Skip to main content

Upgrading a Validator

How to perform an upgrade on a validator

Helm Chart details:

Dynamic YAML Badge

validator: A Helm chart for deploying Chronicle Validator on Kubernetes

— Open in Artifact Hub

Install CRD's​

Starting from Chart Version 0.3.4, tor is deployed using the tor-controller operator, which installs some custom resource definitions. The controller will create a new onion key, which will be persisted as a secret. Please delete your previous secrets containing the tor keys, as they won't be needed. Retrieve the Ghost onion address using kubectl get onion -n <namespace> and notify the Chronicle team of your ETH address and the new Ghost onion address.

If you are running an upgrade from a prior release (< 0.3.4), chances are that Tor Custom Resource Definitions haven't been installed. Helm does not like installing CRD's during a helm upgrade, so we need to manually apply the CRD's like this:

kubectl apply -f https://raw.githubusercontent.com/chronicleprotocol/charts/validator-0.3.6/charts/validator/crds/tor-controller.yaml

It can take a few moments for the tor-controller to be in a ready state, but please make sure its running before upgrading your validator:

kubectl get pods -n tor-controller-system

You should see something like this:

NAME                                                 READY   STATUS    RESTARTS   AGE
tor-controller-controller-manager-6648f44cc8-g6c68 2/2 Running 0 16m

We now have the CRD's deployed (ie kubectl get crds will show the tor custom resource definitions), and our values.yaml updated, we can perform the upgrade:

Upgrading manually (helm upgrade)

Upgrading manually (helm upgrade)​

If you are upgrading from 0.3.x to 0.3.y, simply updating the chart version will suffice:

ssh <SERVER_IP>
su - <FEED_USERNAME>
export FEED_NAME=my-feed

Prepare values​

The values.yaml file is used to configure the validator. The file is generated by the install script, and should be updated to reflect the latest version of the feed chart.

With the latest version of the chart, there are a few changes that need to be made to the values.yaml / generated-values.yaml file:

warning
  • musig is now embedded in the ghost deployment, and all .Values.musig can be removed from the values.yaml file
  • Please remove .Values.ghost.env.CFG_WEB_URL from your values, as this will be dynamically referenced in the Ghost deployment spec.

Please structure your helm values like this:

ghost:
ethConfig:
ethFrom:
existingSecret: '<somesecret>'
key: "ethFrom"
ethKeys:
existingSecret: '<somesecret>'
key: "ethKeyStore"
ethPass:
existingSecret: '<somesecret>'
key: "ethPass"

env:
normal:
CFG_LIBP2P_EXTERNAL_ADDR: '/ip4/1.2.3.4' # public/reachable ip address of node

ethRpcUrl: "https://MY_L1_RPC_URL"
rpcUrl: "https://MY_L1_RPC_URL"

danger

Please ensure your values yaml file is updated to reflect the latest requirements for the validator chart, with the correct values for ethConfig, ethRpcUrl and rpcUrl.

danger

Make sure the TOR crds are installed.

helm repo update
helm upgrade $FEED_NAME -n $FEED_NAME -f $HOME/$FEED_NAME/generated-values.yaml chronicle/validator --version 0.3.6
danger

If upgrading from 0.2.x to 0.3.x, please use the helper script, or manually update your generated-values.yaml as per the steps above

Upgrading using the helper script (upgrade.sh)

Upgrading using upgrade.sh​

warning

Please be aware that the latest helm chart has been renamed from feed to validator. Please use the upgrade.sh script to upgrade your validator to the latest version. This version embeds musig into the ghost pod. The upgrader script will clean up the generated values.yaml file and remove the unecessary musig values.

To simplify the upgrade process, we have created a helper script that will upgrade your validator to the latest version.

This script will attempt to run helm upgrade <feedname> -n <feedname> chronicle/validator on your feed release, with any updated input variables.

caution

Please use the correct FEED_NAME, which should be the same as your helm release name, if deployed using the install.sh script previously

ssh <SERVER_IP>
su - <FEED_USERNAME>
export FEED_NAME=my-feed
danger

Make sure the TOR crds are installed.

Download the latest upgrade.sh​

Get the latest upgrade.sh script:

wget -N https://raw.githubusercontent.com/chronicleprotocol/scripts/main/feeds/k3s-install/upgrade.sh
chmod a+x upgrade.sh
./upgrade.sh
You can set the expected variables in the .env file, or export them as environment variables. If the script fails to find any of these values, it will prompt you for them when running the script.

tip

If kubectl/helm commands fail, please ensure you have $KUBECONFIG set correctly. Take a look here for more detail

Verify the helm release and version​

Verify the chart version has changed and matches what the latest feed version:

helm list -n $FEED_NAME
NAME NAMESPACE REVISION UPDATED STATUS CHART APP VERSION
validator validator 2 2024-08-27 23:44:34.801698 -0300 -03 deployed validator-0.3.6 0.42.2

View all resources created in the namespace​

kubectl get pods,deployment,service,secrets,onion -n demo
NAME READY STATUS RESTARTS AGE
pod/ghost-bc68cd645-tnvl9 1/1 Running 0 65s
pod/ghost-tor-daemon-5bcbbf86b8-5gjh2 1/1 Running 0 5m23s

NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/ghost 1/1 1 1 5m23s
deployment.apps/ghost-tor-daemon 1/1 1 1 5m23s

NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/ghost LoadBalancer 10.43.56.4 172.18.0.3 8000:31288/TCP,9100:30705/TCP,8080:30220/TCP 5m23s
service/ghost-tor-metrics-svc ClusterIP 10.43.74.109 <none> 9035/TCP 5m23s
service/ghost-tor-svc ClusterIP 10.43.63.20 <none> 8888/TCP 5m23s

NAME TYPE DATA AGE
secret/eth-keys Opaque 3 7m50s
secret/ghost-tor-auth tor.k8s.torproject.org/authorized-clients-v3 0 5m23s
secret/ghost-tor-secret tor.k8s.torproject.org/onion-v3 5 5m23s
secret/sh.helm.release.v1.validator.v1 helm.sh/release.v1 1 5m23s
secret/sh.helm.release.v1.validator.v2 helm.sh/release.v1 1 65s

NAME HOSTNAME AGE
onionservice.tor.k8s.torproject.org/ghost mhxzzlxk46vfht7t4mlawtpv6f6uaplh74bwdgchmvmtof7gf45juyqd.onion 5m23s

View pod logs:​

kubectl logs -n demo deployment/ghost
kubectl logs -n demo deployment/ghost-tor-daemon

and you're done!

warning

If you encounter any issues please refer to the Trouble Shooting docs