Celery When a task fails in the body of a chord, and that

The Fix

pip install celery==5.6.0

Based on closed celery/celery issue #8144 · PR/commit linked

Jump to Verify Open PR/Commit

@@ -2234,7 +2234,8 @@ def run(self, header, body, partial_args, app=None, interval=None,
             body.options.update(options)
 
-        bodyres = body.freeze(task_id, root_id=root_id)
+        body_task_id = task_id or uuid()
+        bodyres = body.freeze(body_task_id, group_id=group_id, root_id=root_id)

repro.py

from celery import Celery, group
app = Celery('repro_toy')
app.conf.update(
    broker_url='redis://localhost/0',
    result_backend='redis://localhost/0',
    #task_serializer='pickle',
    #result_serializer='pickle',
    #accept_content=['application/json', 'application/x-python-serialize']
    )

@app.task(bind=True)
def debug_A(self, i):
    print(f'Doing A{i}')

@app.task(bind=True)
def debug_broke(self):
    print(f'OH NO')
    raise Exception("Head Aspload")

@app.task(bind=True)
def debug_finisher(self):
    print(f'OK, We did it all')
    
def probe_results(res):
    print (f"waiting for {res=}")
    res.wait()
    print(f"{res.failed()=}")
    print(f"{res.get()=}")
    
    
def launch(): 
    # This finishes fine 
    g = [debug_A.si(i) for i in range(5)]
    r = group(g) | debug_finisher.si() | debug_A.si(-1)
    res = r.apply_async()
    probe_results(res)

def launch_fail_better(): 
    # This blows up with a nice expected celery.exceptions.ChordError which seems correct
    # When I get reults I get my "Head Aspload" Exception
    g = [debug_A.si(i) for i in range(5)]
    g.append(debug_broke.si())
    r = group(g) | debug_finisher.si()
    res = r.apply_async()
    probe_results(res)

def launch_fail_bad(): 
    # This blows up with TypeError: sequence item 1: expected a bytes-like object, NoneType found
    g = [debug_A.si(i) for i in range(5)]
    g.append(debug_broke.si())
    r = group(g) | debug_finisher.si() | debug_A.si(-1)
    res = r.apply_async()
    probe_results(res)

verify

Re-run the minimal reproduction on your broken version, then apply the fix and re-run.

fix.md

Option A — Upgrade to fixed release\npip install celery==5.6.0\nWhen NOT to use: This fix should not be applied if the task_id is intentionally meant to be None.\n\n

Why This Fix Works in Production

Trigger: When a task fails in the body of a chord, and that chord chains into another task, there another buggy exception raised
Mechanism: The chord body was being frozen with a potentially None task_id, causing error handling to fail
Why the fix works: Fixes a bug where a ValueError was raised when a task_id was empty in a chord with a chain body. (first fixed release: 5.6.0).

Why This Breaks in Prod

The chord body was being frozen with a potentially None task_id, causing error handling to fail
Production symptom (often without a traceback): When a task fails in the body of a chord, and that chord chains into another task, there another buggy exception raised

Proof / Evidence

GitHub issue: #8144
Fix PR: https://github.com/celery/celery/pull/9774
First fixed release: 5.6.0
Reproduced locally: No (not executed)
Last verified: 2026-02-09
Confidence: 0.85
Did this fix it?: Yes (upstream fix exists)
Own content ratio: 0.38

Discussion

High-signal excerpts from the issue thread (symptoms, repros, edge-cases).

Jump to Sources Open on GitHub

“Just tried this: with latest version 5.5.3 the error on your is not TypeError: sequence item 1: expected a bytes-like object, NoneType found but ValueError:…”

@dfm88 · 2025-06-26 · source

“so I guess we can close this issue as well, right?”

@auvipy · 2025-06-26 · source

“Hey @mscout1 :wave:, Thank you for opening an issue”

@open-collective-bot · 2023-03-21 · source

Failure Signature (Search String)

When a task fails in the body of a chord, and that chord chains into another task, there another buggy exception raised
If I leave off the chain, and just run the chord, It does not have this second exception. In that case, I get the original exception from the chord head task when I wait.

Copy-friendly signature

signature.txt

Failure Signature
-----------------
When a task fails in the body of a chord, and that chord chains into another task, there another buggy exception raised
If I leave off the chain, and just run the chord, It does not have this second exception. In that case, I get the original exception from the chord head task when I wait.

Error Message

Signature-only (no traceback captured)

error.txt

Error Message
-------------
When a task fails in the body of a chord, and that chord chains into another task, there another buggy exception raised
If I leave off the chain, and just run the chord, It does not have this second exception. In that case, I get the original exception from the chord head task when I wait.

Minimal Reproduction

repro.py

from celery import Celery, group
app = Celery('repro_toy')
app.conf.update(
    broker_url='redis://localhost/0',
    result_backend='redis://localhost/0',
    #task_serializer='pickle',
    #result_serializer='pickle',
    #accept_content=['application/json', 'application/x-python-serialize']
    )

@app.task(bind=True)
def debug_A(self, i):
    print(f'Doing A{i}')

@app.task(bind=True)
def debug_broke(self):
    print(f'OH NO')
    raise Exception("Head Aspload")

@app.task(bind=True)
def debug_finisher(self):
    print(f'OK, We did it all')
    
def probe_results(res):
    print (f"waiting for {res=}")
    res.wait()
    print(f"{res.failed()=}")
    print(f"{res.get()=}")
    
    
def launch(): 
    # This finishes fine 
    g = [debug_A.si(i) for i in range(5)]
    r = group(g) | debug_finisher.si() | debug_A.si(-1)
    res = r.apply_async()
    probe_results(res)

def launch_fail_better(): 
    # This blows up with a nice expected celery.exceptions.ChordError which seems correct
    # When I get reults I get my "Head Aspload" Exception
    g = [debug_A.si(i) for i in range(5)]
    g.append(debug_broke.si())
    r = group(g) | debug_finisher.si()
    res = r.apply_async()
    probe_results(res)

def launch_fail_bad(): 
    # This blows up with TypeError: sequence item 1: expected a bytes-like object, NoneType found
    g = [debug_A.si(i) for i in range(5)]
    g.append(debug_broke.si())
    r = group(g) | debug_finisher.si() | debug_A.si(-1)
    res = r.apply_async()
    probe_results(res)

What Broke

Chaining tasks after a failed chord results in obscured error messages, leading to confusion.

Why It Broke

The chord body was being frozen with a potentially None task_id, causing error handling to fail

Fix Options (Details)

Option A — Upgrade to fixed release Safe default (recommended)

pip install celery==5.6.0

When NOT to use: This fix should not be applied if the task_id is intentionally meant to be None.

Use when you can deploy the upstream fix. It is usually lower-risk than long-lived workarounds.

Fix reference: https://github.com/celery/celery/pull/9774

First fixed release: 5.6.0

Last verified: 2026-02-09. Validate in your environment.

When NOT to Use This Fix

This fix should not be applied if the task_id is intentionally meant to be None.

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

Capture the exact failing error string in logs and tests so you can reproduce via a minimal script.
Pin production dependencies and upgrade only with a reproducible test that hits the failing path.

Version Compatibility Table

Version	Status
5.6.0	Fixed

Related Issues

No related fixes found.

Cluster: celery:serialization Celery hub Celery best practices All hubs All clusters

Related clusters: Configuration error Data consistency Duplicates

Sources

We don’t republish the full GitHub discussion text. Use the links above for context.

Celery When a task fails in the body of a chord, and that chord chains into another task, there (Fix)