|
| 1 | +/* |
| 2 | +Copyright 2023. |
| 3 | +
|
| 4 | +Licensed under the Apache License, Version 2.0 (the "License"); |
| 5 | +you may not use this file except in compliance with the License. |
| 6 | +You may obtain a copy of the License at |
| 7 | +
|
| 8 | + http://www.apache.org/licenses/LICENSE-2.0 |
| 9 | +
|
| 10 | +Unless required by applicable law or agreed to in writing, software |
| 11 | +distributed under the License is distributed on an "AS IS" BASIS, |
| 12 | +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 13 | +See the License for the specific language governing permissions and |
| 14 | +limitations under the License. |
| 15 | +*/ |
| 16 | + |
| 17 | +package controllers |
| 18 | + |
| 19 | +import ( |
| 20 | + "context" |
| 21 | + "fmt" |
| 22 | + "strconv" |
| 23 | + "time" |
| 24 | + |
| 25 | + "github.com/go-logr/logr" |
| 26 | + networkv1 "github.com/k8snetworkplumbingwg/network-attachment-definition-client/pkg/apis/k8s.cni.cncf.io/v1" |
| 27 | + "github.com/openstack-k8s-operators/lib-common/modules/common" |
| 28 | + "github.com/openstack-k8s-operators/lib-common/modules/common/condition" |
| 29 | + "github.com/openstack-k8s-operators/lib-common/modules/common/helper" |
| 30 | + nad "github.com/openstack-k8s-operators/lib-common/modules/common/networkattachment" |
| 31 | + corev1 "k8s.io/api/core/v1" |
| 32 | + k8s_errors "k8s.io/apimachinery/pkg/api/errors" |
| 33 | + ctrl "sigs.k8s.io/controller-runtime" |
| 34 | + "sigs.k8s.io/controller-runtime/pkg/client" |
| 35 | + "sigs.k8s.io/controller-runtime/pkg/controller/controllerutil" |
| 36 | +) |
| 37 | + |
| 38 | +// FrameworkInstance defines the interface that all test framework CRs must implement |
| 39 | +type FrameworkInstance interface { |
| 40 | + client.Object |
| 41 | + GetConditions() *condition.Conditions |
| 42 | +} |
| 43 | + |
| 44 | +// FrameworkConfig defines framework-specific configuration and behavior |
| 45 | +type FrameworkConfig[T FrameworkInstance] struct { |
| 46 | + // ServiceName for labeling (e.g., "tempest", "tobiko") |
| 47 | + ServiceName string |
| 48 | + |
| 49 | + // NeedsNetworkAttachments indicates if NADs should be handled |
| 50 | + NeedsNetworkAttachments bool |
| 51 | + |
| 52 | + // NeedsConfigMaps indicates if ServiceConfigReadyCondition is needed |
| 53 | + NeedsConfigMaps bool |
| 54 | + |
| 55 | + // GenerateServiceConfigMaps creates framework-specific config maps |
| 56 | + GenerateServiceConfigMaps func(ctx context.Context, r *Reconciler, helper *helper.Helper, instance T, workflowStep int) error |
| 57 | + |
| 58 | + // BuildPod creates the framework-specific pod definition |
| 59 | + BuildPod func(ctx context.Context, r *Reconciler, instance T, labels, annotations map[string]string, workflowStep int) (*corev1.Pod, error) |
| 60 | + |
| 61 | + // GetInitialConditions returns the condition list for a new instance |
| 62 | + GetInitialConditions func() []*condition.Condition |
| 63 | + |
| 64 | + // Field accessors |
| 65 | + GetWorkflowLength func(instance T) int |
| 66 | + GetParallel func(instance T) bool |
| 67 | + GetStorageClass func(instance T) string |
| 68 | + GetNetworkAttachments func(instance T) []string |
| 69 | + GetNetworkAttachmentStatus func(instance T) map[string][]string |
| 70 | + SetNetworkAttachmentStatus func(instance T, status map[string][]string) |
| 71 | + |
| 72 | + GetSpec func(instance T) interface{} // Optional |
| 73 | + GetWorkflowStep func(instance T, step int) interface{} // Optional |
| 74 | +} |
| 75 | + |
| 76 | +// CommonReconcile executes the standard reconciliation workflow using generics |
| 77 | +func CommonReconcile[T FrameworkInstance]( |
| 78 | + ctx context.Context, |
| 79 | + r *Reconciler, |
| 80 | + req ctrl.Request, |
| 81 | + instance T, |
| 82 | + config FrameworkConfig[T], |
| 83 | + Log logr.Logger, |
| 84 | +) (result ctrl.Result, _err error) { |
| 85 | + err := r.Client.Get(ctx, req.NamespacedName, instance) |
| 86 | + if err != nil { |
| 87 | + if k8s_errors.IsNotFound(err) { |
| 88 | + return ctrl.Result{}, nil |
| 89 | + } |
| 90 | + return ctrl.Result{}, err |
| 91 | + } |
| 92 | + |
| 93 | + // Create a helper |
| 94 | + helper, err := helper.NewHelper(instance, r.Client, r.Kclient, r.Scheme, r.Log) |
| 95 | + if err != nil { |
| 96 | + return ctrl.Result{}, err |
| 97 | + } |
| 98 | + |
| 99 | + // Get conditions from instance |
| 100 | + conditions := instance.GetConditions() |
| 101 | + if conditions == nil { |
| 102 | + return ctrl.Result{}, nil // TODO fmt.Errorf("instance does not support conditions") |
| 103 | + } |
| 104 | + |
| 105 | + // Initialize status |
| 106 | + isNewInstance := len(*conditions) == 0 |
| 107 | + if isNewInstance { |
| 108 | + *conditions = condition.Conditions{} |
| 109 | + } |
| 110 | + |
| 111 | + // Save a copy of the condtions so that we can restore the LastTransitionTime |
| 112 | + // when a condition's state doesn't change. |
| 113 | + savedConditions := conditions.DeepCopy() |
| 114 | + |
| 115 | + // Always patch the instance status when exiting this function so we |
| 116 | + // can persist any changes. |
| 117 | + defer func() { |
| 118 | + // update the overall status condition if service is ready |
| 119 | + if conditions.AllSubConditionIsTrue() { |
| 120 | + conditions.MarkTrue(condition.ReadyCondition, condition.ReadyMessage) |
| 121 | + } |
| 122 | + condition.RestoreLastTransitionTimes(conditions, savedConditions) |
| 123 | + if conditions.IsUnknown(condition.ReadyCondition) { |
| 124 | + conditions.Set(conditions.Mirror(condition.ReadyCondition)) |
| 125 | + } |
| 126 | + err := helper.PatchInstance(ctx, instance) |
| 127 | + if err != nil { |
| 128 | + _err = err |
| 129 | + } |
| 130 | + }() |
| 131 | + |
| 132 | + if isNewInstance { |
| 133 | + cl := condition.CreateList(config.GetInitialConditions()...) |
| 134 | + conditions.Init(&cl) |
| 135 | + |
| 136 | + // Register overall status immediately to have an early feedback |
| 137 | + // e.g. in the cli |
| 138 | + return ctrl.Result{}, nil |
| 139 | + } |
| 140 | + |
| 141 | + // Initialize network attachments status if needed |
| 142 | + if config.NeedsNetworkAttachments { |
| 143 | + if config.GetNetworkAttachmentStatus(instance) == nil { |
| 144 | + config.SetNetworkAttachmentStatus(instance, map[string][]string{}) |
| 145 | + } |
| 146 | + } |
| 147 | + |
| 148 | + // Handle service delete |
| 149 | + if !instance.GetDeletionTimestamp().IsZero() { |
| 150 | + Log.Info("Reconciling Service delete") |
| 151 | + controllerutil.RemoveFinalizer(instance, helper.GetFinalizer()) |
| 152 | + Log.Info("Reconciled Service delete successfully") |
| 153 | + return ctrl.Result{}, nil |
| 154 | + } |
| 155 | + |
| 156 | + workflowLength := config.GetWorkflowLength(instance) |
| 157 | + nextAction, workflowStep, err := r.NextAction(ctx, instance, workflowLength) |
| 158 | + |
| 159 | + // Merge workflow step if applicable |
| 160 | + if workflowLength != 0 && workflowStep < workflowLength { |
| 161 | + spec := config.GetSpec(instance) |
| 162 | + workflowStepData := config.GetWorkflowStep(instance, workflowStep) |
| 163 | + MergeSections(spec, workflowStepData) |
| 164 | + } |
| 165 | + |
| 166 | + switch nextAction { |
| 167 | + case Failure: |
| 168 | + return ctrl.Result{}, err |
| 169 | + |
| 170 | + case Wait: |
| 171 | + Log.Info(InfoWaitingOnPod) |
| 172 | + return ctrl.Result{RequeueAfter: RequeueAfterValue}, nil |
| 173 | + |
| 174 | + case EndTesting: |
| 175 | + // All pods created by the instance were completed. Release the lock |
| 176 | + // so that other instances can spawn their pods. |
| 177 | + if lockReleased, err := r.ReleaseLock(ctx, instance); !lockReleased { |
| 178 | + Log.Info(fmt.Sprintf(InfoCanNotReleaseLock, testOperatorLockName)) |
| 179 | + return ctrl.Result{RequeueAfter: RequeueAfterValue}, err |
| 180 | + } |
| 181 | + |
| 182 | + conditions.MarkTrue(condition.DeploymentReadyCondition, condition.DeploymentReadyMessage) |
| 183 | + Log.Info(InfoTestingCompleted) |
| 184 | + return ctrl.Result{}, nil |
| 185 | + |
| 186 | + case CreateFirstPod: |
| 187 | + lockAcquired, err := r.AcquireLock(ctx, instance, helper, config.GetParallel(instance)) |
| 188 | + if !lockAcquired { |
| 189 | + Log.Info(fmt.Sprintf(InfoCanNotAcquireLock, testOperatorLockName)) |
| 190 | + return ctrl.Result{RequeueAfter: RequeueAfterValue}, err |
| 191 | + } |
| 192 | + |
| 193 | + Log.Info(fmt.Sprintf(InfoCreatingFirstPod, workflowStep)) |
| 194 | + |
| 195 | + case CreateNextPod: |
| 196 | + // Confirm that we still hold the lock. This is useful to check if for |
| 197 | + // example somebody / something deleted the lock and it got claimed by |
| 198 | + // another instance. This is considered to be an error state. |
| 199 | + lockAcquired, err := r.AcquireLock(ctx, instance, helper, config.GetParallel(instance)) |
| 200 | + if !lockAcquired { |
| 201 | + Log.Error(err, ErrConfirmLockOwnership, testOperatorLockName) |
| 202 | + return ctrl.Result{RequeueAfter: RequeueAfterValue}, err |
| 203 | + } |
| 204 | + |
| 205 | + Log.Info(fmt.Sprintf(InfoCreatingNextPod, workflowStep)) |
| 206 | + |
| 207 | + default: |
| 208 | + return ctrl.Result{}, ErrReceivedUnexpectedAction |
| 209 | + } |
| 210 | + |
| 211 | + serviceLabels := map[string]string{ |
| 212 | + common.AppSelector: config.ServiceName, |
| 213 | + workflowStepLabel: strconv.Itoa(workflowStep), |
| 214 | + instanceNameLabel: instance.GetName(), |
| 215 | + operatorNameLabel: "test-operator", |
| 216 | + } |
| 217 | + |
| 218 | + workflowStepNum := 0 |
| 219 | + // Create multiple PVCs for parallel execution |
| 220 | + if config.GetParallel(instance) && workflowStep < config.GetWorkflowLength(instance) { |
| 221 | + workflowStepNum = workflowStep |
| 222 | + } |
| 223 | + |
| 224 | + // Create PersistentVolumeClaim |
| 225 | + ctrlResult, err := r.EnsureLogsPVCExists( |
| 226 | + ctx, |
| 227 | + instance, |
| 228 | + helper, |
| 229 | + serviceLabels, |
| 230 | + config.GetStorageClass(instance), |
| 231 | + workflowStepNum, |
| 232 | + ) |
| 233 | + if err != nil { |
| 234 | + return ctrlResult, err |
| 235 | + } else if (ctrlResult != ctrl.Result{}) { |
| 236 | + return ctrlResult, nil |
| 237 | + } |
| 238 | + |
| 239 | + // Generate ConfigMaps if needed |
| 240 | + if config.NeedsConfigMaps { |
| 241 | + if err = config.GenerateServiceConfigMaps(ctx, r, helper, instance, workflowStep); err != nil { |
| 242 | + conditions.Set(condition.FalseCondition( |
| 243 | + condition.ServiceConfigReadyCondition, |
| 244 | + condition.ErrorReason, |
| 245 | + condition.SeverityWarning, |
| 246 | + condition.ServiceConfigReadyErrorMessage, |
| 247 | + err.Error())) |
| 248 | + return ctrl.Result{}, err |
| 249 | + } |
| 250 | + conditions.MarkTrue(condition.ServiceConfigReadyCondition, condition.ServiceConfigReadyMessage) |
| 251 | + } |
| 252 | + // Generate ConfigMaps - end |
| 253 | + |
| 254 | + // Handle network attachments if needed |
| 255 | + var serviceAnnotations map[string]string |
| 256 | + if config.NeedsNetworkAttachments { |
| 257 | + annotations, ctrlResult, err := handleNetworkAttachments( |
| 258 | + ctx, r, instance, helper, serviceLabels, config, workflowStep, conditions, |
| 259 | + ) |
| 260 | + if err != nil || (ctrlResult != ctrl.Result{}) { |
| 261 | + return ctrlResult, err |
| 262 | + } |
| 263 | + serviceAnnotations = annotations |
| 264 | + } |
| 265 | + |
| 266 | + // Build pod |
| 267 | + podDef, err := config.BuildPod(ctx, r, instance, serviceLabels, serviceAnnotations, workflowStep) |
| 268 | + if err != nil { |
| 269 | + return ctrl.Result{}, err |
| 270 | + } |
| 271 | + |
| 272 | + // Create a new pod |
| 273 | + ctrlResult, err = r.CreatePod(ctx, *helper, podDef) |
| 274 | + if err != nil { |
| 275 | + // Release lock on failure |
| 276 | + if lockReleased, lockErr := r.ReleaseLock(ctx, instance); lockReleased { |
| 277 | + return ctrl.Result{RequeueAfter: RequeueAfterValue}, lockErr |
| 278 | + } |
| 279 | + |
| 280 | + conditions.Set(condition.FalseCondition( |
| 281 | + condition.DeploymentReadyCondition, |
| 282 | + condition.ErrorReason, |
| 283 | + condition.SeverityWarning, |
| 284 | + condition.DeploymentReadyErrorMessage, |
| 285 | + err.Error())) |
| 286 | + return ctrlResult, err |
| 287 | + } else if (ctrlResult != ctrl.Result{}) { |
| 288 | + conditions.Set(condition.FalseCondition( |
| 289 | + condition.DeploymentReadyCondition, |
| 290 | + condition.RequestedReason, |
| 291 | + condition.SeverityInfo, |
| 292 | + condition.DeploymentReadyRunningMessage)) |
| 293 | + return ctrlResult, nil |
| 294 | + } |
| 295 | + // Create a new pod - end |
| 296 | + |
| 297 | + return ctrl.Result{}, nil |
| 298 | +} |
| 299 | + |
| 300 | +func handleNetworkAttachments[T FrameworkInstance]( |
| 301 | + ctx context.Context, |
| 302 | + r *Reconciler, |
| 303 | + instance T, |
| 304 | + helper *helper.Helper, |
| 305 | + labels map[string]string, |
| 306 | + config FrameworkConfig[T], |
| 307 | + workflowStep int, |
| 308 | + conditions *condition.Conditions, |
| 309 | +) (map[string]string, ctrl.Result, error) { |
| 310 | + nadList := []networkv1.NetworkAttachmentDefinition{} |
| 311 | + networkAttachments := config.GetNetworkAttachments(instance) |
| 312 | + |
| 313 | + for _, netAtt := range networkAttachments { |
| 314 | + nad, err := nad.GetNADWithName(ctx, helper, netAtt, instance.GetNamespace()) |
| 315 | + if err != nil { |
| 316 | + if k8s_errors.IsNotFound(err) { |
| 317 | + // Since the net-attach-def CR should have been manually created by the user and referenced in the spec, |
| 318 | + // we treat this as a warning because it means that the service will not be able to start. |
| 319 | + r.Log.Info(fmt.Sprintf("network-attachment-definition %s not found", netAtt)) |
| 320 | + conditions.Set(condition.FalseCondition( |
| 321 | + condition.NetworkAttachmentsReadyCondition, |
| 322 | + condition.ErrorReason, |
| 323 | + condition.SeverityWarning, |
| 324 | + condition.NetworkAttachmentsReadyWaitingMessage, |
| 325 | + netAtt)) |
| 326 | + return nil, ctrl.Result{RequeueAfter: time.Second * 10}, nil |
| 327 | + } |
| 328 | + conditions.Set(condition.FalseCondition( |
| 329 | + condition.NetworkAttachmentsReadyCondition, |
| 330 | + condition.ErrorReason, |
| 331 | + condition.SeverityWarning, |
| 332 | + condition.NetworkAttachmentsReadyErrorMessage, |
| 333 | + err.Error())) |
| 334 | + return nil, ctrl.Result{}, err |
| 335 | + } |
| 336 | + |
| 337 | + if nad != nil { |
| 338 | + nadList = append(nadList, *nad) |
| 339 | + } |
| 340 | + } |
| 341 | + |
| 342 | + serviceAnnotations, err := nad.EnsureNetworksAnnotation(nadList) |
| 343 | + if err != nil { |
| 344 | + return nil, ctrl.Result{}, fmt.Errorf("failed create network annotation from %s: %w", |
| 345 | + networkAttachments, err) |
| 346 | + } |
| 347 | + |
| 348 | + // Verify network status if pod exists |
| 349 | + if r.PodExists(ctx, instance, workflowStep) { |
| 350 | + networkReady, networkAttachmentStatus, err := nad.VerifyNetworkStatusFromAnnotation( |
| 351 | + ctx, |
| 352 | + helper, |
| 353 | + networkAttachments, |
| 354 | + labels, |
| 355 | + 1, |
| 356 | + ) |
| 357 | + if err != nil { |
| 358 | + return nil, ctrl.Result{}, err |
| 359 | + } |
| 360 | + |
| 361 | + config.SetNetworkAttachmentStatus(instance, networkAttachmentStatus) |
| 362 | + |
| 363 | + if networkReady { |
| 364 | + conditions.MarkTrue( |
| 365 | + condition.NetworkAttachmentsReadyCondition, |
| 366 | + condition.NetworkAttachmentsReadyMessage) |
| 367 | + } else { |
| 368 | + err := fmt.Errorf("%w: %s", ErrNetworkAttachmentsMismatch, networkAttachments) |
| 369 | + conditions.Set(condition.FalseCondition( |
| 370 | + condition.NetworkAttachmentsReadyCondition, |
| 371 | + condition.ErrorReason, |
| 372 | + condition.SeverityWarning, |
| 373 | + condition.NetworkAttachmentsReadyErrorMessage, |
| 374 | + err.Error())) |
| 375 | + return nil, ctrl.Result{}, err |
| 376 | + } |
| 377 | + } |
| 378 | + |
| 379 | + return serviceAnnotations, ctrl.Result{}, nil |
| 380 | +} |
0 commit comments