Teleport Enterprise Quick Start

Welcome to the Quick Start Guide for Teleport Enterprise.

The goal of this document is to show off the basic capabilities of Teleport. There are three types of services Teleport nodes can run: nodes, proxies and auth servers.

• Auth servers store user accounts and provide authentication and authorization services for every node and every user in a cluster.
• Proxy servers route client connection requests to the appropriate node and serve a Web UI which can also be used to log into SSH nodes. Every client-to-node connection in Teleport must be routed via a proxy.
• Nodes are regular SSH servers, similar to the sshd daemon you may be familiar with. When a node receives a connection request, the request is authenticated through the cluster's auth server.

The teleport daemon runs all three of these services by default. This Quick Start Guide will be using this default behavior to create a cluster and interact with it using Teleport's client-side tools:

Prerequisites

You will need to have access to the customer portal to download the software. You will also need three computers: two servers and one client (probably a laptop) to complete this tutorial. Let's assume the servers have the following DNS names and IPs:

This Quick Start Guide assumes that both servers are running a systemd-based Linux distribution such as Debian, Ubuntu or a RHEL derivative.

Optional: Quickstart using Docker

The instructions below describe how to install Teleport Enterprise directly onto your test system. You can also run Teleport Enterprise using Docker if you don't want to install Teleport Enterprise binaries straight away.

Installing

To start using Teleport Enterprise, you will need to Download the binaries and the license file from the customer portal. After downloading the binary tarball, run:

$ tar -xzf teleport-ent-v5.0.0-rc.2-linux-amd64-bin.tar.gz
$ cd teleport-ent

• Copy teleport and tctl binaries to a bin directory (we suggest /usr/local/bin) on the auth server.
• Copy teleport binary to a bin directory on the node server.
• Copy tsh binary to a bin directory on the client computer.

License File

The Teleport license file contains a X.509 certificate and the corresponding private key in PEM format.

Download the license file from the customer portal and save it as /var/lib/teleport/license.pem on the auth server.

Configuration File

Save the following configuration file as /etc/teleport.yaml on the node.example.com:

teleport:
  auth_token: dogs-are-much-nicer-than-cats
  # you can also use auth server's IP, i.e. "10.1.1.10:3025"
  auth_servers: [ "auth.example.com:3025" ]

  # enable ssh service and disable auth and proxy:
ssh_service:
  enabled: true

auth_service:
  enabled: false
proxy_service:
  enabled: false

Now, save the following configuration file as /etc/teleport.yaml on the auth.example.com:

teleport:
  auth_token: dogs-are-much-nicer-than-cats
  auth_servers: [ "localhost:3025" ]

auth_service:
  # enable the auth service:
  enabled: true

  tokens:
  # this static token is used for other nodes to join this Teleport cluster
  - proxy,node:dogs-are-much-nicer-than-cats
  # this token is used to establish trust with other Teleport clusters
  - trusted_cluster:trains-are-superior-to-cars

  # by default, local authentication will be used with 2FA
  authentication:
      second_factor: otp

  # SSH is also enabled on this node:
ssh_service:
  enabled: "yes"

Systemd Unit File

Next, download the systemd service unit file from examples directory on Github and save it as /etc/systemd/system/teleport.service on both servers.

# run this on both servers:
$ sudo systemctl daemon-reload
$ sudo systemctl enable teleport

Starting

# run this on both servers:
$ sudo systemctl start teleport

Teleport daemon should start and you can use netstat -lptne to make sure that it's listening on TCP/IP ports. On auth.example.com, it should look something like this:

$ auth.example.com ~: sudo netstat -lptne
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address   State       User       PID/Program name
tcp6       0      0 :::3024         LISTEN      0          337/teleport
tcp6       0      0 :::3025         LISTEN      0          337/teleport
tcp6       0      0 :::3080         LISTEN      0          337/teleport
tcp6       0      0 :::3022         LISTEN      0          337/teleport
tcp6       0      0 :::3023         LISTEN      0          337/teleport

and node.example.com should look something like this:

$ node.example.com ~: sudo netstat -lptne
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address   State       User       PID/Program name
tcp6       0      0 :::3022         LISTEN      0          337/teleport

See troubleshooting section at the bottom if something is not working.

Adding Users

This portion of the Quick Start Guide should be performed on the auth server, i.e. on auth.example.com

Every user in a Teleport cluster must be assigned at least one role. By default, Teleport comes with one pre-configured role called "admin". You can see it's definition by executing sudo tctl get roles/admin > admin-role.yaml.

The output will look like this (re-formatted here to use compact YAML representation for brevity):

kind: role
version: v3
metadata:
  name: admin
spec:
  options:
    cert_format: standard
    forward_agent: true
    max_session_ttl: 30h0m0s
    port_forwarding: true
  # allow rules:
  allow:
    logins:
    - '{{internal.logins}}'
    - root
    node_labels:
      '*': '*'
    rules:
    - resources: [role]
      verbs: [list, create, read, update, delete]
    - resources: [auth_connector]
      verbs: [list, create, read, update, delete]
    - resources: [session]
      verbs: [list, read]
    - resources: [trusted_cluster]
      verbs: [list, create, read, update, delete]
  # no deny rules are present, the admin role must have access to everything)
  deny: {}

Pay attention to the allow/logins field in the role definition: by default, this role only allows SSH logins as [email protected].

You probably want to replace "root" with something else. Let's assume there will be a local UNIX account called "admin" on all hosts. In this case you can dump the role definition YAML into admin-role.yaml file and update "allow/logins" to look like this:

allow:
   logins: [admin]

Then send it back into Teleport:

$ sudo tctl create -f admin-role.yaml

Now, lets create a new Teleport user "joe" with "admin" role:

$ sudo tctl users add --roles=admin joe

Signup token has been created and is valid for 1 hours. Share this URL with the user:
https://auth.example.com:3080/web/newuser/22e3acb6a0c2cde22f13bdc879ff9d2a

Share the generated sign-up URL with Joe and let him pick a password and configure the second factor authentication. We recommend Google Authenticator which is available for both Android and iPhone.

Assigning Roles

To update user's roles, dump the user resource into a file:

$ sudo tctl get users/joe > joe.yaml

Edit the YAML file and update the "roles" array. Then, re-insert it back:

$ sudo tctl create -f joe.yaml

Logging In

Joe now has a local account on a Teleport cluster. The local account is good for administrative purposes but regular users of Teleport Enterprise should be using a Single Sign-On (SSO) mechanism.

But first, lets see how Joe can log into the Teleport cluster. He can do this on his client laptop:

$ tsh --proxy=auth.example.com --insecure login --user=joe

Note that "--user=joe" part can be omitted if $USER environment variable is "joe".

Notice that tsh client always needs --proxy flag because all client connections in Teleport always must to go through an SSH proxy, sometimes called an "SSH bastion".

If successful, tsh login command will receive Joe's user certificate and will store it in ~/.tsh/keys/<proxy> directory.

With a certificate in place, Joe can now interact with the Teleport cluster:

# SSH into any host behind the proxy:
$ tsh ssh node.example.com

# See what hosts are available behind the proxy:
$ tsh ls

# Log out (this will remove the user certificate from ~/.tsh)
$ tsh logout

Configuring SSO

The local account is good for administrative purposes but regular users of Teleport Enterprise should be using a Single Sign-On (SSO) mechanism that use SAML or OIDC protocols.

Take a look at the SSH via Single Sign-on chapter to learn the basics of integrating Teleport with SSO providers. We have the following detailed guides for configuring SSO providers:

• Okta
• Active Directory
• One Login
• Github

Any SAML-compliant provider can be configured with Teleport by following the same steps. There are Teleport Enterprise customers who are using Oracle IDM, SailPoint and others.

Run Teleport Enterprise using Docker

We provide pre-built Docker images for every version of Teleport Enterprise. These images are hosted on quay.io.

• All tags under quay.io/gravitational/teleport-ent are Teleport Enterprise images

We currently only offer Docker images for x86_64 architectures.

Pick your image

This table gives an idea of how our image naming scheme works. We offer images which point to a static version of Teleport Enterprise, as well as images which are automatically rebuilt every night. These nightly images point to the latest version of Teleport Enterprise from the three most recent release branches. They are stable, and we recommend their use to easily keep your Teleport Enterprise installation up to date.

For testing, we always recommend that you use the latest release version of Teleport Enterprise, which is currently quay.io/gravitational/teleport-ent:5.0.

Quickstart using docker-compose

The easiest way to start Teleport Enterprise quickly is to use docker-compose with our teleport-ent-quickstart.yml file:

# download the quickstart file from our Github repo
curl -Lso teleport-ent-quickstart.yml https://raw.githubusercontent.com/gravitational/teleport/master/docker/teleport-ent-quickstart.yml

# start teleport quickstart using docker-compose
docker-compose -f teleport-ent-quickstart.yml up

• The docker-compose quickstart will automatically create a config file for you at ./docker/teleport/config/teleport.yaml
• This config is mounted into the container under /etc/teleport/teleport.yaml
• It will also start teleport using this config file, with Teleport's data directory set to ./docker/teleport/data and mounted under /var/lib/teleport
• It will mount your license file (named license.pem) from the current directory into the Docker container
• By default, docker-compose will output Teleport's logs to the console for you to observe.
  • If you would rather run the Teleport container in the background, use docker-compose -f teleport-ent-quickstart.yml up -d
  • You can stop the Teleport container using docker-compose -f teleport-ent-quickstart.yml down

Quickstart using docker run

If you'd prefer to complete these steps manually, here's some sample docker run commands:

# create local config and data directories for teleport, which will be mounted into the container
mkdir -p ~/teleport/config ~/teleport/data

# download your license file from the Gravitational dashboard and put it in the correct directory
# the file needs to be named license.pem
cp ~/downloads/downloaded-license.pem ~/teleport/data/license.pem

# generate a sample teleport config and write it to the local config directory
# this container will write the config and immediately exit - this is expected
docker run --hostname localhost --rm \
  --entrypoint=/bin/sh \
  -v ~/teleport/config:/etc/teleport \
  quay.io/gravitational/teleport-ent:5.0 -c "teleport configure > /etc/teleport/teleport.yaml"

# change the path to the license file in the sample config
sed -i 's_/path/to/license-if-using-teleport-enterprise.pem_/var/lib/teleport/license.pem_g' ~/teleport/config/teleport.yaml

# start teleport with mounted license, config and data directories, plus all ports
docker run --hostname localhost --name teleport \
  -v ~/teleport/config:/etc/teleport \
  -v ~/teleport/data:/var/lib/teleport \
  -p 3023:3023 -p 3025:3025 -p 3080:3080 \
  quay.io/gravitational/teleport-ent:5.0

Creating a Teleport user when using Docker quickstart

To create a user inside your Teleport Enterprise container, use docker exec.

This example command will create a Teleport user called testuser which has the admin role. Feel free to change these to suit your needs.

docker exec teleport tctl users add testuser --roles=admin

When you run this command, Teleport will output a URL which you must open to complete the user signup process:

User testuser has been created but requires a password. Share this URL with the user to complete user setup, link is valid for 1h0m0s:
https://localhost:3080/web/invite/4f2718a52ce107568b191f222ba069f7

NOTE: Make sure localhost:3080 points at a Teleport proxy which users can access.

You can now follow this guide from "Adding Users" onwards to create your user and log into Teleport Enterprise.

Troubleshooting

If Teleport services do not start, take a look at the syslog:

$ sudo journalctl -fu teleport

Usually the error will be reported there. Common reasons for failure are:

• Mismatched tokens, i.e. "auth_token" on the node does not match "tokens/node" value on the auth server.
• Network issues: port 3025 is closed via iptables.
• Network issues: ports 3025 or 3022 are occupied by another process.
• Disk issues: Teleport fails to create /var/lib/teleport because the volume is read-only or not accessible.

Getting Help

If something is not working, please reach out to us by creating a ticket in your customer portal. Customers who have purchased the premium support package can also ping us through your Slack channel.