The Fix
Fixes a race condition in handling health checks for pub-sub by allowing the pubsub's method get_message() to be called without prior subscription.
Based on closed redis/redis-py issue #1720 · PR/commit linked
Production note: This tends to surface only under concurrency. Reproduce with load tests and watch for lock contention/cancellation paths.
@@ -1276,18 +1276,17 @@ def __init__(
self.ignore_subscribe_messages = ignore_subscribe_messages
self.connection = None
+ self.subscribed_event = threading.Event()
# we need to know the encoding options for this connection in order
# to lookup channel and pattern names for callback handlers.
repro.py
#!/usr/bin/env python3
import threading
import time
import redis
def poll(ps):
while True:
message = ps.get_message(timeout=5)
if message is not None:
print(message)
else:
break
def main():
r = redis.Redis.from_url("redis://localhost", health_check_interval=1)
ps = r.pubsub()
ps.subscribe("foo")
poller = threading.Thread(target=poll, args=(ps,))
poller.start()
ps.unsubscribe("foo")
time.sleep(3)
ps.subscribe("baz")
poller.join()
main()
verify
Re-run the minimal reproduction on your broken version, then apply the fix and re-run.
fix.md
Option A — Apply the official fix\nFixes a race condition in handling health checks for pub-sub by allowing the pubsub's method get_message() to be called without prior subscription.\nWhen NOT to use: This fix is not safe if the application relies on immediate unsubscribe confirmations.\n\n
Why This Fix Works in Production
- Trigger: {'type': 'subscribe', 'pattern': None, 'channel': b'foo', 'data': 1}
- Mechanism: Race condition occurs when health checks are issued before unsubscribe confirmation is received
Production impact:
- If left unfixed, failures can be intermittent under concurrency (hard to reproduce; shows up as sporadic 5xx/timeouts).
Why This Breaks in Prod
- Shows up under Python 3.5.1 in real deployments (not just unit tests).
- Race condition occurs when health checks are issued before unsubscribe confirmation is received
- Surfaces as: {'type': 'subscribe', 'pattern': None, 'channel': b'foo', 'data': 1}
Proof / Evidence
- GitHub issue: #1720
- Fix PR: https://github.com/redis/redis-py/pull/1737
- Reproduced locally: No (not executed)
- Last verified: 2026-02-12
- Confidence: 0.70
- Did this fix it?: Yes (upstream fix exists)
- Own content ratio: 0.40
Discussion
High-signal excerpts from the issue thread (symptoms, repros, edge-cases).
“> In this case, the health check will be performed and we will still get a race reading from the socket with the poller thread”
“Hey, Thanks for finding it and bringing it up! I was able to reproduce the bug and to find the RC”
Failure Signature (Search String)
- {'type': 'subscribe', 'pattern': None, 'channel': b'foo', 'data': 1}
Error Message
Stack trace
error.txt
Error Message
-------------
{'type': 'subscribe', 'pattern': None, 'channel': b'foo', 'data': 1}
{'type': 'unsubscribe', 'pattern': None, 'channel': b'foo', 'data': 0}
{'type': 'subscribe', 'pattern': None, 'channel': b'baz', 'data': 1}
{'type': 'subscribe', 'pattern': None, 'channel': b'foo', 'data': 1}
{'type': 'unsubscribe', 'pattern': None, 'channel': b'foo', 'data': 0}
{'type': 'subscribe', 'pattern': None, 'channel': b'baz', 'data': 1}
{'type': 'subscribe', 'pattern': None, 'channel': b'foo', 'data': 1}
{'type': 'unsubscribe', 'pattern': None, 'channel': b'foo', 'data': 0}
{'type': 'subscribe', 'pattern': None, 'channel': b'baz', 'data': 1}
{'type': 'subscribe', 'pattern': None, 'channel': b'foo', 'data': 1}
{'type': 'unsubscribe', 'pattern': None, 'channel': b'foo', 'data': 0}
{'type': 'subscribe', 'pattern': None, 'channel': b'baz', 'data': 1}
{'type': 'subscribe', 'pattern': None, 'channel': b'foo', 'data': 1}
{'type': 'unsubscribe', 'pattern': None, 'channel': b'foo', 'data': 0}
{'type': 'subscribe', 'pattern': None, 'channel': b'baz', 'data': 1}
{'type': 'subscribe', 'pattern': None, 'channel': b'foo', 'data': 1}
{'type': 'unsubscribe', 'pattern': None, 'channel': b'foo', 'data': 0}
{'type': 80, 'pattern': None, 'channel': 79, 'data': 78}
Traceback (most recent call last):
File "/home/bmerry/work/sdp/env3/lib/python3.8/site-packages/redis/connection.py", line 430, in read_from_socket
Exceptio
... (truncated) ...
Minimal Reproduction
repro.py
#!/usr/bin/env python3
import threading
import time
import redis
def poll(ps):
while True:
message = ps.get_message(timeout=5)
if message is not None:
print(message)
else:
break
def main():
r = redis.Redis.from_url("redis://localhost", health_check_interval=1)
ps = r.pubsub()
ps.subscribe("foo")
poller = threading.Thread(target=poll, args=(ps,))
poller.start()
ps.unsubscribe("foo")
time.sleep(3)
ps.subscribe("baz")
poller.join()
main()
Environment
- Python: 3.5.1
What Broke
Health checks may return incorrect responses, causing unexpected behavior in pub-sub messaging.
Why It Broke
Race condition occurs when health checks are issued before unsubscribe confirmation is received
Fix Options (Details)
Option A — Apply the official fix
Fixes a race condition in handling health checks for pub-sub by allowing the pubsub's method get_message() to be called without prior subscription.
When NOT to use: This fix is not safe if the application relies on immediate unsubscribe confirmations.
Fix reference: https://github.com/redis/redis-py/pull/1737
Last verified: 2026-02-12. Validate in your environment.
When NOT to Use This Fix
- This fix is not safe if the application relies on immediate unsubscribe confirmations.
Verify Fix
verify
Re-run the minimal reproduction on your broken version, then apply the fix and re-run.
Did This Fix Work in Your Case?
Quick signal helps us prioritize which fixes to verify and improve.
Prevention
- Add a stress test that runs high-concurrency workloads and fails on thread dumps / blocked locks.
- Enable watchdog dumps in prod (faulthandler, thread dump endpoint) to capture deadlocks quickly.
- Make timeouts explicit and test them (unit + integration) to avoid silent behavior changes.
- Instrument retries (attempt count + reason) and alert on spikes to catch dependency slowdowns.
Related Issues
No related fixes found.
Sources
We don’t republish the full GitHub discussion text. Use the links above for context.