consul/website/content/docs/k8s/installation/multi-cluster/kubernetes.mdx

480 lines
18 KiB
Markdown

---
layout: docs
page_title: Federation Between Kubernetes Clusters
description: >-
Federating multiple Kubernetes clusters.
---
# Federation Between Kubernetes Clusters
-> **1.8.0+:** This feature is available in Consul versions 1.8.0 and higher
~> This topic requires familiarity with [Mesh Gateways](/docs/connect/gateways/mesh-gateway/service-to-service-traffic-datacenters) and [WAN Federation Via Mesh Gateways](/docs/connect/gateways/mesh-gateway/wan-federation-via-mesh-gateways).
-> Looking for a step-by-step guide? Please follow our Learn tutorial: [Secure and Route Service Mesh Communication Across Kubernetes](https://learn.hashicorp.com/tutorials/consul/kubernetes-mesh-gateways).
This page describes how to federate multiple Kubernetes clusters. See [Multi-Cluster Overview](/docs/k8s/installation/multi-cluster)
for more information on use-cases and how it works.
## Primary Datacenter
Consul treats each Kubernetes cluster as a separate Consul datacenter.
In order to federate clusters, one cluster must be designated the
primary datacenter. This datacenter will be
responsible for creating the certificate authority that signs the TLS certificates
Connect uses to encrypt and authorize traffic. It also handles validating global ACL tokens. All other clusters
that are federated are considered secondaries.
#### First Time Installation
If you haven't installed Consul on your cluster, continue reading below. If you've
already installed Consul on a cluster and want to upgrade it to
support federation, see [Upgrading An Existing Cluster](#upgrading-an-existing-cluster).
You will need to use the following `config.yaml` file for your primary cluster,
with the possible modifications listed below.
<CodeBlockConfig filename="config.yaml">
```yaml
global:
name: consul
datacenter: dc1
# TLS configures whether Consul components use TLS.
tls:
# TLS must be enabled for federation in Kubernetes.
enabled: true
federation:
enabled: true
# This will cause a Kubernetes secret to be created that
# can be imported by secondary datacenters to configure them
# for federation.
createFederationSecret: true
acls:
manageSystemACLs: true
# If ACLs are enabled, we must create a token for secondary
# datacenters to replicate ACLs.
createReplicationToken: true
# Gossip encryption secures the protocol Consul uses to quickly
# discover new nodes and detect failure.
gossipEncryption:
autoGenerate: true
connectInject:
# Consul Connect service mesh must be enabled for federation.
enabled: true
controller:
enabled: true
meshGateway:
# Mesh gateways are gateways between datacenters. They must be enabled
# for federation in Kubernetes since the communication between datacenters
# goes through the mesh gateways.
enabled: true
```
</CodeBlockConfig>
Modifications:
1. The Consul datacenter name is `dc1`. The datacenter name in each federated
cluster **must be unique**.
1. ACLs are enabled in the above config file. They can be disabled by setting:
```yaml
global:
acls:
manageSystemACLs: false
createReplicationToken: false
```
ACLs secure Consul by requiring every API call to present an ACL token that
is validated to ensure it has the proper permissions. If you are only testing Consul,
this is not required.
1. Gossip encryption is enabled in the above config file. To disable it, comment
out or delete the `gossipEncryption` key:
```yaml
global:
# gossipEncryption:
# autoGenerate: true
```
Gossip encryption encrypts the communication layer used to discover other
nodes in the cluster and report on failure. If you are only testing Consul,
this is not required.
1. The default mesh gateway configuration
creates a Kubernetes Load Balancer service. If you wish to customize the
mesh gateway, for example using a Node Port service or a custom DNS entry,
see the [Helm reference](/docs/k8s/helm#v-meshgateway) for that setting.
With your `config.yaml` ready to go, follow our [Installation Guide](/docs/k8s/installation/install)
to install Consul on your primary cluster.
-> **NOTE:** You must be using consul-helm 0.21.0+. To update, run `helm repo update`.
#### Upgrading An Existing Cluster
If you have an existing cluster, you will need to upgrade it to ensure it has
the following config:
<CodeBlockConfig filename="config.yaml">
```yaml
global:
tls:
enabled: true
federation:
enabled: true
createFederationSecret: true
acls:
manageSystemACLs: true
createReplicationToken: true
meshGateway:
enabled: true
```
</CodeBlockConfig>
1. `global.tls.enabled` must be `true`. See [Configuring TLS on an Existing Cluster](/docs/k8s/operations/tls-on-existing-cluster)
for more information on safely upgrading a cluster to use TLS.
If you've set `enableAutoEncrypt: true`, this is also supported.
1. `global.federation.enabled` must be set to `true`. This is a new config setting.
1. If using ACLs, you'll already have `global.acls.manageSystemACLs: true`. For the
primary cluster, you'll also need to set `global.acls.createReplicationToken: true`.
This ensures that an ACL token is created that secondary clusters can use to authenticate
with the primary.
1. Mesh Gateways are enabled with the default configuration. The default configuration
creates a Kubernetes Load Balancer service. If you wish to customize the
mesh gateway, see the [Helm reference](/docs/k8s/helm#v-meshgateway) for that setting.
With the above settings added to your existing config, follow the [Upgrading](/docs/k8s/upgrade)
guide to upgrade your cluster and then come back to the [Federation Secret](#federation-secret) section.
-> **NOTE:** You must be using consul-helm 0.21.0+.
#### ProxyDefaults
If you are using consul-helm 0.30.0+ you must also create a [`ProxyDefaults`](/docs/connect/config-entries/proxy-defaults)
resource to configure Consul to use the mesh gateways for service mesh traffic.
```yaml
apiVersion: consul.hashicorp.com/v1alpha1
kind: ProxyDefaults
metadata:
name: global
spec:
meshGateway:
mode: 'local'
```
The `spec.meshGateway.mode` can be set to `local` or `remote`. If set to `local`,
traffic from one datacenter to another will egress through the local mesh gateway.
This may be useful if you prefer all your cross-cluster network traffic to egress
from the same locations.
If set to `remote`, traffic will be routed directly from the pod to the remote mesh gateway
(resulting in one less hop).
Verify that the resource was synced to Consul:
```shell-session
$ kubectl get proxydefaults global
NAME SYNCED AGE
global True 1s
```
Its `SYNCED` status should be `True`.
-> **NOTE:** The `ProxyDefaults` resource can be created in any namespace, but
we recommend creating it in the same namespace that Consul is installed in.
Its name must be `global`.
## Federation Secret
The federation secret is a Kubernetes secret containing information needed
for secondary datacenters/clusters to federate with the primary. This secret is created
automatically by setting:
<CodeBlockConfig highlight="2-3">
```yaml
global:
federation:
createFederationSecret: true
```
</CodeBlockConfig>
After the installation into your primary cluster you will need to export
this secret:
```shell-session
$ kubectl get secret consul-federation --namespace consul --output yaml > consul-federation-secret.yaml
```
!> **Security note:** The federation secret makes it possible to gain
full admin privileges in Consul. This secret must be kept securely, i.e.
it should be deleted from your filesystem after importing it to your secondary
cluster and you should use RBAC permissions to ensure only administrators
can read it from Kubernetes.
~> **Secret doesn't exist?** If you haven't set `global.name` to `consul` then the name of the secret will
be your Helm release name suffixed with `-consul-federation` e.g. `helm-release-consul-federation`. Also ensure you're
using the namespace Consul was installed into.
Now you're ready to import the secret into your secondary cluster(s).
Switch `kubectl` context to your secondary Kubernetes cluster. In this example
our context for our secondary cluster is `dc2`:
```shell-session
$ kubectl config use-context dc2
Switched to context "dc2".
```
And import the secret:
```shell-session
$ kubectl apply --filename consul-federation-secret.yaml
secret/consul-federation configured
```
#### Federation Secret Contents
The automatically generated federation secret contains:
- **Server certificate authority certificate** - This is the certificate authority
used to sign Consul server-to-server communication. This is required by secondary
clusters because they must communicate with the Consul servers in the primary cluster.
- **Server certificate authority key** - This is the signing key for the server certificate
authority. This is required by secondary clusters because they need to create
server certificates for each Consul server using the same certificate authority
as the primary.
!> **Security note:** The certificate authority key would enable an attacker to compromise Consul,
it should be kept securely.
- **Consul server config** - This is a JSON snippet that must be used as part of the server config for secondary datacenters.
It sets:
- [`primary_datacenter`](/docs/agent/config/config-files#primary_datacenter) to the name of the primary datacenter.
- [`primary_gateways`](/docs/agent/config/config-files#primary_gateways) to an array of IPs or hostnames
for the mesh gateways in the primary datacenter. These are the addresses that
Consul servers in secondary clusters will use to communicate with the primary
datacenter.
Even if there are multiple secondary datacenters, only the primary gateways
need to be configured. Upon first connection with a primary datacenter, the
addresses for other secondary datacenters will be discovered.
- **ACL replication token** - If ACLs are enabled, secondary datacenters need
an ACL token in order to authenticate with the primary datacenter. This ACL
token is also used to replicate ACLs from the primary datacenter so that
components in each datacenter can authenticate with one another.
- **Gossip encryption key** - If gossip encryption is enabled, secondary datacenters
need the gossip encryption key in order to be part of the gossip pool.
Gossip is the method by which Consul discovers the addresses and health of other
nodes.
!> **Security note:** This gossip encryption key would enable an attacker to compromise Consul,
it should be kept securely.
## Kubernetes API URL
If ACLs are enabled, you must next determine the Kubernetes API URL for the secondary cluster. The API URL of the primary cluster
must be specified in the config files for all secondary clusters because secondary clusters need
to create global Consul ACL tokens (tokens that are valid in all datacenters) and these tokens can only be created
by the primary datacenter. By setting the API URL, the secondary cluster will configure a [Consul auth method](/docs/security/acl/auth-methods)
in the primary cluster so that components in the secondary cluster can use their Kubernetes ServiceAccount tokens
to retrieve global Consul ACL tokens from the primary.
To determine the Kubernetes API URL, first get the cluster name in your kubeconfig:
```shell-session
$ export CLUSTER=$(kubectl config view -o jsonpath="{.contexts[?(@.name == \"$(kubectl config current-context)\")].context.cluster}")
```
Then get the API URL:
```shell-session
$ kubectl config view -o jsonpath="{.clusters[?(@.name == \"$CLUSTER\")].cluster.server}"
https://<some-url>
```
Keep track of this URL, you'll need it in the next section.
## Secondary Cluster(s)
With the primary cluster up and running, and the [federation secret](#federation-secret) imported
into the secondary cluster, we can now install Consul into the secondary
cluster.
You will need to use the following `config.yaml` file for your secondary cluster(s),
with the modifications listed below.
-> **NOTE: ** You must use a separate Helm config file for each cluster (primary and secondaries) since their
settings are different.
<CodeBlockConfig filename="config-cluster2.yaml">
```yaml
global:
name: consul
datacenter: dc2
tls:
enabled: true
# Here we're using the shared certificate authority from the primary
# datacenter that was exported via the federation secret.
caCert:
secretName: consul-federation
secretKey: caCert
caKey:
secretName: consul-federation
secretKey: caKey
acls:
manageSystemACLs: true
# Here we're importing the replication token that was
# exported from the primary via the federation secret.
replicationToken:
secretName: consul-federation
secretKey: replicationToken
federation:
enabled: true
k8sAuthMethodHost: <kubernetes-api-url>
primaryDatacenter: dc1
gossipEncryption:
secretName: consul-federation
secretKey: gossipEncryptionKey
connectInject:
enabled: true
controller:
enabled: true
meshGateway:
enabled: true
server:
# Here we're including the server config exported from the primary
# via the federation secret. This config includes the addresses of
# the primary datacenter's mesh gateways so Consul can begin federation.
extraVolumes:
- type: secret
name: consul-federation
items:
- key: serverConfigJSON
path: config.json
load: true
```
</CodeBlockConfig>
Modifications:
1. If ACLs are enabled, change the value of `global.federation.k8sAuthMethodHost` to the full URL (including `https://`) of this cluster's
Kubernetes API.
1. `global.federation.primaryDatacenter` must be set to the name of the primary datacenter.
1. The Consul datacenter name for the datacenter in this example is `dc2`. The datacenter name in **each** federated cluster **must be unique**.
1. ACLs are enabled in the above config file. They can be disabled by removing
the whole `acls` block:
```yaml
acls:
manageSystemACLs: false
replicationToken:
secretName: consul-federation
secretKey: replicationToken
```
If ACLs are enabled in one datacenter, they must be enabled in all datacenters
because in order to communicate with that one datacenter ACL tokens are required.
1. Gossip encryption is enabled in the above config file. To disable it, don't
set the `gossipEncryption` key:
```yaml
global:
# gossipEncryption:
# secretName: consul-federation
# secretKey: gossipEncryptionKey
```
If gossip encryption is enabled in one datacenter, it must be enabled in all datacenters
because in order to communicate with that one datacenter the encryption key is required.
1. The default mesh gateway configuration
creates a Kubernetes Load Balancer service. If you wish to customize the
mesh gateway, for example using a Node Port service or a custom DNS entry,
see the [Helm reference](/docs/k8s/helm#v-meshgateway) for that setting.
With your `config.yaml` ready to go, follow our [Installation Guide](/docs/k8s/installation/install)
to install Consul on your secondary cluster(s).
## Verifying Federation
To verify that both datacenters are federated, run the
`consul members -wan` command on one of the Consul server pods:
```shell-session
$ kubectl exec statefulset/consul-server --namespace consul -- consul members -wan
Node Address Status Type Build Protocol DC Segment
consul-server-0.dc1 10.32.4.216:8302 alive server 1.8.0 2 dc1 <all>
consul-server-0.dc2 192.168.2.173:8302 alive server 1.8.0 2 dc2 <all>
consul-server-1.dc1 10.32.5.161:8302 alive server 1.8.0 2 dc1 <all>
consul-server-1.dc2 192.168.88.64:8302 alive server 1.8.0 2 dc2 <all>
consul-server-2.dc1 10.32.1.175:8302 alive server 1.8.0 2 dc1 <all>
consul-server-2.dc2 192.168.35.174:8302 alive server 1.8.0 2 dc2 <all>
```
In this example (run from `dc1`), you can see that this datacenter knows about
the servers in dc2 and that they have status `alive`.
You can also use the `consul catalog services` command with the `-datacenter` flag to ensure
each datacenter can read each other's services. In this example, our `kubectl`
context is `dc1` and we're querying for the list of services in `dc2`:
```shell-session
$ kubectl exec statefulset/consul-server --namespace consul -- consul catalog services -datacenter dc2
consul
mesh-gateway
```
You can switch kubectl contexts and run the same command in `dc2` with the flag
`-datacenter dc1` to ensure `dc2` can communicate with `dc1`.
### Consul UI
We can also use the Consul UI to verify federation.
See [Viewing the Consul UI](/docs/k8s/installation/install#viewing-the-consul-ui)
for instructions on how to view the UI.
~> NOTE: If ACLs are enabled, your kubectl context must be in the primary datacenter
to retrieve the bootstrap token mentioned in the UI documentation.
With the UI open, you'll be able to switch between datacenters via the dropdown
in the top left:
![Consul Datacenter Dropdown](/img/data-center-dropdown.png 'Consul Datacenter Dropdown')
## Next Steps
With your Kubernetes clusters federated, try out using Consul service mesh to
route between services deployed on each cluster by following our Learn tutorial: [Secure and Route Service Mesh Communication Across Kubernetes](https://learn.hashicorp.com/tutorials/consul/kubernetes-mesh-gateways#deploy-microservices).
You can also read our in-depth documentation on [Consul Service Mesh In Kubernetes](/docs/k8s/connect).
If you are still considering a move to Kubernetes, or to Consul on Kubernetes specifically, our [Migrate to Microservices with Consul Service Mesh on Kubernetes](https://learn.hashicorp.com/collections/consul/microservices?utm_source=WEBSITE&utm_medium=WEB_IO&utm_offer=ARTICLE_PAGE&utm_content=DOCS)
collection uses an example application written by a fictional company to illustrate why and how organizations can
migrate from monolith to microservices using Consul service mesh on Kubernetes. The case study in this collection
should provide information valuable for understanding how to develop services that leverage Consul during any stage
of your microservices journey.