User Configurable Password Generation for Secret Engines

»Challenge

Vault's secret engines generate passwords that adhere to a default pattern that may not meet the standards required by your applications or within your organization. You may require passwords with:

• lengths that are smaller or exceed the default length.
• differing charsets
• frequency of characters
• positional requirements for characters
• disallowed repetition
• disallowed words from a dictionary

»Solution

Vault 1.5 introduced support for configurable password generation defined by a password policy. A policy defines the rules and requirements that the password must adhere and can provide that password directly through a new endpoint or within secrets engines.

Secrets engines with support for password policies:

• Active Directory
• Azure
• OpenLDAP
• RabbitMQ

»Prerequisites

• Vault version 1.5 or later; refer to the Getting Started tutorial to install Vault.

• RabbitMQ environment or Docker.

»Start Vault

A Vault server is able to generate passwords that meet the requirements of a configurable password policy. These password policies may also be applied to supported secrets engines.

In another terminal, start a Vault dev server with root as the root token.

$ vault server -dev -dev-root-token-id root

Export an environment variable for the vault CLI to address the Vault server.

$ export VAULT_ADDR=http://0.0.0.0:8200

Login with the root token.

$ vault login root
Success! You are now authenticated. The token information displayed below
is already stored in the token helper. You do NOT need to run "vault login"
again. Future Vault requests will automatically use this token.

Key                  Value
---                  -----
token                root
token_accessor       n9CYvD0GK3iV6nwAOZQAy9Md
token_duration       ∞
token_renewable      false
token_policies       ["root"]
identity_policies    []
policies             ["root"]

The Vault server is ready to explore password policies.

»Start RabbitMQ

RabbitMQ is a message-broker that has a secrets engine that enables Vault to generate user credentials.

In another terminal, start a RabbitMQ server running on port 15672 that has a user named learn_vault with the password hashicorp.

$ docker run --rm --name some-rabbit -p 15672:15672 \
    -e RABBITMQ_DEFAULT_USER=learn_vault \
    -e RABBITMQ_DEFAULT_PASS=hashicorp \
    rabbitmq:3-management

Docker downloads the necessary images and launches a RabbitMQ database container. RabbitMQ is ready for Vault to generate credentials.

»Setup RabbitMQ secrets engine with the default password policy

The RabbitMQ secrets engine generates passwords that adhere to a default password policy.

Enable the RabbitMQ secrets engine at the rabbitmq-default-policy path.

$ vault secrets enable -path rabbitmq-default-policy rabbitmq
Success! Enabled the rabbitmq secrets engine at: rabbitmq-default-policy/

The secrets engine requires configuration so it can communicate with the RabbitMQ server.

Configure the secrets engine to connect to the RabbitMQ server.

$ vault write rabbitmq-default-policy/config/connection \
    connection_uri=http://localhost:15672 \
    username="learn_vault" \
    password="hashicorp"
Success! Data written to: rabbitmq-default-policy/config/connection

The secrets engine is configured to talk to the RabbitMQ server running locally over the HTTP management interface, http://localhost:15672. The user, learn_vault, with its password hashicorp, is an administrator that was generated when RabbitMQ started.

Create a role named example.

$ vault write rabbitmq-default-policy/roles/example \
    vhosts='{"/":{"read": ".*", "write": ".*"}}'
Success! Data written to: rabbitmq-default-policy/roles/example

The example role defines an endpoint for dynamic credential generation. The credentials generated are granted read and write every entity.

Generate credentials from the example role.

$ vault read rabbitmq-default-policy/creds/example
Key                Value
---                -----
lease_id           rabbitmq-default-policy/creds/example/4yCYMMCZ4I4mKBlPsbNNsKdj
lease_duration     768h
lease_renewable    true
password           v0E6LjTeabrtEAz72cbYINNodWSApPS61vGP
username           root-2b3482f7-b522-9997-ab0c-1f3057a396d3

The credentials display the username and password generated. The password generated, v0E6LjTeabrtEAz72cbYINNodWSApPS61vGP, adheres to the default password policy for the secrets engine.

»Define a password policy

Create a policy file named example_policy.hcl.

$ tee example_policy.hcl <<EOF
length=20

rule "charset" {
  charset = "abcdefghijklmnopqrstuvwxyz"
  min-chars = 1
}

rule "charset" {
  charset = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
  min-chars = 1
}

rule "charset" {
  charset = "0123456789"
  min-chars = 1
}

rule "charset" {
  charset = "!@#$%^&*"
  min-chars = 1
}
EOF

The policy is written in HashiCorp Configuration Language (HCL). The length field sets the length of the password returned to 20 charaters. Each rule stanza defines a character set and the minimum number of occurences those characters need to appear in the generated password. These rules are cumulative so each one adds more requirements on the password generated.

Create a Vault password policy named example with the password policy rules defined in example_policy.hcl.

$ vault write sys/policies/password/example policy=@example_policy.hcl
Success! Data written to sys/polices/password/example_policy

This policy can now be accessed directly to generate a password or referenced by its name example when configuring supported secrets engines.

Generate a password from the example password policy.

$ vault read sys/policies/password/example/generate
Key         Value
---         -----
password    #v!RQDHxHunJ1TUmCyys

The password generated adheres to the requirements:

• length of 20 characters
• at least 1 uppercase character
• at least 1 lowercase character
• at least 1 number
• at least 1 symbol

»Setup RabbitMQ secrets engine with a password policy

The RabbitMQ secrets engine may be configured to adopt a password policy.

Enable the RabbitMQ secrets engine at the rabbitmq-with-policy path.

$ vault secrets enable -path rabbitmq-with-policy rabbitmq
Success! Enabled the rabbitmq secrets engine at: rabbitmq-with-policy/

The secrets engine requires configuration so it can communicate with the RabbitMQ server.

Configure the secrets engine to connect to the RabbitMQ server and use the example password policy.

$ vault write rabbitmq-with-policy/config/connection \
    connection_uri=http://localhost:15672 \
    username="learn_vault" \
    password="hashicorp" \
    password_policy="example"
Success! Data written to: rabbitmq-with-policy/config/connection

The same connection information is used to establish the connection with the RabbitMQ server. The difference is that the password_policy has been set to the example policy defined in Define a password policy section.

Create a role named example.

$ vault write rabbitmq-with-policy/roles/example \
    vhosts='{"/":{"read": ".*", "write": ".*"}}'
Success! Data written to: rabbitmq-with-policy/roles/example

The example role defines an endpoint for dynamic credential generation. The credentials generated are granted read and write on every entity.

Generate a login from the example role.

$ vault read rabbitmq-with-policy/creds/example
Key                Value
---                -----
lease_id           rabbitmq-with-policy/creds/example/vfoht23KclCzOuKzlRYeMZZS
lease_duration     768h
lease_renewable    true
password           HtjHFtdy21*iBpNZjLNb
username           root-4efd295b-1a4e-285c-14c8-06b135c5464a

The credentials display the username and password generated. The password generated, HtjHFtdy21*iBpNZjLNb, adheres to the example password policy defined in the secrets engine's configuration.

»Next steps

You configured a RabbitMQ secrets engine that generated credentials with the default password policy. Learn more about the RabbitMQ secrets engine by reading the documentation and the API documentation.

Then you defined a password policy and generated a password. Learn more about by reading the documentation and the API documentation.

Finally, you configured another RabbitMQ secrets engine that generated credentials that used a custom password policy.