ztncui is a web user interface for a standalone ZeroTier network controller.
Screenshots can be seen at key-networks.com/ztncui.
Instructions for installing on Linux from RPM or DEB packges are available at key-networks.com/ztncui.
See github.com/key-networks/ztncui-containerized
Relative directory references below are relative to the cloned ztncui directory.
-
ztncui is a node.js Express application that requires node.js v14.
-
ztncui uses argon2 for password hashing. Argon2 needs the following:
- g++
- node-gyp, which can be installed with:
sudo npm install -g node-gyp
-
ztncui requires ZeroTier One to be installed on the same machine. This will run as the network controller to establish ZeroTier networks.
-
ztncui has been developed on a Linux platform and expects the ZT home directory to be in
/var/lib/zerotier-one
.
git clone https://github.com/key-networks/ztncui
2. Install the node.js packages:
cd ztncui/src
npm install
The app needs to know the zerotier-one authtoken.secret. There are two options:
The user running the ztncui app needs read access to authtoken.secret. This can be achieved with:
sudo usermod -aG zerotier-one username
sudo chmod g+r /var/lib/zerotier-one/authtoken.secret
Where:
- username is the user running the ztncui app
Note that you need to log out and in again to apply the new group membership.
In the root of the ztncui directory, create a .env
file with the content:
ZT_TOKEN=########################
Where:
- ######################## is the token string.
You can also specify in the .env
file a different address for the zerotier-one API (which defaults to localhost:9993):
ZT_ADDR=localhost:9995
Make .env
readable by the user running ztncui only:
chmod 600 .env
The .env
file should make it easier to run ztncui on a non-Linux platform.
To run the server in production mode, add the following to the .env
file (see 3B above):
NODE_ENV=production
Without this, the template engine always re-compiles the pug file when rendering (taking ~200 ms!)
To prevent git from over-writing your password file every time you pull updates from the repository, the etc/passwd file has been added to .gitignore. So you need to copy the default file after the first time you do a git clone. All these things ideally need to be done with a package installer script:
cp -v etc/default.passwd etc/passwd
npm start
This will run the app on TCP port 3000 by default. If port 3000 is already in use, you can specify a different port in the .env
file (see 3B above), e.g.:
HTTP_PORT=3456
To start the app automatically, something like PM2 can be used. Install it with:
sudo npm install -g pm2
Add ztncui as a managed app with:
pm2 start bin/www --name ztncui
To detect the init system:
pm2 startup
PM2 will then give you a command to execute to configure automatic startup of PM2 for your system.
Save the current PM2 process list so that ztncui will restart across reboots:
pm2 save
8. Test access on http://localhost:3000
If the machine has a GUI and GUI web browser, then use it to access the app, otherwise use a text web browser like Lynx or a CLI web browser like curl:
curl http://localhost:3000
You should see the front page of the app (or the raw HTML with curl).
This app listens for HTTP requests on the looback interface (default port 3000). It can be reverse proxied by Nginx (which can proxy the HTTP as HTTPS), or accessed over an SSH tunnel as described below.
The app can be made to listen on all interfaces for HTTP requests by setting HTTP_ALL_INTERFACES in the .env
file, e.g.:
HTTP_ALL_INTERFACES=yes
Note that HTTP traffic is unencrypted, so this should only be done on a secure network, otherwise usernames and passwords will be exposed in plain text over the network.
The app can be made to listen on all interfaces for HTTPS requests by specifying HTTPS_PORT in the .env
file, e.g.:
HTTPS_PORT=3443
The app can be made to listen on a specific interface for HTTPS requests by specifying HTTPS_HOST (the host name or IP address of the interface) in the .env
file, e.g.:
HTTPS_HOST=12.34.56.78
If HTTPS_HOST is not specified, but HTTPS_PORT is specified, then the app will listen for HTTPS requests on all interfaces.
Environment variable | Protocol | Listen On | Port |
---|---|---|---|
[none] | HTTP | localhost | 3000 |
HTTP_PORT | HTTP | localhost | HTTP_PORT |
HTTP_ALL_INTERFACES | HTTP | all interfaces | HTTP_PORT |
HTTPS_PORT | HTTPS | all interfaces | HTTPS_PORT |
HTTPS_HOST | HTTPS | HTTPS_HOST | HTTPS_PORT |
For HTTPS you obviously need a TLS (SSL) certificate and private key pair. There are a few options:
-
By default, if there is no existing TLS certificate and private key pair, the RPM and DEB packages automatically generate a self-signed certificate / private key pair.
-
If you are running directly from source, then generate a self-signed certificate as follows:
cd etc/tls openssl req -x509 -sha256 -nodes -days 365 -newkey rsa:2048 -keyout privkey.pem -out fullchain.pem
Fill in the required details as prompted.
The advantage of this option is that it is quick and easy to generate the certificate / private key pair. The disadvantage is that your web browser will give you a warning that it cannot verify the certificate. You can override this warning and make a temporary exception.
-
Buy a certificate:
You will need to store the private key as
etc/tls/privkey.pem
and the full certificate chain asetc/tls/fullchain.pem
. They need to be in PEM format. -
Get a free certificate from Letsencrypt.org:
a. Install certbot by following the instructions at certbot.eff.org:
i. For "Software" select "None of the above". ii. For "System" select your OS. iii. Follow the instructions to install certbot on your system.
b. Use certbot to generate a certificate in webroot mode from the root of the ztncui directory:
certbot --webroot -w public -d [network_controller_fqdn]
Where [network_controller_fqdn] is the FQDN that resolves back to the address of the machine running the ZeroTier network controller and ztncui.
If certbot runs successfully, it should give you the location of your certificate, which should be something like:
/etc/letsencrypt/live/[network_controller_fqdn]/fullchain.pem
c. Make soft links from etc/tls to the certificate and private key under /etc/letsencrypt/live:
cd etc/tls ln -s /etc/letsencrypt/live/[network_controller_fqdn]/fullchain.pem ln -s /etc/letsencrypt/live/[network_controller_fqdn]/privkey.pem
d. Take note of the options for renewing Letsencrypt certificates and implement an appropriate strategy.
Once you have a certificate at etc/tls/fullchain.pem
and private key at etc/tls/privkey.pem
, you should be able to access ztncui over HTTPS on the port specified by HTTPS_PORT.
An SSH tunnel can be established with:
ssh -f [email protected] -L 3333:localhost:3000 -N
where:
- network.controller.machine is the FQDN of the machine running the ZT network controller and ztncui, and
- user is any user account that you have on that machine.
Once the SSH tunnel has been established, access the ztncui web interface in a web browser on your local machine at: http://localhost:3333
On Windows you can install PuTTY
Open PuTTY and configure as follows:
- Go to Connection -> SSH -> Tunnels.
- Set Source port to 3333
- Set Destination to localhost:3000
- Click on the Add button.
- Go to Session in the Category panel on the left.
- Set Host Name (or IP address) to the FQDN of the machine running the ZT network controller and ztncui.
- Enter a name for the configuration in Saved Sessions and click Save.
- Click the Open button and log into the network controller machine.
Once the SSH tunnel has been established, access the ztncui web interface in a web browser on your local machine at: http://localhost:3333
Once you have access to the web UI of ztncui, log in as user admin with password password.
You will be prompted to change the default password.
It's a good idea to create your own username and delete the default admin account. You can do this by clicking on the Users tab and then the Create user tab. Note that you then have to log out and log in as the new user before you can delete the default admin account.
Click on the Home tab to get to the network controller home page. From there you can click on the Networks tab to see the existing networks configured on the network controller (probably none if you have just set it up).
Click on the Add network tab to create a new ZeroTier network that is controlled by the network controller. Give it a name and click Create Network. You will then be taken back to the Networks page that lists all the networks on the controller.
On the Networks page, click the trash can icon to delete a network. You will be warned that this action cannot be undone. Click the Delete button to confirm the action.
On the Networks page, click the name of the network to rename it.
On the Networks page, click easy setup for the network that you want to auto-configure. Click Generate network address to assign a random network address, or manually enter the network address in CIDR notation. The start and end of the IP assignment pool will be automatically calculated, but these can be manually adjusted. Click Submit to apply the configuration. You should then get a notice that the network setup succeeded.
Note that the easy setup only works for IPv4 at this stage. To set up IPv6, follow the detail link for a network from the Networks page and set up each property manually.
Invite users to join the network with:
sudo zerotier-cli join ################
where ################ is the 16-digit ZeroTier network ID.
Get the user to send you their 10-digit ZeroTier address, which they can get by running:
sudo zerotier-cli status
On the Networks page, click members to see the devices which are trying to join the network. Use the ZeroTier address given to you by the user to identify them and name them appropriately under Member name.
Then check the Authorized checkbox to authorize the user on the network.
If the user's device is online and you click the Refresh button, you should see their IP assignment being populated.
Once two or more members are authorized on the network, they should be able to connect to each other via their assigned IP addresses.
IP assignments can be changed by clicking on the IP address in the members page. Enter an IP address in the managed route subnet and click the + icon. Then delete the old IP address.
Ethernet bridging between virtual and physical networks can be enabled by checking the Active bridge checkbox on the members page.
On the Networks page, click detail to see the detail of a network.
Note that certain properties can be set by clicking on them - e.g.:
- ipAssignmentPools
- name
- routes
- v4AssignMode
- v6AssignMode
Note that editing of certain properties, such as rules and tags, has not been implemented yet. Please feed back on your requirements.
On the network detail page and on the members page, if you click on the member ID, you will end up on the member detail page.
Please give us your feedback... good, bad or ugly. Constructive criticism is welcomed. Please use the contact form at key-networks.com - Thanks :)
Problems with ztncui can be reported using the GitHub issue tracking system. Please use the contact form at key-networks.com to privately report potential vulnerabilities. Thank you.
The ztncui code is open source code, licensed under the GNU GPLv3, and is free to use on those terms. If you are interested in commercial licensing, please contact us via the contact form at key-networks.com .
@lideming for a rework and improvement of the network details page, adding DNS support, peer status/address/latency and other improvements.