impress-2020/deploy/playbooks/setup.yml
Matchu d17263139e Fix pm2 monitoring
Okay huh, while digging a bit into another issue, I found what was wrong with our config and pm2's built-in monitoring! You can't use `yarn start`, because the wrapper script breaks its ability to look inside and see what's happening.

I also removed the compiler flag thing from the `start` script in `package.json`, because I think it's redundant? There's no compilation to be done in a live server.

I think I might remove monit after this? It's nice extra resilience in a sense, but it feels like extra complexity when it's doing the job `pm2` is supposed to do. (And tbh I've almost never heard of nginx crashing, and if it does it's probably a scenario worth investigating by hand.)
2021-11-03 16:46:35 -07:00

243 lines
7.7 KiB
YAML

---
- name: Set up the environment for the impress-2020 app
hosts: webserver
vars:
email_address: "emi@matchu.dev" # TODO: Extract this to personal config?
tasks:
- name: Create web user group
become: yes
group:
name: web
- name: Add current user to web group
become: yes
user:
name: "{{ ansible_user_id }}"
group: web
append: yes
- name: Create the app folder
become: yes
file:
path: /srv/impress-2020
state: directory
# Root and the `web` group may read/write this folder. Everyone else
# may only read it.
group: web
mode: "u=rwx,g=rwx,o=rx"
- name: Add Nodesource apt key
become: yes
apt_key:
id: 9FD3B784BC1C6FC31A8A0A1C1655A0AB68576280
url: https://deb.nodesource.com/gpgkey/nodesource.gpg.key
- name: Add Node v16 apt repository
become: yes
apt_repository:
repo: deb https://deb.nodesource.com/node_16.x focal main
- name: Install Node v16
become: yes
apt:
update_cache: yes
name: nodejs
state: present
- name: Install Yarn
become: yes
npm:
name: yarn
global: yes
- name: Install pm2
become: yes
npm:
name: pm2
global: yes
- name: Create pm2 startup script
# The current user is going to become the pm2 owner of the app server
# process. They'll be able to manage it without `sudo`, including during
# normal deploys, and run `pm2 monit` from their shell to see status.
become: yes
command: "pm2 startup systemd {{ ansible_user_id }} --hp /home/{{ ansible_user_id }}"
- name: Create pm2 ecosystem file
copy:
content: |
module.exports = {
apps: [
{
name: "impress-2020",
cwd: "/srv/impress-2020/current",
// Instead of `yarn start`, we specify the `next` binary
// directly, because it helps pm2 monitor our app correctly.
// https://github.com/vercel/next.js/discussions/10675#discussioncomment-34615
script: "./node_modules/.bin/next",
args: "start --port=3000",
instances: "max",
exec_mode: "cluster",
// We add `app` to the end of the filename, to avoid a pm2
// bug that changes the filename:
// https://github.com/Unitech/pm2/issues/5218#issue-1044210369
pid_file: "/home/{{ ansible_user_id }}/impress-2020-app.pid",
}
]
}
dest: "~/ecosystem.config.js"
# Create a temporary backup file, so we can use it to delete the old
# version of the services. (This is important if e.g. a service is
# removed or renamed, in which case deleting from the *new* config file
# wouldn't include it.)
backup: yes
register: pm2_ecosystem_file
- name: Delete old pm2 services if config file changed
command: "pm2 delete {{ pm2_ecosystem_file.backup_file | quote }}"
when: pm2_ecosystem_file is changed and pm2_ecosystem_file.backup_file is defined
- name: Delete old pm2 config file if it changed
file:
path: "{{ pm2_ecosystem_file.backup_file }}"
state: absent
when: pm2_ecosystem_file is changed and pm2_ecosystem_file.backup_file is defined
- name: Start pm2 services
command: "pm2 start ~/ecosystem.config.js"
- name: Save pm2 startup script
command: pm2 save
- name: Install core snap
become: yes
community.general.snap:
name: core
- name: Install certbot as a snap
become: yes
community.general.snap:
name: certbot
classic: yes
- name: Set up certbot
become: yes
command: "certbot certonly --nginx -n --agree-tos --email {{ email_address }} --domains impress-2020-box.openneo.net"
- name: Install nginx
become: yes
apt:
update_cache: yes
name: nginx
- name: Add impress-2020 config file to nginx
become: yes
copy:
content: |
server {
server_name impress-2020-box.openneo.net;
listen 80;
if ($host = impress-2020-box.openneo.net) {
return 301 https://$host$request_uri;
}
}
server {
server_name impress-2020-box.openneo.net;
listen 443 ssl;
ssl_certificate /etc/letsencrypt/live/impress-2020-box.openneo.net/fullchain.pem;
ssl_certificate_key /etc/letsencrypt/live/impress-2020-box.openneo.net/privkey.pem;
include /etc/letsencrypt/options-ssl-nginx.conf;
ssl_dhparam /etc/letsencrypt/ssl-dhparams.pem;
ssl_session_cache shared:SSL:10m; # https://superuser.com/q/1484466/14127
# TODO: Serve static files directly, instead of through the proxy
location / {
proxy_pass http://127.0.0.1:3000;
}
}
dest: /etc/nginx/sites-enabled/impress-2020
notify:
- Restart nginx
- name: Install monit
become: yes
apt:
update_cache: yes
name: monit
- name: Add monit config file for global settings
become: yes
copy:
content: |
# TODO: Add email monitoring (requires SMTP config)
# This lets us call `sudo monit status` from the command line.
# Without this, the `monit` command can't find the running service.
set pidfile /var/run/monit.pid
# This enables Monit's HTTP server, but only locally, which is
# required for calling `sudo monit status` from the command line.
set httpd port 2812 and
use address localhost
allow localhost
dest: /etc/monit/conf-enabled/global-config
notify:
- Restart monit
- name: Add monit config to watch our pm2 app
become: yes
copy:
content: |
check process impress-2020-as-{{ ansible_user_id }} with pidfile /home/{{ ansible_user_id }}/impress-2020-app-0.pid
start program = "/bin/pm2 start impress-2020" as uid "{{ ansible_user_id }}"
stop program = "/bin/pm2 stop impress-2020" as uid "{{ ansible_user_id }}"
restart program = "/bin/pm2 reload impress-2020" as uid "{{ ansible_user_id }}"
if failed port 3000 protocol http then restart
if 5 restarts within 5 cycles then alert
dest: "/etc/monit/conf-enabled/pm2-as-{{ ansible_user_id }}"
notify:
- Reload monit
- name: Add monit config to watch nginx
become: yes
copy:
content: |
check process nginx with pidfile /var/run/nginx.pid
start program = "/bin/systemctl start nginx"
stop program = "/bin/systemctl stop nginx"
restart program = "/bin/systemctl restart nginx"
if failed host impress-2020-box.openneo.net port 443 type tcpssl protocol http then restart
if 5 restarts within 5 cycles then alert
dest: "/etc/monit/conf-enabled/nginx"
notify:
- Reload monit
- name: Install dependencies for the npm module node-canvas
become: yes
apt:
update_cache: yes
name:
- build-essential
- libcairo2-dev
- libpango1.0-dev
- libjpeg-dev
- libgif-dev
- librsvg2-dev
handlers:
- name: Restart nginx
become: yes
systemd:
name: nginx
state: restarted
- name: Restart monit
become: yes
systemd:
name: monit
state: restarted
- name: Reload monit
become: yes
systemd:
name: monit
state: reloaded