Tutorial: Migrate online from an Azure VM or an on-premises PostgreSQL server to Azure Database for PostgreSQL with the migration service Preview

This article guides you using the Azure portal to migrate your PostgreSQL database from an Azure VM or an on-premises PostgreSQL server to an Azure Database for PostgreSQL. The Azure portal allows you to perform various tasks, including database migration. Following the steps outlined in this tutorial, you can seamlessly transfer your database to Azure and take advantage of its powerful features and scalability.

Configure the migration task

The migration service comes with a simple, wizard-based experience on the Azure portal. Here's how to start:

1. Open your web browser and go to the portal. Enter your credentials to sign in. The default view is your service dashboard.

2. Go to your Azure Database for PostgreSQL Flexible Server target.

3. In the Flexible Server's Overview tab, on the left menu, scroll down to Migration and select it.

   

4. Select the Create button to migrate from an Azure virtual machine (VM) or an on-premises PostgreSQL server to the Flexible Server. If this is your first time using the migration service, an empty grid appears with a prompt to begin your first migration.

   

   If you've already created migrations to your Flexible Server target, the grid contains information about attempted migrations.

5. Select the Create button. Then, you go through a wizard-based series of tabs to create a migration into this Flexible Server target from the PostgreSQL source Server.

Setup

The first tab is the setup tab, where the user initiates the migrations by providing migration details like migration name and source type.

• Migration name is the unique identifier for each migration to this Flexible Server target. This field accepts only alphanumeric characters and doesn't accept any special characters except a hyphen (-). The name can't start with a hyphen and should be unique for a target server. No two migrations to the same Flexible Server target can have the same name.

• Source Server Type — Depending on your PostgreSQL source, you can select Azure VM or on-premises.

• Migration Option allows you to perform validations before triggering a migration. You can pick any of the following options:

  • Validate - Checks your server and database readiness for migration to the target.
  • Migrate - Skips validations and starts migrations.
  • Validate and Migrate—Performs validation before triggering a migration. The migration is triggered only if there are no validation failures.

Choosing the Validate or Validate and Migrate option is always a good practice when performing premigration validations before running the migration. To learn more about the premigration validation, refer to this documentation.

Migration mode allows you to choose the mode for the migration. Offline is the default option.

Select the Next: Connect to source button.

Runtime Server

The Migration Runtime Server is a specialized feature within the migration service in Azure Database for PostgreSQL, designed to act as an intermediary server during migration. It's a separate Azure Database for PostgreSQL - Flexible Server instance that isn't the target server but is used to facilitate the migration of databases from a source environment that is only accessible via a private network.

For more information about the Runtime Server, visit the Migration Runtime Server.

Connect to source

The Connect to Source tab prompts you to give details related to the source selected in the Setup Tab that is the source of the databases.

• Server Name - Provide the Hostname or the IP address of the source PostgreSQL instance
• Port - Port number of the Source server
• Server admin login name - Username of the source PostgreSQL server
• Password - Password of the source PostgreSQL server
• SSL Mode - Supported values are preferred and required. When the SSL at the source PostgreSQL server is OFF, use the SSLMODE=prefer. If the SSL at the source server is ON, use the SSLMODE=require. SSL values can be determined in postgresql.conf file.
• Test Connection - Performs the connectivity test between target and source. Once the connection is successful, users can proceed with the next step. Otherwise, you need to identify the networking issues between the target and source and verify the username/password for the source. Test connection takes a few minutes to establish a connection between the target and source

After the successful test connection, select the Next: Select Migration target

Select migration target

The select migration target tab displays metadata for the Flexible Server target, such as the subscription name, resource group, server name, location, and PostgreSQL version.

• Admin username - Admin username of the target PostgreSQL server
• Password - Password of the target PostgreSQL server
• Custom FQDN/IP (Optional): The custom FQDN/IP field is optional and can be used when the target is behind a custom DNS server or has custom DNS namespaces, making it accessible only via specific FQDNs or IP addresses. For example, this could include entries like flexibleserver.example.com, 198.1.0.2, or a PostgreSQL FQDN such as flexibleserver.postgres.database.chinacloudapi.cn, if the custom DNS server contains the DNS zone postgres.database.chinacloudapi.cn or forwards queries for this zone to 168.63.129.16, where the FQDN is resolved in the Azure public or private DNS zone.
• Test Connection - Performs the connectivity test between target and source. Once the connection is successful, users can proceed with the next step. Otherwise, we need to identify the networking issues between the target and the source and verify the username/password for the target. The test connection takes a few minutes to establish a connection between the target and the source.

After the successful test connection, select the Next: Select Database(s) for Migration

Select databases for migration

Under this tab, a list of user databases is inside the source server selected in the setup tab. You can select and migrate up to eight databases in a single migration attempt. If there are more than eight user databases, the migration process is repeated between the source and target servers for the next set of databases.

After selecting the databases, select the Next: Summary

Summary

The Summary tab summarizes all the source and target details for creating the validation or migration. Review the details and select the start button.

Monitor the migration

After you select the start button, a notification appears in a few seconds, saying that the validation or migration creation is successful. You're redirected automatically to the Migration blade of Flexible Server, which has a new entry for the recently created validation or migration.

The grid that displays the migrations has these columns: Name, Status, Migration mode, Migration type, Source server, Source server type, Databases, **Duration, and Start time. The entries are displayed in the descending order of the start time, with the most recent entry at the top. You can use the refresh button to refresh the status of the validation or migration. Select the migration name in the grid to see the associated details.

When the validation or migration is created, it moves to the InProgress state and PerformingPreRequisiteSteps substate. The workflow takes 2 to 3 minutes to set up the migration infrastructure and network connections.

Migration details

In the Setup tab, we have selected the migration option as Migrate and Validate. In this scenario, validations are performed first before migration starts. After the PerformingPreRequisiteSteps substate is completed, the workflow moves into the substate of Validation in Progress.

• If validation has errors, the migration moves into a Failed state.
• If validation completes without error, the migration starts, and the workflow moves into the substate of Migrating Data.

The validation results are displayed under the Validation tab, and the migration is monitored under the Migration tab.

Some possible migration states:

Migration states

Migration substates

Validation substates

Cutover

If there are both Migrate and Validate and Migrate, completing the online migration requires another step—the user must take a Cutover action. After the copy/clone of the base data is complete, the migration moves to the WaitingForUserAction state and the WaitingForCutoverTrigger substrate. In this state, the user can trigger the cutover from the portal by selecting the migration.

Before initiating cutover, it's important to ensure that:

• Writes to the source are stopped - Latency value is 0 or close to 0. The Latency information can be obtained from the migration details screen as shown below:
• latency value decreases to 0 or close to 0
• The latency value indicates when the target last synced with the source. Writing to the source can be stopped at this point, and a cutover can be initiated. In case there's heavy traffic at the source, it's recommended to stop writes first so that Latency can come close to 0, and then a cutover is initiated.

The Cutover operation applies all pending changes from the Source to the Target and completes the migration. If you trigger a "Cutover" even with nonzero Latency, the replication stops until that point in time. All the data on the source until the cutover point is then applied to the target. If you experience a latency of 15 minutes at the cutover point, all the changed data in the last 15 minutes are applied to the target. The time depends on the backlog of changes occurring in the last 15 minutes. Hence, it's recommended that the latency go to zero or near zero before triggering the cutover.

• The migration moves to the Succeeded state when the Migrating Data substate or the cutover (in Online migration) finishes successfully. If there's a problem at the Migrating Data substate, the migration moves into a Failed state.

Cancel the migration

You can cancel any ongoing validations or migrations. The workflow must be in the InProgress state to be canceled. You can't cancel a validation or migration in the Succeeded or Failed state.

Canceling a validation stops any further validation activity and the validation moves to a Canceled state.

Canceling a migration stops further migration activity on your target server and moves to a Canceled state. It doesn't drop or roll back any changes on your target server. Be sure to drop the databases on your target server that is involved in a canceled migration.

This article explores using the Azure CLI to migrate your PostgreSQL database from an Azure virtual machine or an on-premises PostgreSQL instance to an Azure Database for PostgreSQL. The Azure CLI provides a powerful and flexible command-line interface that allows you to perform various tasks, including database migration. Following the steps outlined in this article, you can seamlessly transfer your database to Azure and take advantage of its powerful features and scalability.

To learn more about Azure CLI with the migration service, visit How to set up Azure CLI for the migration service.

Once the CLI is installed, open the command prompt and log into your Azure account using the below command.

az login

Configure the migration task

To begin the migration, create a JSON file with the migration details. The JSON file contains the following information:

• Edit the below placeholders << >> in the JSON lines and store them in the local machine as <<filename>>.json where the CLI is being invoked. In this tutorial, we have saved the file in C:\migration-CLI\migration_body.json

{
"properties": {
"SourceDBServerResourceId": "<<source hostname or IP address>>:<<port>>@<<username>>",
        "SecretParameters": {
            "AdminCredentials": {
                "SourceServerPassword": "<<Source Password>>",
                "TargetServerPassword": "<<Target Password>>"
            },
			"targetServerUserName": "<<Target username>>"
        },
        "DBsToMigrate": "<<comma separated list of databases in a array like - ["ticketdb","timedb","inventorydb"]>>",
        "OverwriteDBsInTarget": "true",
        "sourceType": "OnPremises",
        "sslMode": "Prefer"
    }
}

• Run the following command to check if any migrations are running. The migration name is unique across the migrations within the Azure Database for PostgreSQL flexible server target.

  az postgres flexible-server migration list --subscription 11111111-1111-1111-1111-111111111111 --resource-group my-learning-rg --name myflexibleserver --filter All
  

• In the above steps, there are no migrations performed so we start with the new migration by running the following command

  az postgres flexible-server migration create --subscription 11111111-1111-1111-1111-111111111111 --resource-group my-learning-rg --name myflexibleserver --migration-name migration1 --migration-mode online --migration-option ValidateAndMigrate --properties "C:\migration-cli\migration_body.json"
  

• Run the following command to initiate the migration status in the previous step. You can check the status of the migration by providing the migration name

  az postgres flexible-server migration show --subscription 11111111-1111-1111-1111-111111111111 --resource-group my-learning-rg --name myflexibleserver --migration-name migration1
  

• The status of the migration progress is shown in the Azure CLI.

• You can also see the status of the Azure Database for PostgreSQL flexible server in the Azure portal.

• You can cancel any ongoing migration attempts using the cancel command. This command stops the particular migration attempt and rolls back all changes on your target server. Here's the CLI command to delete a migration:

  az postgres flexible-server migration update cancel --subscription 11111111-1111-1111-1111-111111111111 --resource-group my-learning-rg --name myflexibleserver --migration-name migration1
  

Cutover

• After the base data migration is complete in online migrations, the migration task moves to the WaitingForCutoverTrigger substate. In this state, the user can trigger the cutover through the CLI using the command below. The cutover can also be triggered from the portal by selecting the migration name in the migration grid.

• You can also initiate the cutover from the Azure portal.

  az postgres flexible-server migration update --subscription 11111111-1111-1111-1111-111111111111 --resource-group my-learning-rg --name myflexibleserver --migration-name migration1 --cutover