Back up Concourse with BOSH Backup and Restore (BBR)

VMware recommends that you perform Concourse backups with BOSH Backup and Restore (BBR). The process involves the following prerequisites:

• A Concourse deployment that includes the BOSH Backup and Restore SDK
• bbr, the BOSH Backup and Restore (BBR) CLI tool
• fly, Concourse's CLI tool

This guide details the process of setting this system up. Once set up, you can follow the steps in the the Performing a Backup section of this guide.

Then, once you've created a valid BBR backup of Concourse, you can use the Performing a restore with BBR guide.

Use the table of contents on the right to navigate this guide.

You can use the bosh releases command to see what you have installed so far. A backup-and-restore-sdk release is required for each Concourse deployment you want to backup from or restore to.

Set up the concourse-bosh-deployment directory on your local machine

Any Concourse deployment you wish to backup from or restore to must have the backup-and-restore-sdk release. If you are starting from scratch on a new workstation, the instructions in this section cover cloning the concourse-bosh-deployment repository to a local directory, so that you can craft a new deployment with the appropriate release and configuration. If you already have a deployment directory with manifest and ops files on your workstation, navigate to that directory before jumping ahead to the Set up the BBR SDK on your BOSH Environment below.

1. Clone the concourse-bosh-deployment repository by running the following command:

   git clone https://github.com/concourse/concourse-bosh-deployment.git
   

2. Move to the concourse-bosh-deployment directory:

   cd concourse-bosh-deployment
   

   All the paths used in this topic are relative to this directory.

3. Checkout the release that corresponds to your working version of Concourse. For example:

   git checkout v5.5.11
   

   git checkout v4.2.4
   

   git checkout v3.13.0
   

   For a list of all Concourse releases, see concourse-bosh-deployment in GitHub.

   Checking out a release rather than a branch means that git produces the following output:

   Note: checking out 'RELEASE'.
   
   You are in 'detached HEAD' state. You can look around, make experimental
   changes and commit them, and you can discard any commits you make in this
   state without impacting any branches by performing another checkout.
   
   If you want to create a new branch to retain commits you create, you may
   do so (now or later) by using -b with the checkout command again. Example:
   
       git checkout -b <new-branch-name>
   
   HEAD is now at HASH... COMMIT-MESSAGE
   

Set up Certificates, Log in, and Alias your BOSH Environment

If you already have your CA certificate and have already logged in to and aliased your BOSH environment, you can skip this section.

1. Get a CA certificate for your BOSH Director.

   1. If you created your BOSH Director manually, retrieve the credentials that were created during setup.

      If you are working with an Operations Manager-deployed BOSH Director, log in to Operations Manager and access the following endpoint in your Tanzu Operations Manager domain:

      https://OPS-MANAGER-DOMAIN/download_root_ca_cert
      

      Where OPS-MANAGER-DOMAIN is the Operations Manager domain.

      This will create a file in your downloads folder called root_ca_certificate.

   2. Copy and rename the root_ca_certificate file into the working directory. For example, on OSX:

      mv ~/Downloads/root_ca_certificate ./ca-cert.crt
      

   This certificate file can be used whenever a CLI command asks for a --ca-cert flag and value.

2. Ensure you are logged in to your BOSH environment with the appropriate BOSH Director credentials. For example, if you saved your CA certificate as ca-cert.crt, run the following command:

   bosh -e BOSH-ENVIRONMENT-IP login --ca-cert=ca-cert.crt
   

   Where BOSH-ENVIRONMENT-IP is your BOSH environment IP address.

   If you set up a BOSH Director directly, the director credentials were returned as a file after the setup process finished.

   If you have an Operations Manager-deployed BOSH Director, you can find the director credentials in the Operations Manager credentials tab or at the following endpoint in your Operations Manager domain:

   http://OPS-MANAGER-DOMAIN/api/v0/deployed/director/credentials/director_credentials
   

   Where OPS-MANAGER-DOMAIN is your Operations Manager domain.

3. Give your environment an alias by running the following command:

   bosh -e BOSH-ENVIRONMENT-IP alias-env ALIAS --ca-cert=ca-cert.crt
   

   Where:

   • BOSH-ENVIRONMENT-IP is your BOSH environment IP address
   • ALIAS is the alias you're creating BOSH environment

   Using an alias for your BOSH environment substantially reduces the keystrokes needed for commands in future. You can use this alias whenever you target this environment by using the -e flag in a BOSH command.

Set up the BBR SDK on your BOSH Environment

This section details the task of uploading the BOSH backup-and-restore-sdk release to a Concourse environment.

1. Upload the backup-and-restore-sdk release by running the following upload-release command:

   bosh \
   -e BOSH-ENVIRONMENT-ALIAS \
   upload-release \
   --sha1 364838c384f2edec80866b4abf2397c4c5d15c62 \
     https://bosh.io/d/github.com/cloudfoundry-incubator/backup-and-restore-sdk-release?v=1.15.1
   

   Where BOSH-ENVIRONMENT-ALIAS is your BOSH environment alias.

   If you run the bosh releases command again, you should see backup-and-restore-sdk in the list.

   An alternative approach is to download the latest release from the BOSH Backup and Restore Release Page and pass in the path to the file in the following format:
   bosh -e BOSH-ENVIRONMENT-ALIAS upload-release PATH-TO-FILE
   

2. If you're backing up version 3.13.0, you need to create and edit a file called backup-atc.yml in ./cluster/operations. Versions 4.2.4 and onwards of the concourse-bosh-deployment repo come with this file already.

   1. For example:

      vim ./cluster/operations/backup-atc.yml
      

   2. Edit this file so that it has the following contents:

      ---
      - type: replace
        path: /releases/-
        value:
          name: backup-and-restore-sdk
          url: https://bosh.io/d/github.com/cloudfoundry-incubator/backup-and-restore-sdk-release?v=((bbr_sdk_version))
          sha1: ((bbr_sdk_sha1))
          version: ((bbr_sdk_version))
      
      - type: replace
        path: /instance_groups/name=web/jobs/-
        value:
          name: bbr-atcdb
          release: concourse
          properties:
            postgresql:
              database: atc
              role:
                name: ((postgres_role))
                password: ((postgres_password))
                postgresql_database: atc
      
      
      - type: replace
        path: /instance_groups/name=web/jobs/-
        value:
          name: database-backup-restorer
          release: backup-and-restore-sdk
      

   3. Save and close the backup-atc.yml file.

3. Open your versions.yml file:

   vim versions.yml
   

4. Update versions.yml by adding lines for the backup-and-restore-sdk's version and SHA1. The two properties you need to define are bbr_sdk_version and bbr_sdk_sha1.

   To deploy backup-and-restore-sdk 1.15.1, copy and paste the highlighted lines into your versions.yml file:

   # this file is partially maintained by CI; the concourse and garden-runc
   # versions and sha1s are automatically bumped, while the rest are preserved
   # as-is.
   #
   # this should make getting started easy while being easy enough to maintain
   # manually. feel free to PR sane defaults along with newly supported
   # infrastructures and such!
   ---
   concourse_version: '5.5.11'
   concourse_sha1: '876a8aef1c790c6e8578a3b7c11c2265072f0565'
   garden_runc_version: '1.13.1'
   garden_runc_sha1: '54cbb89cae1be0708aa056185671665d7f4b2a4f'
   postgres_version: '36'
   postgres_sha1: '3dd10b417b21cfa3257f1cc891e9e46f02fefe16'
   virtualbox_cpi_version: '0.0.12'
   virtualbox_cpi_sha1: 537c77b087eeda859b1eb89553efa6f808214d50
   virtualbox_stemcell_version: '3468.5'
   virtualbox_stemcell_sha1: 25301a41c15caad859b66efc0fd75bacdfc57dab
   bbr_sdk_version: '1.15.1'
   bbr_sdk_sha1: '364838c384f2edec80866b4abf2397c4c5d15c62'
   

   # this file is partially maintained by CI; the concourse and garden-runc
   # versions and sha1s are automatically bumped, while the rest are preserved
   # as-is.
   #
   # this should make getting started easy while being easy enough to maintain
   # manually. feel free to PR sane defaults along with newly supported
   # infrastructures and such!
   ---
   concourse_version: '4.2.4'
   concourse_sha1: '56122002ce49af09b03b8c3a49bdb7578ee2868a'
   garden_runc_version: '1.13.1'
   garden_runc_sha1: '54cbb89cae1be0708aa056185671665d7f4b2a4f'
   postgres_version: '36'
   postgres_sha1: '3dd10b417b21cfa3257f1cc891e9e46f02fefe16'
   virtualbox_cpi_version: '0.0.12'
   virtualbox_cpi_sha1: 537c77b087eeda859b1eb89553efa6f808214d50
   virtualbox_stemcell_version: '3468.5'
   virtualbox_stemcell_sha1: 25301a41c15caad859b66efc0fd75bacdfc57dab
   bbr_sdk_version: '1.15.1'
   bbr_sdk_sha1: '364838c384f2edec80866b4abf2397c4c5d15c62'
   

   # this file is partially maintained by CI; the concourse and garden-runc
   # versions and sha1s are automatically bumped, while the rest are preserved
   # as-is.
   #
   # this should make getting started easy while being easy enough to maintain
   # manually. feel free to PR sane defaults along with newly supported
   # infrastructures and such!
   ---
   concourse_version: '3.13.0'
   concourse_sha1: 'aaa4e4d42adb2293abc79422351ca71ed548f95c'
   garden_runc_version: '1.13.1'
   garden_runc_sha1: '54cbb89cae1be0708aa056185671665d7f4b2a4f'
   postgres_version: '36'
   postgres_sha1: '3dd10b417b21cfa3257f1cc891e9e46f02fefe16'
   virtualbox_cpi_version: '0.0.12'
   virtualbox_cpi_sha1: 537c77b087eeda859b1eb89553efa6f808214d50
   virtualbox_stemcell_version: '3468.5'
   virtualbox_stemcell_sha1: 25301a41c15caad859b66efc0fd75bacdfc57dab
   bbr_sdk_version: '1.15.1'
   bbr_sdk_sha1: '364838c384f2edec80866b4abf2397c4c5d15c62'
   

   If you want to use a different version of backup-and-restore-sdk-release, see backup-and-restore-sdk Release to find the version and SHA1 for your release.

Concourse and BOSH Variables

Deploying Concourse with BOSH requires you to set some variables to fill in details in the manifest that are written as placeholders. These are in the format ((placeholder)). The currently deployed Concourse manifest contains the values that you will use to fill out each of the variables.

The preferred approach is to create a file to define a key/value pair for each required variable, and check it into version control so that it can be safely and easily maintained over time. You might already have a solution for this; if not, follow the steps below.

The following steps describe setting variables that are necessary for deployment in a yaml file. If you prefer to set them on the command line when you run bosh deploy, you can pass them in with the --var KEY=VALUE flag and syntax. For more information about bosh deploy, see Deploy in the BOSH documentation.

1. Run the following command to fetch the manifest and put it in a deployed-manifest.yml file:

   bosh \
   -e BOSH-ENVIRONMENT-ALIAS \
   -d NAME-OF-CONCOURSE-DEPLOYMENT \
   manifest > deployed-manifest.yml
   

   Sending the response to a file is a security precaution that ensures that any included keys are not echoed to your terminal session history. Ensure that you destroy this file after you are done with this backup process.

2. If you do not already have a variable file from a previous deployment, create a file to store your Concourse and BOSH-related environment variables. For example, use vim to create a file called variables.yml by running:

   vim variables.yml
   

   This file will hold all of the variables that BOSH needs to interpolate into references in the manifest file.

   If you have existing variables file, open that file instead.

   The benefit of creating a file for these variables is that you can check this file into your version control, track changes over time, and more easily document and share this information with others in your organization. This also saves you from looking up configuration details every time you need to change your Concourse deployment.

3. Ensure you have the correct key/value pairs in your variables file:

   • If you already have a variables file: Ensure that you have each of the key/value pairs listed in the snippet below. If any are missing, add them in.
   • If you have created a new file: Copy and paste the following snippet to use as a template. Replace each value placeholder in the snippet above as appropriate for your deployment and cloud configuration, using the properties in the deployed-manifest.yml file that created in step 1 of this section.

   ---
   local_user:
       username: CHOOSE-A-USERNAME
       password: CHOOSE-A-PASSWORD
   
   deployment_name: DEPLOYMENT-NAME
   db_persistent_disk_type: PERSISTANT-DISK-TYPE-FOR-DATABASE
   db_vm_type: VM-TYPE-FOR-DATABASE-INSTANCE
   external_url: EXTERNAL-URL
   network_name: NETWORK-NAME
   postgres_password: PASSWORD
   web_ip: WEB-IP
   web_network_name: WEB-NETWORK-NAME
   web_vm_type: VM-TYPE-FOR-WEB-INSTANCE
   worker_vm_type: VM-TYPE-FOR-WORKER-INSTANCE
   

   ---
   local_user:
       username: CHOOSE-A-USERNAME
       password: CHOOSE-A-PASSWORD
   
   deployment_name: DEPLOYMENT-NAME
   db_persistent_disk_type: PERSISTANT-DISK-TYPE-FOR-DATABASE
   db_vm_type: VM-TYPE-FOR-DATABASE-INSTANCE
   external_url: EXTERNAL-URL
   network_name: NETWORK-NAME
   postgres_password: PASSWORD
   web_ip: WEB-IP
   web_network_name: WEB-NETWORK-NAME
   web_vm_type: VM-TYPE-FOR-WEB-INSTANCE
   worker_vm_type: VM-TYPE-FOR-WORKER-INSTANCE
   

   ---
   deployment_name: DEPLOYMENT-NAME
   db_persistent_disk_type: PERSISTANT-DISK-TYPE-FOR-DATABASE
   db_vm_type: VM-TYPE-FOR-DATABASE-INSTANCE
   external_url: EXTERNAL-URL
   network_name: NETWORK-NAME
   postgres_password: PASSWORD
   web_ip: WEB-IP
   web_network_name: WEB-NETWORK-NAME
   web_vm_type: VM-TYPE-FOR-WEB-INSTANCE
   worker_vm_type: VM-TYPE-FOR-WORKER-INSTANCE
   

   Where:

   • DEPLOYMENT-NAME is your Concourse deployment name. You can find this by running:

     bosh -e BOSH-ENVIRONMENT-ALIAS deployments
     

   • NETWORK-NAME is the name subproperty of the network property in your deployed manifest.

   • PASSWORD is the password sub-property of the postgres property in your deployed manifest.

   If your Concourse is deployed with something like CredHub, these variables might be automatically interpolated after instead of during deployment. If you suspect this is the case, you can check is with the following command:

   
    bosh \
    -e BOSH-ENVIRONMENT-ALIAS \
    -d TARGET-CONCOURSE-DEPLOYMENT \
    manifest | grep “(("
    

   This command searches through your deployed manifest for any placeholders that are being interpolated after deployment. If this returns any values, this could indicate that you have something like CredHub involved. Before you proceed, stop to investigate your current deployment, or get in touch with someone who has the necessary context.

4. Save and close the file.

Confirm your Availability Zones

The Concourse manifest has some defaults set for availability zones (AZs). You need to modify these defaults to suit a given environment. To ensure these are set correctly, follow these steps:

1. Retrieve the values of the azs properties in your currently deployed manifest.

   To find these, reference the deployed-manifest.yml file created in the last section, or run the bosh manifest command to re-create it:

   bosh \
   -e BOSH-ENVIRONMENT-ALIAS \
   -d CONCOURSE-DEPLOYMENT-NAME \
   manifest > deployed-manifest.yml
   

2. Open the concourse-bosh-deployment repository's cluster Concourse manifest file:

   vim ./cluster/concourse.yml
   

3. Search through this manifest file to find any azs properties and edit each one of these so that they match the availability zones (azs) in your deployed-manifest.yml file. There should be three in total, one in each instance group.

   For example, if your availability zone is default:

   ...
   instance_groups:
   - name: web
     ...
     azs: [default]
     ...
   
   - name: db
     ...
     azs: [default]
     ...
   
   - name: worker
     ...
     azs: [default]
     ...
   ...
   

4. Save and close the ./cluster/concourse.yml file.

Deploy your Updated Concourse Deployment with BBR

Deploy the new setup to your BOSH environment. The bosh deploy command uses a manifest that acts as a recipe for the deployment that you are creating on your cloud environment. When you run bosh deploy, you need to incorporate the following:

• Your new concourse manifest, for example, ./cluster/concourse.yml
• Your versions file, for example, versions.yml
• Your environment variables, for example, variables.yml
• Any required operations files, including the ./cluster/operations/backup-atc.yml file
• A filename where BOSH can record and access cluster credentials, for example, cluster-creds.yml

Operations files make substitutions in your manifest file to customize your deployment. The example below demonstrates a simple setup with basic authentication.
To change the authentication type, change the -o ./cluster/operations/basic-auth.yml line.

1. Edit and run this snippet, substituting your environment alias and deployment name, along with any additional ops files you want to add.

   For example:

   bosh -e BOSH-ENVIRONMENT-ALIAS deploy -d CONCOURSE-DEPLOYMENT-NAME ./cluster/concourse.yml \
   -l versions.yml \
   -l variables.yml \
   -o ./cluster/operations/backup-atc.yml \
   -o ./cluster/operations/basic-auth.yml \
   -o ./cluster/operations/privileged-http.yml \
   -o ./cluster/operations/static-web.yml \
   --vars-store cluster-creds.yml
   

   When prompted to review the deployment, you should see the addition of the backup-and-restore-sdk and its configuration. If you continue, BOSH prepares the deployment and starts compiling packages for you.

Troubleshooting

If you have CPI errors, try running the deployment script again. In rare cases, the operation can time out while packages are being built, causing a CPI timeout. The deployment should be successful if you re-run it a second time, as by that point the packages are already built.

Install the BOSH Backup and Restore CLI

To install the BBR CLI on your computer, do the following:

1. Check if BBR is already installed by running the following command. If this returns a version number instead of an error, you can skip to the next section.

   bbr -v
   

2. Download the latest release's asset file that corresponds to your operating system from the bosh-backup-and-restore GitHub repository. For example, Apple OSX users should download the darwin-amd64 file.

   There is no Windows release for the BBR CLI. If you are on Windows, you can SSH into a jumpbox where you can use a Linux release of BBR. For instructions, see Installing Bosh Backup and Restore (BBR) in the Cloud Foundry documentation. You can then continue to the Performing a Backup section below.

3. Move the file into your local /bin and rename it to bbr by running the following command:

   mv ~/Downloads/BBR-BINARY-FILE-NAME /usr/local/bin/bbr
   

4. Make the file executable by running the following command:

   chmod +x /usr/local/bin/bbr
   

Performing a Backup

To run a backup with the bbr command, do the following:

1. Retrieve the UAA BBR Client Credentials, username and password, from the following endpoint on your environment:

   https://YOUR-OPS-MANAGER-DOMAIN/api/v0/deployed/director/credentials/uaa_bbr_client_credentials
   

2. Run bbr with the following command. Make sure to use the environment IP address instead of using the environment alias you used with the bosh commands above.

   bbr deployment \
   --target ENVIRONMENT-IP-ADDRESS \
   --deployment TARGET-CONCOURSE-DEPLOYMENT-NAME \
   --username UAA-BBR-USERNAME \
   --password UAA-BBR-PASSWORD \
   --ca-cert CA-CERTIFICATE-FILE \
   backup
   

   A successful backup results in the creation of a directory named after your deployment name and the date and time of the backup.

3. Move the backup directory to a secure location. Note that this backup might include sensitive information, such as passwords in builds, depending on how you have configured your pipelines.

Restoring Concourse with BBR

If you need to restore a backup, see Performing a Restore.

When you are ready to continue with the upgrade process, proceed to upgrade instructions for either BOSH deployments or Helm deployments.