If given a directory, we can construct a reasonable name based on the
executable name, pid and timestamp. Then this is relatively easy to
wire up from kubetest2, if we have an artifacts directory.
If a job's --down was interrupted, the state store can be in an inconsistent state.
This can cause `kops get cluster --full` to fail if certain additional objects are missing, yet `kops get cluster` would succeed.
We omit `--full` to ensure `kops get cluster` succeeds and leak cleanup is reattempted.
Previously we were only not setting --master-count if --master-count was in --create-args.
Now we recognize either flag. This fixes E2E jobs that specify --control-plane-count and kops create cluster ends up having both flags
For consistency with kops create cluster.
It doesn't appear to be used outside this repo, although test-infra
declares KOPS_CONTROL_PLANE_SIZE and that is then passed to our
e2e/scenarios scripts.
Issue #15559
* feat(karpenter): Upgrade to version 0.27.0
Upgrade Karpenter to current last stable version `0.27.0`.
Template have been updated to use the same templates than the Helm chart.
* feat(karpenter): Use AWSNodeTemplate for launchTemplate
To set Launch Templates is deprecated into the provisioner, it is recommends using the `AWSNodeTemplate` to set it.
Ref:
- https://karpenter.sh/v0.27.0/concepts/node-templates/
* feat(karpenter): Enable pruning addon
* Use extra flags in upgrade-ab scenario test
* feat(karpenter): Drop `karpenter` feature flag
* feat(karpenter): Add release note for `1.27`
* feat(karpenter): Upgrade to version 0.27.3
* feat(karpenter): fix template
* feat(karpenter): Upgrade to version 0.27.5
* Update Karpenter documentation with depending kops version
* Delete KOPS_FEATURE_FLAGS from e2e test `run-test`
* Run hack/update-expected.sh
We previously tested that we could not generate 4 zones in a region,
but now us-east-1 has 6 zones, so we need to update the tests to fail
only when even more zones are requested.
We're seeing some logs that don't include any output from kops toolbox
dump, but probably should. Make sure we're including any stderr
output in our logs.
GCE support seems stable now, and we have good clarity at the API
level and how that translates to GCP resources, which was our blocker
previously.
Drop the need for the feature-flag.
Arnaud M.