Flag webhook-timeout provided but not defined for Chart spark-operator-1.1.12

[Taqas]

When pulling the Helm Chart from https://googlecloudplatform.github.io/spark-on-k8s-operator and installing the latest (=spark-operator-1.1.12) I run into flag provided but not defined: -webhook-timeout

Most likely because of this commit.

When I install the Chart from source it works without problems.

helm list
NAME      	NAMESPACE     	REVISION	UPDATED                             	STATUS  	CHART                	APP VERSION
my-release	spark-operator	1       	2021-11-15 00:28:23.264778 +0100 CET	deployed	spark-operator-1.1.12	v1beta2-1.2.3-3.1.1
spark     	spark-operator	2       	2021-11-15 00:29:41.35789 +0100 CET 	deployed	spark-operator-1.1.12	v1beta2-1.2.3-3.1.1

k get pods
NAME                                         READY   STATUS             RESTARTS   AGE
my-release-spark-operator-665d77b858-nr9bf   1/1     Running            0          11m
spark-spark-operator-7cccb4f59f-swr22        0/1     CrashLoopBackOff   7          12m
spark-spark-operator-webhook-init-sb4q5      0/1     Completed          0          10m

I have updated my repo

helm repo add spark-operator https://googlecloudplatform.github.io/spark-on-k8s-operator --force-update
helm repo update

Where is the discrepancy coming from?

[felixng-umh]

same here

[eugen-fried]

+1

[pdrastil]

It's coming from this PR that modified the chart https://github.com/GoogleCloudPlatform/spark-on-k8s-operator/pull/1387 but it seems gcr.docker.prod.walmart.com/spark-operator/spark-operator:latest was not published into Docker registry with this feature. Same applies to other image tags.

[echang0929]

+1

[DiogoReisPinto]

+1

Any idea when this will get fixed?

[duyet]

Currently, I'm working around this by building the spark-operator binary from master and adding the current image: https://github.com/duyet/spark-docker/blob/master/spark-operator/v1beta2-1.2.3-3.1.1-master/Dockerfile

ARG SPARK_IMAGE=gcr.io/spark-operator/spark-operator:v1beta2-1.2.3-3.1.1

# Build spark-operator binary from master branch
FROM golang:1.15.2-alpine as builder
WORKDIR /workspace
RUN apk update && apk add git && \
    git clone --depth 1 https://github.com/GoogleCloudPlatform/spark-on-k8s-operator && \
    cd spark-on-k8s-operator && \
    go mod download && \
    CGO_ENABLED=0 GOOS=linux GOARCH=amd64 GO111MODULE=on go build -a -o /usr/bin/spark-operator main.go

FROM ${SPARK_IMAGE}
USER root
COPY --from=builder /usr/bin/spark-operator /usr/bin/

Quickly test this by using my image when installing the operator:

helm repo add spark-operator https://googlecloudplatform.github.io/spark-on-k8s-operator
helm install <release-name> spark-operator/spark-operator \
  --set image.repository=duyetdev/spark-operator \
  --set image.tag=v1beta2-1.2.3-3.1.1-master

[korjek]

@liyinan926 could you please help with publishing an official docker image that contains the most recent changes?

[autarchprinceps]

Perhaps you should also set the actual image version in the helm chart, then a change in the helm chart would be synchronous with the change in the image, you know like any other helm chart does. Alongside building the image automatically on version tag of course.

[shivakumargowda12]

Is this issue fixed in latest release 1.1.15 ? Because, I am still getting same error with master branch also.

[hobbsh]

Is this issue fixed in latest release 1.1.15 ? Because, I am still getting same error with master branch also.

No, it's still broken. And with this version you still need to manually set the image tag as the one set in appVersion is not found.

[evoosa]

a workaround that worked for me and for @tomarad is to specify both image.tag and version in the command line when running helm install helm install spark-operator spark-operator/spark-operator --namespace data --set sparkJobNamespace=data --set webhook.enable=true --set image.tag=v1beta2-1.2.3-3.1.1 --version=1.1.11

the version is the last version before the webhook.timeout flag was added, as seen in this commit

[apreethi13]

a workaround that worked for me and for @tomarad is to specify both image.tag and version in the command line when running helm install helm install spark-operator spark-operator/spark-operator --namespace data --set sparkJobNamespace=data --set webhook.enable=true --set image.tag=v1beta2-1.2.3-3.1.1 --version=1.1.11

the version is the last version before the webhook.timeout flag was added, as seen in this commit

Bumping down the version from 1.1.13 to 1.1.11 fixed the issue. But would be great if we could have a fix for the latest version.