mirror of
https://github.com/kubevirt/containerized-data-importer.git
synced 2025-06-03 06:30:22 +00:00

There should be no state shared between sync() and updateStatus(). updateStatus() should stand on it's own, and come to it's own conclusions based on what it observes. It is okay if it is "late to the party" and does not observe the latest changes in sync(). It will eventually converge. This is what kubevirt does. Signed-off-by: Arnon Gilboa <agilboa@redhat.com>
430 lines
16 KiB
Go
430 lines
16 KiB
Go
/*
|
|
Copyright 2020 The CDI Authors.
|
|
|
|
Licensed under the Apache License, Version 2.0 (the "License");
|
|
you may not use this file except in compliance with the License.
|
|
You may obtain a copy of the License at
|
|
|
|
http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
Unless required by applicable law or agreed to in writing, software
|
|
distributed under the License is distributed on an "AS IS" BASIS,
|
|
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
See the License for the specific language governing permissions and
|
|
limitations under the License.
|
|
*/
|
|
|
|
package datavolume
|
|
|
|
import (
|
|
"context"
|
|
"fmt"
|
|
"math"
|
|
"strconv"
|
|
|
|
"github.com/go-logr/logr"
|
|
"github.com/pkg/errors"
|
|
|
|
v1 "k8s.io/api/core/v1"
|
|
storagev1 "k8s.io/api/storage/v1"
|
|
"k8s.io/apimachinery/pkg/api/resource"
|
|
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
|
|
"k8s.io/apimachinery/pkg/types"
|
|
"k8s.io/client-go/tools/cache"
|
|
"k8s.io/client-go/tools/record"
|
|
storagehelpers "k8s.io/component-helpers/storage/volume"
|
|
"sigs.k8s.io/controller-runtime/pkg/client"
|
|
"sigs.k8s.io/controller-runtime/pkg/reconcile"
|
|
|
|
cdiv1 "kubevirt.io/containerized-data-importer-api/pkg/apis/core/v1beta1"
|
|
cc "kubevirt.io/containerized-data-importer/pkg/controller/common"
|
|
"kubevirt.io/containerized-data-importer/pkg/util"
|
|
)
|
|
|
|
const (
|
|
// AnnOwnedByDataVolume annotation has the owner DataVolume name
|
|
AnnOwnedByDataVolume = "cdi.kubevirt.io/ownedByDataVolume"
|
|
)
|
|
|
|
// renderPvcSpec creates a new PVC Spec based on either the dv.spec.pvc or dv.spec.storage section
|
|
func renderPvcSpec(client client.Client, recorder record.EventRecorder, log logr.Logger, dv *cdiv1.DataVolume) (*v1.PersistentVolumeClaimSpec, error) {
|
|
if dv.Spec.PVC != nil {
|
|
return dv.Spec.PVC.DeepCopy(), nil
|
|
}
|
|
|
|
if dv.Spec.Storage != nil {
|
|
return pvcFromStorage(client, recorder, log, dv)
|
|
}
|
|
|
|
return nil, errors.Errorf("datavolume one of {pvc, storage} field is required")
|
|
}
|
|
|
|
func pvcFromStorage(client client.Client, recorder record.EventRecorder, log logr.Logger, dv *cdiv1.DataVolume) (*v1.PersistentVolumeClaimSpec, error) {
|
|
storage := dv.Spec.Storage
|
|
pvcSpec := copyStorageAsPvc(log, storage)
|
|
|
|
if dv.Spec.ContentType == cdiv1.DataVolumeArchive {
|
|
if pvcSpec.VolumeMode != nil && *pvcSpec.VolumeMode == v1.PersistentVolumeBlock {
|
|
log.V(1).Info("DataVolume with ContentType Archive cannot have block volumeMode", "namespace", dv.Namespace, "name", dv.Name)
|
|
recorder.Eventf(dv, v1.EventTypeWarning, cc.ErrClaimNotValid, "DataVolume with ContentType Archive cannot have block volumeMode")
|
|
return nil, errors.Errorf("DataVolume with ContentType Archive cannot have block volumeMode")
|
|
}
|
|
volumeMode := v1.PersistentVolumeFilesystem
|
|
pvcSpec.VolumeMode = &volumeMode
|
|
}
|
|
|
|
storageClass, err := cc.GetStorageClassByName(client, storage.StorageClassName)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
if storageClass == nil {
|
|
// Not even default storageClass on the cluster, cannot apply the defaults, verify spec is ok
|
|
if len(pvcSpec.AccessModes) == 0 {
|
|
log.V(1).Info("Cannot set accessMode for new pvc", "namespace", dv.Namespace, "name", dv.Name)
|
|
recorder.Eventf(dv, v1.EventTypeWarning, cc.ErrClaimNotValid, "DataVolume.storage spec is missing accessMode and no storageClass to choose profile")
|
|
return nil, errors.Errorf("DataVolume spec is missing accessMode")
|
|
}
|
|
} else {
|
|
pvcSpec.StorageClassName = &storageClass.Name
|
|
// given storageClass we can apply defaults if needed
|
|
if (pvcSpec.VolumeMode == nil || *pvcSpec.VolumeMode == "") && (len(pvcSpec.AccessModes) == 0) {
|
|
accessModes, volumeMode, err := getDefaultVolumeAndAccessMode(client, storageClass)
|
|
if err != nil {
|
|
log.V(1).Info("Cannot set accessMode and volumeMode for new pvc", "namespace", dv.Namespace, "name", dv.Name, "Error", err)
|
|
recorder.Eventf(dv, v1.EventTypeWarning, cc.ErrClaimNotValid,
|
|
fmt.Sprintf("DataVolume.storage spec is missing accessMode and volumeMode, cannot get access mode from StorageProfile %s", getName(storageClass)))
|
|
return nil, err
|
|
}
|
|
pvcSpec.AccessModes = append(pvcSpec.AccessModes, accessModes...)
|
|
pvcSpec.VolumeMode = volumeMode
|
|
} else if len(pvcSpec.AccessModes) == 0 {
|
|
accessModes, err := getDefaultAccessModes(client, storageClass, pvcSpec.VolumeMode)
|
|
if err != nil {
|
|
log.V(1).Info("Cannot set accessMode for new pvc", "namespace", dv.Namespace, "name", dv.Name, "Error", err)
|
|
recorder.Eventf(dv, v1.EventTypeWarning, cc.ErrClaimNotValid,
|
|
fmt.Sprintf("DataVolume.storage spec is missing accessMode and cannot get access mode from StorageProfile %s", getName(storageClass)))
|
|
return nil, err
|
|
}
|
|
pvcSpec.AccessModes = append(pvcSpec.AccessModes, accessModes...)
|
|
} else if pvcSpec.VolumeMode == nil || *pvcSpec.VolumeMode == "" {
|
|
volumeMode, err := getDefaultVolumeMode(client, storageClass, pvcSpec.AccessModes)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
pvcSpec.VolumeMode = volumeMode
|
|
}
|
|
}
|
|
|
|
requestedVolumeSize, err := resolveVolumeSize(client, dv.Spec, pvcSpec)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
if pvcSpec.Resources.Requests == nil {
|
|
pvcSpec.Resources.Requests = v1.ResourceList{}
|
|
}
|
|
pvcSpec.Resources.Requests[v1.ResourceStorage] = *requestedVolumeSize
|
|
|
|
return pvcSpec, nil
|
|
}
|
|
|
|
func getName(storageClass *storagev1.StorageClass) string {
|
|
if storageClass != nil {
|
|
return storageClass.Name
|
|
}
|
|
return ""
|
|
}
|
|
|
|
func copyStorageAsPvc(log logr.Logger, storage *cdiv1.StorageSpec) *v1.PersistentVolumeClaimSpec {
|
|
input := storage.DeepCopy()
|
|
pvcSpec := &v1.PersistentVolumeClaimSpec{
|
|
AccessModes: input.AccessModes,
|
|
Selector: input.Selector,
|
|
Resources: input.Resources,
|
|
VolumeName: input.VolumeName,
|
|
StorageClassName: input.StorageClassName,
|
|
VolumeMode: input.VolumeMode,
|
|
DataSource: input.DataSource,
|
|
DataSourceRef: input.DataSourceRef,
|
|
}
|
|
|
|
return pvcSpec
|
|
}
|
|
|
|
func getDefaultVolumeAndAccessMode(c client.Client, storageClass *storagev1.StorageClass) ([]v1.PersistentVolumeAccessMode, *v1.PersistentVolumeMode, error) {
|
|
if storageClass == nil {
|
|
return nil, nil, errors.Errorf("no accessMode defined on DV and no StorageProfile")
|
|
}
|
|
|
|
storageProfile := &cdiv1.StorageProfile{}
|
|
err := c.Get(context.TODO(), types.NamespacedName{Name: storageClass.Name}, storageProfile)
|
|
if err != nil {
|
|
return nil, nil, errors.Wrap(err, "cannot get StorageProfile")
|
|
}
|
|
|
|
if len(storageProfile.Status.ClaimPropertySets) > 0 &&
|
|
len(storageProfile.Status.ClaimPropertySets[0].AccessModes) > 0 {
|
|
accessModes := storageProfile.Status.ClaimPropertySets[0].AccessModes
|
|
volumeMode := storageProfile.Status.ClaimPropertySets[0].VolumeMode
|
|
return accessModes, volumeMode, nil
|
|
}
|
|
|
|
// no accessMode configured on storageProfile
|
|
return nil, nil, errors.Errorf("no accessMode defined DV nor on StorageProfile for %s StorageClass", storageClass.Name)
|
|
}
|
|
|
|
func getDefaultVolumeMode(c client.Client, storageClass *storagev1.StorageClass, pvcAccessModes []v1.PersistentVolumeAccessMode) (*v1.PersistentVolumeMode, error) {
|
|
if storageClass == nil {
|
|
// fallback to k8s defaults
|
|
return nil, nil
|
|
}
|
|
|
|
storageProfile := &cdiv1.StorageProfile{}
|
|
err := c.Get(context.TODO(), types.NamespacedName{Name: storageClass.Name}, storageProfile)
|
|
if err != nil {
|
|
return nil, errors.Wrap(err, "cannot get StorageProfile")
|
|
}
|
|
if len(storageProfile.Status.ClaimPropertySets) > 0 {
|
|
volumeMode := storageProfile.Status.ClaimPropertySets[0].VolumeMode
|
|
if len(pvcAccessModes) == 0 {
|
|
return volumeMode, nil
|
|
}
|
|
// check for volume mode matching with given pvc access modes
|
|
for _, cps := range storageProfile.Status.ClaimPropertySets {
|
|
for _, accessMode := range cps.AccessModes {
|
|
for _, pvcAccessMode := range pvcAccessModes {
|
|
if accessMode == pvcAccessMode {
|
|
return cps.VolumeMode, nil
|
|
}
|
|
}
|
|
}
|
|
}
|
|
// if not found return default volume mode for the storage class
|
|
return volumeMode, nil
|
|
}
|
|
|
|
// since volumeMode is optional - > gracefully fallback to k8s defaults,
|
|
return nil, nil
|
|
}
|
|
|
|
func getDefaultAccessModes(c client.Client, storageClass *storagev1.StorageClass, pvcVolumeMode *v1.PersistentVolumeMode) ([]v1.PersistentVolumeAccessMode, error) {
|
|
if storageClass == nil {
|
|
return nil, errors.Errorf("no accessMode defined on DV, no StorageProfile ")
|
|
}
|
|
|
|
storageProfile := &cdiv1.StorageProfile{}
|
|
err := c.Get(context.TODO(), types.NamespacedName{Name: storageClass.Name}, storageProfile)
|
|
if err != nil {
|
|
return nil, errors.Wrap(err, "no accessMode defined on DV, cannot get StorageProfile")
|
|
}
|
|
|
|
if len(storageProfile.Status.ClaimPropertySets) > 0 {
|
|
// check for access modes matching with given pvc volume mode
|
|
defaultAccessModes := []v1.PersistentVolumeAccessMode{}
|
|
for _, cps := range storageProfile.Status.ClaimPropertySets {
|
|
if cps.VolumeMode != nil && pvcVolumeMode != nil && *cps.VolumeMode == *pvcVolumeMode {
|
|
if len(cps.AccessModes) > 0 {
|
|
return cps.AccessModes, nil
|
|
}
|
|
} else if len(cps.AccessModes) > 0 && len(defaultAccessModes) == 0 {
|
|
defaultAccessModes = cps.AccessModes
|
|
}
|
|
}
|
|
// if not found return default access modes for the storage profile
|
|
if len(defaultAccessModes) > 0 {
|
|
return defaultAccessModes, nil
|
|
}
|
|
}
|
|
|
|
// no accessMode configured on storageProfile
|
|
return nil, errors.Errorf("no accessMode defined on StorageProfile for %s StorageClass", storageClass.Name)
|
|
}
|
|
|
|
func resolveVolumeSize(c client.Client, dvSpec cdiv1.DataVolumeSpec, pvcSpec *v1.PersistentVolumeClaimSpec) (*resource.Quantity, error) {
|
|
// resources.requests[storage] - just copy it to pvc,
|
|
requestedSize, found := dvSpec.Storage.Resources.Requests[v1.ResourceStorage]
|
|
|
|
if !found {
|
|
// Storage size can be empty when cloning
|
|
isClone := dvSpec.Source.PVC != nil
|
|
if isClone {
|
|
return &requestedSize, nil
|
|
}
|
|
return nil, errors.Errorf("Datavolume Spec is not valid - missing storage size")
|
|
}
|
|
|
|
// disk or image size, inflate it with overhead
|
|
requestedSize, err := inflateSizeWithOverhead(c, requestedSize.Value(), pvcSpec)
|
|
|
|
return &requestedSize, err
|
|
}
|
|
|
|
// inflateSizeWithOverhead inflates a storage size with proper overhead calculations
|
|
func inflateSizeWithOverhead(c client.Client, imgSize int64, pvcSpec *v1.PersistentVolumeClaimSpec) (resource.Quantity, error) {
|
|
var returnSize resource.Quantity
|
|
|
|
if util.ResolveVolumeMode(pvcSpec.VolumeMode) == v1.PersistentVolumeFilesystem {
|
|
fsOverhead, err := cc.GetFilesystemOverheadForStorageClass(c, pvcSpec.StorageClassName)
|
|
if err != nil {
|
|
return resource.Quantity{}, err
|
|
}
|
|
// Parse filesystem overhead (percentage) into a 64-bit float
|
|
fsOverheadFloat, _ := strconv.ParseFloat(string(fsOverhead), 64)
|
|
|
|
// Merge the previous values into a 'resource.Quantity' struct
|
|
requiredSpace := GetRequiredSpace(fsOverheadFloat, imgSize)
|
|
returnSize = *resource.NewScaledQuantity(requiredSpace, 0)
|
|
} else {
|
|
// Inflation is not needed with 'Block' mode
|
|
returnSize = *resource.NewScaledQuantity(imgSize, 0)
|
|
}
|
|
|
|
return returnSize, nil
|
|
}
|
|
|
|
// GetRequiredSpace calculates space required taking file system overhead into account
|
|
func GetRequiredSpace(filesystemOverhead float64, requestedSpace int64) int64 {
|
|
// the `image` has to be aligned correctly, so the space requested has to be aligned to
|
|
// next value that is a multiple of a block size
|
|
alignedSize := util.RoundUp(requestedSpace, util.DefaultAlignBlockSize)
|
|
|
|
// count overhead as a percentage of the whole/new size, including aligned image
|
|
// and the space required by filesystem metadata
|
|
spaceWithOverhead := int64(math.Ceil(float64(alignedSize) / (1 - filesystemOverhead)))
|
|
return spaceWithOverhead
|
|
}
|
|
|
|
func createStorageProfile(name string,
|
|
accessModes []v1.PersistentVolumeAccessMode,
|
|
volumeMode v1.PersistentVolumeMode) *cdiv1.StorageProfile {
|
|
claimPropertySets := []cdiv1.ClaimPropertySet{{
|
|
AccessModes: accessModes,
|
|
VolumeMode: &volumeMode,
|
|
}}
|
|
return createStorageProfileWithClaimPropertySets(name, claimPropertySets)
|
|
}
|
|
|
|
func createStorageProfileWithClaimPropertySets(name string,
|
|
claimPropertySets []cdiv1.ClaimPropertySet) *cdiv1.StorageProfile {
|
|
return createStorageProfileWithCloneStrategy(name, claimPropertySets, nil)
|
|
}
|
|
|
|
func createStorageProfileWithCloneStrategy(name string,
|
|
claimPropertySets []cdiv1.ClaimPropertySet,
|
|
cloneStrategy *cdiv1.CDICloneStrategy) *cdiv1.StorageProfile {
|
|
|
|
return &cdiv1.StorageProfile{
|
|
ObjectMeta: metav1.ObjectMeta{
|
|
Name: name,
|
|
},
|
|
Status: cdiv1.StorageProfileStatus{
|
|
StorageClass: &name,
|
|
ClaimPropertySets: claimPropertySets,
|
|
CloneStrategy: cloneStrategy,
|
|
},
|
|
}
|
|
}
|
|
|
|
func hasAnnOwnedByDataVolume(obj metav1.Object) bool {
|
|
_, ok := obj.GetAnnotations()[AnnOwnedByDataVolume]
|
|
return ok
|
|
}
|
|
|
|
func getAnnOwnedByDataVolume(obj metav1.Object) (string, string, error) {
|
|
val := obj.GetAnnotations()[AnnOwnedByDataVolume]
|
|
return cache.SplitMetaNamespaceKey(val)
|
|
}
|
|
|
|
func setAnnOwnedByDataVolume(dest, obj metav1.Object) error {
|
|
key, err := cache.MetaNamespaceKeyFunc(obj)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
if dest.GetAnnotations() == nil {
|
|
dest.SetAnnotations(make(map[string]string))
|
|
}
|
|
dest.GetAnnotations()[AnnOwnedByDataVolume] = key
|
|
|
|
return nil
|
|
}
|
|
|
|
func getReconcileResult(result *reconcile.Result) reconcile.Result {
|
|
if result != nil {
|
|
return *result
|
|
}
|
|
return reconcile.Result{}
|
|
}
|
|
|
|
// adapted from k8s.io/kubernetes/pkg/controller/volume/persistentvolume/pv_controller.go
|
|
// checkVolumeSatisfyClaim checks if the volume requested by the claim satisfies the requirements of the claim
|
|
func checkVolumeSatisfyClaim(volume *v1.PersistentVolume, claim *v1.PersistentVolumeClaim) error {
|
|
requestedQty := claim.Spec.Resources.Requests[v1.ResourceName(v1.ResourceStorage)]
|
|
requestedSize := requestedQty.Value()
|
|
|
|
// check if PV's DeletionTimeStamp is set, if so, return error.
|
|
if volume.ObjectMeta.DeletionTimestamp != nil {
|
|
return fmt.Errorf("the volume is marked for deletion %q", volume.Name)
|
|
}
|
|
|
|
volumeQty := volume.Spec.Capacity[v1.ResourceStorage]
|
|
volumeSize := volumeQty.Value()
|
|
if volumeSize < requestedSize {
|
|
return fmt.Errorf("requested PV is too small")
|
|
}
|
|
|
|
// this differs from pv_controller, be loose on storage class if not specified
|
|
requestedClass := storagehelpers.GetPersistentVolumeClaimClass(claim)
|
|
if requestedClass != "" && storagehelpers.GetPersistentVolumeClass(volume) != requestedClass {
|
|
return fmt.Errorf("storageClassName does not match")
|
|
}
|
|
|
|
if CheckVolumeModeMismatches(&claim.Spec, &volume.Spec) {
|
|
return fmt.Errorf("incompatible volumeMode")
|
|
}
|
|
|
|
if !CheckAccessModes(claim, volume) {
|
|
return fmt.Errorf("incompatible accessMode")
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
// CheckVolumeModeMismatches is a convenience method that checks volumeMode for PersistentVolume
|
|
// and PersistentVolumeClaims
|
|
// TODO - later versions of k8s.io/component-helpers have this function
|
|
func CheckVolumeModeMismatches(pvcSpec *v1.PersistentVolumeClaimSpec, pvSpec *v1.PersistentVolumeSpec) bool {
|
|
// In HA upgrades, we cannot guarantee that the apiserver is on a version >= controller-manager.
|
|
// So we default a nil volumeMode to filesystem
|
|
requestedVolumeMode := v1.PersistentVolumeFilesystem
|
|
if pvcSpec.VolumeMode != nil {
|
|
requestedVolumeMode = *pvcSpec.VolumeMode
|
|
}
|
|
pvVolumeMode := v1.PersistentVolumeFilesystem
|
|
if pvSpec.VolumeMode != nil {
|
|
pvVolumeMode = *pvSpec.VolumeMode
|
|
}
|
|
return requestedVolumeMode != pvVolumeMode
|
|
}
|
|
|
|
// CheckAccessModes returns true if PV satisfies all the PVC's requested AccessModes
|
|
// TODO - later versions of k8s.io/component-helpers have this function
|
|
func CheckAccessModes(claim *v1.PersistentVolumeClaim, volume *v1.PersistentVolume) bool {
|
|
pvModesMap := map[v1.PersistentVolumeAccessMode]bool{}
|
|
for _, mode := range volume.Spec.AccessModes {
|
|
pvModesMap[mode] = true
|
|
}
|
|
|
|
for _, mode := range claim.Spec.AccessModes {
|
|
_, ok := pvModesMap[mode]
|
|
if !ok {
|
|
return false
|
|
}
|
|
}
|
|
return true
|
|
}
|
|
|
|
func getReconcileRequest(obj client.Object) reconcile.Request {
|
|
return reconcile.Request{NamespacedName: client.ObjectKeyFromObject(obj)}
|
|
}
|