How to run nginx.service after /vagrant is mounted

问题:

What I want to do?

I'm trying to make nginx load configurations from /vagrant mounted by vagrant automatically.

So I edited nginx.service to make it boot after shared folder mounted, but it not works.

Certainly nginx has been booted after virtualbox-guest-utils.service, however, it seems to be booted before vagrant.mount (/vagrant). Because nginx couldn't load configurations from /vagrant and it works after running command systemctl restart nginx.service manually.

How to run .service unit after auto generated .mount unit started?

Environment

• Vagrant 1.8.1
• Ubuntu Server 15.10 (ubuntu/wily64)
• VirtualBox 5.0.14

systemctl cat nginx.service

Pattern 1 : Not work

# /lib/systemd/system/nginx.service
# Stop dance for nginx
# =======================
#
# ExecStop sends SIGSTOP (graceful stop) to the nginx process.
# If, after 5s (--retry QUIT/5) nginx is still running, systemd takes control
# and sends SIGTERM (fast shutdown) to the main process.
# After another 5s (TimeoutStopSec=5), and if nginx is alive, systemd sends
# SIGKILL to all the remaining processes in the process group (KillMode=mixed).
#
# nginx signals reference doc:
# http://nginx.org/en/docs/control.html
#
[Unit]
Description=A high performance web server and a reverse proxy server
After=network.target

[Service]
Type=forking
PIDFile=/run/nginx.pid
ExecStartPre=/usr/sbin/nginx -t -q -g 'daemon on; master_process on;'
ExecStart=/usr/sbin/nginx -g 'daemon on; master_process on;'
ExecReload=/usr/sbin/nginx -g 'daemon on; master_process on;' -s reload
ExecStop=-/sbin/start-stop-daemon --quiet --stop --retry QUIT/5 --pidfile /run/nginx.pid
TimeoutStopSec=5
KillMode=mixed

[Install]
WantedBy=multi-user.target

# /etc/systemd/system/nginx.service.d/override.conf
[Unit]
Requires=virtualbox-guest-utils.service vagrant.mount
After=virtualbox-guest-utils.service vagrant.mount

Pattern 2 : Not work

# /lib/systemd/system/nginx.service
# Stop dance for nginx
# =======================
#
# ExecStop sends SIGSTOP (graceful stop) to the nginx process.
# If, after 5s (--retry QUIT/5) nginx is still running, systemd takes control
# and sends SIGTERM (fast shutdown) to the main process.
# After another 5s (TimeoutStopSec=5), and if nginx is alive, systemd sends
# SIGKILL to all the remaining processes in the process group (KillMode=mixed).
#
# nginx signals reference doc:
# http://nginx.org/en/docs/control.html
#
[Unit]
Description=A high performance web server and a reverse proxy server
After=network.target

[Service]
Type=forking
PIDFile=/run/nginx.pid
ExecStartPre=/usr/sbin/nginx -t -q -g 'daemon on; master_process on;'
ExecStart=/usr/sbin/nginx -g 'daemon on; master_process on;'
ExecReload=/usr/sbin/nginx -g 'daemon on; master_process on;' -s reload
ExecStop=-/sbin/start-stop-daemon --quiet --stop --retry QUIT/5 --pidfile /run/nginx.pid
TimeoutStopSec=5
KillMode=mixed

[Install]
WantedBy=multi-user.target

# /etc/systemd/system/nginx.service.d/override.conf
[Unit]
RequiresMountsFor=/vagrant

回答1:

I wrote a post how I'm using udev event to restart nginx and php5-fpm.

In short I'm adding /etc/udev/rules.d/50-vagrant-mount.rules with rule:

SUBSYSTEM=="bdi",ACTION=="add",RUN+="/bin/bash /root/.udev-mount-restart-services.sh"

And the script /root/.udev-mount-restart-services.sh is:

sleep 5 # wait for a bit for NFS to make sure resources are available
systemctl restart php5-fpm > /dev/null 2>&1
systemctl restart nginx > /dev/null 2>&1

回答2:

Using the mount.target unit seems to work.

Vagrantfile:

# -*- mode: ruby -*-
# vi: set ft=ruby :

Vagrant.configure("2") do |config|
  config.vm.box = "ubuntu/wily64"
  config.vm.provision :shell, path: "bootstrap.sh"

  # Disabling mounting of the /vagrant directory to make sure ngnix is blocked
  # Comment this out to allow the shared folder
  config.vm.synced_folder ".", "/vagrant", disabled: true
end

bootstrap.sh:

#!/usr/bin/env bash

apt-get update
apt-get install -y nginx

# Make vagrant.mount loosely dependent on nginx.service
sed -i 's/WantedBy=multi-user.target/WantedBy=vagrant.mount/' /lib/systemd/system/nginx.service
systemctl daemon-reload

# Update service symlinks
systemctl disable nginx.service
systemctl enable nginx.service

Test this by running vagrant reload to cause the VM to boot without the /vagrant directory mounted. Log into the VM and see that nginx isn't running:

vagrant@vagrant-ubuntu-wily-64:~$ sudo systemctl status nginx.service 
● nginx.service - A high performance web server and a reverse proxy server
   Loaded: loaded (/lib/systemd/system/nginx.service; enabled; vendor preset: enabled)
   Active: inactive (dead)

Then comment out the Vagrantfile line that disables the shared folder and vagrant reload again. Log into the VM again and see that nginx is running:

vagrant@vagrant-ubuntu-wily-64:~$ systemctl status nginx.service 
● nginx.service - A high performance web server and a reverse proxy server
  Loaded: loaded (/lib/systemd/system/nginx.service; enabled; vendor preset: enabled)
  Active: active (running) since Mon 2016-07-25 04:28:00 UTC; 42s ago
 Process: 1152 ExecStart=/usr/sbin/nginx -g daemon on; master_process on; (code=exited, status=0/SUCCESS)
 Process: 1146 ExecStartPre=/usr/sbin/nginx -t -q -g daemon on; master_process on; (code=exited, status=0/SUCCESS)
Main PID: 1156 (nginx)
  Memory: 7.4M
     CPU: 17ms
  CGroup: /system.slice/nginx.service
          ├─1156 nginx: master process /usr/sbin/nginx -g daemon on; master_process on;
      └─1157 nginx: worker process

回答3:

edit /lib/systemd/system/nginx.service and change WantedBy directive under Install section...

WantedBy=vagrant.mount

You can check if the vagrant.mount event exists via: systemctl list-units vagrant.mount. You should see this:

UNIT          LOAD   ACTIVE SUB     DESCRIPTION
vagrant.mount loaded active mounted /vagrant

LOAD   = Reflects whether the unit definition was properly loaded.
ACTIVE = The high-level unit activation state, i.e. generalization of SUB.
SUB    = The low-level unit activation state, values depend on unit type.

1 loaded units listed. Pass --all to see loaded but inactive units, too.
To show all installed unit files use 'systemctl list-unit-files'.

回答4:

Summarization previous answers and comments in terms of Ansible (if used for provision):

- name: Copy Nginx Systemd unit to /etc/
  copy:
    src: /lib/systemd/system/nginx.service
    dest: /etc/systemd/system/nginx.service
    remote_src: yes

- name: Change Nginx Systemd unit
  lineinfile:
    path: /etc/systemd/system/nginx.service
    regexp: '^WantedBy='
    line: 'WantedBy={{ vagrant_mount_unit }}'
    state: present

- name: Systemd daemon reload
  systemd:
    daemon_reload: yes

- name: Disable Nginx
  service:
    name: nginx
    enabled: no

- name: Enable Nginx
  service:
    name: nginx
    enabled: yes

Replace {{ vagrant_mount_unit }} by Vagrant mount unit name. It depends from mount point, e.g. if mount point is /home/vagrant/app, unit name would be home-vagrant-app.mount.

回答5:

I wanted to drop my own answer here as it did seem each answer has parts of a correct answer and some details can be found in comments only.

In this example I have 2 file system mounts in /var/www/mymount1 and /var/www/mymount2

This part of my Vagrantfile:

config.vm.provision "shell", path: "docs/vagrant/init-vagrant.sh"
config.vm.synced_folder "mymount1", "/var/www/mymount1", type: "nfs"
config.vm.synced_folder "mymount2", "/var/www/mymount2", type: "nfs"

I am placing the script below in /docs/vagrant/init-vagrant.sh as its my provisioner file.

# Copy nginx systemd script and make edits in /etc/systemd/system, as updates might change the original in /lib/systemd/system/nginx.service
cp /lib/systemd/system/nginx.service /etc/systemd/system/
# These tell that nginx should not be started before file system mounts are available
echo WantedBy=var-www-mymount1.mount >> /etc/systemd/system/nginx.service
echo WantedBy=var-www-mymount2.mount >> /etc/systemd/system/nginx.service

# Update service symlinks
systemctl daemon-reload
systemctl disable nginx.service
systemctl enable nginx.service