Fix excessive io around volume cleanup, move cleanup logic to startup…

… rather than run each volume deprovision

Signed-off-by: Cameron McAvoy <[email protected]>

CHANGELOG.md

@@ -6,6 +6,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.4.1] - 2024-08-19
+### Fixed
+- Fixed volume cleanup to only run at startup, reducing io when many mounts are used
+
 ## [0.4.0] - 2024-08-07
 ### Fixed
 - Fixed locking behavior with volumes, terminating pods could fail to be cleaned up due to a race in the locks.

pkg/ccm/cleanup.go

@@ -11,15 +11,13 @@ import (
 	"k8s.io/mount-utils"
 )
 
-func doCleanup(configMap, volumeID, targetPath string) {
+func doCleanup() {
 	start := time.Now()
 	if err := cleanupDataDir(); err != nil {
-		logger.Error(err, fmt.Sprintf("failed to cleanup on node unpublish volume for volume id %q and target path %q", volumeID, targetPath))
-		unpublishErr.WithLabelValues(configMap, "failed to clean up volume data").Inc()
+		logger.Error(err, "failed to cleanup")
 	}
 	if err := cleanupMetadataDir(); err != nil {
-		logger.Error(err, fmt.Sprintf("failed to cleanup metadata on node unpublish volume for volume id %q and target path %q", volumeID, targetPath))
-		unpublishErr.WithLabelValues(configMap, "failed to clean up volume metadata").Inc()
+		logger.Error(err, "failed to cleanup metadata")
 	}
 	cleanupTime.WithLabelValues().Observe(time.Since(start).Seconds())
 }
@@ -32,6 +30,10 @@ func cleanupDataDir() error {
 	dataDir := path.Join(storageDir, "data")
 	dirEntries, err := os.ReadDir(dataDir)
 	if err != nil {
+		if os.IsNotExist(err) {
+			logger.Info("[cleanup] no data dir, skipping cleanup")
+			return nil
+		}
 		return fmt.Errorf("failed to list dir entries for %q: %w", dataDir, err)
 	}
 

pkg/ccm/driver.go

@@ -11,7 +11,6 @@ import (
 	"sync"
 
 	"github.com/container-storage-interface/spec/lib/go/csi"
-
 	"google.golang.org/grpc"
 
 	"k8s.io/client-go/kubernetes"
@@ -100,6 +99,8 @@ func (d *driver) Run() error {
 		return resp, err
 	}
 
+	doCleanup()
+
 	d.srv = grpc.NewServer(grpc.UnaryInterceptor(errHandler))
 	csi.RegisterIdentityServer(d.srv, d)
 	csi.RegisterNodeServer(d.srv, d)

pkg/ccm/node.go

@@ -152,7 +152,6 @@ func (d *driver) NodeUnpublishVolume(ctx context.Context, req *csi.NodeUnpublish
 		unpublishErr.WithLabelValues(configMap, "volume was already unmounted").Inc()
 	}
 	logger.V(2).Info(fmt.Sprintf("node unpublish volume succeeded for volume id %q target path %q", req.VolumeId, req.TargetPath))
-	doCleanup(configMap, req.VolumeId, req.TargetPath)
 	unpublish.WithLabelValues(configMap).Inc()
 	unpublishTime.WithLabelValues(configMap).Observe(time.Since(start).Seconds())
 	return &csi.NodeUnpublishVolumeResponse{}, nil