Manager

Learn shared state with multiprocessing.Manager.

What is multiprocessing.Manager?

multiprocessing.Manager creates Python objects shared between processes. It starts a server process; client processes access objects through a proxy.

Manager().list() gives a shared list and Manager().dict() a shared dict. Every operation on these objects is forwarded to the server process and made thread-safe. Unlike a plain list, a shared list can be safely modified by multiple processes. This is handy for complex shared state; but it is slower than direct communication.

Using a shared list

A Manager list is used like a plain list. .append(), [i] and len() work; but every operation goes to the server process.

Important: print(shared) prints the proxy; to see the real list use list(shared). A with Manager() as m: block automatically closes the server at the end. This prevents resource leaks. Shared objects are powerful; but remember every access has a cost.

Exercise Shared list with Manager

Import with from multiprocessing import Manager. Open a with Manager() as m: block. Inside create a shared list with shared = m.list(). Append values with shared.append("a") and shared.append("b"). Print list(shared).

Show example solution

from multiprocessing import Manager

with Manager() as m:
    shared = m.list()
    shared.append("a")
    shared.append("b")
    print(list(shared))

Test yourself

Match Manager concepts

Manager

Shared object server
m.list()

Shared list
m.dict()

Shared dict
proxy

Proxy access
What does Manager().list() provide?
```
from multiprocessing import Manager
m = Manager()
shared = m.list()
```
- A process object
- A plain Python list
- A list shared between processes
- A queue
Show answer

A list shared between processes — Manager().list() returns a list safe to share between processes.

Complete the Manager call

from multiprocessing import ____
m = Manager()

Show answer

from multiprocessing import Manager
m = Manager()

Next lesson: Summary - Multiprocessing

We Are Enhancing Your Experience

Cookie Preferences

Strictly Necessary Cookies

Analytical Cookies