP [ag1] GRANT CREATE ANY DATABASE;
GO

  15. Create a database for the AG. The example steps use a database named TestDB. If you're using automatic seeding, set the same path for both the data and the log files.

    Before you run the script, update the values for your database.

    • Replace TestDB with the name of your database.

    • Replace <F:\Path> with the path for your database and log files. Use the same path for the database and log files.

    You can also use the default paths.

    To create your database, run the script.

    CREATE DATABASE [TestDB] CONTAINMENT = NONE
    ON
    PRIMARY(NAME = N'TestDB', FILENAME = N'<F:\Path>\TestDB.mdf')
    LOG ON (NAME = N'TestDB_log', FILENAME = N'<F:\Path>\TestDB_log.ldf');
GO

  16. Take a full backup of the database.

  17. If you aren't using automatic seeding, restore the database on the secondary replica (Linux) server. Migrate a SQL Server database from Windows to Linux using backup and restore. Restore the database WITH NORECOVERY on the secondary replica.

  18. Add the database to the AG. Update the example script. Replace TestDB with the name of your database. On the primary replica, run the T-SQL query to add the database to the AG.

    ALTER AG [ag1] ADD DATABASE TestDB;
GO

  19. Verify that the database is getting populated on the secondary replica.

Fail over the primary replica

[!INCLUDE Force failover]

This article reviewed the steps to create a cross-platform AG to support migration or read-scale workloads. It can be used for manual disaster recovery. It also explained how to fail over the AG. A cross-platform AG uses cluster type NONE and doesn't support high availability.

Related content