Cloud Auto-join

As of Consul 0.9.1, retry-join accepts a unified interface using the go-discover library for automatically joining a Consul datacenter using cloud metadata. To use retry-join with a supported cloud provider, specify the configuration on the command line or configuration file as a key=value key=value ... string.

In Consul 0.9.1-0.9.3 the values need to be URL encoded but for most practical purposes you need to replace spaces with + signs.

As of Consul 1.0 the values are taken literally and must not be URL encoded. If the values contain spaces, equals, backslashes or double quotes then they need to be double quoted and the usual escaping rules apply.

  1. $ consul agent -retry-join 'provider=my-cloud config=val config2="some other val" ...'

or via a configuration file:

  1. {
  2. "retry_join": ["provider=my-cloud config=val config2=\"some other val\" ..."]
  3. }

Auto-join with Network Segments

Enterprise

In order to use cloud auto-join with Network Segments, you must reconfigure the Consul agent’s Serf LAN port to match that of the segment you wish to join.

For example, given the following segment configuration on the server agents:

  1. segments = [
  2. {
  3. name = "alpha"
  4. bind = "{{GetPrivateIP}}"
  5. advertise = "{{GetPrivateIP}}"
  6. port = 8303
  7. },
  8. {
  9. name = "beta"
  10. bind = "{{GetPrivateIP}}"
  11. advertise = "{{GetPrivateIP}}"
  12. port = 8304
  13. }
  14. ]

Cloud Auto-join - 图1

server-config.hcl

A Consul client agent wishing to join the “alpha” segment would need to be configured to use port 8303 as its Serf LAN port prior to attempting to join the cluster.

Cloud Auto-join - 图2

Cloud Auto-join - 图3

The following example configuration overrides the default Serf LAN port using the ports.serf_lan configuration option.

  1. ports {
  2. serf_lan = 8303
  3. }

Cloud Auto-join - 图4

client-config.hcl

Provider-specific configurations

The cloud provider-specific configurations are detailed below. This can be combined with static IP or DNS addresses or even multiple configurations for different providers.

In order to use discovery behind a proxy, you will need to set HTTP_PROXY, HTTPS_PROXY and NO_PROXY environment variables per Golang net/http library.

The following sections give the options specific to each supported cloud provider.

Amazon EC2

This returns the first private IP address of all servers in the given region which have the given tag_key and tag_value.

  1. $ consul agent -retry-join "provider=aws tag_key=... tag_value=..."
  1. {
  2. "retry_join": ["provider=aws tag_key=... tag_value=..."]
  3. }
  • provider (required) - the name of the provider (“aws” in this case).
  • tag_key (required) - the key of the tag to auto-join on.
  • tag_value (required) - the value of the tag to auto-join on.
  • region (optional) - the AWS region to authenticate in.
  • addr_type (optional) - the type of address to discover: private_v4, public_v4, public_v6. Default is private_v4. (>= 1.0)
  • access_key_id (optional) - the AWS access key for authentication (see below for more information about authenticating).
  • secret_access_key (optional) - the AWS secret access key for authentication (see below for more information about authenticating).

Authentication & Precedence

  • Static credentials access_key_id=... secret_access_key=...
  • Environment variables (AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY)
  • Shared credentials file (~/.aws/credentials or the path specified by AWS_SHARED_CREDENTIALS_FILE)
  • ECS task role metadata (container-specific).
  • EC2 instance role metadata.

The only required IAM permission is ec2:DescribeInstances, and it is recommended that you make a dedicated key used only to auto-join the datacenter. If the region is omitted it will be discovered through the local instance’s EC2 metadata endpoint.

Microsoft Azure

This returns the first private IP address of all servers in the given region which have the given tag_name and tag_value applied to their virtual NIC in the tenant and subscription, or in the given resource_group of a vm_scale_set for Virtual Machine Scale Sets.

  1. $ consul agent -retry-join "provider=azure tag_key=... tag_value=... tenant_id=... client_id=... subscription_id=... secret_access_key=..."
  1. {
  2. "retry_join": [
  3. "provider=azure tag_name=... tag_value=... tenant_id=... client_id=... subscription_id=... secret_access_key=..."
  4. ]
  5. }
  • provider (required) - the name of the provider (“azure” in this case).
  • tenant_id (required) - the tenant to join machines in.
  • client_id (required) - the client to authenticate with.
  • secret_access_key (required) - the secret client key. NOTE This value often may have an equals sign in its value, especially if generated from the Azure Portal, so is important to wrap in single quotes eg. secret_access_key='fpOfcHQJAQBczjAxiVpeyLmX1M0M0KPBST+GU2GvEN4='

Variables can also be provided by environmental variables:

Use these configuration parameters when using tags:

  • tag_name - the name of the tag to auto-join on.
  • tag_value - the value of the tag to auto-join on.

Use these configuration parameters (instead of tag_name and tag_value) when using Virtual Machine Scale Sets (Consul 1.0.3 and later):

  • resource_group - the name of the resource group to filter on.
  • vm_scale_set - the name of the virtual machine scale set to filter on.

When using tags the only permission needed is Microsoft.Network/networkInterfaces/read. The scope for this permission is the Resource Group that contains the virtual NICs for the Virtual Machines.

When using Virtual Machine Scale Sets the only role action needed is Microsoft.Compute/virtualMachineScaleSets/*/read.

Note: If the Consul datacenter is hosted on Azure, Consul can use Managed Service Identities (MSI) to access Azure instead of an environment variable, shared client id and secret. MSI must be enabled on the VMs or Virtual Machine Scale Sets hosting Consul. It is the preferred configuration since MSI prevents your Azure credentials from being stored in Consul configuration. This feature is supported in Consul 1.7 and above. When using MSI, the tag_key, tag_value and subscription_id need to be supplied for Virtual machines. Be aware that the amount of time that Azure takes for the VMs to detect the MSI permissions can be between a minute to an hour.

Google Compute Engine

This returns the first private IP address of all servers in the given project which have the given tag_value.

  1. $ consul agent -retry-join "provider=gce project_name=... tag_value=..."
  1. {
  2. "retry_join": ["provider=gce project_name=... tag_value=..."]
  3. }
  • provider (required) - the name of the provider (“gce” in this case).
  • tag_value (required) - the value of the tag to auto-join on.
  • project_name (optional) - the name of the project to auto-join on. Discovered if not set.
  • zone_pattern (optional) - the list of zones can be restricted through an RE2 compatible regular expression. If omitted, servers in all zones are returned.
  • credentials_file (optional) - the credentials file for authentication. Note, if you set -config-dir do not store the credentials.json file in the configuration directory as it will be parsed as a config file and Consul will fail to start. See below for more information.

Authentication & Precedence

Discovery requires a GCE Service Account. Credentials are searched using the following paths, in order of precedence.

  • Use credentials from credentials_file, if provided.
  • Use JSON file from GOOGLE_APPLICATION_CREDENTIALS environment variable.
  • Use JSON file in a location known to the gcloud command-line tool.
    • On Windows, this is %APPDATA%/gcloud/application_default_credentials.json.
    • On other systems, $HOME/.config/gcloud/application_default_credentials.json.
  • On Google Compute Engine, use credentials from the metadata server. In this final case any provided scopes are ignored.

IBM SoftLayer

This returns the first private IP address of all servers for the given datacenter with the given tag_value.

  1. $ consul agent -retry-join "provider=softlayer datacenter=... tag_value=... username=... api_key=..."
  1. {
  2. "retry_join": [
  3. "provider=softlayer datacenter=... tag_value=... username=... api_key=..."
  4. ]
  5. }
  • provider (required) - the name of the provider (“softlayer” in this case).
  • `datacenter ((#sl_datacenter)) (required) - the name of the datacenter to auto-join in.
  • tag_value (required) - the value of the tag to auto-join on.
  • username (required) - the username to use for auth.
  • api_key (required) - the api key to use for auth.

Aliyun (Alibaba Cloud)

This returns the first private IP address of all servers for the given region with the given tag_key and tag_value.

  1. $ consul agent -retry-join "provider=aliyun region=... tag_key=consul tag_value=... access_key_id=... access_key_secret=..."
  1. {
  2. "retry_join": [
  3. "provider=aliyun region=... tag_key=consul tag_value=... access_key_id=... access_key_secret=..."
  4. ]
  5. }
  • provider (required) - the name of the provider (“aliyun” in this case).
  • region (required) - the name of the region.
  • tag_key (required) - the key of the tag to auto-join on.
  • tag_value (required) - the value of the tag to auto-join on.
  • access_key_id (required) -the access key to use for auth.
  • access_key_secret (required) - the secret key to use for auth.

The required RAM permission is ecs:DescribeInstances. It is recommended you make a dedicated key used to auto-join.

Digital Ocean

This returns the first private IP address of all servers for the given region with the given tag_name.

  1. $ consul agent -retry-join "provider=digitalocean region=... tag_name=... api_token=..."
  1. {
  2. "retry_join": ["provider=digitalocean region=... tag_name=... api_token=..."]
  3. }
  • provider (required) - the name of the provider (“digitalocean” in this case).
  • region (required) - the name of the region.
  • tag_name (required) - the value of the tag to auto-join on.
  • api_token (required) -the token to use for auth.

Openstack

This returns the first private IP address of all servers for the given region with the given tag_key and tag_value.

  1. $ consul agent -retry-join "provider=os tag_key=consul tag_value=server user_name=... password=... auth_url=..."
  1. {
  2. "retry_join": [
  3. "provider=os tag_key=consul tag_value=server user_name=... password=... auth_url=..."
  4. ]
  5. }
  • provider (required) - the name of the provider (“os” in this case).
  • tag_key (required) - the key of the tag to auto-join on.
  • tag_value (required) - the value of the tag to auto-join on.
  • domain_name (optional) - the name of the domain.
  • domain_id (optional) - the id of the domain.
  • project_id (optional) - the id of the project (tenant id).
  • region (optional) - the name of the region.
  • user_name (optional) - the username to use for auth.
  • password (optional) - the password to use for auth.
  • token (optional) - the token to use for auth.
  • auth_url (optional) - the identity endpoint to use for auth.
  • insecure (optional) - indicates whether the API certificate should not be checked. Any value means true.

The configuration can also be provided by environment variables.

Scaleway

This returns the first private IP address of all servers for the given region with the given tag_name.

  1. $ consul agent -retry-join "provider=scaleway organization=my-org tag_name=consul-server token=... region=..."
  1. {
  2. "retry_join": [
  3. "provider=scaleway organization=my-org tag_name=consul-server token=... region=..."
  4. ]
  5. }
  • provider (required) - the name of the provider (“scaleway” in this case).
  • region (required) - the name of the region.
  • tag_name (required) - the name of the tag to auto-join on.
  • organization (required) - the organization access key to use for auth (equal to access key).
  • token (required) - the token to use for auth.

TencentCloud

This returns the first IP address of all servers for the given region with the given tag_key and tag_value.

  1. $ consul agent -retry-join "provider=tencentcloud region=... tag_key=consul tag_value=... access_key_id=... access_key_secret=..."
  1. {
  2. "retry_join": [
  3. "provider=tencentcloud region=... tag_key=consul tag_value=... access_key_id=... access_key_secret=..."
  4. ]
  5. }
  • provider (required) - the name of the provider (“tencentcloud” in this case).
  • region (required) - The TencentCloud region.
  • tag_key (required) - The tag key to auto-join on.
  • tag_value (required) - The tag value to auto-join on.
  • address_type (optional) - “private_v4” or “public_v4”, default is “private_v4”.
  • access_key_id (required) - The secret id of TencentCloud.
  • access_key_secret (required) - The secret key of TencentCloud.

This required permission to ‘cvm:DescribeInstances’. It is recommended you make a dedicated key used to auto-join the Consul datacenter.

Joyent Triton

This returns the first PrimaryIP addresses for all servers with the given tag_key and tag_value.

  1. $ consul agent -retry-join "provider=triton account=testaccount url=https://us-sw-1.api.joyentcloud.com key_id=... tag_key=consul-role tag_value=server"
  1. {
  2. "retry_join": [
  3. "provider=triton account=testaccount url=https://us-sw-1.api.joyentcloud.com key_id=... tag_key=consul-role tag_value=server"
  4. ]
  5. }
  • provider (required) - the name of the provider (“triton” in this case).
  • account (required) - the name of the account.
  • url (required) - the URL of the Triton api endpoint to use.
  • key_id (required) - the key id to use.
  • tag_key (optional) - the instance tag key to use.
  • tag_value (optional) - the tag value to use.

vSphere

This returns the first private IP address of all servers for the given region with the given tag_name and category_name.

  1. $ consul agent -retry-join "provider=vsphere category_name=consul-role tag_name=consul-server host=... user=... password=... insecure_ssl=[true|false]"
  1. {
  2. "retry-join": [
  3. "provider=vsphere category_name=consul-role tag_name=consul-server host=... user=... password=... insecure_ssl=[true|false]"
  4. ]
  5. }
  • provider (required) - the name of the provider (“vsphere” is the provider here)
  • tag_name (required) - The name of the tag to look up.
  • category_name (required) - The category of the tag to look up.
  • host (required) - The host of the vSphere server to connect to.
  • user (required) - The username to connect as.
  • password (required) - The password of the user to connect to vSphere as.
  • insecure_ssl (optional) - Whether or not to skip SSL certificate validation.
  • timeout (optional) - Discovery context timeout (default: 10m)

Packet

This returns the first private IP address (or the IP address of address type) of all servers with the given project and auth_token.

  1. $ consul agent -retry-join "provider=packet auth_token=token project=uuid url=... address_type=..."
  1. {
  2. "retry-join": [
  3. "provider=packet auth_token=token project=uuid url=... address_type=..."
  4. ]
  5. }
  • provider (required) - the name of the provider (“packet” is the provider here)
  • project (required) - the UUID of packet project
  • auth_token (required) - the authentication token for packet
  • url (optional) - a REST URL for packet
  • address_type (optional) - the type of address to check for in this provider (“private_v4”, “public_v4” or “public_v6”. Defaults to “private_v4”)

Linode

This returns the first private IP address of all servers for the given region with the given tag_name.

  1. $ consul agent -retry-join "provider=linode region=us-east tag_name=consul-server"
  1. {
  2. "retry-join": ["provider=linode region=us-east tag_name=consul-server"]
  3. }
  • provider (required) is linode
  • api_token (required) - The Linode API token to use
  • region (optional) - The Linode region to filter on
  • tag_name (optional) - The tag name to filter on
  • address_type (optional) - the type of address to check for in this provider (“private_v4”, “public_v4” or “public_v6”. Defaults to “private_v4”)

Variables can also be provided by environment variables:

Kubernetes (k8s)

The Kubernetes provider finds the IP addresses of pods with the matching label or field selector. This is useful for non-Kubernetes agents that are joining a server cluster running within Kubernetes.

The pod IP is used by default, which requires that the agent connecting can network to the pod IP. The host_network boolean can be set to true to use the host IP instead, but this requires the agent ports (Gossip, RPC, etc.) to be exported to the host as well.

By default, no port is specified. This causes Consul to use the default gossip port (default behavior with all join requests). The pod may specify the consul.hashicorp.com/auto-join-port annotation to set the port. The value may be an integer or a named port.

  1. $ consul agent -retry-join "provider=k8s label_selector=\"app=consul,component=server\""
  1. {
  2. "retry-join": ["provider=k8s label_selector=..."]
  3. }
  • provider (required) - the name of the provider (“k8s” is the provider here)
  • kubeconfig (optional) - path to the kubeconfig file. If this isn’t set, then in-cluster auth will be attempted. If that fails, the default kubeconfig paths are tried ($HOME/.kube/config).
  • namespace (optional) - the namespace to search for pods. If this isn’t set, it defaults to all namespaces.
  • label_selector (optional) - the label selector for matching pods.
  • field_selector (optional) - the field selector for matching pods.

The Kubernetes token used by the provider needs to have permissions to list pods in the desired namespace.