Disaster Recovery (Geo)

Tier: Premium, Ultimate
Offering: GitLab Self-Managed

Geo replicates your database, your Git repositories, and other assets. Some known issues exist.

[!warning]

Multi-secondary configurations require the complete re-synchronization and re-configuration of all non-promoted secondaries and causes downtime.

After the secondary site is promoted, the primary site is detached entirely. If you wish to restore the primary site, you must add it as a new secondary site.

Secondary sites with selective synchronization enabled

Promoting a secondary site with selective synchronization enabled results in permanent data loss
for all data that was not replicated to that secondary site. For more information, see Promoting a secondary site with selective synchronization enabled.

The `gitlab-cluster.json` file

When you promote a secondary site to a primary site with gitlab-ctl geo promote, the command automatically creates a /etc/gitlab/gitlab-cluster.json file on each node where it executes. In most circumstances, you don't need to manually edit this file.

The gitlab-cluster.json file allows the promotion command to automate configuration changes without modifying /etc/gitlab/gitlab.rb directly. Programmatically editing gitlab.rb is error-prone, so gitlab-cluster.json serves as a machine-managed override layer.

When both files exist, values in gitlab-cluster.json take precedence over the corresponding values in gitlab.rb when gitlab-ctl reconfigure executes. When you run this command, you see a warning similar to:

plaintext

The 'geo_primary_role' is defined in /etc/gitlab/gitlab-cluster.json as 'true' and overrides the setting in the /etc/gitlab/gitlab.rb
The 'geo_secondary_role' is defined in /etc/gitlab/gitlab-cluster.json as 'false' and overrides the setting in the /etc/gitlab/gitlab.rb

This warning is expected after promotion.

File structure

A typical gitlab-cluster.json file looks like:

json

{
  "primary": true,
  "secondary": false,
  "geo_secondary": {
    "enable": false
  }
}

Key	Description
`primary`	When `true`, enables `geo_primary_role`, which configures the node as a Geo primary.
`secondary`	When `true`, enables `geo_secondary_role`, which configures the node as a Geo secondary.
`geo_secondary`	Contains settings related to the Geo secondary configuration, such as the tracking database. `"enable": false` disables secondary-specific services.

The primary and secondary keys map to the geo_primary_role and geo_secondary_role respectively. These roles are a convenience for single-node setups and should not be used in multi-node configurations where individual service roles are configured explicitly in gitlab.rb.

Remove the file

After a successful promotion, you can keep gitlab-cluster.json in place. However, you should remove it in the following situations:

If you bring a demoted primary back as a new secondary site, you must delete gitlab-cluster.json from every Sidekiq, PostgreSQL, Gitaly, and Rails node.
After you update gitlab.rb to set a Geo role (for example, roles(['geo_primary_role'])), and you want gitlab.rb to be the sole configuration source.
After you recover from a partial failover.

See Recovering from a partial failover for details on when the file is manually created during recovery.

To remove the file:

Run these commands:
shell
```
sudo rm /etc/gitlab/gitlab-cluster.json
sudo gitlab-ctl reconfigure
```
In a multi-node setup, repeat these commands on every node in the site.

For technical details about how gitlab-cluster.json interacts with the reconfigure process, see the Omnibus reconfigure documentation.

Promoting a secondary Geo site in single-secondary configurations

While you can't automatically promote a Geo replica and do a failover, you can promote it manually if you have root access to the machine.

This process promotes a secondary Geo site to a primary site. To regain geographic redundancy as quickly as possible, you should add a new secondary site immediately after following these instructions.

Allow replication to finish if possible

If the secondary site is still replicating data from the primary site, follow the planned failover docs as closely as possible in order to avoid unnecessary data loss.

Step 1. Permanently disable the primary site

[!warning] If the primary site goes offline, there may be data saved on the primary site that have not been replicated to the secondary site. This data should be treated as lost if you proceed.

If an outage on the primary site happens, you should do everything possible to avoid a split-brain situation where writes can occur in two different GitLab instances, complicating recovery efforts. So to prepare for the failover, we must disable the primary site.

If you have SSH access:
1. SSH into the primary site to stop and disable GitLab:
  shell
```
sudo gitlab-ctl stop
```
2. Prevent GitLab from starting up again if the server unexpectedly reboots:
  shell
```
sudo systemctl disable gitlab-runsvdir
```
If you do not have SSH access to the primary site, take the machine offline and prevent it from rebooting by any means at your disposal. You might need to:
- Reconfigure the load balancers.
- Change DNS records (for example, point the primary DNS record to the secondary site to stop usage of the primary site).
- Stop the virtual servers.
- Block traffic through a firewall.
- Revoke object storage permissions from the primary site.
- Physically disconnect a machine.
If you plan to update the primary domain DNS record, you may wish to maintain a low TTL to ensure fast propagation of DNS changes.

[!note] The primary site's /etc/gitlab/gitlab.rb file is not copied to the secondary sites automatically during this process. Make sure that you back up the primary's /etc/gitlab/gitlab.rb file, so that you can later restore any needed values on your secondary sites.

Step 2. Promoting a secondary site

Note the following when promoting a secondary:

If the secondary site has been paused, the promotion performs a point-in-time recovery to the last known state. Data that was created on the primary while the secondary was paused is lost.
If the secondary site has been paused and you encouter an ActiveRecord::StatementInvalid: PG::ReadOnlySqlTransaction: ERROR: cannot execute DELETE in a read-only transaction error message during this process, see this knowledge base article: Geo promotion fails with read-only transaction error or timeout after unexpected primary shutdown.
A new secondary should not be added at this time. If you want to add a new secondary, do this after you have completed the entire process of promoting the secondary to the primary.
If you encounter an ActiveRecord::RecordInvalid: Validation failed: Name has already been taken error message during this process, for more information, see this troubleshooting advice.
If you are using separate URLs, you should point the primary domain DNS at the newly promoted site. Otherwise, runners must be registered again with the newly promoted site, and all Git remotes, bookmarks, and external integrations must be updated.
If you are using location-aware DNS, the runners should automatically connect to the new primary after the old primary is removed from the DNS entry.
After the primary site is down, run gitlab-ctl promotion-preflight-checks on the secondary to check the Geo sync status and perform final validation checks.
If you don't expect the runners connected to the previous primary to come back, you should remove them:
- Through the UI:
  1. In the upper-right corner, select Admin.
  2. Select CI/CD > Runners and remove them.
- Using the Runners API.

Promoting a secondary site running on a single node

SSH in to your secondary site and execute:
- To promote the secondary site to primary:
  shell
```
sudo gitlab-ctl geo promote
```
- To promote the secondary site to primary without any further confirmation:
  shell
```
sudo gitlab-ctl geo promote --force
```
Verify you can connect to the newly-promoted primary site using the URL used previously for the secondary site.
If successful, the secondary site is now promoted to the primary site.

When you run gitlab-ctl geo promote, a gitlab-cluster.json file is created on the node. The file overrides Geo role settings in gitlab.rb when you reconfigure.

Step 3. (Optional) Removing the former secondary's tracking database

If you have any geo_secondary[] configuration options enabled in your /etc/gitlab/gitlab.rb file, comment them out or remove them, and then reconfigure GitLab for the changes to take effect.

At this point, your promoted site is the new primary GitLab site. Optionally, if you wish to set up Geo again as a new secondary site, you can bring the old site back as a secondary.

Promoting a secondary site with multiple nodes and a single-secondary site

SSH to every Sidekiq, PostgreSQL, and Gitaly node in the secondary site and run one of the following commands:
- To promote the node on the secondary site to primary:
  shell
```
sudo gitlab-ctl geo promote
```
- To promote the secondary site to primary without any further confirmation:
  shell
```
sudo gitlab-ctl geo promote --force
```
SSH into each Rails node on your secondary site and run one of the following commands:
- To promote the secondary site to primary:
  shell
```
sudo gitlab-ctl geo promote
```
- To promote the secondary site to primary without any further confirmation:
  shell
```
sudo gitlab-ctl geo promote --force
```
Verify you can connect to the newly-promoted primary site using the URL used previously for the secondary site.
If successful, the secondary site is now promoted to the primary site.

When you run gitlab-ctl geo promote, a gitlab-cluster.json file is created on the node. The file overrides Geo role settings in gitlab.rb when you reconfigure.

Promoting a secondary site with a Patroni standby cluster

SSH to every Sidekiq, PostgreSQL, and Gitaly node in the secondary site and run one of the following commands:
- To promote the secondary site to primary:
  shell
```
sudo gitlab-ctl geo promote
```
- To promote the secondary site to primary without any further confirmation:
  shell
```
sudo gitlab-ctl geo promote --force
```
SSH into each Rails node on your secondary site and run one of the following commands:
- To promote the secondary site to primary:
  shell
```
sudo gitlab-ctl geo promote
```
- To promote the secondary site to primary without any further confirmation:
  shell
```
sudo gitlab-ctl geo promote --force
```
Verify you can connect to the newly-promoted primary site using the URL used previously for the secondary site.
If successful, the secondary site is now promoted to the primary site.

Promoting a secondary site with an external PostgreSQL database

The gitlab-ctl geo promote command can be used in conjunction with an external PostgreSQL database. In this case, you must first manually promote the replica database associated with the secondary site:

Promote the replica database associated with the secondary site. This sets the database to read-write. The instructions vary depending on where your database is hosted:

Amazon RDS
Azure PostgreSQL
Google Cloud SQL

For other external PostgreSQL databases, save the following script in your secondary site, for example /tmp/geo_promote.sh, and modify the connection parameters to match your environment. Then, execute it to promote the replica:

shell

#!/bin/bash

PG_SUPERUSER=postgres

# The path to your pg_ctl binary. You may need to adjust this path to match
# your PostgreSQL installation
PG_CTL_BINARY=/usr/lib/postgresql/16/bin/pg_ctl

# The path to your PostgreSQL data directory. You may need to adjust this
# path to match your PostgreSQL installation. You can also run
# `SHOW data_directory;` from PostgreSQL to find your data directory
PG_DATA_DIRECTORY=/etc/postgresql/16/main

# Promote the PostgreSQL database and allow read/write operations
sudo -u $PG_SUPERUSER $PG_CTL_BINARY -D $PG_DATA_DIRECTORY promote

SSH to every Sidekiq, PostgreSQL, and Gitaly node in the secondary site and run one of the following commands:
- To promote the secondary site to primary:
  shell
```
sudo gitlab-ctl geo promote
```
- To promote the secondary site to primary without any further confirmation:
  shell
```
sudo gitlab-ctl geo promote --force
```
SSH into each Rails node on your secondary site and run one of the following commands:
- To promote the secondary site to primary:
  shell
```
sudo gitlab-ctl geo promote
```
- To promote the secondary site to primary without any further confirmation:
  shell
```
sudo gitlab-ctl geo promote --force
```
Verify you can connect to the newly-promoted primary site using the URL used previously for the secondary site.
If successful, the secondary site is now promoted to the primary site.

(Optional) Updating the primary domain DNS record

Update DNS records for the primary domain to point to the secondary site. This removes the need to update all references to the primary domain, for example changing Git remotes and API URLs.

SSH into the secondary site and login as root:
shell
```
sudo -i
```
Update the primary domain's DNS record. After updating the primary domain's DNS records to point to the secondary site, edit /etc/gitlab/gitlab.rb on the secondary site to reflect the new URL:
ruby
```
# Change the existing external_url configuration
external_url 'https://<new_external_url>'
```
[!note] Changing external_url does not prevent access through the old secondary URL, as long as the secondary DNS records are still intact.
Update the secondary's SSL certificate:
- If you use the Let's Encrypt integration, the certificate updates automatically.
- If you had manually set up, the secondary's certificate, copy the certificate from the primary to the secondary. If you don't have access to the primary, issue a new certificate and make sure it contains both the primary and secondary URLs in the subject alternative names. You can check with:
  shell
```
/opt/gitlab/embedded/bin/openssl x509 -noout -dates -subject -issuer \
    -nameopt multiline -ext subjectAltName -in /etc/gitlab/ssl/new-gitlab.new-example.com.crt
```
Reconfigure the secondary site for the change to take effect:
shell
```
gitlab-ctl reconfigure
```
Execute the command below to update the newly promoted primary site URL:
shell
```
gitlab-rake geo:update_primary_node_url
```
This command uses the changed external_url configuration defined in /etc/gitlab/gitlab.rb.
Verify you can connect to the newly promoted primary using its URL. If you updated the DNS records for the primary domain, these changes may not have yet propagated depending on the previous DNS records TTL.

(Optional) Add secondary Geo site to a promoted primary site

To bring a new secondary site online, follow the Geo setup instructions.

Promoting secondary Geo replica in multi-secondary configurations

If you have more than one secondary site and you need to promote one of them, we suggest you follow Promoting a secondary Geo site in single-secondary configurations and after that you also need two extra steps.

Step 1. Prepare the new primary site to serve one or more secondary sites

SSH into the new primary site and login as root:
shell
```
sudo -i
```

Edit /etc/gitlab/gitlab.rb:

ruby

## Enable a Geo Primary role (if you haven't yet)
roles ['geo_primary_role']

##
# Allow PostgreSQL client authentication from the primary and secondary IPs. These IPs may be
# public or VPC addresses in CIDR format, for example ['198.51.100.1/32', '198.51.100.2/32']
##
postgresql['md5_auth_cidr_addresses'] = ['<primary_site_ip>/32', '<secondary_site_ip>/32']

# Every secondary site needs to have its own slot so specify the number of secondary sites you're going to have
# postgresql['max_replication_slots'] = 1 # Set this to be the number of Geo secondary nodes if you have more than one

##
## Disable automatic database migrations temporarily
## (until PostgreSQL is restarted and listening on the private address).
##
gitlab_rails['auto_migrate'] = false

(For more details about these settings you can read Configure the primary server)

Save the file and reconfigure GitLab for the database listen changes and the replication slot changes to be applied:
shell
```
gitlab-ctl reconfigure
```
Restart PostgreSQL for its changes to take effect:
shell
```
gitlab-ctl restart postgresql
```
Re-enable migrations now that PostgreSQL is restarted and listening on the private address.

Edit /etc/gitlab/gitlab.rb and change the configuration to true:
ruby
```
gitlab_rails['auto_migrate'] = true
```
Save the file and reconfigure GitLab:
shell
```
gitlab-ctl reconfigure
```

Step 2. Initiate the replication process

Now we need to make each secondary site listen to changes on the new primary site. To do that you need to initiate the replication process again but this time for another primary site. All the old replication settings are overwritten.

Existing secondary sites will all have populated databases so you may see a message like this:

shell

Found data inside the gitlabhq_production database! If you are sure you are in the secondary server, override with --force

After you have confirmed that you are on the appropriate secondary site, initiate the replication with --force.

[!warning] Using --force causes all existing data in the database on that secondary server to be deleted.

Promoting a secondary Geo cluster in the GitLab Helm chart

When updating a cloud-native Geo deployment, the process for updating any node that is external to the secondary Kubernetes cluster does not differ from the non cloud-native approach. As such, you can always defer to Promoting a secondary Geo site in single-secondary configurations for more information.

The following sections assume you are using the gitlab namespace. If you used a different namespace when setting up your cluster, you should also replace --namespace gitlab with your namespace.

Step 1. Permanently disable the primary cluster

[!warning] If the primary site goes offline, there may be data saved on the primary site that has not been replicated to the secondary site. This data should be treated as lost if you proceed.

If you have access to the primary Kubernetes cluster, connect to it and disable the GitLab webservice and Sidekiq pods:

shell

kubectl --namespace gitlab scale deploy gitlab-geo-webservice-default --replicas=0
kubectl --namespace gitlab scale deploy gitlab-geo-sidekiq-all-in-1-v1 --replicas=0

If you do not have access to the primary Kubernetes cluster, take the cluster offline and prevent it from coming back online by any means at your disposal. You might need to:
- Reconfigure the load balancers.
- Change DNS records (for example, point the primary DNS record to the secondary site to stop usage of the primary site).
- Stop the virtual servers.
- Block traffic through a firewall.
- Revoke object storage permissions from the primary site.
- Physically disconnect a machine.

Step 2. Promote all secondary site nodes external to the cluster

[!warning] If the secondary site has been paused, this performs a point-in-time recovery to the last known state. Data that was created on the primary while the secondary was paused is lost.

For each node (such as PostgreSQL or Gitaly) outside of the secondary Kubernetes cluster using the Linux package, SSH into the node and run one of the following commands:
- To promote the secondary site node external to the Kubernetes cluster to primary:
  shell
```
sudo gitlab-ctl geo promote
```
- To promote the secondary site node external to the Kubernetes cluster to primary without any further confirmation:
  shell
```
sudo gitlab-ctl geo promote --force
```

Find the toolbox pod:

shell

kubectl --namespace gitlab get pods -lapp=toolbox

Promote the secondary:
shell
```
kubectl --namespace gitlab exec -ti gitlab-geo-toolbox-XXX -- gitlab-rake geo:set_secondary_as_primary
```
Environment variables can be provided to modify the behavior of the task. The available variables are:

Name Default value Description
ENABLE_SILENT_MODE false If true, enables Silent Mode before promotion (GitLab 16.4 and later)

Name	Default value	Description
`ENABLE_SILENT_MODE`	`false`	If `true`, enables Silent Mode before promotion (GitLab 16.4 and later)

Step 3. Promote the secondary cluster

Update the existing cluster configuration.

You can retrieve the existing configuration with Helm:
shell
```
helm --namespace gitlab get values gitlab-geo > gitlab.yaml
```
The existing configuration contains a section for Geo that should resemble:
yaml
```
geo:
   enabled: true
   role: secondary
   nodeName: secondary.example.com
   psql:
      host: geo-2.db.example.com
      port: 5431
      password:
         secret: geo
         key: geo-postgresql-password
```
To promote the secondary cluster to a primary cluster, update role: secondary to role: primary.

If the cluster remains as a primary site, you can remove the entire psql section; it refers to the tracking database and is ignored while the cluster is acting as a primary site.

Update the cluster with the new configuration:
shell
```
helm upgrade --install --version <current Chart version> gitlab-geo gitlab/gitlab --namespace gitlab -f gitlab.yaml
```
Verify you can connect to the newly promoted primary using the URL used previously for the secondary.
Success! The secondary has now been promoted to primary.

Step 4. (Optional) Promote the OpenBao HA cluster

If you have GitLab Secrets Manager enabled, complete the following steps to promote the OpenBao High Availability (HA) cluster after promoting the Kubernetes cluster.

Restart OpenBao pods

After the PostgreSQL replica is promoted to primary, restart the OpenBao pods so they reconnect to the database now that it's writable:

shell

kubectl --namespace gitlab rollout restart deployment -l app=openbao

(Optional) Configure JWT authentication

Skip this step if you've updated DNS records for the primary domain to point to the secondary site.

Using recovery keys, connect to OpenBao API to reconfigure JWT authentication for the secondary domain.

For more information, see Geo configuration.

Restore the unseal secret if needed

The unseal key on the secondary cluster must be the same as the one on the primary key, otherwise OpenBao won't be able to unseal the vault on the secondary.

If there's a mismatch, restore the gitlab-openbao-unseal secret on the secondary cluster from your secrets backup, then restart the OpenBao pods:

shell

kubectl --namespace gitlab rollout restart deployment -l app=openbao

Verify OpenBao is functional

Check that all OpenBao pods are running:

shell

kubectl --namespace gitlab get pods -l app=openbao

Test OpenBao integration by running a CI pipeline that uses a Secrets Manager variable.

Troubleshooting

This section was moved to another location.

Secondary sites with selective synchronization enabled

The gitlab-cluster.json file

File structure

Remove the file

Promoting a secondary Geo site in single-secondary configurations

Allow replication to finish if possible

Step 1. Permanently disable the primary site

Step 2. Promoting a secondary site

Promoting a secondary site running on a single node

Step 3. (Optional) Removing the former secondary's tracking database

Promoting a secondary site with multiple nodes and a single-secondary site

Promoting a secondary site with a Patroni standby cluster

Promoting a secondary site with an external PostgreSQL database

(Optional) Updating the primary domain DNS record

(Optional) Add secondary Geo site to a promoted primary site

Promoting secondary Geo replica in multi-secondary configurations

Step 1. Prepare the new primary site to serve one or more secondary sites

Step 2. Initiate the replication process

Promoting a secondary Geo cluster in the GitLab Helm chart

Step 1. Permanently disable the primary cluster

Step 2. Promote all secondary site nodes external to the cluster

Step 3. Promote the secondary cluster

Step 4. (Optional) Promote the OpenBao HA cluster

Restart OpenBao pods

(Optional) Configure JWT authentication

Restore the unseal secret if needed

Verify OpenBao is functional

Troubleshooting

The `gitlab-cluster.json` file