Jack Francis comments

Results 586 comments of


                                            Jack Francis

Support azure vmss flex

@Daniel-Redeploy have you had a chance to determine if we still need to preserve anything from this PR?

metadata.yaml release test gates

Ref: https://cluster-api.sigs.k8s.io/clusterctl/provider-contract.html?highlight=metadata.#metadata-yaml

Allow MachinePools without MachinePoolMachines to exist in workload cluster for clusterctl upgrade e2e test

/assign @willie-yao

Add retry for creatingAzureManager in case of throttled requests

/test pull-cluster-autoscaler-e2e-azure

Add retry for creatingAzureManager in case of throttled requests

/retest

conformance_test.go exit code to 0 event the non last test suite has failures

After an initial investigation, it appears that the cluster-api is itself wrapping kubetest inside a container execution lifecycle. And the above test output suggests that the underlying container, though its...

conformance_test.go exit code to 0 event the non last test suite has failures

@sbueringer that sounds like a lower-level explanation that would produce these outcomes. Let's keep this CAPZ issue open and move our practical energies towards fixing things to https://github.com/kubernetes-sigs/cluster-api/issues/5123

conformance_test.go exit code to 0 event the non last test suite has failures

So I ran a follow-up local test, and got an error during the conformance test: ``` [FAILED] error while waiting for pod svcaccounts-2356/oidc-discovery-validator to be Succeeded or Failed: pod "oidc-discovery-validator"...

conformance_test.go exit code to 0 event the non last test suite has failures

[junit.zip](https://github.com/kubernetes-sigs/cluster-api-provider-azure/files/10574427/junit.zip) cc @CecileRobertMichon

conformance_test.go exit code to 0 event the non last test suite has failures

In which case this would be a testgrid bug? (Or maybe we need to explicitly always only publish one junit file, or maybe we're misnaming the junit file?)