CENM Deployment with Docker, Kubernetes, and Helm charts

CENM docker images are based on azul/zulu-openjdk:8u242

Each CENM service has dedicated docker image. They are designed to be minimal and optimized to run on Kubernetes cluster. They are not designed to run as standalone docker containers. They are stored in Docker Hub:

CENM serviceDocker HubTag
Identity Managercorda/enterprise-identitymanager1.2-zulu-openjdk8u242
Network Mapcorda/enterprise-networkmap1.2-zulu-openjdk8u242
PKI Toolcorda/enterprise-pkitool1.2-zulu-openjdk8u242
Signercorda/enterprise-signer1.2-zulu-openjdk8u242
Notarycorda/notary1.2-zulu-openjdk8u242

All helm charts by default use CENM docker images with tag 1.2-zulu-openjdk8u242.

The following requirements are needed in order to deploy CENM correctly:

  • These charts are supposed to be a reference installation.
  • They allow to configure several variables related to each CENM service.

These charts are compatible with Corda Enterprise Network Manager (CENM) version 1.2. Earlier CENM releases are not supported.

Each CENM service has its own dedicated folder with more detailed documentation.

Helm Chart
Identity Manager
Signer
Network Map
Corda Notary

The charts are currently developed and tested against Azure Kubernetes Service (AKS).

Recommended production specification for CENM components running on separate VMs:

  • 2 CPU cores, 8 GB Memory

All CENM 1.2 services run inside a single, dedicated Kubernetes namespace (the default name is cenm). Each service runs in its own dedicated Kubernetes pod.

The CENM network is bootstrapped with PKI certificates which default to sample X.500 subject names (e.g. Identity Manager certificate subject is “CN=Test Identity Manager Service Certificate, OU=HQ, O=HoldCo LLC, L=New York, C=US”). The subject names of the whole PKI Certificate Hierarchy can be configured in the Signer Helm chart. For more information about Signer helm chart refer to Signer.

There are three ways of bootstrapping a new CENM environment:

  1. scripted (bootstrap.cenm) with allocating new public IP addresses
  2. scripted (bootstrap.cenm) with reusing already allocated public IP addresses
  3. manually run each helm command

In case of initial bootstrap process R3 recommends to use the first method.

The first two bootstrapping methods uses default values stored in values.yaml file in each helm chart. To fully customize new environment use third method.

Note: Identity Manager and Notary require their public IPs to be known in advance as this is required while running PKI Tool (part of Signer helm chart) and Notary registration.

Be aware that it might take a few minutes to allocate new IP address. In case of subsequent bootstrapping it is possible to reuse existing external IP addresses. NetworkMap and Signer have their public IP addresses allocated while bootstrapping and they don’t need to be known in ahead of time.

Public IP addresses are allocated using Kubernetes Service defined with type: LoadBalancer

The following table represents memory requirements for each CENM component based on default value of cordaJarMx from values.yaml in each chart:

ComponentcordaJarMx (GB)max memory for service JVM (-Xmx) (GB)K8s requests (GB)K8s limits (GB)
Identity Manager1cordaJarMxcordaJarMxcordaJarMx + 2
Signer1cordaJarMxcordaJarMxcordaJarMx + 2
Network Map1cordaJarMxcordaJarMxcordaJarMx + 2
Notary3cordaJarMxcordaJarMxcordaJarMx + 2

Note: Kubernetes cluster should have at least 8 GB of free memory available to all CENM services.

Before proceeding you will need the following:

  1. all required dependencies installed
  2. an AKS cluster up and running and access to it from your local machine
  3. a storage class (cenm) and new namespace (cenm) with the correct RBAC permissions.
  4. obtain the helm charts and deployment scripts

Ensure that helm version gives you version field with any value greater than or equal to 3.1.1 - for instance:

version.BuildInfo{Version:"v3.1.2", GitCommit:"afe70585407b420d0097d07b21c47dc511525ac8", GitTreeState:"clean", GoVersion:"go1.13.8"}

You can find the files required in the following steps on the CENM deployment repo.

Run the following instruction once the previous points have been cleared. These examples use the namespace cenm:

Note: edit k8s/cenm.yaml and modify RoleBinding/everything-in-cenm to include your user(s) Azure group

kubectl apply -f k8s/cenm.yaml  # Run as Kubernetes privileged user
export nameSpace=cenm
kubectl config set-context $(kubectl config current-context) --namespace=${nameSpace}

You can verify this with kubectl get ns


To bootstrap new CENM environment with allocating new, external IP run:

cd k8s/helm
./bootstrap.cenm

Note: obtaining IP addresses might take up to 10 minutes

The script exits after all bootstrapping processes on Kubernetes cluster have been started. The process will continue to run on the cluster after the script has exited. You can monitor the completion of the deployment with:

kubectl get pods -o wide

It is possible that external IPs have been already allocated - in this case it is possible to reuse them by specifying their services names:

cd k8s/helm
./bootstrap.cenm -i idman-ip -n notary-ip

There are several helm commands used to bootstrap new CENM environment, each one creates one CENM service (signer, identity manager, etc). They need to be run in the correct order.

cd k8s/helm

# these helm charts trigger public IP allocation
helm install idman-ip idman-ip
helm install notary-ip notary-ip

# run these commands to display allocated public IP addresses:
kubectl get svc --namespace cenm idman-ip --template "{{ range (index .status.loadBalancer.ingress 0) }}{{.}}{{ end }}"   # step 1
kubectl get svc --namespace cenm notary-ip --template "{{ range (index .status.loadBalancer.ingress 0) }}{{.}}{{ end }}"  # step 2

# these helm charts bootstrap all CENM services
helm install signer signer --set idmanPublicIP=[use IP from step 1]
helm install idman idman
helm install notary notary --set notaryPublicIP=[use IP from step 2]
helm install nmap nmap

# run these commands to display allocated public IP addresses for Signer and NetworkMap:
kubectl get svc --namespace cenm signer --template "{{ range (index .status.loadBalancer.ingress 0) }}{{.}}{{ end }}"
kubectl get svc --namespace cenm nmap --template "{{ range (index .status.loadBalancer.ingress 0) }}{{.}}{{ end }}"

To access Identity Manager, Signer, Network Map or Notary use the following instruction:

ssh -p <shell.sshdPort> -l <shell.user> <IP Address>
  • IP addresses are dynamically allocated for each deployment and can be found with kubectl get svc
  • ssh port and user are specified in Helm charts configurations for each service
  • you will be asked a password for each service, which can be found Helm charts configurations for each service

Note: make sure that you are pointing at the correct namespace - check with:

kubectl config current-context && kubectl config view --minify --output 'jsonpath={..namespace}' && echo`)

To configure Corda node to connect to the CENM services, edit the following properties in your node.conf file:

networkServices {
  doormanURL="http://<IDENTITY-MANAGER-IP>:10000"
  networkMapURL="http://<NETWORK-MAP-IP>:10000"
}
  • the doormanURL property is the public IP address and port of the Identity Manager Service.
  • the networkMapURL is the public IP address and port of the Network Map Service.

Note: to obtain public IPs of Identity Manager and Network Map use:

kubectl get svc idman-ip notary-ip

Also upload the network-root-truststore.jks to your Corda node. You can download it locally from CENM Signer using the following command:

kubectl cp <name-space>/<signer-pod>:DATA/trust-stores/network-root-truststore.jks network-root-truststore.jks

Namespace is set to cenm in this deployment, to obtain the pod name for the signer use:

kubectl get pods -o wide`

Truststore password can be found in the signer/files/pki.conf, the default value used in this helm chart: trust-store-password

For more details about joining CENM network see: Joining an existing compatibility zone.

Each CENM service has a dedicated sidecar to displays live logs from the log/ directory.

To display logs use the following command:

kubectl logs -c <logs-container> <pod-name>

The <logs-container> object container determines where the logs will be drawn from. The services write their logs to dedicated display containers where you can get a live view of the logs:

kubectl logs -c logs-auth <pod-name>   //for auth
kubectl logs -c logs-gateway <pod-name>   //for gateway
kubectl logs -c logs-idman <pod-name>   //for identity manager
kubectl logs -c logs-nmap <pod-name>   //for network map
kubectl logs -c logs-signer <pod-name>   //for signer
kubectl logs -c logs-zone <pod-name>   //for zone

For notary, PKI, and HSM, the command is slightly different as logs containers do not exist for these services:

kubectl logs <pod-name>

To display live logs use the following command:

kubectl logs -c -f <logs-container> <pod-name>

Each service stores configuration file in etc/ folder in a pod. I.e.: to display what is in the Identity Manager etc/ folder run these commands:

kubectl exec -it <pod name> -- ls -al etc/
Defaulting container name to main.
Use 'kubectl describe pod/idman-7699c544dc-bq9lr -n cenm' to see all of the containers in this pod.
total 10
drwxrwxrwx 2 corda corda    0 Feb 11 09:29 .
drwxr-xr-x 1 corda corda 4096 Feb 11 09:29 ..
-rwxrwxrwx 1 corda corda 1871 Feb 11 09:29 idman.conf

kubectl exec -it <pod name> -- cat etc/idman.conf
Defaulting container name to main.
Use 'kubectl describe pod/idman-7699c544dc-bq9lr -n cenm6' to see all of the containers in this pod.

address = "0.0.0.0:10000"
database {
...

Default setting used in CENM services configuration values can be altered by:

  1. changing values in values.yaml
  2. preparing another yaml file with new values and passing it with -f flag (i.e.: helm install -f myvalues.yaml idman)
  3. individual parameters passed with --set (such as helm install --set foo=bar idman)
  4. any combination of the above (i.e.: helm install -f myvalues.yaml --set foo=bar idman)

For more information refer to the official helm documentation: https://helm.sh/docs/chart_template_guide/values_files/

There are two ways to destroy CENM environment:

  • destroy the whole environment
  • destroy all CENM objects without deleting allocated external IP addresses
helm delete nmap notary idman signer notary-ip idman-ip

If you run several ephemeral test network in your development cycle you might want to keep your IP addresses to speed up the process:

helm delete nmap notary idman signer

CENM 1.2 allows you to update network parameters without restarting Network Map. Take the following steps to update network parameters:

  • Login to Network Map pod and edit etc/network-parameters-update-example.conf file:
kubectl exec -it [name of nmap pod] bash
vim etc/network-parameters-update-example.conf
[update the file, save and exit]
  • Connect to Network Map ssh console:
ssh -p [nmap ssh port] -l nmap [nmap public IP address]

and run the following commands:

run networkParametersRegistration networkParametersFile: etc/network-parameters-update-example.conf, \
networkTrustStore: DATA/trust-stores/network-root-truststore.jks, \
trustStorePassword: trust-store-password, \
rootAlias: cordarootca

Once the set date/time has passed run the following in the Network Map ssh console:

run flagDay

Note: For the changes to be advertised to the nodes the new network map has to be signed by signer and it is scheduled according to its configuration.

To cancel flag day:

run cancelUpdate

Note: The following files are part of the default deployment:

etc/network-parameters-update-example.conf
DATA/trust-stores/network-root-truststore.jks

Visit CENM official documentation for more information about network parameters:

Was this page helpful?

Thanks for your feedback!

Chat with us

Chat with us on our #docs channel on slack. You can also join a lot of other slack channels there and have access to 1-on-1 communication with members of the R3 team and the online community.

Propose documentation improvements directly

Help us to improve the docs by contributing directly. It's simple - just fork this repository and raise a PR of your own - R3's Technical Writers will review it and apply the relevant suggestions.

We're sorry this page wasn't helpful. Let us know how we can make it better!

Chat with us

Chat with us on our #docs channel on slack. You can also join a lot of other slack channels there and have access to 1-on-1 communication with members of the R3 team and the online community.

Create an issue

Create a new GitHub issue in this repository - submit technical feedback, draw attention to a potential documentation bug, or share ideas for improvement and general feedback.

Propose documentation improvements directly

Help us to improve the docs by contributing directly. It's simple - just fork this repository and raise a PR of your own - R3's Technical Writers will review it and apply the relevant suggestions.