# Automatic renewal of certificates on Ubuntu with Vault AppRole

This guide shows how an unattended Ubuntu server gets certificates from a remote Vault server with the AppRole authentication method, and keeps them fresh. A systemd timer starts a renewal service, the service renews a certificate before it expires, and after a successful renewal it runs a configurable Bash hook that signals the application to reread the certificate or restarts it. The renewal script is written in Python 3 and talks to Vault through the Vault HTTP API, so the server does not need the Vault CLI.

The renewal service is a systemd template, so one server can renew several certificates, each with its own settings. The guide first tests the whole process, including the hook, with a server certificate served by Caddy (section 3). Then it sets up the client certificate for the WIF tests (section 4).

## Prerequisites

- The Vault PKI configured as described in step 1a of the main `README.md`: the `pki_int` mount and the `wif-client` and `web-server` roles (`vault_configure.sh`).
- A Vault server reachable over HTTPS. The examples use `https://vault.koszewscy.waw.pl`.
- Ubuntu Server 24.04 LTS or 26.04 LTS with systemd and `sudo` access.
- The IP address of the server as seen by Vault. The examples use `203.0.113.10`.

The commands are written for bash.

## 1. Configure Vault

Role: **Administrator**

Point the CLI at the server and authenticate with an administrator token:

```bash
export VAULT_ADDR=https://vault.koszewscy.waw.pl
export VAULT_TOKEN=<admin-token>
```

1. Create a policy that allows signing certificate requests with the `wif-client` and `web-server` roles and nothing else. When the policy already exists, the command replaces it, so the file must list all the paths.
2. Enable the AppRole authentication method.
3. Create the `cert-renew-<server_name>` role. It issues batch tokens valid for 10 minutes, the secret ID does not expire so that the server can renew unattended, and both the secret ID and the tokens work only from the IP address of the server.
4. Create the `cert-renew` role. It has the same settings, but the secret ID and the tokens work from any address.
5. Read the role ID of the role chosen for the server.
6. Generate a response-wrapped secret ID of that role that can be unwrapped once within 10 minutes. Deliver the role ID and the wrapping token to the owner of the server over two different channels.

`pki-client.hcl`:

```hcl
path "pki_int/sign/wif-client" {
  capabilities = ["update"]
}

path "pki_int/sign/web-server" {
  capabilities = ["update"]
}
```

```bash
vault policy write pki-client pki-client.hcl

vault auth enable approle

vault write auth/approle/role/cert-renew-<server_name> \
    token_type=batch \
    token_policies=pki-client \
    token_ttl=10m \
    secret_id_ttl=0 \
    secret_id_num_uses=0 \
    secret_id_bound_cidrs=203.0.113.10/32 \
    token_bound_cidrs=203.0.113.10/32

vault write auth/approle/role/cert-renew \
    token_type=batch \
    token_policies=pki-client \
    token_ttl=10m \
    secret_id_ttl=0 \
    secret_id_num_uses=0

vault read -field=role_id auth/approle/role/cert-renew-<server_name>/role-id

vault write -wrap-ttl=10m -f -field=wrapping_token \
    auth/approle/role/cert-renew-<server_name>/secret-id
```

The last two commands use the role of the server. Use `cert-renew` in their paths for a server that has no fixed address.

Generate a new secret ID, deliver it the same way and replace the file on the server periodically.

> **Note:** The `wif-client` role accepts any common name (`allow_any_name=true`). The server can therefore request a certificate with any subject CN, including the CN that other workloads rely on as the WiF subject. Restrict the role (`allowed_domains`, `allow_bare_domains`, `enforce_hostnames`) when the common name of the server must be fixed.

## 2. Prepare the server

Role: **Server administrator**

### Install the tools

Python 3 handles the HTTP calls and the JSON answers of the Vault API with its standard library only, and `openssl` creates the key and the CSR. Python 3 is `python3` 3.12 on Ubuntu 24.04 and 3.14 on Ubuntu 26.04. The commands do nothing when the packages are already installed:

```bash
sudo apt update
sudo apt install -y python3 openssl
```

### Store the credentials

1. Create the directories for the configuration and for the certificates.
2. Create the credential files with the access limited to root before writing into them.
3. Write the role ID.
4. Store the secret ID with the script. It asks for the wrapping token without echo, so the token does not appear in the process list or in the shell history. The configuration file is not read in this step, so pass the address of Vault. The wrapping token can be unwrapped only once. When the unwrap fails, treat the wrapping token as compromised and ask for a new one.

```bash
sudo install -d -m 0750 /etc/vault-cert /var/lib/vault-cert
sudo install -m 0600 /dev/null /etc/vault-cert/role_id
sudo install -m 0600 /dev/null /etc/vault-cert/secret_id

echo '<role-id>' | sudo tee /etc/vault-cert/role_id > /dev/null
sudo VAULT_ADDR=https://vault.koszewscy.waw.pl /usr/local/sbin/vault-cert-renew --unwrap
```

### Install the renewal script

The renewal script is `vault_cert_renew.py` in the repository. Install it at `/usr/local/sbin/vault-cert-renew`:

```bash
sudo install -m 0755 vault_cert_renew.py /usr/local/sbin/vault-cert-renew
```

The key is generated on the server and never leaves it. The script exits without contacting Vault while the certificate is valid for more than `RENEW_BEFORE_DAYS`, replaces the files only when the new certificate matches the new key, and runs the hook after a successful renewal. The Vault token stays in memory and the batch token expires on its own. A failed Vault call prints the status code and the answer of Vault to the journal.

Test the AppRole authentication with the `--login-test` option. The script logs in with the stored role ID and secret ID and prints the type, the policies and the lifetime of the token it received, without issuing a certificate. The configuration file is not read in this step, so pass the address of Vault. The expected result is a `batch` token with the `pki-client` policy valid for 600 seconds:

```bash
sudo VAULT_ADDR=https://vault.koszewscy.waw.pl /usr/local/sbin/vault-cert-renew --login-test
```

A `permission denied` answer means that the role ID or the secret ID is wrong, or that the request comes from an address outside the CIDR of the role.

### Configure the renewal

Each certificate has an instance name and a configuration file `/etc/vault-cert/<instance>.env`. This guide uses the instances `caddy` for the server certificate of the test site (section 3) and `wif` for the client certificate of the WIF tests (section 4).

`/etc/vault-cert/caddy.env`:

```text
VAULT_ADDR=https://vault.koszewscy.waw.pl
CERT_CN=web.example.com
CERT_DIR=/var/lib/vault-cert/caddy
SIGN_PATH=pki_int/sign/web-server
HOOK=/etc/vault-cert/caddy-hook.sh
```

`/etc/vault-cert/wif.env`:

```text
VAULT_ADDR=https://vault.koszewscy.waw.pl
CERT_CN=test-client
CERT_DIR=/var/lib/vault-cert/wif
RENEW_BEFORE_DAYS=30
CERT_TTL=8760h
```

The `caddy` instance sets the signing path of the `web-server` role and a hook. The `wif` instance uses the defaults for the signing path and has no hook.

The settings are:

| Variable | Default | Meaning |
|----------|---------|---------|
| `VAULT_ADDR` | none | Address of the Vault server. |
| `VAULT_CACERT` | none | CA certificate for the TLS connection, needed only when the server certificate is issued by a private CA. |
| `CERT_CN` | none, required | Subject CN of the certificate. |
| `CERT_DIR` | `/var/lib/vault-cert` | Directory of `client_cert.pem` and `client_key.pem`. |
| `ROLE_ID_FILE` | `/etc/vault-cert/role_id` | File with the AppRole role ID. |
| `SECRET_ID_FILE` | `/etc/vault-cert/secret_id` | File with the AppRole secret ID. |
| `SIGN_PATH` | `pki_int/sign/wif-client` | Vault path used to sign the CSR. |
| `RENEW_BEFORE_DAYS` | `30` | The certificate is renewed when it expires in fewer days. |
| `CERT_TTL` | `8760h` | Requested lifetime of the new certificate. |
| `HOOK` | none | Script run after a successful renewal. |

The hook runs as root after the new files are in place. It gets the paths of the certificate and the key in the `CERT_FILE` and `KEY_FILE` environment variables. A failing hook fails the service, the failure is visible in the journal and in `systemctl status`. Leave `HOOK` unset to renew without any action. When `HOOK` names a file that does not exist or is not executable, the renewal still succeeds and the script logs a warning and skips the hook.

> **Note:** The certificate is already renewed when the hook runs. When the hook fails, the next timer run finds a valid certificate and does not run the hook again. Fix the hook and run it by hand with the environment variables set, or remove `client_cert.pem` to force a new renewal.

### Create the systemd units

The directory `systemd` in the repository contains the service `vault-cert-renew@.service` and the timer `vault-cert-renew@.timer`. The service runs the renewal script once with the settings of the instance. The timer starts it four times a day. `Persistent=true` runs a missed check after the server was powered off, and `RandomizedDelaySec` spreads the load when many servers share the configuration. Install the units:

```bash
sudo install -m 0644 systemd/vault-cert-renew@.service /etc/systemd/system/vault-cert-renew@.service
sudo install -m 0644 systemd/vault-cert-renew@.timer /etc/systemd/system/vault-cert-renew@.timer
sudo systemctl daemon-reload
```

## 3. Test the renewal and the hook with Caddy

Role: **Server administrator**

Caddy serves a test page over TLS with a server certificate issued by the `web-server` role. The hook copies the renewed certificate to Caddy and reloads it, so the serial number of the served certificate shows whether the hook works.

### Install Caddy, the test site and the hook

Install Caddy from the Ubuntu repository (the `universe` component):

```bash
sudo apt install -y caddy
```

The directory `web-test` in the repository contains these files:

- `index.html`, the test page.
- `Caddyfile`, the configuration of a site on port 443 that uses the certificate files `/etc/caddy/server_cert.pem` and `/etc/caddy/server_key.pem`.
- `caddy-hook.sh`, the hook that copies the renewed files to those paths and reloads Caddy, or starts it when it is not running. The reload makes Caddy read the certificate files again.

Install them. The hook goes to the path of `HOOK` in `caddy.env`, and the certificate subject `CERT_CN` in that file is the name of the site.
```bash
sudo install -d -m 0755 /var/www/web-test
sudo install -m 0644 web-test/index.html /var/www/web-test/index.html
sudo install -m 0644 web-test/Caddyfile /etc/caddy/Caddyfile
sudo install -m 0750 web-test/caddy-hook.sh /etc/vault-cert/caddy-hook.sh
```

### Run the renewal and check the result

1. Run the first renewal by hand and read its log. The log shows the login, the signing and the hook.

    ```bash
    sudo systemctl start vault-cert-renew@caddy.service
    sudo journalctl -u vault-cert-renew@caddy.service --no-pager
    ```

2. Check the issued certificate. It has the subject and the SAN `web.example.com` and the extended key usage `TLS Web Server Authentication`.

    ```bash
    sudo openssl x509 -in /var/lib/vault-cert/caddy/client_cert.pem -noout -subject -issuer -serial -enddate -ext subjectAltName,extendedKeyUsage
    ```

3. Request the page and read the serial number of the certificate that Caddy serves. It equals the serial number from the previous step. The `-k` option skips the verification of the certificate chain.

    ```bash
    curl -k https://localhost/
    openssl s_client -connect localhost:443 -servername web.example.com < /dev/null 2> /dev/null | openssl x509 -noout -serial -enddate
    ```

4. Force a renewal by removing the certificate, and read the serial number again. It differs from the previous one, so Caddy serves the new certificate and the hook works.

    ```bash
    sudo rm /var/lib/vault-cert/caddy/client_cert.pem
    sudo systemctl start vault-cert-renew@caddy.service
    openssl s_client -connect localhost:443 -servername web.example.com < /dev/null 2> /dev/null | openssl x509 -noout -serial -enddate
    ```

A run that finds a valid certificate logs `nothing to do`, a renewal logs the new expiry date and the hook. To run the renewal on schedule, enable the timer of the instance as in section 4.

## 4. Renew the WIF client certificate

Role: **Server administrator**

The client certificate for the WIF tests is issued by the `wif-client` role and needs no hook. Its settings are in `/etc/vault-cert/wif.env`.

1. Run the first renewal by hand and read its log.

    ```bash
    sudo systemctl start vault-cert-renew@wif.service
    sudo journalctl -u vault-cert-renew@wif.service --no-pager
    ```

2. Check the certificate.

    ```bash
    sudo openssl x509 -in /var/lib/vault-cert/wif/client_cert.pem -noout -subject -issuer -enddate
    ```

3. Enable the timer.

    ```bash
    sudo systemctl enable --now vault-cert-renew@wif.timer
    ```

4. Check the timer schedule.

    ```bash
    systemctl list-timers 'vault-cert-renew@*'
    ```

## Next steps

1. Create the credentials configuration `cred.json` on the server for the certificate and the key managed by the renewal service. The names come from `env.sh`.

    ```bash
    gcloud iam workload-identity-pools create-cred-config \
        "projects/${PROJECT_NUMBER}/locations/global/workloadIdentityPools/${POOL_ID}/providers/${PROVIDER_ID}" \
        --service-account="${SA_EMAIL}" \
        --credential-cert-path=/var/lib/vault-cert/wif/client_cert.pem \
        --credential-cert-private-key-path=/var/lib/vault-cert/wif/client_key.pem \
        --output-file=cred.json
    ```

2. Activate the service account with `gcloud auth login --cred-file=cred.json`.
3. Run the WIF tests.

Example test commands are in `test_wif.sh`.