Fail whole prebackup if command exit != 0

This will fix issues were problems with pre-backup go undected. K8up
will now completely fail the whole whole prebackup pod if the commands
exist with a code != 0.

Also a small issue with concurrent `restic init` was fixed.

Signed-off-by: Simon Beck <[email protected]>

Author: Kidswiss

cmd/operator/main.go

@@ -41,6 +41,7 @@ var (
 			&cli.StringFlag{Destination: &cfg.Config.FileExtensionAnnotation, Name: "fileextensionannotation", EnvVars: []string{"BACKUP_FILEEXTENSIONANNOTATION"}, Value: "k8up.io/file-extension", Usage: "set the annotation name where the file extension is stored for backup commands"},
 
 			&cli.IntFlag{Destination: &cfg.Config.GlobalKeepJobs, Hidden: true, Name: "globalkeepjobs", EnvVars: []string{"BACKUP_GLOBALKEEPJOBS"}, Value: -1, DefaultText: "unlimited", Usage: "set the number of old jobs to keep when cleaning up, applies to all job types"},
+			&cli.IntFlag{Destination: &cfg.Config.GlobalBackoffLimit, Name: "global-backoff-limit", EnvVars: []string{"BACKUP_GLOBAL_BACKOFF_LIMIT"}, Value: 6, Usage: "set the backoff limit for all backup jobs"},
 			&cli.IntFlag{Destination: &cfg.Config.GlobalFailedJobsHistoryLimit, Name: "global-failed-jobs-history-limit", EnvVars: []string{"BACKUP_GLOBAL_FAILED_JOBS_HISTORY_LIMIT"}, Value: 3, Usage: "set the number of old, failed jobs to keep when cleaning up, applies to all job types"},
 			&cli.IntFlag{Destination: &cfg.Config.GlobalSuccessfulJobsHistoryLimit, Name: "global-successful-jobs-history-limit", EnvVars: []string{"BACKUP_GLOBAL_SUCCESSFUL_JOBS_HISTORY_LIMIT"}, Value: 3, Usage: "set the number of old, successful jobs to keep when cleaning up, applies to all job types"},
 			&cli.IntFlag{Destination: &cfg.Config.GlobalConcurrentArchiveJobsLimit, Name: "global-concurrent-archive-jobs-limit", EnvVars: []string{"BACKUP_GLOBAL_CONCURRENT_ARCHIVE_JOBS_LIMIT"}, DefaultText: "unlimited", Usage: "set the limit of concurrent archive jobs"},

e2e/definitions/annotated-subject/deployment-error.yaml

@@ -0,0 +1,49 @@
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: annotated-subject-deployment
+  namespace: k8up-e2e-subject
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      app: subject
+  template:
+    metadata:
+      labels:
+        app: subject
+      annotations:
+        k8up.io/backupcommand: 'invalid'
+        k8up.io/backupcommand-container: subject-container
+    spec:
+      containers:
+      - image: busybox
+        imagePullPolicy: IfNotPresent
+        name: dummy-container-blocking-first-position
+        command:
+        - "/bin/sh"
+        - "-c"
+        - "sleep infinity"
+      - name: subject-container
+        image: quay.io/prometheus/busybox:latest
+        imagePullPolicy: IfNotPresent
+        args:
+        - sh
+        - -c
+        - |
+          sleep infinity
+        securityContext:
+          runAsUser: $ID
+        volumeMounts:
+        - name: volume
+          mountPath: /data
+        env:
+        - name: BACKUP_FILE_CONTENT
+          value: ""
+        - name: BACKUP_FILE_NAME
+          value: ""
+      volumes:
+      - name: volume
+        persistentVolumeClaim:
+          claimName: subject-pvc

e2e/definitions/operator/deploy.yaml

@@ -9,3 +9,5 @@ k8up:
   envVars:
     - name: K8UP_DEBUG
       value: "true"
+    - name: BACKUP_GLOBAL_BACKOFF_LIMIT
+      value: "2"

e2e/definitions/operator/values.yaml

@@ -170,6 +170,15 @@ given_an_annotated_subject() {
 	echo "✅  The annotated subject is ready"
 }
 
+given_a_broken_annotated_subject() {
+	require_args 2 ${#}
+
+	kubectl apply -f definitions/pv/pvc.yaml
+	yq e '.spec.template.spec.containers[1].securityContext.runAsUser='$(id -u)' ' definitions/annotated-subject/deployment-error.yaml | kubectl apply -f -
+
+	echo "✅  The annotated subject is ready"
+}
+
 given_an_annotated_subject_pod() {
 	require_args 2 ${#}
 
@@ -457,6 +466,20 @@ wait_until() {
 	kubectl -n "${ns}" wait --timeout 5m --for "condition=${condition}" "${object}"
 }
 
+wait_for_until_jsonpath() {
+	require_args 3 ${#}
+
+	local object condition ns
+	object=${1}
+	until=${2}
+	jsonpath=${3}
+	ns=${NAMESPACE=${DETIK_CLIENT_NAMESPACE}}
+
+	echo "Waiting for '${object}' in namespace '${ns}' to become '${condition}' ..."
+	kubectl -n "${ns}" wait --timeout "${until}" --for "${jsonpath}" "${object}"
+}
+
+
 expect_file_in_container() {
 	require_args 4 ${#}
 

e2e/lib/k8up.bash

@@ -0,0 +1,31 @@
+#!/usr/bin/env bats
+
+load "lib/utils"
+load "lib/detik"
+load "lib/k8up"
+
+# shellcheck disable=SC2034
+DETIK_CLIENT_NAME="kubectl"
+# shellcheck disable=SC2034
+DETIK_CLIENT_NAMESPACE="k8up-e2e-subject"
+# shellcheck disable=SC2034
+DEBUG_DETIK="true"
+
+@test "Annotated app, When creating a backup, Then expect Error" {
+	expected_content="expected content: $(timestamp)"
+	expected_filename="expected_filename.txt"
+
+	given_a_running_operator
+	given_a_clean_ns
+	given_s3_storage
+	given_a_broken_annotated_subject "${expected_filename}" "${expected_content}"
+
+	kubectl apply -f definitions/secrets
+	yq e '.spec.podSecurityContext.runAsUser='$(id -u)'' definitions/backup/backup.yaml | kubectl apply -f -
+
+	try "at most 10 times every 5s to get backup named 'k8up-backup' and verify that '.status.started' is 'true'"
+	verify_object_value_by_label job 'k8up.io/owned-by=backup_k8up-backup' '.status.active' 1 true
+
+	wait_for_until_jsonpath backup/k8up-backup 2m 'jsonpath={.status.conditions[?(@.type=="Completed")].reason}=Failed'
+
+}

e2e/test-12-annotated-failure.bats

@@ -1,6 +1,6 @@
 module github.com/k8up-io/k8up/v2
 
-go 1.21
+go 1.23
 
 require (
 	github.com/firepear/qsplit/v2 v2.5.0

go.mod

@@ -18,6 +18,7 @@ import (
 	corev1 "k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/labels"
 	"k8s.io/apimachinery/pkg/types"
+	"k8s.io/utils/ptr"
 	controllerruntime "sigs.k8s.io/controller-runtime"
 	"sigs.k8s.io/controller-runtime/pkg/client"
 )
@@ -236,7 +237,7 @@ func (b *BackupExecutor) startBackup(ctx context.Context) error {
 	}
 
 	index := 0
-	for _, batchJob := range backupJobs {
+	for name, batchJob := range backupJobs {
 		_, err = controllerruntime.CreateOrUpdate(ctx, b.Generic.Config.Client, batchJob.job, func() error {
 			mutateErr := job.MutateBatchJob(ctx, batchJob.job, b.backup, b.Generic.Config, b.Client)
 			if mutateErr != nil {
@@ -262,16 +263,17 @@ func (b *BackupExecutor) startBackup(ctx context.Context) error {
 			}
 			// each job sleeps for index seconds to avoid concurrent restic repository creation. Not the prettiest way but it works and a repository
 			// is created only once usually.
-			if index > 0 {
+			if name == "prebackup" || index != 0 {
 				batchJob.job.Spec.Template.Spec.Containers[0].Env = append(batchJob.job.Spec.Template.Spec.Containers[0].Env, corev1.EnvVar{
 					Name:  "SLEEP_DURATION",
-					Value: (time.Duration(index) * time.Second).String(),
+					Value: (3 * time.Second).String(),
 				})
 			}
 			b.backup.Spec.AppendEnvFromToContainer(&batchJob.job.Spec.Template.Spec.Containers[0])
 			batchJob.job.Spec.Template.Spec.Volumes = append(batchJob.job.Spec.Template.Spec.Volumes, batchJob.volumes...)
 			batchJob.job.Spec.Template.Spec.Volumes = append(batchJob.job.Spec.Template.Spec.Volumes, utils.AttachTLSVolumes(b.backup.Spec.Volumes)...)
 			batchJob.job.Spec.Template.Spec.Containers[0].VolumeMounts = append(b.newVolumeMounts(batchJob.volumes), b.attachTLSVolumeMounts()...)
+			batchJob.job.Spec.BackoffLimit = ptr.To(int32(cfg.Config.GlobalBackoffLimit))
 
 			batchJob.job.Spec.Template.Spec.Containers[0].Args = b.setupArgs()
 

operator/backupcontroller/executor.go

@@ -51,6 +51,7 @@ type Configuration struct {
 	GlobalKeepJobs                   int
 	GlobalFailedJobsHistoryLimit     int
 	GlobalSuccessfulJobsHistoryLimit int
+	GlobalBackoffLimit               int
 	GlobalRepoPassword               string
 	GlobalRestoreS3AccessKey         string
 	GlobalRestoreS3Bucket            string

operator/cfg/config.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"fmt"
 	"io"
+	"os"
 	"strings"
 
 	"github.com/firepear/qsplit/v2"
@@ -73,6 +74,8 @@ func PodExec(pod BackupPod, log logr.Logger) (*ExecData, error) {
 
 		if err != nil {
 			execLogger.Error(err, "streaming data failed", "namespace", pod.Namespace, "pod", pod.PodName)
+			// we just completely hard fail the whole backup pod
+			os.Exit(1)
 			return
 		}
 	}()