Save 10-15% Register for HashiConf and save big when you buy 2+ tickets Get your passes
Boundary
Self-managed deployment

Upgrade and database migration

  • 13min
  • |
  • Boundary
  • Terraform

This tutorial provides step-by-step instructions to safely upgrade Boundary to its latest version, including database backup and data migration.

When to upgrade

While new features are added to Boundary the backing database schema must be updated to support these features. Database migration is a simple process that is performed prior to upgrading Boundary on controller instances.

Even though the migrate command is resilient to errors due to utilizing PostgreSQL transactions, it is still highly recommended that you follow good database practices and backup prior to migrating because the migrate command does not allow reverting back to an older version. Versions of Boundary greater than 0.1.5 will not let you start a new controller until the database migration is complete, but will direct you to use the boundary database migrate command.

Read the release notes to learn if there are any version specific notes related to your current Boundary version or any versions between your current version and the intended upgrade version.

Boundary database migration workflow

  1. Backup Postgres database
  2. Stop all running controllers
  3. Perform database migration
  4. Restart controllers
  5. Upgrade Workers

Prerequisites

Deploy the demo environment

The demo environment provided for this tutorial includes a Docker-Compose cluster that deploys:

  • A Boundary 0.11.2 controller server
  • A Postgres database

The Terraform Boundary Provider is also used in this tutorial to easily provision resources using Docker, and must be available in your PATH when deploying the demo environment.

Boundary components

A worker server is not used in this example, but the process outlined in this tutorial applies to workers as well.

To learn more about the various Boundary components, refer back to the Start a Development Environment tutorial.

The lab environment can be downloaded or cloned from the following Github repository:

https://github.com/hashicorp-education/learn-boundary-upgrade

  1. In your terminal, clone the repository to get the example files locally:

    $ git clone git@github.com:hashicorp-education/learn-boundary-upgrade.git

  2. Move into the learn-boundary-upgrade folder.

    $ cd learn-boundary-upgrade

  3. Ensure that you are in the correct directory by listing its contents.

    $ ls -R1
README.md
compose
run
terraform

./compose:
boundary.hcl
docker-compose.yml

./terraform:
main.tf

    The repository contains the following files:

    • run: A script used to deploy and tear down the Docker-Compose configuration.
    • compose/docker-compose.yml: The Docker-Compose configuration file describing how to provision and network the database and the controller.
    • compose/boundary.hcl: The controller configuration file used when running Boundary.
    • terraform/main.tf: The Terraform provisioning instructions using the Boundary provider.

  4. This tutorial makes it easy to launch the test environment with the run script.

    $ ./run all
~/learn-boundary-upgrade/compose ~/learn-boundary-upgrade
Creating boundary_db_1 ... done
Creating boundary_db-init_1 ... done
Creating boundary_controller_1 ... done
~/learn-boundary-upgrade
~/learn-boundary-upgrade/terraform ~/learn-boundary-upgrade

Initializing the backend...

Initializing provider plugins...
- Reusing previous version of hashicorp/boundary from the dependency lock file
- Installing hashicorp/boundary v1.1.4...
- Installed hashicorp/boundary v1.1.4 (signed by HashiCorp)

Terraform has been successfully initialized!

You may now begin working with Terraform. Try running "terraform plan" to see
any changes that are required for your infrastructure. All Terraform commands
should now work.

If you ever set or change modules or backend configuration for Terraform,
rerun this command to reinitialize your working directory. If you forget, other
commands will detect it and remind you to do so if necessary.
boundary_scope.global: Creating...
boundary_scope.global: Creation complete after 0s [id=global]
boundary_scope.org: Creating...
boundary_role.global_anon_listing: Creating...
boundary_scope.org: Creation complete after 0s [id=o_s8On2vNvek]
boundary_auth_method.password: Creating...
boundary_scope.project: Creating...
boundary_role.org_anon_listing: Creating...
boundary_scope.project: Creation complete after 0s [id=p_rPNIMwdKR6]
boundary_host_catalog.databases: Creating...
boundary_auth_method.password: Creation complete after 0s [id=ampw_nvdhLojZ5L]
boundary_account.user["user3"]: Creating...
boundary_account.user["user2"]: Creating...
boundary_account.user["user1"]: Creating...
boundary_host_catalog.databases: Creation complete after 1s [id=hcst_1OlyrFWz49]
boundary_host.localhost: Creating...
boundary_account.user["user3"]: Creation complete after 1s [id=apw_8NeMx1Zlue]
boundary_account.user["user1"]: Creation complete after 2s [id=apw_aoxvyDgUHE]
boundary_account.user["user2"]: Creation complete after 2s [id=apw_KEGlxo60hU]
boundary_user.user["user1"]: Creating...
boundary_user.user["user3"]: Creating...
boundary_user.user["user2"]: Creating...
boundary_role.global_anon_listing: Creation complete after 2s [id=r_6EkHHA90LO]
boundary_host.localhost: Creation complete after 1s [id=hst_sIUKY2KyFX]
boundary_host_set.local: Creating...
boundary_user.user["user1"]: Creation complete after 2s [id=u_MQrWfzZfVj]
boundary_user.user["user3"]: Creation complete after 2s [id=u_HFaZjpc3Qf]
boundary_user.user["user2"]: Creation complete after 2s [id=u_7p18aQAw8n]
boundary_role.proj_admin: Creating...
boundary_role.org_admin: Creating...
boundary_host_set.local: Creation complete after 2s [id=hsst_EAIm8i6sQk]
boundary_target.ssh: Creating...
boundary_target.postgres: Creating...
boundary_role.org_anon_listing: Creation complete after 4s [id=r_rROmpodNgH]
boundary_target.postgres: Creation complete after 2s [id=ttcp_vF60hdPacR]
boundary_target.ssh: Creation complete after 2s [id=ttcp_NSWrkMNOxO]
boundary_role.org_admin: Creation complete after 3s [id=r_EYtyxVVV8q]
boundary_role.proj_admin: Creation complete after 3s [id=r_ykBlbeDPyd]

Apply complete! Resources: 19 added, 0 changed, 0 destroyed.

    The generated scope ids or user details can be viewed by inspecting the terraform/terraform.tfstate file.

    You can tear down the environment at any time by executing ./run cleanup.

  5. To verify that you deployed everything correctly, print the running docker containers and notice the ones named with the prefix "boundary".

    $ docker ps
CONTAINER ID   IMAGE                      COMMAND                  CREATED         STATUS                     PORTS                              NAMES
f8af936f1818   hashicorp/boundary:0.11.2   "/bin/boundary serve…"   6 minutes ago   Up 6 minutes (healthy)     0.0.0.0:9200-9202->9200-9202/tcp   boundary_controller_1
ab73470b49fd   postgres                   "docker-entrypoint.s…"   6 minutes ago   Up 6 minutes (healthy)     0.0.0.0:5432->5432/tcp             boundary_db_1

Prepare for an upgrade

Before performing the Boundary upgrade it is recommended to backup the existing Postgres database. This is done manually using the Postgres CLI. Note that your database instance must be running Postgres 11 or greater.

Versions of Boundary greater than 0.1.5 will notify the user if a database migration needs to be performed. This occurs when the database schema is updated to reflect new and updated features. It's important to note that running an earlier version of Boundary after performing a database migration is not supported, and careful backup and verification is encouraged to prevent data loss.

Backup Postgres database

If you're interested in examining the database schema before running a backup, you can do so by accessing the database on port 5432 with the username and password postgres.

To create a backup, run pg_dump on the boundary_db_1 container as the root user and output the database to a file like boundary_backup.sql.

$ docker exec -it --user=root boundary-db-1 pg_dump -h localhost \
       -p 5432 -U postgres -d boundary -f boundary_backup.sql

The backup file boundary_backup.sql will be present in the current working directory on the boundary_db_1 container. You can verify that the backup was performed by using grep to search for the schema names in the backup file.

$ docker exec boundary-db-1 grep 'CREATE TABLE' boundary_backup.sql | awk -F'.' '{print $2}' | tr -d '('
auth_account
auth_method
auth_password_account
auth_password_argon2_conf
auth_password_argon2_cred
auth_password_conf
auth_password_credential
auth_password_method
auth_token
db_test_car
db_test_rental
db_test_scooter
db_test_user
host
host_catalog
host_set
iam_group
iam_group_member_user
iam_user
iam_group_role
iam_role
iam_user_role
iam_role_grant
iam_scope
iam_scope_global
iam_scope_org
iam_scope_project
iam_scope_type_enm
kms_database_key
kms_database_key_version
kms_oplog_key
kms_oplog_key_version
kms_root_key
kms_root_key_version
kms_session_key
kms_session_key_version
kms_token_key
kms_token_key_version
oplog_entry
oplog_metadata
oplog_ticket
recovery_nonces
schema_migrations
server
session
session_connection
session_connection_closed_reason_enm
session_connection_state
session_connection_state_enm
session_state
session_state_enm
session_termination_reason_enm
static_host
static_host_catalog
static_host_set
static_host_set_member
target
target_tcp
target_host_set
wh_date_dimension
wh_host_dimension
wh_session_accumulating_fact
wh_session_connection_accumulating_fact
wh_time_of_day_dimension
wh_user_dimension

Note

This tutorial does NOT use best practices for backing up a Postgres database. The method outlined here creates a logical backup of the database in a human-readable format, and does not offer point-in-time recovery (PITR). For production scenarios, use of an automated backup process to perform PITR physical backups is highly recommended. One such solution would be Backup and Recovery Manager, Barman. For more information refer to the Postgres Backup Documentation.

Upgrade the Boundary controllers

Note

Currently Boundary only supports offline database migration, so there will be necessary downtime while the migration is performed.

  1. Stop all running controllers

    Controllers need to be upgraded to the latest version of Boundary, and must be stopped while the upgrade is performed. In practice, this might look like logging in to the running controller and stopping the boundary server process. For this Docker example, we must kill the process by stopping the controller container and started a new one with the upgraded version of Boundary installed.

    $ docker stop boundary-controller-1

  2. Upgrade the controllers

    In this tutorial the controllers are upgraded by launching an upgraded Boundary controller container. The docker run command below will place you in an interactive shell session where you can run the Boundary CLI directly.

    Note

    Do not exit the interactive session until the tutorial is complete. Open a new shell session if you want to return to your local machine.

    $ docker run -it \
   --network boundary_default \
   --volume "${PWD}/compose/:/boundary/" \
   -p "9200:9200" -p "9201:9201" -p "9202:9202" \
   --env "BOUNDARY_PG_URL=postgresql://postgres:postgres@db/boundary?sslmode=disable" \
   --name boundary_controller_upgrade \
   --entrypoint /bin/sh \
   hashicorp/boundary:0.12.0

    • The network is the running Docker-Compose network the database is on: boundary_default

    • The volume is the path to the local compose/ files to be copied onto the new container: "${PWD}/compose/:/boundary/"

    • The p options are the published (forwarded) ports the controller requires: 9200:9200, 9201:9201, 9202:9202

    • The env is the exported environment variable: "BOUNDARY_PG_URL=postgresql://postgres:postgres@db/boundary?sslmode=disable"

    • The name is the name we can reference the container with: boundary_controller_upgrade

    • The entrypoint is the command we should run on the container to start a shell session: /bin/sh

    • hashicorp/boundary:0.12.0 is the name of the newer image containing an upgraded version of Boundary

  3. Check the version of Boundary

    $ boundary version

Version information:
  Git Revision:        034f27ec8528251ce23b5afbf3851c48b0c54bec
  Version Number:      0.12.0

  4. Attempt to restart Boundary

    Now that the new version of Boundary is available, try restarting the Boundary server.

    $ boundary server -config boundary/boundary.hcl
Database schema must be updated to use this version. Run 'boundary database
migrate' to update the database. NOTE: Boundary does not currently support
live migration; ensure all controllers are shut down before running the
migration command.

    Versions of Boundary greater than 0.1.5 will display this message when the new version requires a schema update for the database, and will prevent Boundary from starting. Be aware upgrades prior to this version will not show this message.

    Note

    If the boundary/ directory or the boundary/boundary.hcl files are missing from the container it implies that the docker run command was executed from the wrong directory. In this case you will need to:

    • exit the running container and run docker rm -f boundary_controller_upgrade
    • Navigate back to the learn-boundary-upgrade folder
    • Proceed back to the Upgrade the controllers section above and execute docker run again

Perform database migration

With the Postgres database successfully backed up the migration can safely be performed.

  1. Run the migration

    $ boundary database migrate -config boundary/boundary.hcl
Migrations successfully run.

  2. Restart Boundary

    $ boundary server -config boundary/boundary.hcl
==> Boundary server configuration:

            [Recovery] AEAD Type: aes-gcm
               [Root] AEAD Type: aes-gcm
         [Worker-Auth] AEAD Type: aes-gcm
                           Cgo: disabled
Controller Public Cluster Addr: 0.0.0.0:9201
                     Listener 1: tcp (addr: "0.0.0.0:9200", cors_allowed_headers: "[]", cors_allowed_origins: "[serve://boundary]", cors_enabled: "true", max_request_duration: "1m30s", purpose: "api")
                     Listener 2: tcp (addr: "0.0.0.0:9201", max_request_duration: "1m30s", purpose: "cluster")
                     Listener 3: tcp (addr: "0.0.0.0:9202", max_request_duration: "1m30s", purpose: "proxy")
                     Log Level: info
                           Mlock: supported: true, enabled: false
                        Version: Boundary v0.12.0
                     Version Sha: 034f27ec8528251ce23b5afbf3851c48b0c54bec

==> Boundary server started! Log data will stream in below:

    Boundary should now be up and running. You can open a new shell session to test as needed, or visit the Boundary Web UI in a browser tab on http://localhost:9200, logging into the primary scope as user1 and the password password or with any of the user info in the terraform/terraform.tfstate file.

Upgrade the Boundary workers

This tutorial doesn't include a worker, but it is recommended that the same upgrade process is performed on any worker instances when upgrading a controller. While workers do not directly connect to database instances, upgrading workers and controllers to the same version of Boundary ensures continued protocol compatibility and communication.

To upgrade your workers:

  1. Stop the Boundary service on the worker instance
  2. Replace the Boundary binary
  3. Restart the Boundary service

Recommended approach for High Availability

This tutorial covers a simple example of what to expect when performing an upgrade and migration. For non-dev environments and High-Availability considerations, be sure to familiarize yourself with the HA architecture recommendations and controller configurations.

Cleanup and teardown

  1. Stop and destroy the upgraded controller

    The upgraded controller container is running independently of Docker-Compose, and must be stopped and removed directly. It is not necessary exit the shell session you have running in the background to destroy the container.

    $ docker rm -f boundary_controller_upgrade

  2. Cleanup the database instance

    The remaining database container and network resources can be cleaned up using the provided run script.

    $ ./run cleanup

    Check your work with a quick docker ps and ensure there are no more containers with the boundary_ prefix leftover. If unexpected containers still exist, execute the docker rm -f command from above on each to remove them.

Help and reference

Collection Overview

Self-managed deployment

 Next

Worker-aware targets