Generate ssl check list directly from letsencrypt variables

This autogenerates the list of ssl domains for the ssl-cert-check tool
directly from the letsencrypt list.

The first step is the install-certcheck role that replaces the
puppet-ssl_cert_check module that does the same.  The reason for this
is so that during gate testing we can test this on the test
bridge.openstack.org server, and avoid adding another node as a
requirement for this test.

letsencrypt-request-certs is updated to set a fact
letsencrypt_certcheck_domains for each host that is generating a
certificate.  As described in the comments, this defaults to the first
host specified for the certificate and the listening port can be
indicated (if set, this new port value is stripped when generating
certs as is not necessary for certificate generation).

The new letsencrypt-config-certcheck role runs and iterates all
letsencrypt hosts to build the final list of domains that should be
checked.  This is then extended with the
letsencrypt_certcheck_additional_domains value that covers any hosts
using certificates not provisioned by letsencrypt using this
mechanism.

These additional domains are pre-populated from the openstack.org
domains in the extant check file, minus those openstack.org domain
certificates we are generating via letsencrypt (see
letsencrypt-create-certs/handlers/main.yaml).  Additionally, we
update some of the certificate variables in host_vars that are
listening on port !443.

As mentioned, bridge.openstack.org is placed in the new certcheck
group for gate testing, so the tool and config file will be deployed
to it.  For production, cacti is added to the group, which is where
the tool currently runs.  The extant puppet installation is disabled,
pending removal in a follow-on change.

Change-Id: Idbe084f13f3684021e8efd9ac69b63fe31484606

inventory/groups.yaml

@@ -28,6 +28,8 @@ groups:
 #  backup-server:
 #    - backup[0-9]*.opendev.org
   cacti: cacti[0-9]*.open*.org
+  certcheck:
+    - cacti[0-9]*.open*.org
   cloud-launcher:
     - bridge.openstack.org
   codesearch:

manifests/site.pp

@@ -28,7 +28,9 @@ node /^health\d*\.openstack\.org$/ {
 # Node-OS: xenial
 node /^cacti\d+\.open.*\.org$/ {
   $group = "cacti"
-  include openstack_project::ssl_cert_check
+  # NOTE(ianw) 2020-05 : disabled pending removal, migrated to
+  # ansible.
+  #  include openstack_project::ssl_cert_check
   class { 'openstack_project::cacti':
     cacti_hosts => hiera_array('cacti_hosts'),
     vhost_name  => 'cacti.openstack.org',

playbooks/group_vars/certcheck.yaml

@@ -0,0 +1,17 @@
+letsencrypt_certcheck_additional_domains:
+  - ask.openstack.org 443
+  - ethercalc.openstack.org 443
+  - etherpad.openstack.org 443
+  - firehose.openstack.org 8883
+  - git.openstack.org 443
+  - openstackid-dev.openstack.org 443
+  - openstackid.org 443
+  - refstack.openstack.org 443
+  - review.openstack.org 443
+  - storyboard.openstack.org 443
+  - survey.openstack.org 443
+  - static.openstack.org 443
+  - translate.openstack.org 443
+  - wiki.openstack.org 443
+  - www.openstack.org 443
+  - zuul.openstack.org 443

playbooks/host_vars/gitea01.opendev.org.yaml

@@ -1,4 +1,4 @@
 letsencrypt_certs:
   gitea01-main:
-    - gitea01.opendev.org
+    - gitea01.opendev.org:3000
     - opendev.org

playbooks/host_vars/gitea02.opendev.org.yaml

@@ -1,4 +1,4 @@
 letsencrypt_certs:
   gitea02-main:
-    - gitea02.opendev.org
+    - gitea02.opendev.org:3000
     - opendev.org

playbooks/host_vars/gitea03.opendev.org.yaml

@@ -1,4 +1,4 @@
 letsencrypt_certs:
   gitea03-main:
-    - gitea03.opendev.org
+    - gitea03.opendev.org:3000
     - opendev.org

playbooks/host_vars/gitea04.opendev.org.yaml

@@ -1,4 +1,4 @@
 letsencrypt_certs:
   gitea04-main:
-    - gitea04.opendev.org
+    - gitea04.opendev.org:3000
     - opendev.org

playbooks/host_vars/gitea05.opendev.org.yaml

@@ -1,4 +1,4 @@
 letsencrypt_certs:
   gitea05-main:
-    - gitea05.opendev.org
+    - gitea05.opendev.org:3000
     - opendev.org

playbooks/host_vars/gitea06.opendev.org.yaml

@@ -1,4 +1,4 @@
 letsencrypt_certs:
   gitea06-main:
-    - gitea06.opendev.org
+    - gitea06.opendev.org:3000
     - opendev.org

playbooks/host_vars/gitea07.opendev.org.yaml

@@ -1,4 +1,4 @@
 letsencrypt_certs:
   gitea07-main:
-    - gitea07.opendev.org
+    - gitea07.opendev.org:3000
     - opendev.org

playbooks/host_vars/gitea08.opendev.org.yaml

@@ -1,4 +1,4 @@
 letsencrypt_certs:
   gitea08-main:
-    - gitea08.opendev.org
+    - gitea08.opendev.org:3000
     - opendev.org

playbooks/host_vars/insecure-ci-registry01.opendev.org.yaml

@@ -1,4 +1,4 @@
 letsencrypt_certs:
   insecure-ci-registry01-main:
-    - insecure-ci-registry01.opendev.org
+    - insecure-ci-registry01.opendev.org:5000
     - insecure-ci-registry.opendev.org

playbooks/roles/install-certcheck/README.rst

@@ -0,0 +1,24 @@
+Install ssl-cert-check
+
+Installs the ssl-cert-check tool and a cron job to check the freshness
+of the SSL certificates for the configured domains daily.
+
+**Role Variables**
+
+.. zuul:rolevar:: ssl_cert_check_domain_list
+   :default: /var/lib/certcheck/domainlist
+
+   The list of domains to check
+
+.. zuul:rolevar:: ssl_cert_check_days
+   :default: 30
+
+   Warn about certificates who have less than this number of days to
+   expiry.
+
+.. zuul:rolevar:: ssl_cert_check_email
+   :default: root
+
+   The email to send reports to
+
+

playbooks/roles/install-certcheck/defaults/main.yaml

@@ -0,0 +1,3 @@
+cert_check_domain_list: /var/lib/certcheck/domainlist
+cert_check_days: 30
+cert_check_email: root

playbooks/roles/install-certcheck/tasks/main.yaml

@@ -0,0 +1,32 @@
+- name: Ensure dependencies
+  package:
+    name:
+      - openssl
+      - bsd-mailx
+
+- name: Ensure certcheck user
+  user:
+    name: certcheck
+    comment: User for SSL validation
+
+- name: Ensure certcheck config directory
+  file:
+    state: directory
+    path: '{{ cert_check_domain_list | dirname }}'
+    owner: certcheck
+    group: certcheck
+    mode: 0755
+
+- name: Pull latest ssl-cert-check from git
+  git:
+    repo: 'https://github.com/Matty9191/ssl-cert-check'
+    dest: /opt/ssl-cert-check
+
+- name: Install cron job
+  cron:
+    user: certcheck
+    name: 'Run certcheck'
+    state: present
+    job: "/opt/ssl-cert-check/ssl-cert-check -a -q -f {{ cert_check_domain_list }} -x {{ cert_check_days }} -e {{ cert_check_email }}"
+    hour: 12
+    minute: 04

playbooks/roles/letsencrypt-config-certcheck/README.rst

@@ -0,0 +1,27 @@
+Generate SSL check list
+
+This role automatically generates a list of domains for the
+certificate validation checks.  This ensures our certificates are
+valid and are being renewed as expected.
+
+This role must run after ``letsencrypt-request-certs`` role, as that
+builds the ``letsencrypt_certcheck_domains`` variable for each
+host and certificate.  It must also run on a host that has already run
+the ``install-certcheck`` role.
+
+**Role Variables**
+
+.. zuul:rolevar:: letsencrypt_certcheck_domain_list
+   :default: /var/lib/certcheck/ssldomains
+
+   The ssl-cert-check domain configuration file to write.  See also
+   the ``install-certcheck`` role.
+
+.. zuul:rolevar:: letsencrypt_certcheck_additional_domains
+   :default: []
+
+   A list of additional domains to check for hosts not using the
+   ``letsencrypt-*`` roles.  Each entry should be in the format
+   ``hostname port``.
+
+

playbooks/roles/letsencrypt-config-certcheck/defaults/main.yaml

@@ -0,0 +1 @@
+letsencrypt_certcheck_domain_list: /var/lib/certcheck/ssldomains

playbooks/roles/letsencrypt-config-certcheck/tasks/main.yaml

@@ -0,0 +1,17 @@
+- name: Make domain list
+  set_fact:
+    letsencrypt_certcheck_domains: []
+
+- name: Build SSL domain list
+  set_fact:
+    letsencrypt_certcheck_domains: '{{ letsencrypt_certcheck_domains }} + {{ hostvars[item]["letsencrypt_certcheck_domains" ] }}'
+  with_inventory_hostnames:
+    - letsencrypt:!disabled
+
+- name: Write configuration file
+  template:
+    dest: '{{ letsencrypt_certcheck_domain_list }}'
+    src: ssldomains.j2
+    owner: certcheck
+    group: certcheck
+    mode: 0644

playbooks/roles/letsencrypt-config-certcheck/templates/ssldomains.j2

@@ -0,0 +1,6 @@
+{% for domain in letsencrypt_certcheck_domains %}
+{{ domain }}
+{% endfor %}
+{% for domain in letsencrypt_certcheck_additional_domains %}
+{{ domain }}
+{% endfor %}

playbooks/roles/letsencrypt-create-certs/tasks/acme.yaml

@@ -1,6 +1,6 @@
 - name: 'Build arguments for letsencrypt acme.sh driver for: {{ item.key }}'
   set_fact:
-    acme_args: '"{% for domain in item.value %}-d {{ domain }} {% endfor %}"'
+    acme_args: '"{% for domain in item.value %}-d {{ domain.split(":")[0] }} {% endfor %}"'
 
 - name: 'Run acme.sh driver for {{ item.key }} certificate issue'
   shell:
@@ -12,4 +12,4 @@
     LETSENCRYPT_STAGING: '{{ "1" if letsencrypt_use_staging else "0" }}'
   notify: 'letsencrypt updated {{ item.key }}'
 
-# Keys generated!
+# Keys generated!

playbooks/roles/letsencrypt-request-certs/README.rst

@@ -62,3 +62,12 @@ provision process.
      _acme-challenge.hostname01.opendev.org.  IN   CNAME  acme.opendev.org.
      _acme-challenge.hostname.opendev.org.    IN   CNAME  acme.opendev.org.
      _acme-challenge.foo.opendev.org.         IN   CNAME  acme.opendev.org.
+
+   The hostname in the first entry for each certificate will be
+   registered with the ``letsencrypt-config-certcheck`` for periodic
+   freshness tests; from the example above, ``hostname01.opendev.org``
+   and ``foo.opendev.org`` would be checked.  By default this will
+   check on port 443; if the certificate is actually active on another
+   port you can specify it after a colon;
+   e.g. ``foo.opendev.org:5000`` would indicate this host listens with
+   this certificate on port 5000.

playbooks/roles/letsencrypt-request-certs/tasks/acme.yaml

@@ -2,7 +2,7 @@
   set_fact:
     # NOTE(ianw): note the domains are passed in one string (between
     # ") as it makes argument parsing a little easier in the driver.sh
-    acme_args: '"{% for domain in cert.value %}-d {{ domain }} {% endfor %}"'
+    acme_args: '"{% for domain in cert.value %}-d {{ domain.split(":")[0] }} {% endfor %}"'
 
 - name: Run acme.sh driver for certificate issue
   shell:

playbooks/roles/letsencrypt-request-certs/tasks/main.yaml

@@ -1,10 +1,6 @@
 - set_fact:
     acme_txt_required: []
 
-- name: Show cert list
-  debug:
-    var: letsencrypt_certs
-
 # Handle multiple certs for a single host; like
 #
 # letsencrypt_certs:
@@ -22,5 +18,25 @@
   loop_control:
     loop_var: cert
 
-- debug:
+- name: Create ssl check domain list
-    var: acme_txt_required
+  # For each generated certificate get the first entry as the domain
+  # to run the certificate validation tests against.  If it specifies
+  # a port explicitly, use that, otherwise assume 443.
+  #
+  # Later in ssl-check role, the final certificate validation list is
+  # generated by walking the letsencrypt_certcheck_domains variable
+  # for each host in the letsencrypt group.
+  set_fact:
+    letsencrypt_certcheck_domains: >-
+      {%- set d = [] -%}
+      {%- for cert in letsencrypt_certs.keys() -%}
+       {%- for host in letsencrypt_certs[cert] -%}
+        {%- if loop.first -%}
+         {%- if not ":" in host -%}
+           {%- set host = host+":443" -%}
+         {%- endif -%}
+         {%- set d = d.append(host.replace(":"," ")) -%}
+        {% endif %}
+       {% endfor %}
+      {% endfor %}
+      {{- d -}}

playbooks/service-letsencrypt.yaml

@@ -1,6 +1,9 @@
 # This needs to happen in order.  letsencrypt hosts export their TXT
 # authentication records which is installed onto adns1, and then the
 # hosts verify to issue/renew keys
+- hosts: "certcheck:!disabled"
+  roles:
+    - install-certcheck
 - hosts: "letsencrypt:!disabled"
   name: "Base: deploy and renew certificates"
   roles:
@@ -14,3 +17,6 @@
   name: "Create certs"
   roles:
     - letsencrypt-create-certs
+- hosts: "certcheck:!disabled"
+  roles:
+    - letsencrypt-config-certcheck

playbooks/zuul/templates/gate-groups.yaml.j2

@@ -5,6 +5,9 @@ groups:
   docker:
     - bionic-docker
 
+  certcheck:
+    - bridge.openstack.org
+
   letsencrypt:
     - letsencrypt01.opendev.org
     - letsencrypt02.opendev.org

playbooks/zuul/templates/host_vars/letsencrypt01.opendev.org.yaml.j2

@@ -1,7 +1,7 @@
 letsencrypt_certs:
   letsencrypt01-main-service:
-    - letsencrypt01.opendev.org
+    - letsencrypt01.opendev.org:5000
     - letsencrypt.opendev.org
     - alias.opendev.org
   letsencrypt01-other-service:
-    - someotherservice.opendev.org
+    - someotherservice.opendev.org

testinfra/test_letsencrypt.py

@@ -15,6 +15,7 @@
 import pytest
 
 testinfra_hosts = ['adns-letsencrypt.opendev.org',
+                   'bridge.openstack.org',
                    'letsencrypt01.opendev.org',
                    'letsencrypt02.opendev.org']
 
@@ -138,3 +139,25 @@ def test_acme_sh_config(host):
     config = host.file('/root/.acme.sh/account.conf')
     assert config.exists
     assert config.contains("^ACCOUNT_EMAIL='le-test@opendev.org'")
+
+def test_certcheck_config(host, zuul_data):
+    if host.backend.get_hostname() != 'bridge.openstack.org':
+        pytest.skip()
+
+    if zuul_data['extra']['zuul']['job'] != 'system-config-run-letsencrypt':
+        pytest.skip()
+
+    domainlist = host.file('/var/lib/certcheck/ssldomains')
+
+    # TODO(ianw): figure out a flag or something from the
+    # system-config-run-letsencrypt test so that we can assert this
+    # file exists only in that case.
+    if not domainlist.exists:
+        pytest.skip()
+
+    assert domainlist.exists
+    assert domainlist.user == 'certcheck'
+    # from variables
+    assert domainlist.contains('^letsencrypt01.opendev.org 5000')
+    # from extra list; may need to change if list is modified
+    assert domainlist.contains('^wiki.openstack.org 443')

zuul.d/system-config-run.yaml

@@ -189,6 +189,9 @@
         - playbooks/service-nameserver.yaml
         - playbooks/service-letsencrypt.yaml
     host-vars:
+      bridge.openstack.org:
+        host_copy_output:
+          '/var/lib/certcheck': logs
       letsencrypt01.opendev.org:
         host_copy_output:
           '/var/log/acme.sh': logs