• Steps
• Upgrade the bundled PostgreSQL to version 12 (optional)
• Upgrade the bundled PostgreSQL chart
• Upgrade steps for 4.0 release
  • Prepare the existing database
  • Delete existing PostgreSQL data
  • Upgrade GitLab
  • Restore the Database
  • Troubleshooting 4.0 release upgrade process
• Upgrade steps for 3.0 release
  • Prepare the existing database
  • Prepare the cluster database secrets
  • Delete existing services
  • Upgrade GitLab
  • Restore the Database
  • Troubleshooting 3.0 release upgrade process

Upgrade Guide

Before upgrading your GitLab installation, you need to check the changelog corresponding to the specific release you want to upgrade to and look for any release notes that might pertain to the new GitLab chart version.

We also recommend that you take a backup first. Also note that you need to provide all values using helm upgrade --set key=value syntax or -f values.yml instead of using --reuse-values because some of the current values might be deprecated.

Mappings between chart versioning and GitLab versioning can be found here.

Steps

The following are the steps to upgrade GitLab to a newer version:

1. Check the change log for the specific version you would like to upgrade to
2. Go through deployment documentation step by step

3. Extract your previous --set arguments with

   helm get values gitlab > gitlab.yaml
   

4. Decide on all the values you need to set
5. If you would like to use the GitLab Operator go through the steps outlined in Operator installation

6. Perform the upgrade, with all --set arguments extracted in step 4

   helm upgrade gitlab gitlab/gitlab \
     --version <new version> \
     -f gitlab.yaml \
     --set gitlab.migrations.enabled=true \
     --set ...
   

Upgrade the bundled PostgreSQL to version 12 (optional)

Upgrading to PostgreSQL 12 for GitLab 13.x is optional. PostgreSQL 12 is supported by GitLab 13.4 and later. PostgreSQL 12 will become the minimum required PostgreSQL version in GitLab 14.0, scheduled for April 2021. PostgreSQL 12 brings significant performance improvements.

To upgrade the bundled PostgreSQL to version 12, the following steps are required:

1. Prepare the existing database.
2. Delete existing PostgreSQL data.
3. Update the postgresql.image.tag value to 12.4.0 and reinstall the chart to create a new PostgreSQL 12 database.
4. Restore the database.

Upgrade the bundled PostgreSQL chart

As part of the 4.0.0 release of this chart, we upgraded the bundled PostgreSQL chart from 7.7.0 to 8.9.4. This is not a drop in replacement. Manual steps need to be performed to upgrade the database. The steps have been documented in the 4.0 upgrade steps.

As part of the 3.0.0 release of this chart, we upgraded the bundled PostgreSQL chart from 0.11.0 to 7.7.0. This is not a drop in replacement. Manual steps need to be performed to upgrade the database. The steps have been documented in the 3.0 upgrade steps.

Upgrade steps for 4.0 release

The 4.0.0 release requires manual steps in order to perform the upgrade.

If you are using the bundled PostgreSQL, the best way to perform this upgrade is to back up your old database, and restore into a new database instance.

Prepare the existing database

Note the following:

• If you are not using the bundled PostgreSQL chart (postgresql.install is false), you do not need to perform this step.
• If you have multiple charts installed in the same namespace. It may be necessary to pass the Helm release name to the database-upgrade script as well. Replace bash -s STAGE with bash -s -- -r RELEASE STAGE in the example commands provided later.
• If you installed a chart to a namespace other than your kubectl context’s default, you must pass the namespace to the database-upgrade script. Replace bash -s STAGE with bash -s -- -n NAMESPACE STAGE in the example commands provided later. This option can be used along with -r RELEASE. You can set the context’s default namespace by running kubectl config set-context --current --namespace=NAMESPACE, or using kubens from kubectx.

The pre stage will create a backup of your database using the backup-utility script in the task-runner pod, which gets saved to the configured s3 bucket (MinIO by default):

# GITLAB_RELEASE should be the version of the chart you are installing, starting with 'v': v4.0.0
curl -s https://gitlab.com/gitlab-org/charts/gitlab/raw/${GITLAB_RELEASE}/scripts/database-upgrade | bash -s pre

Delete existing PostgreSQL data

The 4.0 release updates the default bundled PostgreSQL version from 10.9.0 to 11.7.0. Since the data format has changed, upgrading requires removing the existing PostgreSQL StatefulSet before upgrading to the 4.0 release. The StatefulSet will be recreated in the next step.

kubectl delete statefulset RELEASE-NAME-postgresql
kubectl delete pvc data-RELEASE_NAME-postgresql-0

Upgrade GitLab

Upgrade GitLab following our standard procedure, with the following additions of:

If you are using the bundled PostgreSQL, disable migrations using the following flag on your upgrade command:

1. --set gitlab.migrations.enabled=false

We will perform the migrations for the Database in a later step for the bundled PostgreSQL.

Restore the Database

Note the following:

• If you are not using the bundled PostgreSQL chart (postgresql.install is false), you do not need to perform this step.
• You’ll need to be using Bash 4.0 or above to run the script successfully as it requires the use of bash associative arrays.

1. Wait for the upgrade to complete for the task-runner pod. RELEASE_NAME should be the name of the GitLab release from helm list

   kubectl rollout status -w deployment/RELEASE_NAME-task-runner
   

2. After the task-runner pod is deployed successfully, run the post steps:

   This step will do the following:

   1. Set replicas to 0 for the webservice, sidekiq, and gitlab-exporter deployments. This will prevent any other application from altering the database while the backup is being restored.
   2. Restore the database from the backup created in the pre stage.
   3. Run database migrations for the new version.
   4. Unpause the deployments from the first step.

   # GITLAB_RELEASE should be the version of the chart you are installing, starting with 'v': v4.0.0
   curl -s https://gitlab.com/gitlab-org/charts/gitlab/raw/${GITLAB_RELEASE}/scripts/database-upgrade | bash -s post
   

Troubleshooting 4.0 release upgrade process

• If you see any failure during the upgrade, it may be useful to check the description of gitlab-upgrade-check pod for details:

  kubectl get pods -lrelease=RELEASE,app=gitlab
  kubectl describe pod <gitlab-upgrade-check-pod-full-name>
  

Upgrade steps for 3.0 release

The 3.0.0 release requires manual steps in order to perform the upgrade.

If you are using the bundled PostgreSQL, the best way to perform this upgrade is to backup your old database, and restore into a new database instance. We’ve automated some of the steps, as an alternative, you can perform the steps manually.

Prepare the existing database

Note the following:

• If you are not using the bundled PostgreSQL chart (postgresql.install is false), you do not need to perform this step.
• If you have multiple charts installed in the same namespace. It may be necessary to pass the Helm release name to the database-upgrade script as well. Replace bash -s STAGE with bash -s -- -r RELEASE STAGE in the example commands provided later.
• If you installed a chart to a namespace other than your kubectl context’s default, you must pass the namespace to the database-upgrade script. Replace bash -s STAGE with bash -s -- -n NAMESPACE STAGE in the example commands provided later. This option can be used along with -r RELEASE. You can set the context’s default namespace by running kubectl config set-context --current --namespace=NAMESPACE, or using kubens from kubectx.

The pre stage will create a backup of your database using the backup-utility script in the task-runner pod, which gets saved to the configured s3 bucket (MinIO by default):

# GITLAB_RELEASE should be the version of the chart you are installing, starting with 'v': v3.0.0
curl -s https://gitlab.com/gitlab-org/charts/gitlab/-/raw/${GITLAB_RELEASE}/scripts/database-upgrade  | bash -s pre

Prepare the cluster database secrets

NOTICE: If you are not using the bundled PostgreSQL chart (postgresql.install is false):

• If you have supplied global.psql.password.key, you do not need to perform this step.
• If you have supplied global.psql.password.secret, additionally set global.psql.password.key to the name of your existing key to bypass this step.

The secret key for the application database key is changing from postgres-password, to postgresql-password. Use one of the two steps described below to update your database password secret:

1. If you’d like to use an auto-generated PostgreSQL password, delete the existing secret to allow the upgrade to generate a new password for you. RELEASE-NAME should be the name of the GitLab release from helm list:

   # Create a local copy of the old secret in case we need to restore the old database
   kubectl get secret RELEASE-NAME-postgresql-password -o yaml > postgresql-password.backup.yaml
   # Delete the secret so a new one can be created
   kubectl delete secret RELEASE-NAME-postgresql-password
   

2. If you want to use the same password, edit the secret, and change the key from postgres-password to postgresql-password. Additionally, we need a secret for the superuser account. Add a key for that user postgresql-postgres-password:

   # Encode the superuser password into base64
   echo SUPERUSER_PASSWORD | base64
   kubectl edit secret RELEASE-NAME-postgresql-password
   # Make the appropriate changes in your EDITOR window
   

Delete existing services

The 3.0 release updates an immutable field in the NGINX Ingress, this requires us to first delete all the services before upgrading. You can see more details in our troubleshooting documentation, under Immutable Field Error, spec.clusterIP.

1. Remove all affected services. RELEASE_NAME should be the name of the GitLab release from helm list:

   kubectl delete services -lrelease=RELEASE_NAME
   

Upgrade GitLab

Upgrade GitLab following our standard procedure, with the following additions of:

If you are using the bundled PostgreSQL, disable migrations using the following flag on your upgrade command:

1. --set gitlab.migrations.enabled=false

We will perform the migrations for the Database in a later step for the bundled PostgreSQL.

Restore the Database

Note the following:

• If you are not using the bundled PostgreSQL chart (postgresql.install is false), you do not need to perform this step.
• You’ll need to be using Bash 4.0 or above to run the script successfully as it requires the use of bash associative arrays.

1. Wait for the upgrade to complete for the task-runner pod. RELEASE_NAME should be the name of the GitLab release from helm list

   kubectl rollout status -w deployment/RELEASE_NAME-task-runner
   

2. After the task-runner pod is deployed successfully, run the post steps:

   This step will do the following:

   1. Set replicas to 0 for the webservice, sidekiq, and gitlab-exporter deployments. This will prevent any other application from altering the database while the backup is being restored.
   2. Restore the database from the backup created in the pre stage.
   3. Run database migrations for the new version.
   4. Unpause the deployments from the first step.

   # GITLAB_RELEASE should be the version of the chart you are installing, starting with 'v': v3.0.0
   curl -s https://gitlab.com/gitlab-org/charts/gitlab/-/raw/${GITLAB_RELEASE}/scripts/database-upgrade | bash -s post
   

Troubleshooting 3.0 release upgrade process

• Make sure that you are using Helm 2.14.3 or >= 2.16.1 due to the bug in 2.15.x.

• If you see any failure during the upgrade, it may be useful to check the description of gitlab-upgrade-check pod for details:

  kubectl get pods -lrelease=RELEASE,app=gitlab
  kubectl describe pod <gitlab-upgrade-check-pod-full-name>
  

• You may face the error below when running helm upgrade:

  Error: kind ConfigMap with the name "gitlab-gitlab-shell-sshd" already exists in the cluster and wasn't defined in the previous release.
  Before upgrading, please either delete the resource from the cluster or remove it from the chart
  Error: UPGRADE FAILED: kind ConfigMap with the name "gitlab-gitlab-shell-sshd" already exists in the cluster and wasn't defined in the previous release.
  Before upgrading, please either delete the resource from the cluster or remove it from the chart
  

  The error message can also mention other configmaps like gitlab-redis-health, gitlab-redis-headless, etc. To fix it, make sure that the services were removed as mentioned in the upgrade steps for 3.0 release. After that, also delete the configmaps shown in the error message with: kubectl delete configmap <configmap-name>.