module 'google.api_core' has no attribute 'gapic_v1'

[ubaierbhat]

Since last week my DataFlow runs are failing because of this error. All I see in the logs is the following error

Stack trace

AttributeError: module 'google.api_core' has no attribute 'gapic_v1'
at PublisherClient (/usr/local/lib/python3.7/site-packages/google/cloud/pubsub_v1/gapic/publisher_client.py:245)
at <module> (/usr/local/lib/python3.7/site-packages/google/cloud/pubsub_v1/gapic/publisher_client.py:75)
at <module> (/usr/local/lib/python3.7/site-packages/google/cloud/pubsub_v1/publisher/client.py:29)
at <module> (/usr/local/lib/python3.7/site-packages/google/cloud/pubsub_v1/publisher/__init__.py:17)
at <module> (/usr/local/lib/python3.7/site-packages/google/cloud/pubsub_v1/__init__.py:18)
at <module> (/usr/local/lib/python3.7/site-packages/etl/util.py:13)
at <module> (/usr/local/lib/python3.7/site-packages/etl/extract/m/a.py:10)
at <module> (/usr/local/lib/python3.7/site-packages/etl/extract/m/a.py:6)
at find_class (/usr/local/lib/python3.7/site-packages/dill/_dill.py:474)
at load (/usr/local/lib/python3.7/site-packages/dill/_dill.py:305)
at loads (/usr/local/lib/python3.7/site-packages/dill/_dill.py:317)
at loads (/usr/local/lib/python3.7/site-packages/apache_beam/internal/pickler.py:283)
at __init__ (/usr/local/lib/python3.7/site-packages/dataflow_worker/workercustomsources.py:144)
at get_work_items (/usr/local/lib/python3.7/site-packages/dataflow_worker/workitem.py:131)
at run (/usr/local/lib/python3.7/site-packages/dataflow_worker/batchworker.py:807)

Environment details

python 3.7

$ pip freeze | grep google
google-api-core==1.20.0
google-api-helper==0.3.1
google-api-python-client==1.9.2
google-apitools==0.5.31
google-auth==1.16.1
google-auth-httplib2==0.0.3
google-auth-oauthlib==0.4.1
google-cloud-bigquery==1.24.0
google-cloud-bigtable==1.0.0
google-cloud-core==1.3.0
google-cloud-datastore==1.7.4
google-cloud-dlp==0.13.0
google-cloud-language==1.3.0
google-cloud-pubsub==1.0.2
google-cloud-spanner==1.13.0
google-cloud-videointelligence==1.13.0
google-cloud-vision==0.42.0
google-resumable-media==0.5.1
googleapis-common-protos==1.52.0
grpc-google-iam-v1==0.12.3

[plamut]

@ubaierbhat Could you please provide a bit more info, e.g. versions installed etc.? That would make diagnosing the issue easier, thanks!

(you can take GitHub issue bug report template as a guidance)

[ubaierbhat]

@plamut I have updated the description. the builds started failing since 5th June. The versions were set dynamically so I can't find out which versions were run before 5th.

[ubaierbhat]

@plamut regression in google-api-core is causing this problem googleapis/python-api-core#47

This issue can be moved google-api-python-client (as latest version 1.9.* requires google-api-core >= 1.18.0)

the temporary solution is to pin google-api-python-client to an older version

'google-api-core==1.17.0',
'google-api-python-client==1.8.4'

[plamut]

@ubaierbhat Thanks for additional research, I only found the time to come back at this today.

It indeed seems like an external issue, but since a workaround exists, I'll classify it as P2.

By the way - google-api-core 1.17.0 suffers from another regression (fixed in the most recent version) that breaks the streaming pull, thus the users that use that feature should instead temporarily pin google-api-core to an even older version, e.g. 1.16.0.

[meredithslota]

The root cause was fixed in googleapis/python-api-core#47 so it should be (famous last words) safe to unpin versions as per @busunkim96's comment on June 18th: googleapis/python-api-core#47 (comment)

@ubaierbhat Can you try bumping the version for google-api-core and seeing if that fixes the issue?

[bayesfactor]

I had to downgrade to 1.16.0 to resolve the import error

[meredithslota]

Removing the "external" label since the external issue is fixed.

[plamut]

@ubaierbhat @bayesfactor The external issue has been fixed, can any of you two confirm that by upgrading the dependencies to their latest versions indeed fixes the issue?

Thanks!

[meredithslota]

I'm going to go ahead and close this for now, but please feel free to reopen if the fix does not resolve your issue!

[ubaierbhat]

@meredithslota thanks and sorry, wasn't able to test because of busy schedule. I will test and report if the issue still exists (hopefully won't need to 🤞 )