teleport/docs/quickstart.md

Quick Start

Introduction

Welcome to the Teleport Quick Start Guide. 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.

• An auth server is the core of a cluster. Auth servers store user accounts and offer
  authentication and authorization service for every node and every user in a cluster.
• Nodes are regular SSH nodes, similar to the sshd daemon you are probably used to. When a node receives a connection request, it authenticates it via the cluster's auth server.
• Proxies route client connection requests to the appropriate node and serve a Web UI which can also be used to login into SSH nodes. Every client-to-node connection in Teleport must be routed via a proxy.

The teleport daemon runs all 3 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:

Tool	Description
tctl	Cluster administration tool used to invite nodes to a cluster and manage user accounts.
tsh	Similar in principle to OpenSSH's ssh. Used to login into remote SSH nodes, list and search for nodes in a cluster, securely upload/download files, etc.
browser	You can use your web browser to login into any Teleport node by opening https://<proxy-host>:3080.

Installing

Gravitational Teleport natively runs on any modern Linux distribution and OSX. You can download pre-built binaries from here or you can build it from source.

Starting Teleport

Let's start Teleport on a single-node. First, create a directory for Teleport to keep its data. By default it's /var/lib/teleport. Then start teleport daemon:

mkdir -p /var/lib/teleport
teleport start

At this point you should see Teleport print listening IPs of all 3 services into the console. Congratulations - You are running Teleport!

Creating Users

Teleport users are defined on a cluster level, and every Teleport user must be associated with a list of machine-level OS usernames it can authenticate as during a login. This list is called "user mappings".

If you do not specify the mappings, the new Teleport user will be assigned a mapping with the same name. Let's create a Teleport user with the same name as the OS user:

> tctl users add $USER

Signup token has been created. Share this URL with the user:
https://localhost:3080/web/newuser/96c85ed60b47ad345525f03e1524ac95d78d94ffd2d0fb3c683ff9d6221747c2

tctl prints a sign-up URL for you to visit and complete registration. Open this link in a browser, install Google Authenticator on your phone, set up 2nd factor authentication and pick a password.

Having done that, you will be presented with a Web UI where you will see your machine and will be able to log into it using web-based terminal.

Login

Let's login using the tsh command line tool:

tsh --proxy=localhost localhost

Notice that tsh client always needs --proxy flag because all client connections in Teleport have to go via a proxy sometimes called an "SSH bastion".

Adding Nodes to Cluster

Let's add another node to your cluster. Let's assume the other node can be reached by hostname "luna". tctl command below will create a single-use token for a node to join and will print instructions for you to follow:

> tctl nodes add

The invite token: n92bb958ce97f761da978d08c35c54a5c
Run this on the new node to join the cluster:
teleport start --roles=node --token=n92bb958ce97f761da978d08c35c54a5c --auth-server=10.0.10.1

Start teleport daemon on "luna" as shown above, but make sure to use the proper --auth-server IP to point back to your localhost.

Once you do that, "luna" will join the cluster. To verify, type this on your localhost:

> tsh --proxy=localhost ls

Node Name     Node ID                     Address            Labels
---------     -------                     -------            ------
localhost     xxxxx-xxxx-xxxx-xxxxxxx     10.0.10.1:3022     
luna          xxxxx-xxxx-xxxx-xxxxxxx     10.0.10.2:3022     

Using Node Labels

Notice the "Labels" column in the output above. It is currently not populated. Teleport lets you apply static or dynamic labels to your nodes. As the cluster grows and nodes assume different roles, labels will help to find the right node quickly.

Let's see labels in action. Stop teleport on "luna" and restart it with the following command:

teleport start --roles=node --auth-server=10.0.10.1 --nodename=db --labels "location=virginia,arch=[1h:/bin/uname -m]"

Notice a few things here:

• We did not use --token flag this time, because "luna" is already a member of the cluster.
• We renamed "luna" to "db" because this machine is running a database. This name only exists within Teleport, the actual hostname has not changed.
• We assigned a static label "location" to this host and set it to "virginia".
• We also assigned a dynamic label "arch" which will evaluate /bin/uname -m command once an hour and assign the output to this label value.

Let's take a look at our cluster now:

> tsh --proxy=localhost ls

Node Name     Node ID                     Address            Labels
---------     -------                     -------            ------
localhost     xxxxx-xxxx-xxxx-xxxxxxx     10.0.10.1:3022     
db            xxxxx-xxxx-xxxx-xxxxxxx     10.0.10.2:3022     location=virginia,arch=x86_64

Let's use the newly created labels to filter the output of tsh ls and ask to show only nodes located in Virginia:

> tsh --proxy=localhost ls location=virginia

Node Name     Node ID                     Address            Labels
---------     -------                     -------            ------
db            xxxxx-xxxx-xxxx-xxxxxxx     10.0.10.2:3022     location=virginia,arch=x86_64

Labels can be used with the regular ssh command too. This will execute ls -l / command on all servers located in Virginia:

> tsh --proxy=localhost ssh location=virginia ls -l /

Sharing SSH Sessions with Colleagues

Suppose you are trying to troubleshoot a problem on a node. Sometimes it makes sense to ask another team member for help. Traditionally this could be done by letting them know which node you're on, having them SSH in, start a terminal multiplexer like screen and join a session there.

Teleport makes this a bit more convenient. Let's login into "db" and ask Teleport for your current session status:

> tsh --proxy=teleport.example.com ssh db
luna > teleport status

User ID    : joe, logged in as joe from 10.0.10.1 43026 3022
Session ID : 7645d523-60cb-436d-b732-99c5df14b7c4
Session URL: https://teleport.example.com:3080/web/sessions/7645d523-60cb-436d-b732-99c5df14b7c4

You can share the Session URL with a colleague in your organization. Assuming that teleport.example.com is your company's Teleport proxy, she will be able to join and help you troubleshoot the problem on "db" in her browser.

Also, people can join your session via CLI assuming they have Teleport installed. They just have to run:

> tsh --proxy=teleport.example.com join 7645d523-60cb-436d-b732-99c5df14b7c4

NOTE: for this to work, both of you must have proper user mappings allowing you access db under the same OS user.

Inviting Colleagues to your Laptop

Sometimes you may want to temporarily share the terminal on your own laptop (if you trust your guests, of course). First, you will have to start teleport with --roles=node in a separate Terminal:

> teleport start --proxy=teleport.example.com

... then you will need to start a local SSH session by logging into localhost and asking for a session ID:

> tsh --proxy=teleport.example.com ssh localhost
localhost> teleport status

Now you can invite someone into your localhost session. They will need to have a proper user mapping, of course, to be allowed to join your session. To disconnect, shut down teleport daemon or simply exit the tsh session.

Running in Production

We hope this Guide helped you to quickly set up Teleport to play with on localhost. For production environments we strongly recommend the following:

• Install HTTPS certificates for every Teleport proxy.
• Run Teleport auth on isolated servers. The auth service can run in a highly available (HA) configuration.
• Use a configuration file instead of command line flags because it gives you more flexibility, for example for configuring HA clusters.