[
https://issues.apache.org/jira/browse/BEAM-14014?focusedWorklogId=768997&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-768997
]
ASF GitHub Bot logged work on BEAM-14014:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 11/May/22 11:01
Start Date: 11/May/22 11:01
Worklog Time Spent: 10m
Work Description: tvalentyn commented on code in PR #17244:
URL: https://github.com/apache/beam/pull/17244#discussion_r870146064
##########
sdks/python/apache_beam/io/gcp/gcsfilesystem.py:
##########
@@ -38,6 +40,15 @@ class GCSFileSystem(FileSystem):
CHUNK_SIZE = gcsio.MAX_BATCH_OPERATION_SIZE # Chuck size in batch operations
GCS_PREFIX = 'gs://'
+ def __init__(self, pipeline_options):
+ super().__init__(pipeline_options)
+ if isinstance(pipeline_options, PipelineOptions):
+ gcs_options = pipeline_options.view_as(GoogleCloudOptions)
+ impersonate_service_account = gcs_options.impersonate_service_account
Review Comment:
+1 to Kenn's comment. This looks strange to me as well. Can we make pipeline
options a parameter of get_service_credentials() and adjust credential
generation and usages of this method accordingly? Given that:
- credential is a global singleton
- pipeline options are also global,
you could clear the gcs_options.impersonate_service_account within the code
that creates the credential singleton.
Issue Time Tracking
-------------------
Worklog Id: (was: 768997)
Time Spent: 4h 40m (was: 4.5h)
> Support impersonation credentials in Dataflow runner.
> -----------------------------------------------------
>
> Key: BEAM-14014
> URL: https://issues.apache.org/jira/browse/BEAM-14014
> Project: Beam
> Issue Type: Improvement
> Components: runner-dataflow
> Reporter: Valentyn Tymofieiev
> Assignee: Ryan Thompson
> Priority: P2
> Time Spent: 4h 40m
> Remaining Estimate: 0h
>
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
