consul/website/source/intro/getting-started/agent.html.markdown

layout	page_title	sidebar_current	description
intro	Run the Agent	gettingstarted-agent	The Consul agent can run in either server or client mode. Each datacenter must have at least one server, though a cluster of 3 or 5 servers is recommended. A single server deployment is highly discouraged in production as data loss is inevitable in a failure scenario.

Run the Consul Agent

After Consul is installed, the agent must be run. The agent can run either in server or client mode. Each datacenter must have at least one server, though a cluster of 3 or 5 servers is recommended. A single server deployment is highly discouraged as data loss is inevitable in a failure scenario.

All other agents run in client mode. A client is a very lightweight process that registers services, runs health checks, and forwards queries to servers. The agent must be run on every node that is part of the cluster.

For more detail on bootstrapping a datacenter, see this guide.

Starting the Agent

For simplicity, we'll start the Consul agent in development mode for now. This mode is useful for bringing up a single-node Consul environment quickly and easily. It is not intended to be used in production as it does not persist any state.

-$ consul agent -dev
==> Starting Consul agent...
==> Starting Consul agent RPC...
==> Consul agent running!
           Version: 'v0.7.0'
         Node name: 'Armons-MacBook-Air'
        Datacenter: 'dc1'
            Server: true (bootstrap: false)
       Client Addr: 127.0.0.1 (HTTP: 8500, HTTPS: -1, DNS: 8600, RPC: 8400)
      Cluster Addr: 127.0.0.1 (LAN: 8301, WAN: 8302)
    Gossip encrypt: false, RPC-TLS: false, TLS-Incoming: false
             Atlas: <disabled>

==> Log data will now stream in as it occurs:

    2016/09/15 10:21:10 [INFO] raft: Initial configuration (index=1): [{Suffrage:Voter ID:127.0.0.1:8300 Address:127.0.0.1:8300}]
    2016/09/15 10:21:10 [INFO] raft: Node at 127.0.0.1:8300 [Follower] entering Follower state (Leader: "")
    2016/09/15 10:21:10 [INFO] serf: EventMemberJoin: Armons-MacBook-Air 127.0.0.1
    2016/09/15 10:21:10 [INFO] serf: EventMemberJoin: Armons-MacBook-Air.dc1 127.0.0.1
    2016/09/15 10:21:10 [INFO] consul: Adding LAN server Armons-MacBook-Air (Addr: tcp/127.0.0.1:8300) (DC: dc1)
    2016/09/15 10:21:10 [INFO] consul: Adding WAN server Armons-MacBook-Air.dc1 (Addr: tcp/127.0.0.1:8300) (DC: dc1)
    2016/09/15 10:21:13 [DEBUG] http: Request GET /v1/agent/services (180.708µs) from=127.0.0.1:52369
    2016/09/15 10:21:13 [DEBUG] http: Request GET /v1/agent/services (15.548µs) from=127.0.0.1:52369
    2016/09/15 10:21:17 [WARN] raft: Heartbeat timeout from "" reached, starting election
    2016/09/15 10:21:17 [INFO] raft: Node at 127.0.0.1:8300 [Candidate] entering Candidate state in term 2
    2016/09/15 10:21:17 [DEBUG] raft: Votes needed: 1
    2016/09/15 10:21:17 [DEBUG] raft: Vote granted from 127.0.0.1:8300 in term 2. Tally: 1
    2016/09/15 10:21:17 [INFO] raft: Election won. Tally: 1
    2016/09/15 10:21:17 [INFO] raft: Node at 127.0.0.1:8300 [Leader] entering Leader state
    2016/09/15 10:21:17 [INFO] consul: cluster leadership acquired
    2016/09/15 10:21:17 [DEBUG] consul: reset tombstone GC to index 3
    2016/09/15 10:21:17 [INFO] consul: New leader elected: Armons-MacBook-Air
    2016/09/15 10:21:17 [INFO] consul: member 'Armons-MacBook-Air' joined, marking health alive
    2016/09/15 10:21:17 [INFO] agent: Synced service 'consul'

As you can see, the Consul agent has started and has output some log data. From the log data, you can see that our agent is running in server mode and has claimed leadership of the cluster. Additionally, the local member has been marked as a healthy member of the cluster.

~> Note for OS X Users: Consul uses your hostname as the default node name. If your hostname contains periods, DNS queries to that node will not work with Consul. To avoid this, explicitly set the name of your node with the -node flag.

Cluster Members

If you run consul members in another terminal, you can see the members of the Consul cluster. We'll cover joining clusters in the next section, but for now, you should only see one member (yourself):

$ consul members
Node                Address            Status  Type    Build     Protocol  DC
Armons-MacBook-Air  172.20.20.11:8301  alive   server  0.6.1dev  2         dc1

The output shows our own node, the address it is running on, its health state, its role in the cluster, and some version information. Additional metadata can be viewed by providing the -detailed flag.

The output of the members command is based on the gossip protocol and is eventually consistent. That is, at any point in time, the view of the world as seen by your local agent may not exactly match the state on the servers. For a strongly consistent view of the world, use the HTTP API as it forwards the request to the Consul servers:

$ curl localhost:8500/v1/catalog/nodes
[{"Node":"Armons-MacBook-Air","Address":"127.0.0.1","TaggedAddresses":{"lan":"127.0.0.1","wan":"127.0.0.1"},"CreateIndex":4,"ModifyIndex":110}]

In addition to the HTTP API, the DNS interface can be used to query the node. Note that you have to make sure to point your DNS lookups to the Consul agent's DNS server which runs on port 8600 by default. The format of the DNS entries (such as "Armons-MacBook-Air.node.consul") will be covered in more detail later.

$ dig @127.0.0.1 -p 8600 Armons-MacBook-Air.node.consul
...

;; QUESTION SECTION:
;Armons-MacBook-Air.node.consul.	IN	A

;; ANSWER SECTION:
Armons-MacBook-Air.node.consul.	0 IN	A	127.0.0.1

Stopping the Agent

You can use Ctrl-C (the interrupt signal) to gracefully halt the agent. After interrupting the agent, you should see it leave the cluster and shut down.

By gracefully leaving, Consul notifies other cluster members that the node left. If you had forcibly killed the agent process, other members of the cluster would have detected that the node failed. When a member leaves, its services and checks are removed from the catalog. When a member fails, its health is simply marked as critical, but it is not removed from the catalog. Consul will automatically try to reconnect to failed nodes, allowing it to recover from certain network conditions, while left nodes are no longer contacted.

Additionally, if an agent is operating as a server, a graceful leave is important to avoid causing a potential availability outage affecting the consensus protocol. See the guides section for details on how to safely add and remove servers.

Next Steps

Your simple Consul cluster is up and running. Let's give it some services!