Skip to content

feature: Periodically clean up invalid blade #53

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Nov 17, 2020
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
Expand Up @@ -41,6 +41,9 @@ spec:
{{- if .Values.daemonset.enable }}
- '--daemonset-enable'
{{- end }}
{{- if .Values.remove.blade.interval }}
- '--remove-blade-interval={{ .Values.remove.blade.interval }}'
{{- end }}
imagePullPolicy: {{ .Values.operator.pullPolicy }}
env:
- name: WATCH_NAMESPACE
Expand Down
6 changes: 5 additions & 1 deletion deploy/helm/chaosblade-operator-for-v2/values.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -21,4 +21,8 @@ webhook:
enable: true

daemonset:
enable: true
enable: true

remove:
blade:
interval: 72h
Original file line number Diff line number Diff line change
Expand Up @@ -41,6 +41,9 @@ spec:
{{- if .Values.daemonset.enable }}
- '--daemonset-enable'
{{- end }}
{{- if .Values.remove.blade.interval }}
- '--remove-blade-interval={{ .Values.remove.blade.interval }}'
{{- end }}
imagePullPolicy: {{ .Values.operator.pullPolicy }}
env:
- name: WATCH_NAMESPACE
Expand Down
6 changes: 5 additions & 1 deletion deploy/helm/chaosblade-operator-for-v3/values.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -22,4 +22,8 @@ webhook:
enable: true

daemonset:
enable: true
enable: true

remove:
blade:
interval: 72h
68 changes: 67 additions & 1 deletion pkg/controller/chaosblade/controller.go
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,10 @@ import (
"context"
"encoding/json"
"fmt"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/types"
"sigs.k8s.io/controller-runtime/pkg/client"
"time"

"github.com/operator-framework/operator-sdk/pkg/k8sutil"
"github.com/sirupsen/logrus"
Expand Down Expand Up @@ -49,7 +53,14 @@ const chaosbladeFinalizer = "finalizer.chaosblade.io"
// Add creates a new ChaosBlade Controller and adds it to the Manager. The Manager will set fields on the Controller
// and Start it when the Manager is Started.
func Add(mgr manager.Manager) error {
return add(mgr, newReconciler(mgr))
if err := add(mgr, newReconciler(mgr)); err != nil {
return err
}
// add periodically clean up blade ticker
return mgr.Add(manager.RunnableFunc(func(s <-chan struct{}) error {
startPeriodicallyCleanUpBlade(mgr)
return nil
}))
}

// newReconciler returns a new reconcile.Reconciler
Expand Down Expand Up @@ -97,6 +108,61 @@ func add(mgr manager.Manager, rcb *ReconcileChaosBlade) error {
return nil
}

// if blade status is destroying
func startPeriodicallyCleanUpBlade(mgr manager.Manager) {
go func() {
cli := mgr.GetClient()
duration, err := time.ParseDuration(chaosblade.RemoveBladeInterval)
if err != nil {
logrus.Errorf("parse interval error: %v, use default interval: %s", err, chaosblade.DefaultRemoveBladeInterval)
duration, err = time.ParseDuration(chaosblade.DefaultRemoveBladeInterval)
chaosblade.RemoveBladeInterval = chaosblade.DefaultRemoveBladeInterval
if err != nil {
logrus.Fatalf("start periodically clean up blade, ticker error: %v", err)
}
}
// first clean up
periodicallyCleanUpBlade(cli, duration)

// ticker clean up
ticker := time.NewTicker(time.Second * time.Duration(duration.Seconds()))
logrus.Infof("start periodically clean up blade ticker, interval: %s", chaosblade.RemoveBladeInterval)
for range ticker.C {
periodicallyCleanUpBlade(cli, duration)
}
}()
}

func periodicallyCleanUpBlade(cli client.Client, interval time.Duration) {
results := &v1alpha1.ChaosBladeList{}
if err := cli.List(context.TODO(), results, &client.ListOptions{}); err != nil {
logrus.Errorf("periodically clean up, list blade error: %v", err)
}
logrus.Infof("periodically clean up blade, blade size: %d", len(results.Items))
for _, item := range results.Items {
if item.DeletionTimestamp == nil {
continue
}
sub := time.Now().Sub(item.DeletionTimestamp.Time)
if item.Status.Phase == v1alpha1.ClusterPhaseDestroying && sub.Seconds() > interval.Seconds() {
logrus.Infof("periodically clean up blade %s, deletion time: %s", item.Name, item.DeletionTimestamp.String())
// patch blade
if err := cli.Patch(context.TODO(),
&v1alpha1.ChaosBlade{
TypeMeta: metav1.TypeMeta{
APIVersion: "chaosblade.io/v1alpha1",
Kind: "ChaosBlade",
},
ObjectMeta: metav1.ObjectMeta{Name: item.Name},
},
client.RawPatch(types.MergePatchType, []byte(`{"metadata":{"finalizers":[]}}`)),
); err != nil {
logrus.Errorf("patch blade: %s, error: %v", item.Name, err)
}
}
}
}

// blank assignment to verify that ReconcileChaosBlade implements reconcile.Reconciler
var _ reconcile.Reconciler = &ReconcileChaosBlade{}

Expand Down
11 changes: 7 additions & 4 deletions pkg/runtime/chaosblade/chaosblade.go
Original file line number Diff line number Diff line change
Expand Up @@ -23,10 +23,11 @@ import (
)

var (
ImageRepository string
Version string
PullPolicy string
DaemonsetEnable bool
ImageRepository string
Version string
PullPolicy string
DaemonsetEnable bool
RemoveBladeInterval string
)

const (
Expand All @@ -38,6 +39,7 @@ const (
)

const DaemonsetPodName = "chaosblade-tool"
const DefaultRemoveBladeInterval = "72h"

var DaemonsetPodLabels = map[string]string{
"app": "chaosblade-tool",
Expand All @@ -64,6 +66,7 @@ func init() {
f.StringVar(&ImageRepository, "chaosblade-image-repository", "chaosbladeio/chaosblade-tool", "Image repository of chaosblade tool")
f.StringVar(&PullPolicy, "chaosblade-image-pull-policy", "IfNotPresent", "Pulling policy of chaosblade image, default value is IfNotPresent.")
f.BoolVar(&DaemonsetEnable, "daemonset-enable", false, "Deploy chaosblade daemonset to resolve chaos experiment environment of network, default value is false.")
f.StringVar(&RemoveBladeInterval, "remove-blade-interval", DefaultRemoveBladeInterval, "Periodically clean up blade state is destroying, default value is 24h.")
}

func FlagSet() *pflag.FlagSet {
Expand Down