[
https://issues.apache.org/jira/browse/BEAM-10308?focusedWorklogId=457410&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-457410
]
ASF GitHub Bot logged work on BEAM-10308:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 10/Jul/20 21:47
Start Date: 10/Jul/20 21:47
Worklog Time Spent: 10m
Work Description: TheNeuralBit commented on pull request #12196:
URL: https://github.com/apache/beam/pull/12196#issuecomment-656906406
Here are some answers to the questions from
https://beam.apache.org/contribute/release-guide/#review-cherry-picks:
> Is this a regression from a previous release? (If no, fix could go to a
newer version.)
No, it's a bug, but it existed in previous releases.
> Is this a new feature or related to a new feature? (If yes, fix could go
to a new version.)
Not a new feature, a bugfix for a >1 year old feature.
> Would this impact production workloads for users? (E.g. if this is a
direct runner only fix it may not need to be a cherry pick.)
Yes, python users may want to use external transforms in production, and
they could be blocked by this bug.
> What percentage of users would be impacted by this issue if it is not
fixed? (E.g. If this is predicted to be a small number it may not need to be a
cherry pick.)
Hard to say. I'm not sure how many users are trying out external transforms,
and of those that are it's possible they won't ever encounter this bug.
> Would it be possible for the impacted users to skip this version? (If
users could skip this version, fix could go to a newer version.)
Not sure this applies since there isn't a "working" version for users to
stay on and skip 2.23.0.
To be honest I think this situation is somewhat of a gap in the guidance
there. There may be cases where its worth delaying a release for a severe
bugfix even if it's a longstanding issue and not a regression, and that's not
addressed in our release guide. We could draw a line somewhere: e.g. maybe data
loss/incorrect result bugs can delay, crashes can not?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Issue Time Tracking
-------------------
Worklog Id: (was: 457410)
Time Spent: 6h (was: 5h 50m)
> Component id assignement is not consistent across PipelineContext instances
> ---------------------------------------------------------------------------
>
> Key: BEAM-10308
> URL: https://issues.apache.org/jira/browse/BEAM-10308
> Project: Beam
> Issue Type: Bug
> Components: cross-language, sdk-py-core
> Reporter: Brian Hulette
> Assignee: Brian Hulette
> Priority: P1
> Fix For: 2.24.0
>
> Time Spent: 6h
> Remaining Estimate: 0h
>
> The "unique ref" ids used in PipelineContext are generated on the fly, which
> can cause us to get a different id for the same component in different
> contexts.
> This becomes a problem when ExternalTransform is used, because it creates its
> own pipeline context for expansion. So its possible the component ids in the
> expansion request will actually refer to an entirely different component when
> the pipeline is finally assembled for execution.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
