ConsistentRead tries 10 times

We've seen clusters where 3 attempts were not enough. Bumping to 10. The
slowdown should be negligible and it will reduce retry attempts in the
upper layers of kubelet.

staging/src/k8s.io/mount-utils/mount_helper_unix.go

@@ -32,7 +32,7 @@ const (
 	// At least number of fields per line in /proc/<pid>/mountinfo.
 	expectedAtLeastNumFieldsPerMountInfo = 10
 	// How many times to retry for a consistent read of /proc/mounts.
-	maxListTries = 3
+	maxListTries = 10
 )
 
 // IsCorruptedMnt return true if err is about corrupted mount point