193ee88ac0
smart_update() currently uses the label 'app.kubernetes.io/instance: zuul' to select the scheduler pod(s) when attempting to kick a smart update in the pod, however, the pods are created with 'app.kubernetes.io/instance: my-instance-name', which will cause us to not find any scheduler pods to update. Instead, we should use the correct label expression for the query. Once smart_update() was finding the scheduler pod correctly, pod_exec() fails on an unnecessary call to load_kube_config(). Removing this enables the pod_exec() - which is only ever used by smart_update() - to succeed. Change-Id: I57e4e3dacac1df24a941dd23da6344b4243a0bed |
||
|---|---|---|
| build | ||
| deploy | ||
| doc | ||
| helm/zuul-operator | ||
| playbooks/zuul-operator-functional | ||
| tools | ||
| zuul_operator | ||
| .git-blame-ignore-revs | ||
| .gitignore | ||
| .gitreview | ||
| .zuul.yaml | ||
| LICENSE | ||
| Makefile | ||
| README.md | ||
| noxfile.py | ||
| requirements.txt | ||
| setup.cfg | ||
| setup.py | ||
| tox.ini | ||
README.md
Zuul Operator
Build the image
$ make image
Install the operator
$ make install
kubectl apply -f deploy/crds/zuul-ci_v1alpha2_zuul_crd.yaml -f deploy/rbac.yaml -f deploy/operator.yaml
Look for operator pod and check it's output
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
zuul-operator-c64756f66-rbdmg 2/2 Running 0 3s
$ kubectl logs zuul-operator-c64756f66-rbdmg
[...]
{"level":"info","ts":1554197305.5853095,"logger":"cmd","msg":"Go Version: go1.10.3"}
{"level":"info","ts":1554197305.5854425,"logger":"cmd","msg":"Go OS/Arch: linux/amd64"}
{"level":"info","ts":1554197305.5854564,"logger":"cmd","msg":"Version of operator-sdk: v0.6.0"}
{"level":"info","ts":1554197305.5855,"logger":"cmd","msg":"Watching namespace.","Namespace":"default"}
[...]
Usage
$ kubectl apply -f - <<EOF
apiVersion: operator.zuul-ci.org/v1alpha2
kind: Zuul
metadata:
name: example-zuul
spec:
EOF
zuul.zuul-ci.org/example-zuul created