v1.9.0-alpha.3

[enisoc]

According to the timeline, we should cut v1.9.0-alpha.3 on Wed, Nov 15.

This is a tracker for burning down release-blocking test failures and recording release cut status.

[spiffxp]

Swept through sig-release-master-blocking, tied open issues from 1.9.0-alpha.2 to here. It looks like there's something GKE related going on that's masking any more specific test cases at the moment.

[spiffxp]

I am inclined to suggest removing the gke jobs from release-master-blocking considering that they're all still failing. Will have a PR out later

[spiffxp]

Commentary on the jobs that are failing as of this morning:

• gce-scale-correctness looks to be legit failures or flakes
• gce-scale-performance failed due to resource exhaustion in us-east-1a, so technically speaking we don't have coverage on the latest commit; the previous run passed (ref: ci-kubernetes-e2e-gce-scale-* jobs cannot create 64-core master kubernetes#55777)
• soak-gce-gci is a job I'd like to see removed from release-master-blocking

[enisoc]

@spiffxp Are there any in-flight PRs that hope to fix scale, or is it still under investigation? If it won't be fixed today, I propose going forward with the alpha cut as is. Any objection?

[enisoc]

@dashpole is going to start the alpha.3 cut soon.

This looks like the best recent build on master:

* build                               #7216   #808    [06:41 11/15]
* (--buildversion=v1.9.0-alpha.2.808+1ced91f2011b6d)
- test-go                             #11957  #808    PASSED
- verify-master                       #16059  #808    PASSED
- node-kubelet                        #8501   #808    PASSED
- e2e-gci-gce                         #17770  #808    PASSED
- e2e-gci-gke                         #18364  #808    PASSED
- e2e-kops-aws                        #19674  #808    PASSED
- e2e-gce-gpu                         #2280   #808    PASSED
- e2e-gce-device-plugin-gpu           #1114   #808    PASSED
- e2e-gke-device-plugin-gpu           #604    #808    PASSED
- e2e-gke-device-plugin-gpu-p100      #14     #792    PASSED
- e2e-gci-gce-slow                    #12045  #808    PASSED
- e2e-gci-gke-slow                    #12395  #808    PASSED
- e2e-gci-gce-serial                  #2520   #808    PASSED
- e2e-gci-gke-serial                  #4312   #808    PASSED
- e2e-gci-gce-alpha-features          #17488  #808    PASSED
- e2e-gci-gce-audit                   #4081   #808    PASSED
- e2e-gci-gce-scalability             #7989   #808    PASSED
- e2e-gce-scale-correctness           No Good Runs    SKIPPING
- e2e-gce-scale-performance           --      --      FAILED
- e2e-gci-gce-ingress                 #14333  #808    PASSED
- e2e-gci-gke-ingress                 #16610  #808    PASSED
- e2e-gci-gce-reboot                  #17477  #808    PASSED
- e2e-gci-gke-reboot                  #10100  #808    PASSED
- soak-gce-gci                        No Good Runs    SKIPPING
- e2e-kubeadm-gce                     #8560   #808    PASSED
- periodic-kubernetes-e2e-debs-pushed No Good Runs    SKIPPING

[dashpole]

mock release went just fine. Running with --no-mock now

[dashpole]

/close
v1.9.0-alpha.3 is out!

[porridge]

FWIW, retry of e2e-gce-scale-performance succeeded: kubernetes/kubernetes#55777 (comment)