The main purpose of this guide is to summarize all the steps one has to take to get Node.js and MongoDB up and running on a Ubuntu web server. It is not 100% complete but it covers hopefully the most things you need to know to run different Node.js applications behind nginx, using MongoDB as a database.
If you find anything you can't agree with, please let me know. Send me an email or just a pull request and I will see if I am able to fix this ;-).
This guide contains some details regarding security that won't necessarily matter if you only want to run a single website on your server. In that case take a look at "Setting up Ubuntu Server for running a single Website with Node.js and MongoDB".
The make tools are essential to build some npm packages and other stuff. So it’s generally a good idea to install them early.
sudo apt-get install gcc make build-essential
(If the installation of build-essential fails, see Misc/Missing packages section).
sudo apt-get install nginx
Once the setup of nginx is complete, you should be able to call http://{server_ip} and see the default page with the “Welcome to nginx!” headline.
Also make sure the server starts automatically after booting the system (Should be enabled by default):
sudo update-rc.d nginx defaults
wget -qO- https://raw.githubusercontent.com/creationix/nvm/v0.14.0/install.sh | bash
# Refresh Path
source ~/.profile
# Use latest Node.JS version
nvm install v0.11.13
# Make it default
nvm use default v0.11.13
sudo npm install bower -g
PM2 manages the different node applications running on this server.
sudo npm install pm2 -g
Glances can be used to monitor the overall state of the server.
sudo apt-get install python-pip build-essential python-dev
sudo pip install Glances
sudo apt-get install lm-sensors
sudo pip install PySensors
sudo apt-get install git
sudo apt-get install zip
For a detailed explanation see http://docs.mongodb.org/manual/tutorial/install-mongodb-on-ubuntu/.
sudo apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv 7F0CEB10
echo 'deb http://downloads-distro.mongodb.org/repo/ubuntu-upstart dist 10gen' | sudo tee /etc/apt/sources.list.d/mongodb.list
sudo apt-get update
sudo apt-get install -y mongodb-org
echo "mongodb-org hold" | sudo dpkg --set-selections
echo "mongodb-org-server hold" | sudo dpkg --set-selections
echo "mongodb-org-shell hold" | sudo dpkg --set-selections
echo "mongodb-org-mongos hold" | sudo dpkg --set-selections
echo "mongodb-org-tools hold" | sudo dpkg --set-selections
sudo vi /etc/mongod.conf
Set:
auth = true
bind_ip = 127.0.0.1
Open the Mongo Console with mongo.
use admin
db.createUser({ user: "<username>",
pwd: "<password>",
roles: [ "userAdminAnyDatabase",
"dbAdminAnyDatabase",
"readWriteAnyDatabase"
]})
sudo service mongod restart
mongo -u admin -p {adminpassword} --authenticationDatabase admin
Add a MongoDB user for the website
use website-com
db.createUser(
{
user: "website-com",
pwd: "{newpassword}",
roles: [
{ role: "readWrite", db: "website-com" }
]})
Websites are organized as follows:
The main identifier is the domain name of a website, while dots are replaced by dashes.
| Directory | Path |
|---|---|
| Single website’s root | /var/www/xyz-com |
| Single website’s Git repository | /var/www/xyz-com/repo |
| Single website’s web root | /var/www/xyz-com/www |
Examples:
- blog.xyz.de => blog-xyz-de
- website.com => website-com
Use the domain name for the user’s name:
sudo adduser website-com
In /var/www execute:
sudo chown website-com website-com –R
In /var/www/website-com/repo run
git init --bare
This hook is used to deploy changes made to the master repository. It can be customized for each website depending on the specific needs.
Go to /var/www/website-com/repo/hooks and create a new file called “post-receive”:
vi post-receive
Add the following commands to it:
#!/bin/bash
# Make this executable: chmod +x post-receive
PREPARATION_DIR="/var/www/website-com/repo/$(uuidgen)"
WEBSITE_ROOT="/var/www/website-com/www"
PM2_APP_NAME="website-com"
echo "Deployment started"
read oldrev newrev branch
if [[ $branch =~ .*/master$ ]];
then
echo "Master received. Deploying to production..."
# Creates a temporary working directory
mkdir $PREPARATION_DIR
# Checks out the master from the repository
GIT_WORK_TREE="$PREPARATION_DIR" git checkout -f
# Installing all npm and bower modules/packages
cd $PREPARATION_DIR
npm install --allow-root
bower install --allow-root
# Removes all files in the Website's root
cd $WEBSITE_ROOT
rm -rf *
# Copies all files over
cd $PREPARATION_DIR
cp -r . $WEBSITE_ROOT
# Restart the Website via PM2
pm2 restart $PM2_APP_NAME
# Removes the preparation directory
rm -R $PREPARATION_DIR
else
echo "$branch successfully received. Nothing to do: only the master branch may be deployed on this server."
fi
echo "Deployment finished"
Remember the value of PM2_APP_NAME and use it as an identifier for your pm2 application later.
After saving, make the script executable:
chmod +x post-receive
By cloning your deployment repository on your local development machine and pushing the first version to the master branch, you should receive a fresh version at /var/www/website-com/www (Check with ls –l).
git clone ssh://{username}@{ipaddress}/var/www/website-com/repo website-com
Make sure the user you’re connecting with has the necessary rights to run the Git repository. It’s recommended to connect with the user you have just created before to run the website, because he/she has the necessary access rights.
Call visudo and add the following line:
{username} ALL=(ALL:ALL) ALL
To start pm2 with the system:
pm2 startup ubuntu
PM2 will tell you, you have to run this command as root, and print the full command to execute, for example:
sudo env PATH=$PATH:/home/{username}/.nvm/v0.10.32/bin pm2 startup ubuntu -u {username}
Call visudo again and remove the line you did just add before.
cd /var/www/website-com/www/
pm2 start app.js --name "website-com"
Note: The PM2 instance is now running in the context if your website's user. This is especially useful to limit file access, but you can't see any other PM2 process or applications as you could see if running all websites under one single user.
If everything works PM2 reponds with Process {nameofstarting.js} launched. Wait a few seconds and use
pm2 list
for a fresh status update. For more information see
pm2 help
We’re using a single configuration, which can be found at:
sudo vi /etc/nginx/sites-available/default
nginx is running as a reverse proxy to handle all the public stuff on port 80 for us. It then passes all the traffic we want to to our node application.
server {
listen 80;
server_name your-domain.com;
location / {
proxy_pass http://localhost:{YOUR_PORT};
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection 'upgrade';
proxy_set_header Host $host;
proxy_cache_bypass $http_upgrade;
}
}
After saving the configuration, use
sudo service nginx reload
to tell the server it should use it.
As I am currently using a Hetzner server, I am also able to use 100 GB backup space which they offer free of charge. But anyway: If you're not hosting your server at Hetzner this may also apply to your actual hosting environment.
| Directory | Path |
|---|---|
| Backup Root | /var/backup |
| Website Backups | /var/backup/www |
| MongoDB Backups | /var/backup/mongo |
| nginx Backups | /var/backup/nginx |
There needs to be a user with a backup role to create backups of all databases:
use admin
db.createUser(
{
user: "backup”,
pwd: "{newpassword}",
roles: [{ role: "backup", db: "admin" }]
}
)
Save the following shell script as /var/backup/create-backup-for-mongo and make it executable:
#!/bin/bash
echo "Mongo Backup started"
ADMIN_USERNAME="backup"
ADMIN_PASSWORD="{password}"
BACKUP_TARGET_ROOT="/var/backup/mongo"
CURRENT_BACKUP_TARGET="$BACKUP_TARGET_ROOT/$(uuidgen)"
# Remove all but the latest 30 backups
cd $BACKUP_TARGET_ROOT
rm -rf `ls -t | tail -n +30`
# Back up all the databases to a new directory
mongodump -u $ADMIN_USERNAME -p $ADMIN_PASSWORD -o $CURRENT_BACKUP_TARGET --authenticationDatabase admin
zip -r "$(uuidgen).zip" $CURRENT_BACKUP_TARGET
rm -rf $CURRENT_BACKUP_TARGET
echo "Mongo Backup finished"
Create the script /var/backup/create-backup-for-www and make it executable:
#!/bin/bash
echo "WWW Backup started"
BACKUP_SOURCE="/var/www"
BACKUP_TARGET_ROOT="/var/backup/www"
CURRENT_BACKUP_TARGET="$BACKUP_TARGET_ROOT/$(uuidgen)"
cd $BACKUP_TARGET_ROOT
rm -rf `ls -t | tail -n +30`
# Back up all the websites to a new directory
rsync -a -E -c --stats $BACKUP_SOURCE $CURRENT_BACKUP_TARGET
zip -r "$(uuidgen).zip" $CURRENT_BACKUP_TARGET
rm -rf $CURRENT_BACKUP_TARGET
echo "WWW Backup finished"
Create the script /var/backup/create-backup-for-nginx and make it executable:
#!/bin/bash
echo "nginx Backup started"
BACKUP_SOURCE="/etc/nginx"
BACKUP_TARGET_ROOT="/var/backup/nginx"
CURRENT_BACKUP_TARGET="$BACKUP_TARGET_ROOT/$(uuidgen)"
# Remove all but the latest 30 backups
cd $BACKUP_TARGET_ROOT
rm -rf `ls -t | tail -n +30`
rsync -a -E -c --stats $BACKUP_SOURCE $CURRENT_BACKUP_TARGET
zip -r "$(uuidgen).zip" $CURRENT_BACKUP_TARGET
rm -rf $CURRENT_BACKUP_TARGET
echo "nginx Backup finished"
sudo apt-get install cifs-utils
This will mount the backup server as a local drive and it will reconnect on each reboot (it’s a single line added to fstab).
sudo vi /etc/fstab
//u100445.your-backup.de/backup /mnt/backup-server cifs iocharset=utf8,rw,credentials=/var/backup/backup-credentials.txt,uid={localusername},gid={localusergroupid},file_mode=0660,dir_mode=0770 0 0
(If you are not aware of the group id, just type id and it will be displayed for the current user (in case you want to mount the drive in his/her context)).
vi /var/backup/backup-credentials.txt
Set:
username={username}
password={password}
Create a script that combines all backup actions and that finally transfers everything from the current backup folder to the backup server. Save that script as /var/backup/create-and-transfer-backups and make it executable.
#!/bin/bash
echo "Global Backup started"
#Important: use absolute paths to be independent of the user context
/var/backup/create-backup-for-mongo
/var/backup/create-backup-for-www
/var/backup/create-backup-for-nginx
rsync -avzE --delete --stats /var/backup /mnt/backup-server
echo "Global Backup finished"
sudo vi /etc/crontab
Set:
# m h dom mon dow user command
10 14 * * * root bash /var/backup/create-and-transfer-backups
(Runs the backup every day at 2:10 pm.)
| Action | Command |
|---|---|
| Change User name | passwd |
| Grant root priviliges | visudo, add: {username} ALL=(ALL:ALL) ALL |
| System UpdateS: Fetching the list of available updates | sudo apt-get update |
| System UpdateS: Strictly upgrading the current packages | sudo apt-get upgrade |
| System UpdateS: Installing updates (new ones) | sudo apt-get dist-upgrade |
| Reboot | sudo reboot |
| Remove a directory and its sub-directories recursively | rm –R {directoryname} |
| Remove everything, files and directories | sudo rm -rf * |
| Show current network settings | ifconfig |
| Make rudi the owner of a directory | chown -R rudi {directoryname} |
| Start a system service | sudo service {servicename} start |
| Restart a system service | sudo service {servicename} restart |
| Stop a system service | sudo service {servicename} stop |
| Reload a system service | sudo service {servicename} reload |
| See who’s the owner of a process | ps aux |
| chmod Details | http://serverfault.com/a/357109 & http://linuxcommand.org/lts0070.php |
| Grant a Role in MongoDB afterwards | db.grantRolesToUser("username", [{ role: "readWrite", db: "dbname" }]) |
| Make a shell script executable | chmod +x {scriptname} |
Create (or edit) a file named config in ~/.ssh/ (User's root directory)
Host username
HostName xxx.xxx.xxx.xxx
User username
Now just type ssh username in terminal.
If some packages are not available via apt-get, take a look at the sources configuration:
sudo vi /etc/apt/sources.list
If the extras and the archive repositories are commented out, just remove the comments.