How to set up Vault

The steps below assume you already have a Hashicorp Vault instance running. For more details, see the Hashicorp Vault documentation.

To configure Confluence to work with HashiCorp Vault:

1. Create a secret in your HashiCorp Vault instance.

2. Create a policy with permission to read your secret.

3. Authenticate Confluence with Vault.

4. Add the Vault configuration data to the <home-directory>/confluence.cfg.xml file.

These steps are explained in more detail below.

Step 1: Create a secret in your HashiCorp Vault instance

If you haven’t created a secret in the KV V2 Secret Engine of your Vault instance before, take a look at the Hashicorp Vault documentation for more information.

This secret must contain a single value for your JDBC password.

Step 2: Create a policy with permission to read your secret

If you need detailed instructions on creating a policy in Vault, see the Hashicorp Vault documentation. The details below provide additional information from the Confluence perspective.

To retrieve your secret from the Vault, Confluence must have a policy with the read permission.

Below is a sample Vault policy with permission to read a secret in the KV V2 Secret Engine.

path "secret/data/sample/secret" {
  capabilities = ["read"]
}

In the sample path above, there are three components:

If the previous policy is located in ./sample_policy.hcl, this command will create the policy on the server:

vault policy write sample_policy ./sample_policy.hcl

Step 3: Authenticate Confluence with Vault

You can choose to authenticate with a token, or, if you’re using a Kubernetes environment, with the Kubernetes auth method. Both methods are described below.

Authenticate with a token

The information below assumes you’re familiar with creating a Vault token. Refer to the HashiCorp Vault documentation for more information and token options.

1. Create a new token using the command:

   vault token create -policy=sample_policy

2. To confirm that your token and policy allow access to the secret, run the commands:
   

   export VAULT_TOKEN=<YOUR_TOKEN>
   vault kv get -mount=secret sample/secret

3. You should see the following output:

   ====== Secret Path ======
   secret/data/sample/secret
   
   ======= Metadata =======
   Key         Value
   ---         -----
   ~~~~        ~~~~~
   
   ====== Data ======
   Key         Value
   ---         -----
   ~~~~        ~~~~~

   If you don’t see the output above, refer to the Hashicorp documentation to troubleshoot the issue.
   To complete the process, an environment variable associated with the token must be present on Confluence.

4. Define the environment variable SECRET_STORE_VAULT_TOKEN in the context of the Confluence instance. A simple way to do this is to add the following line to the ~/.bashrc file for the user running Confluence:

   export SECRET_STORE_VAULT_TOKEN=<YOUR_TOKEN>

Authenticate using Kubernetes Service Account Token

If Confluence is operating within a Kubernetes environment, you can leverage the Kubernetes auth method. This method uses a Kubernetes Service Account Token to confirm the identity of the pod that runs Confluence and to grant the appropriate access.

Refer to the Hashicorp Vault documentation for more information on how to set up Kubernetes auth method in your Vault instance. Make sure you have enabled Kubernetes auth method on your Vault server before you start the steps below.

You will also need to set some environment variables in the following steps. The table below describes these.

1. If you used custom path to create a Kubernetes auth method, replace kubernetes in the CLI command in the following step with your path name.

2. Define a role to link the auth method with the sample_policy you created with the following command:

   vault write auth/kubernetes/role/<YOUR_NEW_ROLE_NAME> \
       bound_service_account_names=<YOUR_PRODUCT_SERVICE_ACCOUNT_NAME> \
       bound_service_account_namespaces=<YOUR_PRODUCT_SERVICE_NAMESPACE> \
       policies=sample_policy

3. Ensure that your Confluence pod has access to the secret.
   Currently, Vault CLI doesn’t offer support for logging in with Kubernetes auth method, but you can log in to retrieve client token using HTTP API and then use this generated token to test for access.

4. If you can’t retrieve the secret with the generated token, refer to Hashicorp’s documentation to troubleshoot the issue.

5. Refer to the table at the start of these steps to set the following environment variables for Confluence:

• SECRET_STORE_VAULT_KUBE_AUTH_ROLE

• SECRET_STORE_VAULT_KUBE_AUTH_PATH (optional)

• SECRET_STORE_VAULT_KUBE_AUTH_JWT_PATH (optional)

Step 4: Add the Vault configuration data to confluence.cfg.xml

Vault is configured via a JSON object that is added to the <home-directory>/confluence.cfg.xml file. The JSON configuration object has a number of fields. Make sure you refer to the following table for details on each of these properties.

1. In the Confluence home directory, back up the confluence.cfg.xml file. Move the backup file to a safe place outside of your Confluence server.

2. In the confluence.cfg.xml file, add or modify the jdbc.password.decrypter.classname property to contain:

   com.atlassian.secrets.store.vault.VaultSecretStore

3. In the confluence.cfg.xml file, add or modify the hibernate.connection.password property to contain your JSON configuration object. Use the table at the start of these steps for further information on these fields.
   Here is an example of how it might look:
   

   {"mount": "secret", "path": "sample/secret", "key": "password", "endpoint": "https://127.0.0.1:8200"}

4. Restart Confluence