Issue with Airflow 1.10.13, 2020-11-24 release

[jain6968]

Apache Airflow version: composer-1.11.2-airflow-1.10.6

Kubernetes version (if you are using kubernetes) (use kubectl version): 1.16.13
apiVersion: kubelet.config.k8s.io/v1beta1

Environment: Using compute engine

1. composer-1.11.2-airflow-1.10.6
2. Python – 3.6
3. 'dbt==0.17.0',
4. 'google-cloud-storage',
5. 'google-cloud-secret-manager==1.0.0',
6. 'protobuf==3.12.2'

• Cloud provider or hardware configuration:
  gcp compute engine

• OS (e.g. from /etc/os-release):
  Debian , Intel Broadwell

• Kernel (e.g. uname -a):
  n1-standard-1 (1 vCPU, 3.75 GB memory)

• Install tools:
  python sdk 3.6.0

• Others:
  apache beam , dbt

What happened:

Below is the configuration which worked till December 1st

1. composer-1.11.2-airflow-1.10.6
2. Python – 3.6
3. 'dbt==0.17.0',
4. 'google-cloud-storage',
5. 'google-cloud-secret-manager==1.0.0',
6. 'protobuf==3.12.2'

With the above configuration we are observing below error.
ERROR: Cannot install dbt and protobuf==3.12.2 because these package versions have conflicting dependencies.\n\nThe conflict is caused by:\n The user requested protobuf==3.12.2\n dbt-bigquery 0.17.0 depends on protobuf<3.12 and >=3.6.0\n\nTo fix this you could try to:\n1. loosen the range of package versions you've specified\n2. remove package versions to allow pip attempt to solve the dependency conflict\n\nERROR: ResolutionImpossible

Post receiving this error message we have tried many version changes to resolve the above issue. Below listed are few of the fixes applied and errors we have observed for those fixes.

1. Fix applied: downgraded protobuf version from 3.12.2 to 3.11.3 as per the above error
   Error: Requested docutils<0.16,>=0.10 from https://files.pythonhosted.org/packages/d4/12/6c3fd74a590c7327c98cae008c11d536029fa9cd7924de477e8cb8804186/docutils-0.15.1-post1.tar.gz#sha256=f33ddb723332c6d6b6d99731ee1fc0c35eb4044a2df5cca1c64c8aa78eaf22cb (from botocore<1.15,>=1.5.0->dbt-redshift==0.17.0->dbt==0.17.0) has different version in metadata: '0.15.1'

2. Fix applied: Specified docutils version to 0.15.2 as to fix the above error
   Error: SyntaxError: Missing parentheses in call to 'print'. Did you mean print("Generating %s..." % output)?\n ----------------------------------------\nERROR: Command errored out with exit status 1: python setup.py egg_info Check the logs for full command output.

3. Fix applied: Tried upgrading python version from 3.6 to 3.8.
   Error: The path python3.8 (from --python=python3.8) does not exist.

4. Fix applied: Removed package versioning for above libraries to allow pip attempt to solve the dependency conflict.
   Error: Command errored out with exit status 1:\n command: /tmp/venv__vuo821/bin/python3.6 -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'/tmp/pip-install-ghz6thkq/protobuf_a5303ec5c2df4986bb9b05948ae30052/setup.py'"'"'; file='"'"'/tmp/pip-install-ghz6thkq/protobuf_a5303ec5c2df4986bb9b05948ae30052/setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(file);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, file, '"'"'exec'"'"'))' egg_info --egg-base /tmp/pip-pip-egg-info-gi38587q\n cwd: /tmp/pip-install-ghz6thkq/protobuf_a5303ec5c2df4986bb9b05948ae30052/\n Complete output (6 lines):\n Traceback (most recent call last):\n File "", line 1, in \n File "/tmp/pip-install-ghz6thkq/protobuf_a5303ec5c2df4986bb9b05948ae30052/setup.py", line 52\n print "Generating %s..." % output\n ^\n SyntaxError: Missing parentheses in call to 'print'. Did you mean print("Generating %s..." % output)?\n ----------------------------------------\nERROR: Command errored out with exit status 1: python setup.py egg_info Check the logs for full command output

5. Fix applied: Upgraded dbt to 0.17.2
   Error: Received SIGTERM. Terminating subprocesses.

6. Fix applied: Tried downgrading setuptools version to 49.6.0
   Error: SyntaxError: Missing parentheses in call to 'print'. Did you mean print("Generating %s..." % output)?\n ------------------------------- ---------\nERROR: Command errored out with exit status 1: python setup.py egg_info Check the logs for full command output.

7. Fix applied: upgraded Composer and Airflow versions to composer-1.13.2-airflow-1.10.9
   Error: Got similar errors with above configuration combinations

What you expected to happen:

We expect the DAG to complete and run the python script without any dependency issues.

How to reproduce it:
with Airflow version composer-1.11.2-airflow-1.10.6 and environment configs.

Error Details from logs:

ERROR: Cannot install dbt and protobuf==3.12.2 because these package versions have conflicting dependencies.\n\nThe conflict is caused by:\n The user requested protobuf==3.12.2\n dbt-bigquery 0.17.0 depends on protobuf<3.12 and >=3.6.0\n\nTo fix this you could try to:\n1. loosen the range of package versions you've specified\n2. remove package versions to allow pip attempt to solve the dependency conflict\n\nERROR: ResolutionImpossible"

Anything else we need to know:

Apart from the above fixes we have tried other combinations of packages and specifying package versions mentioned in https://pip.pypa.io/en/latest/user_guide/#fixing-conflicting-dependencies
Unfortunately none of those fixes worked.

[boring-cyborg]

Thanks for opening your first issue here! Be sure to follow the issue template!

[jain6968]

Have followed the issue template, just expecting response / resolution asap as we are in mid of release!

[potiuk]

Have followed the issue template, just expecting response / resolution asap as we are in mid of release!

I think you should post in slack (see airflow.apache.org - community page). This is a better place to ask such questions. Maybe someone with similar experience can help you. We are just in the middle of releasing Airflow 2.0 so do not expect a lot of attention from the committers.

It is generally expected that you won't be able to install some packages due to conflicts - Airflow has > 420 dependencies and it's only recently (1.10.14 which is sue tomorrow and 2.0.0rc which is due today) we managed to get consistent set of constraints for our tests (and even that excludes apache-beam for now). And if you add your own packages, you are a bit on your own there as we cannot really help with those and there is no guarantee those deps will work.

In this particular case, I have no high hopes though - it's pretty much unsolvable as you have conflicting dependencies. You cannot install protobuf protobuf==3.12.2 because you have the google libraries that require lower version.

You can always use PythonVirtualEnv operator https://airflow.apache.org/docs/apache-airflow/stable/_modules/airflow/operators/python_operator.html if you need to run particular python operation as suggested by Composer team : https://cloud.google.com/composer/docs/how-to/using/installing-python-dependencies