Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

RedisSession fails on Redis Cluster: add_items/clear_session raise AttributeError ('ClusterPipeline' has no 'raise_first_error')

Open
  • Is something wrong?
#5,322 2 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
35/100
Issue type
Bug
Clarity
Mostly clear
Activity status
Active
Tech stack
python, redis
Domain
backend, databases, distributed-systems

Research direction

Start with _write_items_attempt in src/agents/extensions/memory/redis_session.py and reproduce the failure using the provided Redis Cluster setup. Investigate how add_items and clear_session handle ClusterPipeline and the keys' cluster slots; the issue outlines alternative approaches but does not choose one. Done means both operations work with RedisCluster while preserving session writes and clearing.

Written by the indexing model from the issue text.

Description

Please read this first
  • Have you read the docs? Yes
  • Have you searched for related issues? Yes. #5202 moved clear_session into the same watched transaction; I found no issue covering Redis Cluster.
Describe the bug

RedisSession fails on a Redis Cluster client (redis.asyncio.cluster.RedisCluster). add_items (since 0.20) and clear_session (latest) raise AttributeError before any write lands, so a session backed by Redis Cluster cannot store history.

Two things combine:

  1. _write_items_attempt hooks pipeline internals that only the standalone redis.asyncio.client.Pipeline has: it wraps pipe.raise_first_error and pipe.parse_response and swaps pipe.connection_pool (src/agents/extensions/memory/redis_session.py:483 on main). redis-py's ClusterPipeline has neither raise_first_error nor parse_response, and has per-node pools.
  2. The write is a WATCH/MULTI transaction over {prefix}:{id}, {prefix}:{id}:messages and {prefix}:{id}:counter, which don't share a hash tag, so they land in different cluster slots (e.g. 13192 vs 7329). Redis Cluster only runs a transaction within one slot, so adding the missing hooks would not be enough on its own.

Up to 0.14.5, add_items used a plain pipeline, which redis-py's cluster client splits per key, and it worked on a cluster.

Debug information
  • Agents SDK version: 0.23.1 (latest release); same code on main. First broken release we hit: 0.20.0 (0.14.5 works)
  • Related library versions: redis 8.1.0
  • Python version: 3.12
  • Operating system: macOS 15 (also seen on Linux in Kubernetes)
  • Model and model provider: not involved (no model call)
  • Does the issue reproduce with the latest Agents SDK release? Yes
  • Does the issue occur consistently or intermittently? Consistently
Traceback (most recent call last):
  ...
  File "agents/extensions/memory/redis_session.py", line 483, in _write_items_attempt
    raise_first_error = pipe.raise_first_error
AttributeError: 'ClusterPipeline' object has no attribute 'raise_first_error'
Repro steps

Start a local 3-node cluster:

for p in 7101 7102 7103; do redis-server --port $p --cluster-enabled yes --cluster-config-file nodes-$p.conf --daemonize yes --save ""; done
redis-cli --cluster create 127.0.0.1:7101 127.0.0.1:7102 127.0.0.1:7103 --cluster-replicas 0

import asyncio
from redis.asyncio.cluster import RedisCluster
from agents.extensions.memory.redis_session import RedisSession

async def main() -> None:
    client = RedisCluster(host="127.0.0.1", port=7101)
    session = RedisSession("s1", redis_client=client, key_prefix="repro")
    await session.add_items([{"role": "user", "content": "hi"}])   # AttributeError
    await session.clear_session()                                  # AttributeError (latest)

asyncio.run(main())

On 0.23.1 against the cluster: add_items FAIL, clear_session FAIL, get_items OK, pop_item OK. On 0.14.5: all four OK.

Expected behavior

RedisSession works with RedisCluster, as it did through 0.14.5. Either hash-tag the session's keys into one slot and use a transaction path that ClusterPipeline supports, or fall back to a non-transactional pipeline when the client is a cluster. Our workaround meanwhile is a subclass that restores the 0.14.5 add_items.

Dominant language
Python
Stars
30k
Forks
4.9k
Avg merge
18h 13m
Merged PRs (30d)
215

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from openai/openai-agents-python

All issues in openai/openai-agents-python

Similar issues

More Python issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.