mqllr / storj-storagenode-chart

Chart Helm for storagenode STORJ

STORJ Storagenode

Helm chart to install a storj storagenode.

Prerequisites

Identity

The storage node needs an identity previously generated https://documentation.storj.io/dependencies/identity.

Storage

This helm chart use a StatefulSet object with a volumeClaimTemplates. If the storagenode run on a baremetal machine without a PersistentVolume controller (https://kubernetes.io/docs/concepts/storage/persistent-volumes/#types-of-persistent-volumes), you can use the no-provisioner:

---
kind: StorageClass
apiVersion: storage.k8s.io/v1
metadata:
  name: local-storage
provisioner: kubernetes.io/no-provisioner
volumeBindingMode: WaitForFirstConsumer

---
apiVersion: v1
kind: PersistentVolume
metadata:
  name: local-pv-storagenode1
spec:
  capacity:
    storage: 2Ti
  accessModes:
  - ReadWriteOnce
  persistentVolumeReclaimPolicy: Retain
  storageClassName: local-storage
  local:
    path: /data/disk1    # -> Where your disk is locally mounted
  nodeAffinity:
    required:
      nodeSelectorTerms:
      - matchExpressions:
        - key: kubernetes.io/hostname
          operator: In
          values:
            - mysweetmachine   # -> Set your node hostname

Sysctl Configuration for UDP

If you have service.quic enabled (the default), you will need to update the net.core.rmem_max sysctl value, or the storj storagenode will complain in the logs. Doing this on your worker node(s) will pass the setting through to the storj pod. Note that this may impact other pods as well, since it is being done at the node level. Please see https://docs.storj.io/node/dependencies/quic-requirements/linux-configuration-for-udp/ for full details and instructions.

Configuration

Parameter	Description	Default	Required
config.wallet	ERC20 address for payment	nil	yes
config.email	Email address used by Storj	nil	yes
config.address	(domain|ip):port for external communication	nil	yes
config.storage	Storage size allocated	1TB	no
service.storagenode.type	Service type for the storagenode	NodePort	no
service.storagenode.annotations	Service annotations	{}	no
service.storagenode.loadBalancerIP	Secify a Load balancer IP if the provider allow you	null	no
service.storagenode.port	Service port for the storagenode	28967	no
service.storagenode.nodePort	Node port to expose for the storagenode	""	no
service.stats.enabled	Expose the node's Dashboard	true	no
service.stats.type	Service type for the dashboard	ClusterIP	no
service.stats.port	Service port for the dashboard	14002	no
service.quic.enabled	Expose the storagenode's UDP port for quic	true	no
service.quic.type	Service type for the storagenode's UDP port	NodePort	no
service.quic.loadBalancerIP	Secify a Load balancer IP if the provider allow you	null	no
service.quic.port	Service port for the storagenode's UDP port	28967	no
service.quic.nodePort	Node port to expose for the storagenode's UDP port	""	no
replicaCount	Number of replica	1	no
podAnnotations	Annotations for the pod	{}	no
podSecurityContext	Custom security context	{}	no
nodeSelector	Node labels for pod assignment	{}	no
tolerations	Node taints to tolerate	{}	no
affinity	Pod affinity	{}	no
imagePullSecrets		{}	no
storagenode.image.repository	Image name	storjlabs/storagenode:latest	no
storagenode.image.pullPolicy	Container pull policy	Always	no
storagenode.securtyContext	Custom security context for container	{}	no
storagenode.resources	Resources request and limit YAML	{}	no
identity.externalSecret.secretName	Specify the secretName	""	yes
persistence.enabled	Create a persistence volume	true	no
persistence.annotations	Persistent volume claim annotation	{}	no
persistence.volumeClaimTemplate	Persistent volume claim YAML	{}	no
metrics.enabled	Start the container storj-exporter	true	no
metrics.image.repository	Image name	anclrii/storj-exporter:1.0.10	no
metrics.image.pullPolicy	Container pull policy	IfNotPresent	no
metrics.securtyContext	Custom security context for container	{}	no
metrics.resources	Resources request and limit YAML	{}	no
ingress.enabled	If true, an ingress object will be created to expose the stat dashboard only	false	no
ingress.annotations	Annotations for the ingress	{}	no
ingress.hosts	Ingress hostname	[]	no
ingress.tls	Ingress TLS	[]	no

Installing the Chart

1. Add the helm repo

helm repo add storj-storagenode-chart https://mqllr.github.io/storj-storagenode-chart

2. Create your identity

Create a secret in kubernetes with your node identity

curl -L https://github.com/MqllR/storj-storagenode-chart/releases/download/identity/identity-to-kube-secret-amd64 -o identity-to-kube-secret
./identity-to-kube-secret -secret-name "storj-identity-node1" | kubectl apply -f -

3. Install the chart

helm install node storj-storagenode-chart/storj-storagenode --set [email protected],config.wallet=0xdfca4035b9f16c40b558218d1bedc08590fe28d4,config.address=mydomain.net:28967,identity.externalSecret.secretName="storj-identity-node1"

storj-storagenode-chart's Issues

How to run the setup process?

According to this https://docs.storj.io/node/setup/cli/storage-node#setting-up-the-storage-node
looks like I need to set env:SETUP=true first before running the container. I think I'm having crashloops because of this. Any advice on how to get this going? or otherwise I'm happy to make a PR to add another variable to set env variables.

Error: Error starting master database on storagenode: group:
--- stat config/storage/blobs: no such file or directory
--- stat config/storage/temp: no such file or directory
--- stat config/storage/garbage: no such file or directory
--- stat config/storage/trash: no such file or directory

Ingress error on k3s v1.22.7+k3s1

I am getting this error after enabling ingress in chart:

Helm install failed: unable to build kubernetes objects from release manifest: unable to recognize "": no matches for kind "Ingress" in version "networking.k8s.io/v1beta1"

Exporter errors cause pod to Crashloop

Errors like anclrii/Storj-Exporter#57 cause the whole pod to go into a crashloop.

I think this is preventing the node from being online just because the exporter has an issue. Can the exporter be broken out into it's own pod so the uptime of the node is not impacted by the availability of the exporter, or am I misunderstanding the impact here?