Migration of etcd-io/etcd GitHub Actions to Prow
What should be cleaned up or changed: As a part of the migration, following GitHub Actions running as presubmits/postsubmits/periodics may need to be migrated to Prow:
main :
- [ ] contrib.yaml
- [ ] coverage.yaml
- [ ] fuzzing.yaml
- [x] govuln.yaml
- [ ] grpcproxy.yaml
- [ ] measure-testgrid-flakiness.yaml
- [ ] release.yaml
release-3.5 :
- [ ] functional.yaml
- [ ] govuln.yaml
- [ ] grpcproxy.yaml
- [ ] release.yaml
- [ ] static-analysis.yaml
release-3.4 :
- [ ] govuln.yaml
- [ ] release.yaml
- [ ] static-analysis.yaml
Provide any links for context: Parent Issue ref: https://github.com/kubernetes/k8s.io/issues/6102
This is an initial list of GitHub workflows I have identified for the migration, please review the list if anything needs to be added/removed.
/sig etcd
/assign
cc @jmhbnz @ahrtr @serathius
Also link to https://github.com/kubernetes/k8s.io/issues/6102
I wasn't aware of this task. Watching now.
The Kubernetes project currently lacks enough contributors to adequately respond to all issues.
This bot triages un-triaged issues according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue as fresh with
/remove-lifecycle stale - Close this issue with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
/remove-lifecycle stale
The Kubernetes project currently lacks enough contributors to adequately respond to all issues.
This bot triages un-triaged issues according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue as fresh with
/remove-lifecycle stale - Close this issue with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
/remove-lifecycle stale
@ArkaSaha30, the completion track is not up to date. But I don't have permission to edit your description. Should I create a new ticket to track the completion? So I can update it? Let me know.
Refer to https://github.com/kubernetes/test-infra/issues/32754#issuecomment-2942746295
As discussed at today's community meeting. We have all build GitHub workflows in release 3.4 and 3.5. In 3.6 and main, we have a build-all target. However, there is duplication with the release tests, as these are built for all supported architectures and operating systems.
We'll be removing release-3.5's test-smoke job. The Makefile target is the following:
test-smoke:
$(info log-file: test-$(TEST_SUFFIX).log)
PASSES="fmt bom dep build unit" ./test.sh 2<&1 | tee test-$(TEST_SUFFIX).log
! grep "FAIL:" test-$(TEST_SUFFIX).log
All the passes it runs are duplicated:
-
fmt: It's replaced by verify -
bom: It's part of verify -
dep: It's part of verify -
build: Covered already by build and by release tests -
unit: We already have a unit test job
Refer to https://github.com/kubernetes/test-infra/issues/32754#issuecomment-3014660991
X-posting https://github.com/etcd-io/etcd/pull/20116#issuecomment-2950067337, as we're having issues running 3.4 e2e tests on the Prow infra.
@joshjms, I can think of two solutions:
- We run the tests on the Prow infra using a different image, which will require a different job configuration for release-3.4 e2e jobs
- We keep this job as a GitHub action/workflow
Given the timeline to give support to this version, I lean towards option 2.
That would mean:
- Keep the
tests.yamlfile, simplify it, so it only runs e2e tests- Revert https://github.com/kubernetes/test-infra/pull/34916
@jmhbnz, thoughts?
Edit: We're backporting the e2e tests change to release-3.4.
And now, with the activity, my comment keeps getting hidden. So, copying https://github.com/kubernetes/test-infra/issues/32754#issuecomment-2942746295 to the bottom to keep it visible.
Because I don't have access to update the issue. I'll start my own completion matrix below:
main:
- [x] contrib.yaml
- [x] coverage.yaml
- [x] fuzzing.yaml
- [x] govuln.yaml
- [x] grpcproxy.yaml
- [ ] measure-testgrid-flakiness.yaml - Doesn't apply. Check with Siyuan if we still need this workflow.
- [x] release.yaml
release-3.5:
- [x] e2e.yaml
- [x] #35053
- [x] https://github.com/etcd-io/etcd/pull/20315
- [x] ~functional.yaml~: Just remove the workflow (refer to https://github.com/etcd-io/etcd/pull/20021#issuecomment-2918549192) - https://github.com/etcd-io/etcd/pull/20315
- [x] govuln.yaml
- [x] #35050
- [x] etcd-io/etcd#20172
- [x] https://github.com/etcd-io/etcd/pull/20315
- [x] grpcproxy.yaml
- [x] etcd-io/etcd#20197
- [x] #35044
- [x] etcd-io/etcd#20236
- [x] release.yaml
- [x] etcd-io/etcd#20168
- [x] etcd-io/etcd#20293
- [x] #35098
- [x] https://github.com/etcd-io/etcd/pull/20315
- [x] static-analysis.yaml
- [x] https://github.com/kubernetes/test-infra/pull/34782
- [x] Implement missing verify targets: etcd-io/etcd#20046
- [x] etcd-io/etcd#20099
- [x] tests.yaml
- [x] https://github.com/etcd-io/etcd/pull/20315
- [x] etcd-io/etcd#20198 is blocking integration tests and the rest of the progress.
- [x] ~amd64 smoke~: Delete this job (and the arm64 one). Refer to https://github.com/kubernetes/test-infra/issues/32754#issuecomment-2920762844
- [x] amd64 integration, 1 2, and 4 CPUs
- [x] #34983
- [x] #34989
- [x] arm64/amd64/i386 unit
- [x] #35000
- [x] #35011
- [x] ~all-build (cross-compiling): Probably not needed, as covered by release tests.~
release-3.4:
- [x] govuln.yaml
- [x] etcd-io/etcd#19810
- [x] #34737
- [x] etcd-io/etcd#19881
- [x] release.yaml
- [x] etcd-io/etcd#19815
- [x] #34740
- [x] etcd-io/etcd#19877
- [x] static-analysis.yaml
- [x] etcd-io/etcd#19911
- [x] https://github.com/etcd-io/etcd/pull/20011
- [x] tests.yaml
- [x] etcd-io/etcd#20116
- [x] amd64 integration, 1 2, and 4 CPUs
- [x] 4 CPUs: #34843
- [x] 2, 1 CPUs: #34877
- [x] ~amd64 functional~: Just remove the workflow (refer to https://github.com/etcd-io/etcd/pull/20021#issuecomment-2918549192)
- [x] etcd-io/etcd#20021
- [x] ~#34849~
- [x] amd64 unit 4 CPUs, 386 unit 1 CPU - #34883
- [x] amd64 grcpproxy
- [x] etcd-io/etcd#20063
- [x] #34891
- [x] amd64 e2e
- [x] etcd-io/etcd#20136
- [x] #34916
- [x] #34955
- [x] ~all-build (cross-compiling): Probably not needed, as covered by release tests.~
With all tasks completed, we have successfully migrated the Prow repository to etcd. Thanks for all your work, @joshjms and @abdurrehman107 :)
/close
@ivanvc: Closing this issue.
In response to this:
With all tasks completed, we have successfully migrated the Prow repository to etcd. Thanks for all your work, @joshjms and @abdurrehman107 :)
/close
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.