Deploy a multi-zone global control plane
Prerequisites
To set up a multi-zone deployment we will need to:
Usage
Set up the global control plane
The global control plane must run on a dedicated cluster (unless using “Universal on Kubernetes” mode), and cannot be assigned to a zone.
The global control plane on Kubernetes must reside on its own Kubernetes cluster, to keep its resources separate from the resources the zone control planes create during synchronization.
Run:
kumactl install control-plane \
--set "controlPlane.mode=global" \
| kubectl apply -f -
Before using Kuma with helm, please follow these steps to configure your local helm repo.
helm install --create-namespace --namespace kuma-system \
--set "controlPlane.mode=global" \
kuma kuma/kuma
Find the external IP and port of the kuma-global-zone-sync
service in the kuma-system
namespace:
kubectl get services -n kuma-system
NAMESPACE NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kuma-system kuma-global-zone-sync LoadBalancer 10.105.9.10 35.226.196.103 5685:30685/TCP 89s
kuma-system kuma-control-plane ClusterIP 10.105.12.133 <none> 5681/TCP,443/TCP,5676/TCP,5677/TCP,5678/TCP,5679/TCP,5682/TCP,5653/UDP 90s
By default, it’s exposed on port 5685. In this example the value is 35.226.196.103:5685
. You pass this as the value of <global-kds-address>
when you set up the zone control planes.
Running global control plane in “Universal on Kubernetes” mode means using PostgreSQL as storage instead of Kubernetes.
It means that failover / HA / reliability characteristics will change.
Please read Kubernetes and
PostgreSQL docs for more details.
Before using Kuma with helm, please follow these steps to configure your local helm repo and learn the reference helm configuration values.yaml
.
-
Define Kubernetes secrets with database sensitive information
apiVersion: v1
kind: Secret
metadata:
name: your-secret-name
type: Opaque
data:
POSTGRES_DB: ...
POSTGRES_HOST_RW: ...
POSTGRES_USER: ...
POSTGRES_PASSWORD: ...
-
Create a values.yaml
file with: controlPlane.environment=universal
and controlPlane.mode=global
in the chart (values.yaml
).
-
Set controlPlane.secrets
with database sensitive information
# ...
secrets:
postgresDb:
Secret: your-secret-name
Key: POSTGRES_DB
Env: KUMA_STORE_POSTGRES_DB_NAME
postgresHost:
Secret: your-secret-name
Key: POSTGRES_HOST_RW
Env: KUMA_STORE_POSTGRES_HOST
postgrestUser:
Secret: your-secret-name
Key: POSTGRES_USER
Env: KUMA_STORE_POSTGRES_USER
postgresPassword:
Secret: your-secret-name
Key: POSTGRES_PASSWORD
Env: KUMA_STORE_POSTGRES_PASSWORD
-
Optionally set postgres
with TLS settings
# ...
# Postgres' settings for universal control plane on k8s
postgres:
# -- Postgres port, password should be provided as a secret reference in "controlPlane.secrets"
# with the Env value "KUMA_STORE_POSTGRES_PASSWORD".
# Example:
# controlPlane:
# secrets:
# - Secret: postgres-postgresql
# Key: postgresql-password
# Env: KUMA_STORE_POSTGRES_PASSWORD
port: "5432"
# TLS settings
tls:
# -- Mode of TLS connection. Available values are: "disable", "verifyNone", "verifyCa", "verifyFull"
mode: disable # ENV: KUMA_STORE_POSTGRES_TLS_MODE
# -- Whether to disable SNI the postgres `sslsni` option.
disableSSLSNI: false # ENV: KUMA_STORE_POSTGRES_TLS_DISABLE_SSLSNI
# -- Secret name that contains the ca.crt
caSecretName:
# -- Secret name that contains the client tls.crt, tls.key
secretName:
-
Run helm install
helm install kuma \
--create-namespace \
--skip-crds \
--namespace kuma-system \
--values values.yaml \
kuma/kuma
-
Find the external IP and port of the kuma-global-zone-sync
service in the kuma-system
namespace:
kubectl get services -n kuma-system
NAMESPACE NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kuma-system kuma-global-zone-sync LoadBalancer 10.105.9.10 35.226.196.103 5685:30685/TCP 89s
kuma-system kuma-control-plane ClusterIP 10.105.12.133 <none> 5681/TCP,443/TCP,5676/TCP,5677/TCP,5678/TCP,5679/TCP,5682/TCP,5653/UDP 90s
In this example the value is 35.226.196.103:5685
. You pass this as the value of <global-kds-address>
when you set up the zone control planes.
When running the global control plane in Universal mode, a database must be used to persist state for production deployments.
Ensure that migrations have been run against the database prior to running the global control plane.
-
Set up the global control plane, and add the global
environment variable:
KUMA_MODE=global \
KUMA_ENVIRONMENT=universal \
KUMA_STORE_TYPE=postgres \
KUMA_STORE_POSTGRES_HOST=<postgres-host> \
KUMA_STORE_POSTGRES_PORT=<postgres-port> \
KUMA_STORE_POSTGRES_USER=<postgres-user> \
KUMA_STORE_POSTGRES_PASSWORD=<postgres-password> \
KUMA_STORE_POSTGRES_DB_NAME=<postgres-db-name> \
kuma-cp run
Set up the zone control planes
You need the following values to pass to each zone control plane setup:
zone
– the zone name. An arbitrary string. This value registers the zone control plane with the global control plane.
kds-global-address
– the external IP and port of the global control plane.
On each zone control plane, run:
kumactl install control-plane \
--set "controlPlane.mode=zone" \
--set "controlPlane.zone=<zone-name>" \
--set "ingress.enabled=true" \
--set "controlPlane.kdsGlobalAddress=grpcs://<global-kds-address>:5685" \
| kubectl apply -f -
Before using Kuma with helm, please follow these steps to configure your local helm repo.
helm install --create-namespace --namespace kuma-system \
--set "controlPlane.mode=zone" \
--set "controlPlane.zone=<zone-name>" \
--set "ingress.enabled=true" \
--set "controlPlane.kdsGlobalAddress=grpcs://<global-kds-address>:5685" \
kuma kuma/kuma
where controlPlane.zone
is the same value for all zone control planes in the same zone.
Add --set egress.enabled=true
to list of arguments if you want to deploy optional
Zone Egress
.
When running the zone control plane in Universal mode, a database must be used to persist state for production deployments.
Ensure that migrations have been run against the database prior to running the zone control plane.
-
On each zone control plane, run:
KUMA_MODE=zone \
KUMA_MULTIZONE_ZONE_NAME=<zone-name> \
KUMA_ENVIRONMENT=universal \
KUMA_STORE_TYPE=postgres \
KUMA_STORE_POSTGRES_HOST=<postgres-host> \
KUMA_STORE_POSTGRES_PORT=<postgres-port> \
KUMA_STORE_POSTGRES_USER=<postgres-user> \
KUMA_STORE_POSTGRES_PASSWORD=<postgres-password> \
KUMA_STORE_POSTGRES_DB_NAME=<postgres-db-name> \
KUMA_MULTIZONE_ZONE_GLOBAL_ADDRESS=grpcs://<global-kds-address>:5685 \
kuma-cp run
where KUMA_MULTIZONE_ZONE_NAME
is the same value for all zone control planes in the same zone.
-
Generate the zone proxy token:
To register the zone ingress and zone egress with the zone control plane, we need to generate a token first
kumactl generate zone-token --zone=<zone-name> --scope egress --scope ingress > /tmp/zone-token
You can also generate the token
with the REST API
.
Alternatively, you could generate separate tokens for ingress and egress.
-
Create an ingress
data plane proxy configuration to allow kuma-cp
services to be exposed for cross-zone communication:
echo "type: ZoneIngress
name: ingress-01
networking:
address: 127.0.0.1 # address that is routable within the zone
port: 10000
advertisedAddress: 10.0.0.1 # an address which other zones can use to consume this zone-ingress
advertisedPort: 10000 # a port which other zones can use to consume this zone-ingress" > ingress-dp.yaml
-
Apply the ingress config, passing the IP address of the zone control plane to cp-address
:
kuma-dp run \
--proxy-type=ingress \
--cp-address=https://<kuma-cp-address>:5678 \
--dataplane-token-file=/tmp/zone-token \
--dataplane-file=ingress-dp.yaml
If zone-ingress is running on a different machine than zone-cp you need to
copy CA cert file from zone-cp (located in ~/.kuma/kuma-cp.crt
) to somewhere accessible by zone-ingress (e.g. /tmp/kuma-cp.crt
).
Modify the above command and provide the certificate path in --ca-cert-file
argument.
kuma-dp run \
--proxy-type=ingress \
--cp-address=https://<kuma-cp-address>:5678 \
--dataplane-token-file=/tmp/zone-token \
--ca-cert-file=/tmp/kuma-cp.crt \
--dataplane-file=ingress-dp.yaml
-
Optional: if you want to deploy zone egress
Create a ZoneEgress
data plane proxy configuration to allow kuma-cp
services
to be configured to proxy traffic to other zones or external services through
zone egress:
echo "type: ZoneEgress
name: zoneegress-01
networking:
address: 127.0.0.1 # address that is routable within the zone
port: 10002" > zoneegress-dataplane.yaml
-
Apply the egress config, passing the IP address of the zone control plane to cp-address
:
kuma-dp run \
--proxy-type=egress \
--cp-address=https://<kuma-cp-address>:5678 \
--dataplane-token-file=/tmp/zone-token \
--dataplane-file=zoneegress-dataplane.yaml
Verify control plane connectivity
If your global control plane runs on Kubernetes, you’ll need to configure your kumactl
like so:
# forward traffic from local pc into global control plane in the cluster
kubectl -n kuma-system port-forward svc/kuma-control-plane 5681:5681 &
# configure control plane for kumactl
kumactl config control-planes add \
--name global-control-plane \
--address http://localhost:5681 \
--skip-verify
You can run kumactl get zones
, or check the list of zones in the web UI for the global control plane, to verify zone control plane connections.
When a zone control plane connects to the global control plane, the Zone
resource is created automatically in the global control plane.
The Zone Ingress tab of the web UI also lists zone control planes that you
deployed with zone ingress.
Ensure mTLS is enabled on the multi-zone meshes
mTLS is mandatory to enable cross-zone service communication.
mTLS can be configured in your mesh configuration as indicated in the mTLS section.
This is required because Kuma uses the Server Name Indication field, part of the TLS protocol, as a way to pass routing information cross zones.
Cross-zone communication details
For this example we will assume we have a service running in a Kubernetes zone exposing a kuma.io/service
with value echo-server_echo-example_svc_1010
.
The following examples are running in the remote zone trying to access the previously mentioned service.
To view the list of service names available, run:
kubectl get serviceinsight all-services-default -oyaml
apiVersion: kuma.io/v1alpha1
kind: ServiceInsight
mesh: default
metadata:
name: all-services-default
spec:
services:
echo-server_echo-example_svc_1010:
dataplanes:
online: 1
total: 1
issuedBackends:
ca-1: 1
status: online
The following are some examples of different ways to address echo-server
in the
echo-example
Namespace
in a multi-zone mesh.
To send a request in the same zone, you can rely on Kubernetes DNS and use the usual Kubernetes hostnames and ports:
curl http://echo-server:1010
Requests are distributed round-robin between zones.
You can use
locality-aware load balancing
to keep requests in the same zone.
To send a request to any zone, you can
use the generated kuma.io/service
and Kuma DNS:
curl http://echo-server_echo-example_svc_1010.mesh:80
Kuma DNS also supports RFC 1123 compatible names, where underscores are replaced with dots:
curl http://echo-server.echo-example.svc.1010.mesh:80
SERVICE STATUS DATAPLANES
echo-service_echo-example_svc_1010 Online 1/1
To consume the service in a Universal deployment without transparent proxy add the following outbound to your
dataplane configuration
:
outbound:
- port: 20012
tags:
kuma.io/service: echo-server_echo-example_svc_1010
From the data plane running you will now be able to reach the service using localhost:20012
.
Alternatively, if you configure
transparent proxy
you can just call echo-server_echo-example_svc_1010.mesh
without defining an outbound
section.
For security reasons it’s not possible to customize the kuma.io/service
in Kubernetes.
If you want to have the same service running on both Universal and Kubernetes make sure to align the Universal’s data plane inbound to have the same kuma.io/service
as the one in Kubernetes or leverage
TrafficRoute
.
Delete a zone
To delete a Zone
we must first shut down the corresponding Kuma zone control plane instances. As long as the Zone CP is running this will not be possible, and Kuma returns a validation error like:
zone: unable to delete Zone, Zone CP is still connected, please shut it down first
When the Zone CP is fully disconnected and shut down, then the Zone
can be deleted. All corresponding resources (like Dataplane
and DataplaneInsight
) will be deleted automatically as well.
kubectl delete zone zone-1
kumactl delete zone zone-1
Disable a zone
Change the enabled
property value to false
in the global control plane:
apiVersion: kuma.io/v1alpha1
kind: Zone
metadata:
name: zone-1
spec:
enabled: false
type: Zone
name: zone-1
spec:
enabled: false
With this setting, the global control plane will stop exchanging configuration with this zone.
As a result, the zone’s ingress from zone-1 will be deleted from other zone and traffic won’t be routed to it anymore.
The zone will show as Offline in the GUI and CLI.