Make sure to handle FaultyDisks in listing ops (#6204)

Continuing from PR 157ed65c35

Our posix.go implementation did not handle  I/O errors
properly on the disks, this led to situations where
top-level callers such as ListObjects might return early
without even verifying all the available disks.

This commit tries to address this in Kubernetes, drbd/nbd based
persistent volumes which can disconnect under load and
result in the situations with disks return I/O errors.

This commit also simplifies listing operation, listing
never returns any error. We can avoid this since we pretty
much ignore most of the errors anyways. When objects are
accessed directly we return proper errors.

cmd/fs-v1.go

@@ -968,13 +968,15 @@ func (fs *FSObjects) DeleteObject(ctx context.Context, bucket, object string) er
 // is a leaf or non-leaf entry.
 func (fs *FSObjects) listDirFactory(isLeaf isLeafFunc) listDirFunc {
 	// listDir - lists all the entries at a given prefix and given entry in the prefix.
-	listDir := func(bucket, prefixDir, prefixEntry string) (entries []string, delayIsLeaf bool, err error) {
+	listDir := func(bucket, prefixDir, prefixEntry string) (entries []string, delayIsLeaf bool) {
+		var err error
 		entries, err = readDir(pathJoin(fs.fsPath, bucket, prefixDir))
 		if err != nil {
-			return nil, false, err
+			logger.LogIf(context.Background(), err)
+			return
 		}
 		entries, delayIsLeaf = filterListEntries(bucket, prefixDir, entries, prefixEntry, isLeaf)
-		return entries, delayIsLeaf, nil
+		return entries, delayIsLeaf
 	}
 
 	// Return list factory instance.