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

126 lines
5.1 KiB
Markdown
Raw Normal View History

2014-02-08 00:41:03 +00:00
---
layout: "intro"
page_title: "Run the Agent"
sidebar_current: "gettingstarted-agent"
---
# Run the Consul Agent
After Consul is installed, the agent must be run. The agent can either run
in a server or client mode. Each datacenter must at least one server, and
a recommended 3 or 5. A single server deployment is _**highly**_ discouraged
as data loss is inevitable in a failure scenario. [This guide](/docs/guides/bootstrapping.html)
covers bootstrapping a new datacenter. All other agents run in client mode, which
is a very lightweight process that registers services, runs health checks,
and forwards queries to servers. The agent must be run for every node that
will be part of the cluster.
2014-02-08 00:41:03 +00:00
## Starting the Agent
For simplicity, we'll run a single Consul agent in server mode right now:
2014-02-08 00:41:03 +00:00
```
$ ./bin/consul agent -server -bootstrap -data-dir /tmp/consul
==> WARNING: Bootstrap mode enabled! Do not enable unless necessary
==> WARNING: It is highly recommended to set GOMAXPROCS higher than 1
==> Starting Consul agent...
==> Starting Consul agent RPC...
==> Consul agent running!
2014-04-11 23:23:16 +00:00
Node name: 'Armons-MacBook-Air'
Datacenter: 'dc1'
Server: true (bootstrap: true)
Client Addr: 127.0.0.1 (HTTP: 8500, DNS: 8600, RPC: 8400)
Cluster Addr: 10.1.10.38 (LAN: 8301, WAN: 8302)
2014-02-08 00:41:03 +00:00
==> Log data will now stream in as it occurs:
2014-04-11 23:23:16 +00:00
[INFO] serf: EventMemberJoin: Armons-MacBook-Air.local 10.1.10.38
[INFO] raft: Node at 10.1.10.38:8300 [Follower] entering Follower state
[INFO] consul: adding server for datacenter: dc1, addr: 10.1.10.38:8300
[ERR] agent: failed to sync remote state: rpc error: No cluster leader
[WARN] raft: Heartbeat timeout reached, starting election
[INFO] raft: Node at 10.1.10.38:8300 [Candidate] entering Candidate state
[INFO] raft: Election won. Tally: 1
[INFO] raft: Node at 10.1.10.38:8300 [Leader] entering Leader state
[INFO] consul: cluster leadership acquired
[INFO] consul: New leader elected: Armons-MacBook-Air
[INFO] consul: member 'Armons-MacBook-Air' joined, marking health alive
2014-02-08 00:41:03 +00:00
```
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.
2014-02-08 00:41:03 +00:00
## Cluster Members
If you run `consul members` in another terminal, you can see the members of
the Consul cluster. You should only see one member (yourself). We'll cover
2014-02-08 00:41:03 +00:00
joining clusters in the next section.
```
$ consul members
Armons-MacBook-Air 10.1.10.38:8301 alive role=consul,dc=dc1,vsn=1,vsn_min=1,vsn_max=1,port=8300,bootstrap=1
2014-02-08 00:41:03 +00:00
```
This command, along with many others, communicates with a running Consul
agent via an internal RPC protocol. When starting the Consul agent, you
2014-02-08 00:41:03 +00:00
may have noticed that it tells you the "RPC addr". This is the address
that commands such as `consul members` use to communicate with the agent.
2014-02-08 00:41:03 +00:00
By default, RPC listens only on loopback, so it is inaccessible outside
of your machine for security reasons.
If you're changed the default RPC address, you'll have to specify
2014-02-08 00:41:03 +00:00
an `-rpc-addr` to both the agent and any commands so that it doesn't
collide with other agents.
It is important to note that the output of the `members` command is
generated by from the [gossip protocol](/docs/internals/gossip.html),
and is eventually consistent. Consul uses this to maintain a strongly
consistent catalog of nodes that can be queried using the [HTTP API](/docs/agent/http.html):
```
$ curl localhost:8500/v1/catalog/nodes
[{"Node":"Armons-MacBook-Air","Address":"10.1.10.38"}]
```
Alternatively, the [DNS interface](/docs/agent/dns.html) could be used:
```
$ dig @127.0.0.1 -p 8600 Armons-MacBook-Air.node.consul
; <<>> DiG 9.8.3-P1 <<>> @127.0.0.1 -p 8600 Armons-MacBook-Air.node.consul
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 63911
;; flags: qr aa rd; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 0
;; WARNING: recursion requested but not available
;; QUESTION SECTION:
;Armons-MacBook-Air.node.consul. IN A
;; ANSWER SECTION:
Armons-MacBook-Air.node.consul. 0 IN A 10.1.10.38
```
2014-02-08 00:41:03 +00:00
## 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 gracefully
and shut down.
By gracefully leaving, Consul notifies other cluster members that the
2014-02-08 00:41:03 +00:00
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,
it's services and checks are removed from the catalog. When a member fails,
it's health is simply marked as critical, but is not removed from the catalog.
Consul will automatically try to reconnect to _failed_ nodes, which allows 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](/docs/internals/consensus.html).
See the [guides section](/docs/guides/index.html) to safely add and remove servers.
2014-02-08 00:41:03 +00:00