Closed CharlesQQ closed 1 year ago
Thanks @CharlesQQ Have you tested it? Does it solve the issue addressed by #1946?
Have you tested it? Does it solve the issue addressed by https://github.com/karmada-io/karmada/issues/1946?
yes,I test it, karmada search can fetch cluster's resource in pull mode successfully, but sometimes, when restart karmada, karmada-search-controller might print log, whatever in push or pull mode:
I0614 15:13:55.311227 27 controller.go:170] Cluster member2 is not registered
W0614 15:56:54.931730 27 cache.go:96] SingleClusterInformerManager for cluster(member2) is nil.
I think it might aggregated-apiserver restarted and not started completely when karamda-search already started; so is might be necessary aggregated-apiserver running healthy when restarting karmada-search?
/cc @RainbowMango
Sorry @CharlesQQ , I missed this PR, so busy these days. I'll look at by this week.
Generally looks good to me. cc @RainbowMango to check.
/assign
@XiShanYongYe-Chang can you figure out why unit test failed, i run unit test successfully
=== RUN TestObtainCredentialsFromMemberCluster/report_secret_enabled credential_test.go:145: ObtainCredentialsFromMemberCluster() error = failed to get serviceAccount secret for impersonation from cluster(member1), error: failed to get serviceAccount secret, error: timed out waiting for the condition, wantErr false --- FAIL: TestObtainCredentialsFromMemberCluster (60.00s) --- FAIL: TestObtainCredentialsFromMemberCluster/disable_report_secret (30.00s) --- FAIL: TestObtainCredentialsFromMemberCluster/report_secret_enabled (30.00s)
i run successfully
=== RUN TestObtainCredentialsFromMemberCluster --- PASS: TestObtainCredentialsFromMemberCluster (4.01s) === RUN TestObtainCredentialsFromMemberCluster/disable_report_secret --- PASS: TestObtainCredentialsFromMemberCluster/disable_report_secret (2.00s) === RUN TestObtainCredentialsFromMemberCluster/report_secret_enabled --- PASS: TestObtainCredentialsFromMemberCluster/report_secret_enabled (2.01s) PASS
Occasional timeout.
I can't locate the error directly
/test ?
@CharlesQQ: Cannot trigger testing until a trusted user reviews the PR and leaves an /ok-to-test
message.
Let me try on my side.
Just re-triggered it as it works on my side:
go test -race -failfast -v -count=10 ./pkg/...
why unit test for ObtainCredentialsFromMemberCluster
successfully on my side, but failed on CI Workflow
Mhmm... Something seems wrong.
because of WaitForServiceAccountSecretCreation
logic changed
cc @XiShanYongYe-Chang
why CI workflow not be triggered?
why CI workflow not be triggered?
Maybe at that time, the ci stopped working, and I found that the other one pr was in the same state. We need to retrigger it.
Thanks for your hard work.
Generally looks good to me.
But why the cluster-provider
, --cluster-region
and --cluster-zone
are still in this PR?
But why the cluster-provider, --cluster-region and --cluster-zone are still in this PR?
fixed, why e2e test failed?
Message: failed to create a ClusterClient: Secret "karmada-member1" not found
I'm not sure, but please rebase your code with the master, and push again.
It's maybe not an occasional mistake:
Status:
Conditions:
Last Transition Time: 2022-07-06T05:31:55Z
Message: failed to create a ClusterClient: Secret "karmada-member1" not found
Reason: StatusCollectionFailed
Status: False
Type: Ready
Events: <none>
For the Push
mode cluster, we need to create both of those two secrets by default.
Great! Seems we are ready to move this forward now.
Please help update the release-note
part of the PR description since the flag name has changed.
And, before merging this patch, I want to confirm if this fixed the issue addressed by #1946, can you help to confirm that? (Please test with the latest patch)
And, before merging this patch, I want to confirm if this fixed the issue addressed by https://github.com/karmada-io/karmada/issues/1946, can you help to confirm that?
yes, I test it,and karamda-search can fetch resource from cluster in pull mode sueecessfully, but if restart karamda and karmada-aggregated-apiserver not healthy, log print following message:
I0707 11:10:53.182584 25 controller.go:170] Cluster member1 is not registered
I0707 11:10:53.182596 25 controller.go:170] Cluster member2 is not registered
Did anything change in the force push?
Did anything change in the force push?
yes,has code conflict in agent/app/options, because of last merge add new flag ProfileOpts
/approve
I can see the provider
/region
/zone
logic still in this PR.
Given this PR has been delayed for a long time, let's merge it now.
But if we are not going to introduce the provider
/region
/zone
logs, please remember to remove the redundant logic before v1.3. @XiShanYongYe-Chang
[APPROVALNOTIFIER] This PR is APPROVED
This pull-request has been approved by: RainbowMango
The full list of commands accepted by this bot can be found here.
The pull request process is described here
But if we are not going to introduce the
provider
/region
/zone
logs, please remember to remove the redundant logic before v1.3. @XiShanYongYe-Chang
Ok~
Signed-off-by: charlesQQ charles_ali@qq.com
What type of PR is this? /kind feature
What this PR does / why we need it: Allowed karmada-agent report secret for Pull mode cluster
Which issue(s) this PR fixes: Part of https://github.com/karmada-io/karmada/issues/1946
Special notes for your reviewer:
Does this PR introduce a user-facing change?: