Unable to deploy more than one cluster with Operator

[ognjen-it]

Hi everyone,
I deployed arangodb-operator-crd and arangodb-operator to manage my ArangoDB cluster.
When I deploy the first cluster, it was working. However, when I tried to deploy the second one, I sow next error message:

DBG Failed to fetch deployment health error="JWT Secret is missing" component=deployment deployment=ognjenarangodb-cluster operator-id=jb7p9

If I create the secret manually, the operator will show the next error (service does not exist):

2021-07-20T08:46:52Z DBG Failed to set number of servers error="Put \"https://ognjenarangodb-cluster.arangodb-operator.svc:8529/_admin/cluster/numberOfServers\": dial tcp: lookup ognjenarangodb-cluster.arangodb-operator.svc on 169.254.25.10:53: no such host" component=deployment deployment=ognjenarangodb-cluster operator-id=jb7p9
2021-07-20T08:46:52Z DBG Cluster update failed error="Put \"https://ognjenarangodb-cluster.arangodb-operator.svc:8529/_admin/cluster/numberOfServers\": dial tcp: lookup ognjenarangodb-cluster.arangodb-operator.svc on 169.254.25.10:53: no such host" component=deployment deployment=ognjenarangodb-cluster operator-id=jb7p9

How I deployed arangodb-operator-crd?

helm -n arangodb-operator install arangodb-operator-crd https://github.com/arangodb/kube-arangodb/releases/download/1.1.10/kube-arangodb-crd-1.1.10.tgz

How I deployed arangodb-operator?

helm -n arangodb-operator install ts https://github.com/arangodb/kube-arangodb/releases/download/1.1.10/kube-arangodb-1.1.10.tgz

How I deployed cluster?

apiVersion: "database.arangodb.com/v1alpha"
kind: "ArangoDeployment"
metadata:
  name: "ognjenarangodb-cluster"
  namespace: arangodb-operator
spec:
  mode: Cluster
  environment: Production
  image: "arangodb/arangodb:3.7.10"
  imagePullPolicy: IfNotPresent
  agents:
    count: 3
    args:
      - --log.level=debug
    resources:
      requests:
        storage: 4Gi
    storageClassName: rook-ceph-block
  dbservers:
    count: 3
    resources:
      requests:
        storage: 8Gi
    storageClassName: rook-ceph-block
  coordinators:
    count: 3

Kubernetes version:
v1.18.8

[ajanikow]

Hello!

This error should be ignored, health inspector started before cluster has been created.

Fix for this will be in next patch release, there is problem with ArangoMember - 2 deployments in same NS just collide with each other.

Expected version: 1.2.1 - already in testing

Best Regards,
Adam.

[ognjen-it]

@ajanikow
Thank you for your fast answare!

I have one more question/suggestion/problem.
If I deploy arangodb-operator in the arangodb-operator-ns namespace, I cannot deploy clusters in different namespaces.
Do you know if it's possible to do with the current version of arangodb-operator? :)

[ajanikow]

Hello!

Issue was closed by PR by mistake.

Currently it is not supported due to RBAC. We deploy 1 Operator per namespace, what allows us to manage operator version in separated way per namespace (so, for example, stage).

[ognjen-it]

Thank you!