ステージングインスタンスのセットアップ

別の分離された環境に GitHub Enterprise Server インスタンスを設定し、そのインスタンスを使って変更の検証とテストを行うことができます。

この記事の内容

About staging instances

GitHub recommends that you set up a separate environment to test backups, updates, or changes to the configuration for your GitHub Enterprise Server instance. This environment, which you should isolate from your production systems, is called a staging environment.

For example, to protect against loss of data, you can regularly validate the backup of your production instance. You can regularly restore the backup of your production data to a separate GitHub Enterprise Server instance in a staging environment. On this staging instance, you could also test the upgrade to the latest feature release of GitHub Enterprise Server.

Tip

You may reuse your existing GitHub Enterprise license file as long as the staging instance is not used in a production capacity.

Considerations for a staging environment

To thoroughly test GitHub Enterprise Server and recreate an environment that's as similar to your production environment as possible, consider the external systems that interact with your instance. For example, you may want to test the following in your staging environment.

  • Authentication, especially if you use an external authentication provider like SAML
  • Integration with an external ticketing system
  • Integration with a continuous integration server
  • External scripts or software that use the GitHub Enterprise Server APIs
  • External SMTP server for email notifications

Setting up a staging instance

You can set up a staging instance from scratch and configure the instance however you like. For more information, see Setting up a GitHub Enterprise Server instance and Configuring GitHub Enterprise.

Alternatively, you can create a staging instance that reflects your production configuration by restoring a backup of your production instance to the staging instance.

  1. Back up your production instance.
  2. Set up a staging instance.
  3. Configure GitHub Actions.
  4. Configure GitHub Packages.
  5. Restore your production backup.
  6. Review the instance's configuration.
  7. Apply the instance's configuration.

1. Back up your production instance

If you want to test changes on an instance that contains the same data and configuration as your production instance, back up the data and configuration from the production instance using GitHub Enterprise Server Backup Utilities. For more information, see Configuring backups on your instance.

Warning

If you use GitHub Actions or GitHub Packages in production, your backup will include your production configuration for external storage. To avoid potential loss of data by writing to your production storage from your staging instance, you must configure each feature in steps 3 and 4 before you restore your backup.

2. Set up a staging instance

Set up a new instance to act as your staging environment. You can use the same guides for provisioning and installing your staging instance as you did for your production instance. For more information, see Setting up a GitHub Enterprise Server instance.

If you plan to restore a backup of your production instance, continue to the next step. Alternatively, you can configure the instance manually and skip the following steps.

Warning

Restoring backups with GitHub Actions will not succeed on a non-configured instance. To enable GitHub Actions, an instance with a hostname configured is required. For more information, see Configuring the hostname for your instance.

3. Configure GitHub Actions

Optionally, if you use GitHub Actions on your production instance, configure the feature on the staging instance before restoring your production backup. If you don't use GitHub Actions, skip to 1. Configure GitHub Packages.

Warning

If you don't configure GitHub Actions on the staging instance before restoring your production backup, your staging instance will use your production instance's external storage, which could result in loss of data. We strongly recommended that you use different external storage for your staging instance. For more information, see Using a staging environment.

  1. SSH into the staging instance. For more information, see "Accessing the administrative shell (SSH)."

    Shell
    ssh -p 122 admin@HOSTNAME

  2. To configure the staging instance to use an external storage provider for GitHub Actions, enter one of the following commands.

    • Azure Blob Storage:

      Shell
      ghe-config secrets.actions.storage.blob-provider "azure"

    • Amazon S3:

      Shell
      ghe-config secrets.actions.storage.blob-provider "s3"

    • Google Cloud Storage:

      Shell
      ghe-config secrets.actions.storage.blob-provider "gcs"

  3. Configure the external storage connection by entering the following commands, replacing the placeholder values with actual values for your connection.

    • Azure Blob Storage:

      Shell
      ghe-config secrets.actions.storage.azure.connection-string "CONNECTION STRING"

    • Amazon S3:

      Shell
      ghe-config secrets.actions.storage.s3.bucket-name "S3 BUCKET NAME"
ghe-config secrets.actions.storage.s3.service-url "S3 SERVICE URL"
ghe-config secrets.actions.storage.s3.access-key-id "S3 ACCESS KEY ID"
ghe-config secrets.actions.storage.s3.access-secret "S3 ACCESS SECRET"

      Optionally, to force path-style addressing for S3, also enter the following command.

      Shell
      ghe-config secrets.actions.storage.s3.force-path-style true

    • Google Cloud Storage:

      Shell
      ghe-config secrets.actions.storage.gcs.service-url "SERVICE URL"
ghe-config secrets.actions.storage.gcs.bucket-name "BUCKET NAME"
ghe-config secrets.actions.storage.gcs.access-key-id "HMAC ACCESS ID"
ghe-config secrets.actions.storage.gcs.access-secret "HMAC SECRET"

  4. To prepare to enable GitHub Actions on the staging instance, enter the following command.

    Shell
    ghe-config app.actions.enabled true

  5. To apply the configuration changes, enter the following command.

    Shell
    ghe-config-apply

4. Configure GitHub Packages

Optionally, if you use GitHub Packages on your production instance, configure the feature on the staging instance before restoring your production backup. If you don't use GitHub Packages, skip to 1. Restore your production backup.

Warning

If you don't configure GitHub Packages on the staging instance before restoring your production backup, your staging instance will use your production instance's external storage, which could result in loss of data. We strongly recommended that you use different external storage for your staging instance.

  1. Review the backup you will restore to the staging instance.

    • If you took the backup with GitHub Enterprise Server Backup Utilities 3.5 or later, the backup includes the configuration for GitHub Packages. Continue to the next step.
    • If you took the backup with GitHub Enterprise Server Backup Utilities 3.4 or earlier, configure GitHub Packages on the staging instance. For more information, see Getting started with GitHub Packages for your enterprise.

  2. SSH into the staging instance. For more information, see "Accessing the administrative shell (SSH)."

    Shell
    ssh -p 122 admin@HOSTNAME

  3. Configure the external storage connection by entering the following commands, replacing the placeholder values with actual values for your connection.

    • Azure Blob Storage:

      Shell
      ghe-config secrets.packages.blob-storage-type "azure"
ghe-config secrets.packages.azure-container-name "AZURE CONTAINER NAME"
ghe-config secrets.packages.azure-connection-string "CONNECTION STRING"

    • Amazon S3:

      Shell
      ghe-config secrets.packages.blob-storage-type "s3"
ghe-config secrets.packages.service-url "S3 SERVICE URL"
ghe-config secrets.packages.s3-bucket "S3 BUCKET NAME"
ghe-config secrets.packages.aws-access-key "S3 ACCESS KEY ID"
ghe-config secrets.packages.aws-secret-key "S3 ACCESS SECRET"

  4. To prepare to enable GitHub Packages on the staging instance, enter the following command.

    Shell
    ghe-config app.packages.enabled true

5. Restore your production backup

Use the ghe-restore command to restore the rest of the data from the backup. For more information, see Configuring backups on your instance.

If the staging instance is already configured and you want to overwrite settings, certificate, and license data, add the -c option to the command. For more information about the option, see Using the backup and restore commands in the GitHub Enterprise Server Backup Utilities documentation.

6. Review the instance's configuration

To access the staging instance using the same hostname, update your local hosts file to resolve the staging instance's hostname by IP address by editing the /etc/hosts file in macOS or Linux, or the C:\Windows\system32\drivers\etc file in Windows.

Note

Your staging instance must be accessible from the same hostname as your production instance. Changing the hostname for your GitHub Enterprise Server instance is not supported. For more information, see Configuring the hostname for your instance.

Then, review the staging instance's configuration in the Management Console. For more information, see Administering your instance from the web UI.

Warning

If you configured GitHub Actions or GitHub Packages for the staging instance, to avoid overwriting production data, ensure that the external storage configuration in the Management Console does not match your production instance.

7. Apply the instance's configuration

To apply the configuration from the Management Console, click Save settings.

Bringing a staging instance back online

You may want to power off a staging instance to save costs and power it back on when needed.

An instance can stay offline for 60 days as of the latest patch release of this version, increased from 7 days.

If you bring the instance back online within the allowed offline time period, GitHub Enterprise Server instantiates successfully. If the instance stays offline for longer than the allowed period, GitHub Enterprise Server fails to instantiate successfully, and an error message with the text server has been offline for more than the configured server_rejoin_age_max may appear in the system logs. See About system logs.

If the instance is stuck in the error state, you can run these commands to recover.

sudo mv /data/user/consul/server_metadata.json /data/user/consul/server_metadata.json.bak
ghe-config-apply

Further reading