harbor-cluster-operator
harbor-cluster-operator copied to clipboard
Kubernetes operator for deploying and managing Harbor and its dependent services (database, cache and storage) in a scalable and high-available way
``` 2020-10-19T11:23:00.842Z ERROR controller Reconciler error {"reconcilerGroup": "goharbor.io", "reconcilerKind": "HarborCluster", "controller": "harborcluster", "name": "sz-harbor-cluster", "namespace": "sample", "error": "Harbor.goharbor.io \"sz-harbor-cluster-harbor\" is invalid: spec.components.core.cacheSecret: Required value"} github.com/go-logr/zapr.(*zapLogger).Error /go/pkg/mod/github.com/go-logr/[email protected]/zapr.go:132 sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).reconcileHandler /go/pkg/mod/sigs.k8s.io/[email protected]/pkg/internal/controller/controller.go:237 sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).processNextWorkItem /go/pkg/mod/sigs.k8s.io/[email protected]/pkg/internal/controller/controller.go:209...
In v0.5.0 there's a known issue that sometimes when deploy harbor, that pod sz-harbor-cluster-harbor-registry-68974d455d-27c5t container `registry` may failed to start, and in log it will report parse /etc/registry/config.yaml failed or...
Some of the `Cluster` and `ClusterRole` are put under the namespace, this isn't correct and will cause some failure when deploying the operator in some kinds of Kubernetes environment https://github.com/goharbor/harbor-cluster-operator/blob/master/manifests/all-in-one.yaml#L2302
change dnsNames something like this long, harbor installation will fail `a22bfe0fb89664d67a8dda658ed65028-08c77fe69bd59516.abc.ap-southeast-1.amazonaws1.com` ![image](https://user-images.githubusercontent.com/188115/95834918-2a2ca800-0d70-11eb-97ea-5f8bd3097202.png) however if change to a shorter name like `ap-southeast-1.amazonaws1.com`, it will be success. It seems there's some length...
All pods' name begins with the name of the harbor cluster excepted redis and postgres. Can we make the pod name of redis and postgres also begins with the name...
We can follow the interface defined in the harbor operator. This can be the second step of merging the two operators.
Please include the component name for the pod of postgres so that it's easy to know the components used by the harbor. ![image](https://user-images.githubusercontent.com/5634205/93410784-3b4adc00-f8cc-11ea-83e0-c0662e07b99d.png)
We only need two ingresses for the harbor instance, one for the (core, portal), and one for notary. ![image](https://user-images.githubusercontent.com/5634205/93416050-85858a80-f8d7-11ea-8b11-ae837e3c450a.png)
Seems that the command tries to delete the resources which are already removed by deleting the namespaces ```shell yinw-a01:harbor-cluster-operator yinw$ kubectl delete -f manifests/all-in-one.yaml namespace "harbor-cluster-operator-system" deleted namespace "harbor-operator-system" deleted...
https://book.kubebuilder.io/reference/generating-crd.html#scale