kepler-operator
kepler-operator copied to clipboard
sustainable-computing-io
.spec.volumes: field not declared in schema
Attempting to deploy, with
Following the docs to deploy the operator, I ran :
$ make deploy OPERATOR_IMG=quay.io/sustainable_computing_io/kepler-operator:v1alpha1
./hack/tools.sh kustomize
✅ kustomize matching v3.8.7 already installed
{Version:kustomize/v3.8.7 GitCommit:ad092cc7a91c07fdf63a2e4b7f13fa588a39af4f BuildDate:2020-11-11T23:14:14Z GoOs:linux GoArch:amd64}
./hack/tools.sh controller-gen
✅ controller-gen matching Version: v0.12.1 already installed
Version: v0.12.1
/home/jtaleric/code/kepler-operator/tmp/bin/controller-gen object:headerFile="hack/boilerplate.go.txt" paths="./pkg/..."
/home/jtaleric/code/kepler-operator/tmp/bin/controller-gen rbac:roleName=manager-role crd webhook paths="./pkg/..." output:crd:artifacts:config=config/crd/bases
/home/jtaleric/code/kepler-operator/tmp/bin/kustomize build config/crd | \
kubectl apply --server-side --force-conflicts -f -
customresourcedefinition.apiextensions.k8s.io/keplerinternals.kepler.system.sustainable.computing.io serverside-applied
customresourcedefinition.apiextensions.k8s.io/keplers.kepler.system.sustainable.computing.io serverside-applied
/home/jtaleric/code/kepler-operator/tmp/bin/kustomize build config/default | \
sed -e "s|<OPERATOR_IMG>|quay.io/sustainable_computing_io/kepler-operator:v1alpha1|g" \
-e "s|<KEPLER_IMG>|quay.io/sustainable_computing_io/kepler:release-0.7.2|g" \
| tee tmp/deploy.yaml | \
kubectl apply --server-side --force-conflicts -f -
namespace/kepler-operator-system serverside-applied
customresourcedefinition.apiextensions.k8s.io/keplerinternals.kepler.system.sustainable.computing.io serverside-applied
customresourcedefinition.apiextensions.k8s.io/keplers.kepler.system.sustainable.computing.io serverside-applied
serviceaccount/kepler-operator-controller-manager serverside-applied
role.rbac.authorization.k8s.io/kepler-operator-leader-election-role serverside-applied
clusterrole.rbac.authorization.k8s.io/kepler-operator-manager-role serverside-applied
clusterrole.rbac.authorization.k8s.io/kepler-operator-metrics-reader serverside-applied
clusterrole.rbac.authorization.k8s.io/kepler-operator-proxy-role serverside-applied
rolebinding.rbac.authorization.k8s.io/kepler-operator-leader-election-rolebinding serverside-applied
clusterrolebinding.rbac.authorization.k8s.io/kepler-operator-manager-rolebinding serverside-applied
clusterrolebinding.rbac.authorization.k8s.io/kepler-operator-proxy-rolebinding serverside-applied
service/kepler-operator-controller-manager-metrics-service serverside-applied
service/kepler-operator-webhook-service serverside-applied
servicemonitor.monitoring.coreos.com/kepler-operator-controller-manager-metrics-monitor serverside-applied
mutatingwebhookconfiguration.admissionregistration.k8s.io/kepler-operator-mutating-webhook-configuration serverside-applied
validatingwebhookconfiguration.admissionregistration.k8s.io/kepler-operator-validating-webhook-configuration serverside-applied
Error from server: failed to create typed patch object (kepler-operator-system/kepler-operator-controller; apps/v1, Kind=Deployment): .spec.volumes: field not declared in schema
make: *** [Makefile:246: deploy] Error 1
The Deployment "kepler-operator-controller" is invalid: spec.template.spec.containers[0].volumeMounts[0].name: Not found: "cert"
@vprashar2929 so, make fresh seems to deploy a kind cluster? I am trying to use make deploy install on a existing cluster?
I want to do the same thing, on a bare metal (VM-based) cluster, deployed with kubespray, kubeadm.
I already have prometheus operator deployed, as well.
@elafontaine We will address this post by merging this. In the meantime, as a workaround, I suggest you can use operator-sdk to install operator on k8s
make tools # making sure that all the necessary tools are installed locally including operator-sdk
./tmp/bin/operator-sdk olm install --verbose --timeout 5m # installs olm on k8s cluster
./tmp/bin/operator-sdk run bundle quay.io/sustainable_computing_io/kepler-operator-bundle:0.13.0 --install-mode AllNamespaces --skip-tls -n openshift-operators # deploys operator using latest released bundle
The last command did not execute correctly for me...
FATA[0021] Failed to run bundle: create catalog: error creating catalog source: namespaces "openshift-operators" not found
I'm was trying to avoid dealing with the openshift related tooling, but to keep using the operators as they're very useful :).
Anyway, removing the last part of the command (the namespace option) I was able to deploy, but the kepler-operator-controlller never started up correctly.
./tmp/bin/operator-sdk run bundle quay.io/sustainable_computing_io/kepler-operator-bundle:0.13.0 --install-mode AllNamespaces
2024-07-01T18:50:27Z ERROR controller-runtime.source.EventHandler if kind is a CRD, it should be installed before calling Start{"kind": "SecurityContextConstraints.security.openshift.io", "error": "failed to get restmapping: no matches for kind \"SecurityContextConstraints\" in group \"security.openshift.io\""}
sigs.k8s.io/controller-runtime/pkg/internal/source.(*Kind).Start.func1.1
/go/pkg/mod/sigs.k8s.io/[email protected]/pkg/internal/source/kind.go:63
k8s.io/apimachinery/pkg/util/wait.loopConditionUntilContext.func2
/go/pkg/mod/k8s.io/[email protected]/pkg/util/wait/loop.go:87
k8s.io/apimachinery/pkg/util/wait.loopConditionUntilContext
/go/pkg/mod/k8s.io/[email protected]/pkg/util/wait/loop.go:88
k8s.io/apimachinery/pkg/util/wait.PollUntilContextCancel
/go/pkg/mod/k8s.io/[email protected]/pkg/util/wait/poll.go:33
sigs.k8s.io/controller-runtime/pkg/internal/source.(*Kind).Start.func1
/go/pkg/mod/sigs.k8s.io/[email protected]/pkg/internal/source/kind.go:56
Ah, my bad you also need SCC CRD to be added. Can you also run these commands?
kubectl apply --force -f ./hack/crds
kubectl create namespace openshift-config-managed
kubectl create -f hack/monitoring/rbac
./tmp/bin/operator-sdk olm install --verbose --timeout 5m
./tmp/bin/operator-sdk run bundle quay.io/sustainable_computing_io/kepler-operator-bundle:0.13.0 --install-mode AllNamespaces --skip-tls -n operators --timeout 5m
so I'm sorry, I didn't saw your message and I tried to start from the make cluster-up instead... That did not work either as the "make run" isn't passing.
2024-07-03T10:37:30-04:00 INFO Wait completed, proceeding to shutdown the manager
2024-07-03T10:37:30-04:00 ERROR setup problem running manager {"error": "open /var/folders/p8/5wb68xbn4w9d4zzqn41gc9yw0000gp/T/k8s-webhook-server/serving-certs/tls.crt: no such file or directory"}
main.main
/Users/elafonta/kepler/kepler-operator/cmd/manager/main.go:195
runtime.main
/Users/elafonta/go/pkg/mod/golang.org/[email protected]/src/runtime/proc.go:267
exit status 1
make: *** [run] Error 1
I think the error may be related as it's a problem with the TLS volume again...
that's correct make run will not run as by default it looks for webhooks certs that are not present on the cluster. In this case, I suggest you can use make run ENABLE_WEBHOOKS=false
ok, I've been continuing to try to deploy the operator, but it does require the cert-manager and only running the make run ENABLE_WEBHOOKS=false seems to work, but the problem is that it keeps a process spawned up and I would much prefer to run an operator image pre-bundled. I'm going to try to figure out a way to keep working on my side, but I would very much prefer to be able to run "make cluster-up" and get a kind cluster with the operator already deployed.
Can you help me understand why is the certificate signing a requirement for kepler operator ?
#418 now addresses this issue. Now you can run Kepler operator on the vanilla k8s cluster: https://github.com/sustainable-computing-io/kepler-operator?tab=readme-ov-file#run-kepler-operator-on-vanilla-kubernetes @elafontaine @jtaleric can you try deploying the same Again apologies for late response 😅
prefer to be able to run "make cluster-up" and get a kind cluster with the operator already deployed.
@elafontaine Yes you can use make fresh which will create a Kind cluster with all the dependencies and will also build, deploy the Operator on the cluster
Can you help me understand why is the certificate signing a requirement for kepler operator ?
Operator make use of webhooks which need to communicate securely over HTTPS. That is why we need valid TLS certs to establish the connection