zitadel/operator/crtlcrd/controller.go

75 lines
1.9 KiB
Go
Raw Normal View History

fix: operator picks (#1463) * feat(crd): add crd mode for operators (#1329) * feat(operator): add base for zitadel operator * fix(operator): changed pipeline to release operator * fix(operator): fmt with only one parameter * fix(operator): corrected workflow job name * fix(zitadelctl): added restore and backuplist command * fix(zitadelctl): scale for restore * chore(container): use scratch for deploy container * fix(zitadelctl): limit image to scratch * fix(migration): added migration scripts for newer version * fix(operator): changed handling of kubeconfig in operator logic * fix(operator): changed handling of secrets in operator logic * fix(operator): use new version of zitadel * fix(operator): added path for migrations * fix(operator): delete doublets of migration scripts * fix(operator): delete subpaths and integrate logic into init container * fix(operator): corrected path in dockerfile for local migrations * fix(operator): added migrations for cockroachdb-secure * fix(operator): delete logic for ambassador module * fix(operator): added read and write secret commands * fix(operator): correct and align operator pipeline with zitadel pipeline * fix(operator): correct yaml error in operator pipeline * fix(operator): correct action name in operator pipeline * fix(operator): correct case-sensitive filename in operator pipeline * fix(operator): upload artifacts from buildx output * fix(operator): corrected attribute spelling error * fix(operator): combined jobs for operator binary and image * fix(operator): added missing comma in operator pipeline * fix(operator): added codecov for operator image * fix(operator): added codecov for operator image * fix(testing): code changes for testing and several unit-tests (#1009) * fix(operator): usage of interface of kubernetes client for testing and several unit-tests * fix(operator): several unit-tests * fix(operator): several unit-tests * fix(operator): changed order for the operator logic * fix(operator): added version of zitadelctl from semantic release * fix(operator): corrected function call with version of zitadelctl * fix(operator): corrected function call with version of zitadelctl * fix(operator): add check output to operator release pipeline * fix(operator): set --short length everywhere to 12 * fix(operator): zitadel setup in job instead of exec with several unit tests * fix(operator): fixes to combine newest zitadel and testing branch * fix(operator): corrected path in Dockerfile * fix(operator): fixed unit-test that was ignored during changes * fix(operator): fixed unit-test that was ignored during changes * fix(operator): corrected Dockerfile to correctly use env variable * fix(operator): quickfix takeoff deployment * fix(operator): corrected the clusterrolename in the applied artifacts * fix: update secure migrations * fix(operator): migrations (#1057) * fix(operator): copied migrations from orbos repository * fix(operator): newest migrations * chore: use cockroach-secure * fix: rename migration * fix: remove insecure cockroach migrations Co-authored-by: Stefan Benz <stefan@caos.ch> * fix: finalize labels * fix(operator): cli logging concurrent and fixe deployment of operator during restore * fix: finalize labels and cli commands * fix: restore * chore: cockroachdb is always secure * chore: use orbos consistent-labels latest commit * test: make tests compatible with new labels * fix: default to sa token for start command * fix: use cockroachdb v12.02 * fix: don't delete flyway user * test: fix migration test * fix: use correct table qualifiers * fix: don't alter sequence ownership * fix: upgrade flyway * fix: change ownership of all dbs and tables to admin user * fix: change defaultdb user * fix: treat clientid status codes >= 400 as errors * fix: reconcile specified ZITADEL version, not binary version * fix: add ca-certs * fix: use latest orbos code * fix: use orbos with fixed race condition * fix: use latest ORBOS code * fix: use latest ORBOS code * fix: make migration and scaling around restoring work * fix(operator): move zitadel operator * chore(migrations): include owner change migration * feat(db): add code base for database operator * fix(db): change used image registry for database operator * fix(db): generated mock * fix(db): add accidentally ignored file * fix(db): add cockroachdb backup image to pipeline * fix(db): correct pipeline and image versions * fix(db): correct version of used orbos * fix(db): correct database import * fix(db): go mod tidy * fix(db): use new version for orbos * fix(migrations): include migrations into zitadelctl binary (#1211) * fix(db): use statik to integrate migrations into binary * fix(migrations): corrections unit tests and pipeline for integrated migrations into zitadelctl binary * fix(migrations): correction in dockerfile for pipeline build * fix(migrations): correction in dockerfile for pipeline build * fix(migrations): dockerfile changes for cache optimization * fix(database): correct used part-of label in database operator * fix(database): correct used selectable label in zitadel operator * fix(operator): correct lables for user secrets in zitadel operator * fix(operator): correct lables for service test in zitadel operator * fix: don't enable database features for user operations (#1227) * fix: don't enable database features for user operations * fix: omit database feature for connection info adapter * fix: use latest orbos version * fix(crd): corrected logic to get database connection and other info * fix(crd): corrected yaml tags and start for zitadel operator * fix(crd): move some dependencies and use consistent structure * fix(crd): corrected unit-tests * fix(crd): corrected main files for debug starts * chore(pipeline): use correct version for zitadelctl build * fix(crd): correct calculating of current db state for zitadel operator * fix(crd): use binary version for deployment of crd mode operators * fix(crd): add gitops attribute for reconciling * fix(crd): corrected crd with newest version * fix(migration): collect cleanup functions and only use them if all jobs are successful * fix(zitadelctl): import gcp auth to connect to gke cluster * feat: Add read and writesecret options for crd mode (#1435) * fix: don't require orbconfig for crd mode * test: pass * fix(zitadelctl): import gcp auth to connect to gke cluster * feat: add read and writesecret option for crd mode * test: fix * fix: make all crd secrets writable * fix: use in-cluster configs for in-cluster operators * chore: remove unnecessary debug files Co-authored-by: Stefan Benz <stefan@caos.ch> * fix: Crdoperatormerge review (#1385) * fix: don't require orbconfig for crd mode * test: pass * fix(zitadelctl): import gcp auth to connect to gke cluster * fix: ensure caos-system namespace * fix: apply orbconfig at takeoff * docs: improve help for creating an orbconfig * docs: describe orbconfig properties * docs: add --gitops to help message example * fix(pipeline): correct upload of artifacts in dev releases * test: pass Co-authored-by: Stefan Benz <stefan@caos.ch> * fix(test): corrected falsely merged tests * chore: update orbos library * fix: only handle exactly named and namespaced crd resource * fix: print errors, check correct crd namespace * fix: validate bucket secret * chore: compile * fix(operator): corrected secret handling when unused secrets are not defined * fix(operator): corrected handling of jobs * fix: dont print logs when readsecret path is provided * fix(operator): corrected handling of jobs and sort for mounted volumes * fix(operator): sort for volumes * fix(operator): change orboos import to newest release Co-authored-by: Florian Forster <florian@caos.ch> Co-authored-by: Elio Bischof <eliobischof@gmail.com> (cherry picked from commit fa9bd5a8e7a5032b21272552fbc4cdf1630db049) * fix(operator): Standard timeout handling (#1458) * fix: always use standard time.Duration * fix: give backup and restore more time * fix: give backup and restore jobs more time (cherry picked from commit 7468b7d1e818102dcd58dbb6d3647f49277a4e65) * fix go mod Co-authored-by: Stefan Benz <46600784+stebenz@users.noreply.github.com> Co-authored-by: Elio Bischof <eliobischof@gmail.com>
2021-03-24 09:31:19 +00:00
package crtlcrd
import (
"github.com/caos/orbos/mntr"
"github.com/caos/orbos/pkg/kubernetes"
databasev1 "github.com/caos/zitadel/operator/api/database/v1"
zitadelv1 "github.com/caos/zitadel/operator/api/zitadel/v1"
"github.com/caos/zitadel/operator/crtlcrd/database"
"github.com/caos/zitadel/operator/crtlcrd/zitadel"
"github.com/pkg/errors"
"k8s.io/apimachinery/pkg/runtime"
clientgoscheme "k8s.io/client-go/kubernetes/scheme"
ctrl "sigs.k8s.io/controller-runtime"
)
const (
Database = "database"
Zitadel = "zitadel"
)
var (
scheme = runtime.NewScheme()
)
func init() {
_ = clientgoscheme.AddToScheme(scheme)
_ = databasev1.AddToScheme(scheme)
_ = zitadelv1.AddToScheme(scheme)
}
func Start(monitor mntr.Monitor, version, metricsAddr string, features ...string) error {
cfg := ctrl.GetConfigOrDie()
mgr, err := ctrl.NewManager(cfg, ctrl.Options{
Scheme: scheme,
MetricsBindAddress: metricsAddr,
Port: 9443,
LeaderElection: false,
LeaderElectionID: "9adsd12l.caos.ch",
})
if err != nil {
return errors.Wrap(err, "unable to start manager")
}
k8sClient := kubernetes.NewK8sClientWithConfig(monitor, cfg)
for _, feature := range features {
switch feature {
case Database:
if err = (&database.Reconciler{
ClientInt: k8sClient,
Monitor: monitor,
Scheme: mgr.GetScheme(),
Version: version,
}).SetupWithManager(mgr); err != nil {
return errors.Wrap(err, "unable to create controller")
}
case Zitadel:
if err = (&zitadel.Reconciler{
ClientInt: k8sClient,
Monitor: monitor,
Scheme: mgr.GetScheme(),
Version: version,
}).SetupWithManager(mgr); err != nil {
return errors.Wrap(err, "unable to create controller")
}
}
}
if err := mgr.Start(ctrl.SetupSignalHandler()); err != nil {
return errors.Wrap(err, "problem running manager")
}
return nil
}