How to implement zookeeper distributed lock (file system, notification mechanism)?

With zookeeper The consistency file system , The problem of locks becomes easy . There are two types of lock services , One is to keep exclusive , The other is to control the timing .

For the first category , We will zookeeper On the one znode Think of it as a lock , adopt createznode The way to achieve . All clients are going to create /distribute_lock node , Finally, the client successfully created owns the lock . Delete the self created distribute_lock The node releases the lock .

For the second category , /distribute_lock Already exist in advance , All clients create temporary sequence number directory nodes under it , And choose master equally , Get the lock with the smallest number , Use up delete , Convenient in turn .