Checklist

• I have read the relevant section in the
  contribution guide
  on reporting bugs.
• I have checked the issues list
  for similar or identical bug reports.
• I have checked the pull requests list
  for existing proposed fixes.
• I have checked the commit log
  to find out if the bug was already fixed in the master branch.
• I have included all related issues and possible duplicate issues
  in this issue (If there are none, check this box anyway).

Mandatory Debugging Information

• I have included the output of celery -A proj report in the issue.
  (if you are not able to do this, then at least specify the Celery
  version affected).
• I have verified that the issue exists against the master branch of Celery.
• I have included the contents of pip freeze in the issue.
• I have included all the versions of all the external dependencies required
  to reproduce this bug.

Optional Debugging Information

• I have tried reproducing the issue on more than one Python version
  and/or implementation.
• I have tried reproducing the issue on more than one message broker and/or
  result backend.
• I have tried reproducing the issue on more than one version of the message
  broker and/or result backend.
• I have tried reproducing the issue on more than one operating system.
• I have tried reproducing the issue on more than one workers pool.
• I have tried reproducing the issue with autoscaling, retries,
  ETA/Countdown & rate limits disabled.
• I have tried reproducing the issue after downgrading
  and/or upgrading Celery and its dependencies.

Related Issues and Possible Duplicates

Related Issues

• Subsequent groups within a chain fail Subsequent groups within a chain fail #3585
• Nested group(chain(group)) fails Nested group(chain(group)) fails #5496
• group([chain(group(task(), task()), group(...)), chain(group(...), group(...))]) construct never finishes group([chain(group(task(), task()), group(...)), chain(group(...), group(...))]) construct never finishes #2354
• Canvas with group-task-task-group does not work Canvas with group-task-task-group does not work #5354

Possible Duplicates

• Nested group-chain-group structure never succeeds Nested group-chain-group structure never succeeds #2573

Environment & Settings

Celery version:

software -> celery:4.3.0 (rhubarb) kombu:4.5.0 py:3.6.5
            billiard:3.6.0.0 redis:3.2.1
platform -> system:Linux arch:64bit
            kernel version:4.9.0-9-amd64 imp:CPython
loader   -> celery.loaders.app.AppLoader
settings -> transport:redis results:redis:///

broker_url: 'redis://localhost:6379//'
result_backend: 'redis:///'

Steps to Reproduce

Required Dependencies

• Minimal Python Version: 3.6
• Minimal Celery Version: 4.3
• Minimal Kombu Version: 4.5.0
• Minimal Broker Version: N/A or Unknown
• Minimal Result Backend Version: Redis
• Minimal OS and/or Kernel Version: Debian Stretch
• Minimal Broker Client Version: N/A or Unknown
• Minimal Result Backend Client Version: N/A or Unknown

Python Packages

amqp==2.4.2
billiard==3.6.0.0
celery==4.3.0
celery-eternal==0.1.1
celery-singleton==0.1.3
certifi==2018.4.16
kombu==4.5.0
pytz==2019.1
redis==3.2.1
vine==1.3.0

Other Dependencies

Minimally Reproducible Test Case

from time import sleep

from celery import Celery
from celery import group

app = Celery('tasks', broker = 'redis://')
app.conf['result_backend'] = app.conf.broker_url

@app.task
def prod(x, y):
    return x*y

@app.task
def subtract(args):
    return args[0]-args[1]

@app.task(shared=False)
def identity(args):
    """Identity task (returns its input)."""
    return args

x = (
    group( prod.s(1, 1), prod.s(2, 1) )
    |
    identity.s()
    |
    subtract.s()
    |
    group( prod.s(5), prod.s(6) )
)
r = x.delay()
sleep(10)
print(r.waiting())

# Another case
x = (
    group( prod.s(1, 1), prod.s(2, 1) )
    |
    subtract.si((3,4))
    |
    subtract.si((4,3))
    |
    subtract.si((6,5))
    |
    group( prod.si(5, 6), prod.si(6, 5) )
)
r = x.delay()
sleep(10)
print(r.waiting())

# However this works
# x = (
#    group( prod.s(1, 1), prod.s(2, 1) )
#    |
#    subtract.si((3,4))
#    |
#    group( prod.si(5, 6), prod.si(6, 5) )
# )
# r = x.delay()
# print([p.result for p in r.results])
# prints [30, 30]

# The adding more than one component in between 2 groups
# in a chain is causing the canvas to wait right before the final group

Expected Behavior

Above code snippet should return a list of 2 values
([5, 6] in for first case, [30, 30] for second case)

Actual Behavior

The code never never returns, keeps waiting

Celery worker logs for the first case mentioned

[2019-05-10 21:12:42,361: INFO/MainProcess] Connected to redis://localhost:6379//
[2019-05-10 21:12:42,366: INFO/MainProcess] mingle: searching for neighbors
[2019-05-10 21:12:43,380: INFO/MainProcess] mingle: all alone
[2019-05-10 21:12:43,393: INFO/MainProcess] celery@debian ready.
[2019-05-10 21:13:43,557: INFO/MainProcess] Received task: celery_task.prod[e11748a0-dca2-41fb-b799-a2ae7879c9fd]  
[2019-05-10 21:13:43,559: INFO/MainProcess] Received task: celery_task.prod[6ec1c88f-cca7-417c-8ff1-eacd9bdef7a1]  
[2019-05-10 21:13:43,566: INFO/ForkPoolWorker-2] Task celery_task.prod[6ec1c88f-cca7-417c-8ff1-eacd9bdef7a1] succeeded in 0.0064055299990286585s: 2
[2019-05-10 21:13:43,585: INFO/ForkPoolWorker-1] Task celery_task.prod[e11748a0-dca2-41fb-b799-a2ae7879c9fd] succeeded in 0.02549504399939906s: 1
[2019-05-10 21:13:43,586: INFO/MainProcess] Received task: celery_task.identity[87148533-8da1-4fc6-bbfe-608470cb3c1a]  
[2019-05-10 21:13:43,588: INFO/ForkPoolWorker-1] Task celery_task.identity[87148533-8da1-4fc6-bbfe-608470cb3c1a] succeeded in 0.0011925399994652253s: [2, 1]
[2019-05-10 21:13:43,588: INFO/MainProcess] Received task: celery_task.subtract[e907af43-a886-45ea-bfe3-f709c8267e5d]  
[2019-05-10 21:13:43,589: INFO/ForkPoolWorker-1] Task celery_task.subtract[e907af43-a886-45ea-bfe3-f709c8267e5d] succeeded in 0.00028916799965372775s: 1

Considering the first case mentioned in the code snippet, you can see that value of the task before the second group, subtract, is returned but the last group (consisting 2 prods) is never received, even after waiting many minutes. (I put the sleep(10) just as a proxy for some wait)