TcpFakeServer holding thread open after application executes when using Redis TimeSeries

[fisher60]

Describe the bug
Using a Redis timeseries with TcpFakeServer results in the application hanging open due to multiple threads spawning/running in the background.

To Reproduce
Steps to reproduce the behavior:

1. Follow the documentation here to start a server running in a thread.
2. Connect to redis and create a redis TimeSeries client.
3. run timeseries.add()
4. application logic finishes, application is held open due to background threads

Expected behavior
It is expected that once the application finishes (i.e the timeseries.add() executes), the thread will also exit. It seems that multiple, non-daemon threads, are spawned during this action.

Desktop (please complete the following information):

• OS:Widnows 11
• Python 3.13.7
• redis-py version:2.31.0

dependencies = [
   "kraken-client>=0.1.0",
   "websockets>=12.0",
   "colorlog>=6.7.0",
   "opentelemetry-api>=1.35.0",
   "opentelemetry-exporter-otlp>=1.35.0 ",
   "opentelemetry-instrumentation-logging>=0.56b0",
   "opentelemetry-sdk>=1.35.0",
   "redis>=6.4.0",
   "mock>=5.1.0",
   "pandas>=2.1.3",
   "pydantic>=2.5.2",
]
[tool.pdm.dev-dependencies]
dev = [
   "setuptools>=68.2.2",
   "pre-commit>=3.5.0",
   "ruff>=0.1.11",
   "fakeredis>=2.31.0"
]

Additional context
Example to reproduce:

from threading import Thread
import time
from fakeredis import TcpFakeServer
from redis import Redis

from market_gather.datastream import Ticker


REDIS_IP = "127.0.0.1"
REDIS_PORT = 6379

REDIS_URL = f"redis://{REDIS_IP}:{REDIS_PORT}"

server_address = (REDIS_IP, REDIS_PORT)
server = TcpFakeServer(server_address, server_type="redis")
t = Thread(target=server.serve_forever, daemon=True)
t.start()

print("fake redis started")

redis: Redis = Redis.from_url(REDIS_URL)

print("creating timeseries")
timeseries = redis.ts()

timestamp = round(time.time() * 1000)

timeseries.add(
    "ticker:test:thing",
    timestamp,
    0.5,
    duplicate_policy="LAST",
    labels={"symbol": "test", "type": "thing"},
)

print("complete")

Output from above

python .\redis_test.py
fake redis started
creating timeseries
complete

[cunla]

You are not closing the thread you created.

Use server.shutdown() - that should do the trick. To be extra sure:

server.shutdown() # Stop the serve_forever loop
server.server_close() # Close the server socket
t.join() # Wait for the server thread to finish

I updated the docs to include this information

[fisher60]

Hello, this does not resolve my issue.

I have amended my example with your instructions. This results in the application hanging indefinitely and the line print("complete") becoming unreachable.

from threading import Thread
import time
from fakeredis import TcpFakeServer
from redis import Redis

from market_gather.datastream import Ticker


REDIS_IP = "127.0.0.1"
REDIS_PORT = 6379

REDIS_URL = f"redis://{REDIS_IP}:{REDIS_PORT}"

server_address = (REDIS_IP, REDIS_PORT)
server = TcpFakeServer(server_address, server_type="redis")
t = Thread(target=server.serve_forever, daemon=True)
t.start()

print("fake redis started")

redis: Redis = Redis.from_url(REDIS_URL)

print("creating timeseries")
timeseries = redis.ts()

timestamp = round(time.time() * 1000)

timeseries.add(
    "ticker:test:thing",
    timestamp,
    0.5,
    duplicate_policy="LAST",
    labels={"symbol": "test", "type": "thing"},
)

server.shutdown() # Stop the serve_forever loop
server.server_close() # Close the server socket
t.join() # Wait for the server thread to finish

print("complete")

Result:

fake redis started
creating timeseries

[cunla]

I see, you want the thread to end when the process ends.. that's a good call - Let me check