Merge pull request ceph#114 from openshift-cherrypick-robot/cherry-pi…

…ck-113-to-release-4.12

[release-4.12] Sync devel branch with upstream

go.mod

@@ -26,7 +26,7 @@ require (
 	github.com/stretchr/testify v1.8.0
 	golang.org/x/crypto v0.0.0-20220214200702-86341886e292
 	golang.org/x/sys v0.0.0-20220209214540-3681064d5158
-	google.golang.org/grpc v1.47.0
+	google.golang.org/grpc v1.48.0
 	google.golang.org/protobuf v1.28.0
 	k8s.io/api v0.24.2
 	k8s.io/apimachinery v0.24.2

go.sum

@@ -1674,8 +1674,8 @@ google.golang.org/grpc v1.37.0/go.mod h1:NREThFqKR1f3iQ6oBuvc5LadQuXVGo9rkm5ZGrQ
 google.golang.org/grpc v1.38.0/go.mod h1:NREThFqKR1f3iQ6oBuvc5LadQuXVGo9rkm5ZGrQdJfM=
 google.golang.org/grpc v1.40.0/go.mod h1:ogyxbiOoUXAkP+4+xa6PZSE9DZgIHtSpzjDTB9KAK34=
 google.golang.org/grpc v1.41.0/go.mod h1:U3l9uK9J0sini8mHphKoXyaqDA/8VyGnDee1zzIUK6k=
-google.golang.org/grpc v1.47.0 h1:9n77onPX5F3qfFCqjy9dhn8PbNQsIKeVU04J9G7umt8=
-google.golang.org/grpc v1.47.0/go.mod h1:vN9eftEi1UMyUsIF80+uQXhHjbXYbm0uXoFCACuMGWk=
+google.golang.org/grpc v1.48.0 h1:rQOsyJ/8+ufEDJd/Gdsz7HG220Mh9HAhFHRGnIjda0w=
+google.golang.org/grpc v1.48.0/go.mod h1:vN9eftEi1UMyUsIF80+uQXhHjbXYbm0uXoFCACuMGWk=
 google.golang.org/protobuf v1.22.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
 google.golang.org/protobuf v1.23.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
 google.golang.org/protobuf v1.23.1-0.20200526195155-81db48ad09cc/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=

internal/rbd/errors.go

@@ -42,6 +42,4 @@ var (
 	ErrMissingImageNameInVolID = errors.New("rbd image name information can not be empty in volID")
 	// ErrDecodeClusterIDFromMonsInVolID is returned when mons hash decoding on migration volID.
 	ErrDecodeClusterIDFromMonsInVolID = errors.New("failed to get clusterID from monitors hash in volID")
-	// ErrUnHealthyMirroredImage is returned when mirrored image is not healthy.
-	ErrUnHealthyMirroredImage = errors.New("mirrored image is not healthy")
 )

internal/rbd/replicationcontrollerserver.go

@@ -581,16 +581,6 @@ func (rs *ReplicationServer) PromoteVolume(ctx context.Context,
 		}
 	}
 
-	mirrorStatus, err := rbdVol.getImageMirroringStatus()
-	if err != nil {
-		return nil, status.Error(codes.Internal, err.Error())
-	}
-
-	err = checkHealthyPrimary(ctx, rbdVol, mirrorStatus)
-	if err != nil {
-		return nil, status.Error(codes.Internal, err.Error())
-	}
-
 	var mode librbd.ImageMirrorMode
 	mode, err = getMirroringMode(ctx, req.GetParameters())
 	if err != nil {
@@ -620,31 +610,6 @@ func (rs *ReplicationServer) PromoteVolume(ctx context.Context,
 	return &replication.PromoteVolumeResponse{}, nil
 }
 
-// checkHealthyPrimary checks if the image is a healhty primary or not.
-// healthy primary image will be in up+stopped state, for states other
-// than this it returns an error message.
-func checkHealthyPrimary(ctx context.Context, rbdVol *rbdVolume, mirrorStatus *librbd.GlobalMirrorImageStatus) error {
-	localStatus, err := mirrorStatus.LocalStatus()
-	if err != nil {
-		// LocalStatus can fail if the local site status is not found in
-		// mirroring status. Log complete sites status to debug why getting
-		// local status failed
-		log.ErrorLog(ctx, "mirroring status is %+v", mirrorStatus)
-
-		return fmt.Errorf("failed to get local status: %w", err)
-	}
-
-	if !localStatus.Up || localStatus.State != librbd.MirrorImageStatusStateStopped {
-		return fmt.Errorf("%s %w. State is up=%t, state=%q",
-			rbdVol,
-			ErrUnHealthyMirroredImage,
-			localStatus.Up,
-			localStatus.State)
-	}
-
-	return nil
-}
-
 // DemoteVolume extracts the RBD volume information from the
 // volumeID, If the image is present, mirroring is enabled and the
 // image is in promoted state it will demote the volume as secondary.

internal/rbd/replicationcontrollerserver_test.go

@@ -279,75 +279,3 @@ func TestCheckVolumeResyncStatus(t *testing.T) {
 		})
 	}
 }
-
-func Test_checkHealthyPrimary(t *testing.T) {
-	t.Parallel()
-	rbdVol := &rbdVolume{
-		rbdImage: rbdImage{
-			RbdImageName: "test",
-			Pool:         "test-pool",
-		},
-	}
-	tests := []struct {
-		name         string
-		ctx          context.Context
-		rbdVol       *rbdVolume
-		mirrorStatus *librbd.GlobalMirrorImageStatus
-		wantErr      bool
-	}{
-		{
-			name:   "when image is in up+stopped state",
-			ctx:    context.TODO(),
-			rbdVol: rbdVol,
-			mirrorStatus: &librbd.GlobalMirrorImageStatus{
-				SiteStatuses: []librbd.SiteMirrorImageStatus{
-					{
-						MirrorUUID: "",
-						State:      librbd.MirrorImageStatusStateStopped,
-						Up:         true,
-					},
-				},
-			},
-			wantErr: false,
-		},
-		{
-			name:   "when image is in up+error state",
-			ctx:    context.TODO(),
-			rbdVol: rbdVol,
-			mirrorStatus: &librbd.GlobalMirrorImageStatus{
-				SiteStatuses: []librbd.SiteMirrorImageStatus{
-					{
-						MirrorUUID: "",
-						State:      librbd.MirrorImageStatusStateError,
-						Up:         false,
-					},
-				},
-			},
-			wantErr: true,
-		},
-		{
-			name:   "when image is in up+replaying state",
-			ctx:    context.TODO(),
-			rbdVol: rbdVol,
-			mirrorStatus: &librbd.GlobalMirrorImageStatus{
-				SiteStatuses: []librbd.SiteMirrorImageStatus{
-					{
-						MirrorUUID: "",
-						State:      librbd.MirrorImageStatusStateReplaying,
-						Up:         true,
-					},
-				},
-			},
-			wantErr: true,
-		},
-	}
-	for _, tt := range tests {
-		ts := tt
-		t.Run(ts.name, func(t *testing.T) {
-			t.Parallel()
-			if err := checkHealthyPrimary(ts.ctx, ts.rbdVol, ts.mirrorStatus); (err != nil) != ts.wantErr {
-				t.Errorf("checkHealthyPrimary() error = %v, wantErr %v", err, ts.wantErr)
-			}
-		})
-	}
-}

scripts/rook.sh

@@ -86,6 +86,9 @@ function deploy_rook() {
 		check_ceph_cluster_health
 	fi
 
+	# Make sure Ceph Mgr is running
+	check_ceph_mgr
+
 	# Check if CephFileSystem is empty
 	if ! kubectl_retry -n rook-ceph get cephfilesystems -oyaml | grep 'items: \[\]' &>/dev/null; then
 		check_mds_stat
@@ -166,6 +169,22 @@ function check_ceph_cluster_health() {
 	echo ""
 }
 
+function check_ceph_mgr() {
+	for ((retry = 0; retry <= ROOK_DEPLOY_TIMEOUT; retry = retry + 5)); do
+		echo "Waiting for Ceph Mgr... ${retry}s" && sleep 5
+
+		MGR_POD=$(kubectl_retry -n rook-ceph get pods -l app=rook-ceph-mgr -o jsonpath='{.items[0].metadata.name}')
+		MGR_POD_STATUS=$(kubectl_retry -n rook-ceph get pod "$MGR_POD" -ojsonpath='{.status.phase}')
+		[[ "$MGR_POD_STATUS" = "Running" ]] && break
+	done
+
+	if [ "$retry" -gt "$ROOK_DEPLOY_TIMEOUT" ]; then
+		echo "[Timeout] Ceph Mgr is not running (timeout)"
+		return 1
+	fi
+	echo ""
+}
+
 function check_mds_stat() {
 	for ((retry = 0; retry <= ROOK_DEPLOY_TIMEOUT; retry = retry + 5)); do
 		FS_NAME=$(kubectl_retry -n rook-ceph get cephfilesystems.ceph.rook.io -ojsonpath='{.items[0].metadata.name}')