Replace fixed sleep with a polling loop that waits for the deployment
namespace to be fully deleted. Without this, the start command fails
with 403 Forbidden because k8s rejects resource creation in a
namespace that is still terminating.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Recreating a kind cluster in the same CI run fails due to stale
etcd/certs and cgroup detection issues. Use --skip-cluster-management
to reuse the existing cluster, and --delete-volumes to clear PVs so
fresh PVCs can bind on restart.
The volume retention semantics are preserved: bind-mount host path
data survives (filesystem is old), provisioner volumes are fresh
(PVs were deleted).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
kubectl commands that query jobs or pod specs exit non-zero when the
resource doesn't exist yet. Under set -e, a bare command substitution
like var=$(kubectl ...) aborts the script silently. Add || true so
the polling loop and assertion logic can handle failures gracefully.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
deploy init already writes 'secrets: {}' into the spec file. The test
was appending a second secrets block via heredoc, which ruamel.yaml
rejects as a duplicate key. Use sed to replace the empty value instead.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace the fixed `sleep 20` with a polling loop that waits for
`kind get clusters` to report no clusters. The previous approach
was flaky on CI runners where Docker takes longer to tear down
cgroup hierarchies after `kind delete cluster`.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Use --skip-cluster-management to avoid destroying and recreating the
kind cluster during the stop/start volume retention test. The second
kind create fails on some CI runners due to cgroups detection issues.
Use --delete-volumes to clear PVs so fresh PVCs can bind on restart.
Bind-mount data survives on the host filesystem; provisioner volumes
are recreated fresh.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add a job compose file for the test stack and extend the k8s deploy
test to verify new features:
- Namespace isolation: pod exists in laconic-{id}, not default
- Stack labels: app.kubernetes.io/stack label set on pods
- Job completion: test-job runs to completion (status.succeeded=1)
- Secrets: spec secrets: key results in envFrom secretRef on pod
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
In kind, when we bind-mount a host directory it is first mounted into the kind container at /mnt, then into the pod at the desired location.
We accidentally picked this up for full-blown k8s, and were creating volumes at /mnt. This changes the behavior for both kind and regular k8s so that bind mounts are only allowed if a fully-qualified path is specified. If no path is specified at all, a default storageClass is assumed to be present, and the volume managed by a provisioner.
Eg, for kind, the default provisioner is: https://github.com/rancher/local-path-provisioner
```
stack: test
deploy-to: k8s-kind
config:
test-variable-1: test-value-1
network:
ports:
test:
- '80'
volumes:
# this will be bind-mounted to a host-path
test-data-bind: /srv/data
# this will be managed by the k8s node
test-data-auto:
configmaps:
test-config: ./configmap/test-config
```
Reviewed-on: https://git.vdb.to/cerc-io/stack-orchestrator/pulls/741
Co-authored-by: Thomas E Lackey <telackey@bozemanpass.com>
Co-committed-by: Thomas E Lackey <telackey@bozemanpass.com>
Prathamesh Musale
Thomas E Lackey
David Boreham