fix: make sure scan iterator commands are always issued to the same replica

[rdelcampog]

Pull Request check-list

Please make sure to review and check all of these items:

• Do tests and lints pass with this change?
• Do the CI tests pass with this change (enable it first in your forked repo and wait for the github action build to finish)? Yes!
• Is the new or changed code fully tested?
• Is a documentation update included (if this change modifies existing APIs, or introduces new ones)? bugfix, not applicable
• Is there an example added to the examples folder (if applicable)? bugfix, not applicable

NOTE: these things are not required to open a PR and can be done
afterwards / while the PR is open.

Description of change

Fixes #3197. Inspired on @agnesnatasya pull request here.

cc/ @dmaier-redislabs @gerzse as both participated in the other PR.

[petyaslavova]

Hi @rdelcampog, thank you for your contribution! We will review your change soon.

[Copilot]

Pull Request Overview

This PR implements a fix to ensure scan iterator commands (SCAN, SSCAN, HSCAN, ZSCAN) always use the same replica connection throughout their iteration when using Sentinel-managed connections, addressing issue #3197.

Key changes made:

• Added iter_req_id tracking to maintain connection consistency for scan operations
• Modified scan iterator methods to generate unique IDs and clean up after completion
• Enhanced connection pools to support replica affinity during scan iterations

Reviewed Changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
redis/sentinel.py	Added _iter_req_connections tracking and enhanced get_connection() with replica affinity logic
redis/asyncio/sentinel.py	Added async version of replica affinity tracking for AsyncSentinelConnectionPool
redis/commands/core.py	Modified all scan iterator methods to generate unique iter_req_id and ensure cleanup
redis/client.py	Added cleanup logic when SCAN commands complete (cursor returns to 0)
redis/asyncio/client.py	Added async version of cleanup logic for completed SCAN commands
tests/test_sentinel_managed_connection.py	Comprehensive test coverage for replica consistency behavior
tests/test_asyncio/test_sentinel_managed_connection.py	Async version of replica consistency tests

Comments suppressed due to low confidence (1)

redis/asyncio/sentinel.py:1

• The pytestmark = pytest.mark.asyncio line was removed but this is needed for async tests to run properly. This line should be restored to ensure all async test functions in this file are automatically marked with the asyncio pytest marker.

import asyncio

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Using next() on rotate_slaves() assumes it returns an iterator, but rotate_slaves() returns a generator from proxy.rotate_slaves(). This should handle the StopIteration exception properly or use a different approach to get a single replica address.

Suggested change

	try:
	replica_address = next(self.rotate_slaves())
	connection.connect_to(replica_address)
	# Track this replica for future requests with this iter_req_id
	self._iter_req_connections[iter_req_id] = replica_address
	except (SlaveNotFoundError, StopIteration):
	replica_address = next(self.rotate_slaves(), None)
	if replica_address is not None:
	connection.connect_to(replica_address)
	# Track this replica for future requests with this iter_req_id
	self._iter_req_connections[iter_req_id] = replica_address
	else:

[Copilot]

Using __anext__() directly on an async generator is not the recommended approach. Consider using anext() builtin function instead: replica_address = await anext(self.rotate_slaves()).

[Copilot]

The imports random and time are placed inside the method rather than at the top of the file. These should be moved to the top-level imports for better code organization and consistency.

[Copilot]

The imports random and time are placed inside the method rather than at the top of the file. These should be moved to the top-level imports for better code organization and consistency.

[Copilot]

The variable pool is not defined in this scope. It should be referencing conn.connection_pool or self.connection_pool to access the cleanup method.

Suggested change

	if hasattr(pool, "cleanup"):
	pool.cleanup(iter_req_id)
	if hasattr(conn.connection_pool, "cleanup"):
	conn.connection_pool.cleanup(iter_req_id)

[Copilot]

The variable pool is not defined in this scope. It should be referencing conn.connection_pool or self.connection_pool to access the cleanup method.