Closed 2hangchen closed 2 years ago
I think it seems not suitable for this cleanup. If we consider it from repeated execution, after the removing, there's only one line in the function, further considered, shall we cancel the installCRDs
function.
On another angle, we usually package code to a function to achieve a specific outcome. I suppose the original of installCRDs
maybe wanna package the process of all the resources. If so, we should think about all the dependencies, including the namespace, RBAC, and so on.
On another angle, we usually package code to a function to achieve a specific outcome. I suppose the original of
installCRDs
maybe wanna package the process of all the resources. If so, we should think about all the dependencies, including the namespace, RBAC, and so on.
I agree with you. By the way ,is there any planning for scripts for deployed offine?
By the way ,is there any planning for scripts for deployed offine?
I am not very clear about deployment offline, as far as know, the pulling image needs internet, but you can save the images to local and change the image URL in YAML files.
By the way ,is there any planning for scripts for deployed offine?
I am not very clear about deployment offline, as far as know, the pulling image needs internet, but you can save the images to local and change the image URL in YAML files.
ok,thx
Is there anything we need to address in this PR?
I suggest we may cancel the installCRDs
function first. Use directly kubectl kustomize "${REPO_ROOT}/charts/_crds" | kubectl apply -f -
in the code. If necessary later, we can package all the Karmada resources code to one function. How about it @2hangchen
I suggest we may cancel the
installCRDs
function first. Use directlykubectl kustomize "${REPO_ROOT}/charts/_crds" | kubectl apply -f -
in the code. If necessary later, we can package all the Karmada resources code to one function. How about it @2hangchen
I think it is fine.
/assign @lfbear
Don't know if this pr is blocked by failing test, just re-triggered.
I suggest we may cancel the
installCRDs
function first. Use directlykubectl kustomize "${REPO_ROOT}/charts/_crds" | kubectl apply -f -
in the code. If necessary later, we can package all the Karmada resources code to one function. How about it @2hangchenI think it is fine.
@2hangchen Do you have time to finish this recently? :)
Kind ping @2hangchen @lfbear
/lgtm
/remove-lgtm
@2hangchen Could you merge the master branch and solve the conflict.
@2hangchen Could you merge the master branch and solve the conflict.
I'm sorry. I just saw it. I'll deal with it now.
/retest
@2hangchen: Cannot trigger testing until a trusted user reviews the PR and leaves an /ok-to-test
message.
/ok-to-test
[BeforeSuite] BeforeSuite
/home/runner/work/karmada/karmada/test/e2e/suite_test.go:81
Unexpected error:
<*errors.errorString | 0xc000585e50>: {
s: "cluster member1 not ready",
}
cluster member1 not ready
occurred
/home/runner/work/karmada/karmada/test/e2e/framework/cluster.go:60
Manually re-triggered.
It seems not the problem with the e2e test.
I checked the code again and found that namespace creating can not remove in the function installCRDs
because it will create a namespace in the karmada-apiserver
cluster and another (line:129 in hack/deploy-karmada.sh) will create a namespace in the karmada-host
cluster.
/remove-approve
[APPROVALNOTIFIER] This PR is NOT APPROVED
This pull-request has been approved by: To complete the pull request process, please ask for approval from lfbear after the PR has been reviewed.
The full list of commands accepted by this bot can be found here.
It seems not the problem with the e2e test. I checked the code again and found that namespace creating can not remove in the function
installCRDs
because it will create a namespace in thekarmada-apiserver
cluster and another (line:129 in hack/deploy-karmada.sh) will create a namespace in thekarmada-host
cluster.
@lfbear I get it. I got it wrong
What type of PR is this?
What this PR does / why we need it: /kind cleanup
Special notes for your reviewer: This command has been executed before (line:118)