Couchbase cluster upgrade to 2.3 failing

Couchbase Server Kubernetes
1

Hi,

I am trying to upgrade our couchbase clusters to 2.3 from 2.1 and keep on seeing the error:
spec.version in body cannot be updated

I have tried to copy as much the same yaml configs as we used to deploy 2.1.

My sample Yaml:

cluster:
image: “couchbase/couchbase-server:7.0.3”
recoveryPolicy: PrioritizeDataIntegrity
security:
adminSecret: auth-dev-cluster-couchbase-cluster
rbac:
selector:
matchLabels:
cluster: dev-cluster-couchbase-cluster
securityContext:
fsGroup: 1000
runAsNonRoot: true
runAsUser: 1000
networking:
exposeAdminConsole: true
adminConsoleServices:
- data
exposedFeatures:
- client
- xdcr
exposedFeatureTrafficPolicy: Local
exposedFeatureServiceType: NodePort
networkPlatform: Istio
adminConsoleServiceType: NodePort
xdcr: {}
monitoring:
prometheus:
image: “couchbase/couchbase-exporter:1.0.3”
enabled: true
backup:
managed: true
image: couchbase/operator-backup:1.2.0
serviceAccountName: backup-dev-cluster-couchbase-cluster
logging:
logRetentionCount: 20
logRetentionTime: 604800s
cluster:
autoFailoverMaxCount: 3
autoFailoverOnDataDiskIssues: true
autoFailoverOnDataDiskIssuesTimePeriod: 2m0s
autoFailoverTimeout: 2m0s
dataServiceMemoryQuota: 5Gi
indexServiceMemoryQuota: 1Gi
searchServiceMemoryQuota: 1Gi
eventingServiceMemoryQuota: 256Mi
analyticsServiceMemoryQuota: 1Gi
indexStorageSetting: plasma
indexer:
storageMode: plasma
autoCompaction:
databaseFragmentationThreshold:
percent: 25
size: 1Gi
viewFragmentationThreshold:
percent: 25
size: 1Gi
tombstonePurgeInterval: 48h0m0s
timeWindow: {}
buckets:
selector:
matchLabels:
cluster: dev-cluster-couchbase-cluster
enablePreviewScaling: true
hibernationStrategy: Immediate
servers:
data:
autoscaleEnabled: true
pod:
metadata: {}
spec:
containers:
resources:
limits:
cpu: “2”
memory: 6Gi
requests:
memory: 2Gi
services:
- data
size: 5
volumeMounts:
default: default
data: data
index_query:
autoscaleEnabled: true
pod:
metadata: {}
spec:
containers:
resources:
limits:
cpu: “2”
memory: 2Gi
requests:
memory: 500Mi
services:
- index
- query
size: 2
volumeMounts:
default: default
index: index
default:
autoscaleEnabled: true
pod:
metadata: {}
spec:
containers:
env:
envFrom: null
resources:
limits:
cpu: “2”
memory: 2Gi
requests:
memory: 500Mi
size: 1
services:
- search
- eventing
volumeMounts:
default: default
volumeClaimTemplates:

  • metadata:
    name: data
    spec:
    accessModes:
    • ReadWriteOnce
      resources:
      requests:
      storage: 5Gi
      storageClassName: pure-block
  • metadata:
    name: index
    spec:
    accessModes:
    • ReadWriteOnce
      resources:
      requests:
      storage: 5Gi
      storageClassName: pure-block
  • metadata:
    name: default
    spec:
    accessModes:
    • ReadWriteOnce
      resources:
      requests:
      storage: 5Gi
      storageClassName: pure-block
2

Hello!

How are you attempting to upgrade operator? Is it via Helm?

Thanks,

Justin Ashworth

3

yes… operator and DAC upgraded fine but cluster is what is throwing error

4

That’s weird, we don’t have a spec.version field. Are you sure it is the DAC or Operator that is rejecting your yaml?

5

This is the full error:

Error: UPGRADE FAILED: cannot patch “dev-cluster-couchbase-cluster” with kind CouchbaseCluster: admission webhook “couchbase-admission-controller-couchbase-admission-controller.default.svc” denied the request: validation failure list:
spec.version in body cannot be updated

Let me know if you want me to run it with debug

6

Can you do helm show chart couchbase/couchbase-operator and let me know the output?

7

apiVersion: v2
appVersion: 2.5.0
description: A Helm chart to deploy the Couchbase Autonomous Operator for easily deploying,
managing, and maintaining Couchbase Clusters. Couchbase Server is a NoSQL document
database with a distributed architecture for performance, scalability, and availability.
It enables developers to build applications easier and faster by leveraging the
power of SQL with the flexibility of JSON.
home: https://couchbase.com
icon: https://couchbase-partners.github.io/helm-charts/assets/img/CB_Logomark_220.png
keywords:

And this is what is currently deployed:

list
NAME NAMESPACE REVISION UPDATED STATUS CHART APP VERSION
dev-cluster cb 78 2023-09-07 16:24:03.4779279 -0400 EDT failed couchbase-operator-2.3.002 2.3.0
operator cb 1 2023-09-06 17:07:21.5719857 -0400 EDT deployed couchbase-operator-2.3.002 2.3.0

8

Did you perform a helm upgrade on the dev-cluster release?

9

Thats what we keep getting error on:

helm -n cb upgrade dev-cluster couchbase/couchbase-operator --version 2.3.002 -f dev-cluster.yaml --set cluster.security.password=<>

Error: UPGRADE FAILED: cannot patch “dev-cluster-couchbase-cluster” with kind CouchbaseCluster: admission webhook “couchbase-admission-controller-couchbase-admission-controller.default.svc” denied the request: validation failure list:
spec.version in body cannot be updated

10

Did you update the crds via:

helm pull couchbase/couchbase-operator
tar -xf couchbase-operator-2.3.002.tbz
kubectl apply -f couchbase-operator/crds/couchbase.crds.yaml
11

Nope… Do I need to?

And if yes, what needs to be updated?

12

You can find the documentation for updating a release here
https://docs.couchbase.com/operator/2.3/helm-managing-guide.html#upgrading-the-autonomous-operator

The issue appears to be a 2.1 CRD with a 2.3 Operator/DAC, which would cause issues.

1 Like
13

That did the trick. I updated the crds and cluster upgraded.
Thanks for your help

14

Great! Thanks for reaching out!