Make e2e run in HA mode

[yaron2]

Do not merge before #2185.

Closes #2188

[codecov]

Codecov Report

Merging #2189 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #2189   +/-   ##
=======================================
  Coverage   46.26%   46.26%           
=======================================
  Files          69       69           
  Lines        6078     6078           
=======================================
  Hits         2812     2812           
  Misses       3008     3008           
  Partials      258      258           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9518288...495bfb5. Read the comment docs.

[tcnghia]

/ok-to-test

[dapr-bot]

Found the available test cluster - dapr-aks-e2e-07 for linux. Please check the build status.

[dapr-bot]

Found the available test cluster - dapr-aks-e2e-08 for windows. Please check the build status.

[yaron2]

I think the 30s timeout for helm install needs to increase, I've seen the KiND tests timeout when installing.

/cc @youngbupark

[youngbupark]

@yaron2 Once all e2e tests are green, then I will cut 0.11.1 release

cc/ @mchmarny @msfussell @orizohar @yolember

[yaron2]

I think the 30s timeout for helm install needs to increase, I've seen the KiND tests timeout when installing.

/cc @youngbupark

Also, I think as @tcnghia mentioned, we need to increase the node size / add nodes.. Pods might be pending.

[dapr-bot]

Congrats! All end-to-end tests have passed on linux. Thanks for your contribution!

[dapr-bot]

End-to-end tests failed on windows. Please check the build logs

[youngbupark]

I think the 30s timeout for helm install needs to increase, I've seen the KiND tests timeout when installing.
/cc @youngbupark

Also, I think as @tcnghia mentioned, we need to increase the node size / add nodes.. Pods might be pending.

helm install \
	dapr --namespace=dapr-tests --wait --timeout 5m0s\
	--set global.ha.enabled=true --set-string global.tag=dev-linux-amd64 --set-string global.registry=localhost:5000/dapr --set global.logAsJson=true --set global.daprControlPlaneOs=linux ./charts/dapr

5 minutes timeout. So it may be because of node size.

[youngbupark]

/ok-to-test

[yaron2]

I think the 30s timeout for helm install needs to increase, I've seen the KiND tests timeout when installing.
/cc @youngbupark

Also, I think as @tcnghia mentioned, we need to increase the node size / add nodes.. Pods might be pending.

helm install \
	dapr --namespace=dapr-tests --wait --timeout 5m0s\
	--set global.ha.enabled=true --set-string global.tag=dev-linux-amd64 --set-string global.registry=localhost:5000/dapr --set global.logAsJson=true --set global.daprControlPlaneOs=linux ./charts/dapr

5 minutes timeout. So it may be because of node size.

Yeah. The pods are probably pending, as helm install times out on KinD.

[dapr-bot]

Found the available test cluster - dapr-aks-e2e-05 for linux. Please check the build status.

[youngbupark]

windows e2e test is failed. restarting test.

[dapr-bot]

Found the available test cluster - dapr-aks-e2e-06 for windows. Please check the build status.

[yaron2]

End-to-end tests failed on windows. Please check the build logs

Also, do we know why Windows tests are interminantly failing on the scheduled runs?

https://github.com/dapr/dapr/actions?query=workflow%3Adapr-test

[youngbupark]

pubsub test failed in windows e2e test. I hope this was just an intermittent failure.

Failed Platform.addApps(), deployment "pubsub-publisher" is not in desired state, received: &Deployment{ObjectMeta:{      0 0001-01-01 00:00:00 +0000 UTC <nil> <nil> map[] map[] [] []  []},Spec:DeploymentSpec{Replicas:nil,Selector:nil,Template:{{      0 0001-01-01 00:00:00 +0000 UTC <nil> <nil> map[] map[] [] []  []} {[] [] [] []  <nil> <nil>  map[]   <nil>  false false false <nil> nil []   nil  [] []  <nil> nil [] <nil> <nil> <nil> map[] []}},Strategy:DeploymentStrategy{Type:,RollingUpdate:nil,},MinReadySeconds:0,RevisionHistoryLimit:nil,Paused:false,ProgressDeadlineSeconds:nil,},Status:DeploymentStatus{ObservedGeneration:0,Replicas:0,UpdatedReplicas:0,AvailableReplicas:0,UnavailableReplicas:0,Conditions:[]DeploymentCondition{},ReadyReplicas:0,CollisionCount:nil,},}: etcdserver: request timed out2020/10/03 18:47:08 Running teardown...
FAIL	github.com/dapr/dapr/tests/e2e/pubsub	28.030s
Enter TestMain

[youngbupark]

I think the 30s timeout for helm install needs to increase, I've seen the KiND tests timeout when installing.
/cc @youngbupark

Also, I think as @tcnghia mentioned, we need to increase the node size / add nodes.. Pods might be pending.

helm install \
	dapr --namespace=dapr-tests --wait --timeout 5m0s\
	--set global.ha.enabled=true --set-string global.tag=dev-linux-amd64 --set-string global.registry=localhost:5000/dapr --set global.logAsJson=true --set global.daprControlPlaneOs=linux ./charts/dapr

5 minutes timeout. So it may be because of node size.

Yeah. The pods are probably pending, as helm install times out on KinD.

Turned off HA for KinD test for now. we can revisit it later. let's see how aks cluster tests go.

[tcnghia]

If it is difficult for the KinD workflows to support HA testing I think we could turn on the HA testing for AKS workflows first. That allows this PR to be checked in early leading to more testing/baking time (through scheduled workflows) before we cut the patch release.

Independently in that time window, we could try turning on the HA testing for KinD workflows.

[dapr-bot]

Congrats! All end-to-end tests have passed on linux. Thanks for your contribution!

[dapr-bot]

Congrats! All end-to-end tests have passed on windows. Thanks for your contribution!