Highly available Vault Enterprise disaster recovery clusters with integrated storage (Raft)

Important Note: This chart is not compatible with Helm 2. Please use Helm 3.6+ with this chart.

The following is an example of creating a disaster recovery cluster using Vault Helm.

For more information on Disaster Recovery, see the official documentation.

For license configuration refer to Running Vault Enterprise.

Primary cluster

First, create the primary cluster:

helm install vault-primary hashicorp/vault \
  --set='server.image.repository=hashicorp/vault-enterprise' \
  --set='server.image.tag=1.18.1-ent' \
  --set='server.ha.enabled=true' \
  --set='server.ha.raft.enabled=true'

Next, initialize and unseal vault-primary-0 pod:

kubectl exec -ti vault-primary-0 -- vault operator init
kubectl exec -ti vault-primary-0 -- vault operator unseal

Finally, join the remaining pods to the Raft cluster and unseal them. The pods will need to communicate directly so we'll configure the pods to use the internal service provided by the Helm chart:

kubectl exec -ti vault-primary-1 -- vault operator raft join http://vault-primary-0.vault-primary-internal:8200
kubectl exec -ti vault-primary-1 -- vault operator unseal

kubectl exec -ti vault-primary-2 -- vault operator raft join http://vault-primary-0.vault-primary-internal:8200
kubectl exec -ti vault-primary-2 -- vault operator unseal

To verify if the Raft cluster has successfully been initialized, run the following.

First, login using the root token on the vault-primary-0 pod:

kubectl exec -ti vault-primary-0 -- vault login

Next, list all the raft peers:

$ kubectl exec -ti vault-primary-0 -- vault operator raft list-peers

Node                                    Address                        State       Voter
----                                    -------                        -----       -----
a1799962-8711-7f28-23f0-cea05c8a527d    vault-primary-0.vault-primary-internal:8201    leader      true
e6876c97-aaaa-a92e-b99a-0aafab105745    vault-primary-1.vault-primary-internal:8201    follower    true
4b5d7383-ff31-44df-e008-6a606828823b    vault-primary-2.vault-primary-internal:8201    follower    true

Secondary cluster

With the primary cluster created, next create a secondary cluster and enable disaster recovery replication.

helm install vault-secondary hashicorp/vault \
  --set='server.image.repository=hashicorp/vault-enterprise' \
  --set='server.image.tag=1.18.1-ent' \
  --set='server.ha.enabled=true' \
  --set='server.ha.raft.enabled=true'

Next, initialize and unseal vault-secondary-0 pod:

kubectl exec -ti vault-secondary-0 -- vault operator init
kubectl exec -ti vault-secondary-0 -- vault operator unseal

Finally, join the remaining pods to the Raft cluster and unseal them. The pods will need to communicate directly so we'll configure the pods to use the internal service provided by the Helm chart:

kubectl exec -ti vault-secondary-1 -- vault operator raft join http://vault-secondary-0.vault-secondary-internal:8200
kubectl exec -ti vault-secondary-1 -- vault operator unseal

kubectl exec -ti vault-secondary-2 -- vault operator raft join http://vault-secondary-0.vault-secondary-internal:8200
kubectl exec -ti vault-secondary-2 -- vault operator unseal

To verify if the Raft cluster has successfully been initialized, run the following.

First, login using the root token on the vault-secondary-0 pod:

kubectl exec -ti vault-secondary-0 -- vault login

Next, list all the raft peers:

$ kubectl exec -ti vault-secondary-0 -- vault operator raft list-peers

Node                                    Address                        State       Voter
----                                    -------                        -----       -----
a1799962-8711-7f28-23f0-cea05c8a527d    vault-secondary-0.vault-secondary-internal:8201    leader      true
e6876c97-aaaa-a92e-b99a-0aafab105745    vault-secondary-1.vault-secondary-internal:8201    follower    true
4b5d7383-ff31-44df-e008-6a606828823b    vault-secondary-2.vault-secondary-internal:8201    follower    true

Enable disaster recovery replication on primary

With the initial clusters setup, we can now configure them for disaster recovery replication.

First, on the primary cluster, enable replication:

kubectl exec -ti vault-primary-0 -- vault write -f sys/replication/dr/primary/enable primary_cluster_addr=https://vault-primary-active:8201

Next, create a token the secondary cluster will use to configure replication:

kubectl exec -ti vault-primary-0 -- vault write sys/replication/dr/primary/secondary-token id=secondary

The token in the output will be used when configuring the secondary cluster.

Enable disaster recovery replication on secondary

Using the token created in the last step, enable disaster recovery replication on the secondary:

kubectl exec -ti vault-secondary-0 -- vault write sys/replication/dr/secondary/enable token=<TOKEN FROM PRIMARY>

Last, delete the remainder secondary pods and unseal them using the primary unseal token after Kubernetes reschedules them:

kubectl delete pod vault-secondary-1
kubectl exec -ti vault-secondary-1 -- vault operator unseal <PRIMARY UNSEAL TOKEN>

kubectl delete pod vault-secondary-2
kubectl exec -ti vault-secondary-2 -- vault operator unseal <PRIMARY UNSEAL TOKEN>