[
https://issues.apache.org/jira/browse/BEAM-5299?focusedWorklogId=166647&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-166647
]
ASF GitHub Bot logged work on BEAM-5299:
----------------------------------------
Author: ASF GitHub Bot
Created on: 15/Nov/18 21:08
Start Date: 15/Nov/18 21:08
Worklog Time Spent: 10m
Work Description: kennknowles edited a comment on issue #7012: Revert
"Revert "Revert "[BEAM-5299] Define max timestamp for global w…
URL: https://github.com/apache/beam/pull/7012#issuecomment-439190285
Thank you for tagging me here. This is an important issue.
Our rollback first policy does not apply to a change that passes all of
Beam's postcommit tests. It does apply to _Beam's_ postcommit suites for just
one runner. The purpose of rapid rollback is foremost to restore test signal
and not to disrupt the work of other contributors.
But...
- We have at least three examples of runners where there are probably tests
outside the Beam repo: Dataflow, Samza runner, and IBM Streams.
- We also may have users that try running their production loads against
Beam `master` branch to learn early whether the next release will break them.
We should also respect these other sources of information, balancing haste
with fairness and transparency. If nothing public and owned/blessed by Beam is
broken, then you need to explain the issue and justify it publicly. Certainly
you should contact the author(s) and reviewer(s), which you did here anyhow.
In this case, it sounds like a postcommit did fail, so it is OK to rollback
with only that as the explanation. You should include a Jira with pointer to
the failed run, and also just link it on the PR. So that's missing in this
particular case, and the PR description refers to internal tests so it is
confusing.
I agree with @tweise that we need to draw an extremely sharp distinction
between Beam's own tests and whatever testing a user or vendor might do on
their own.
I suggest that dev@ is the best place to continue this if a deeper
discussion is desired.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on 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: 166647)
Time Spent: 15h 50m (was: 15h 40m)
> Define max global window as a shared value in protos like URN enums.
> --------------------------------------------------------------------
>
> Key: BEAM-5299
> URL: https://issues.apache.org/jira/browse/BEAM-5299
> Project: Beam
> Issue Type: Improvement
> Components: beam-model, sdk-go, sdk-java-core, sdk-py-core
> Reporter: Luke Cwik
> Assignee: Maximilian Michels
> Priority: Minor
> Labels: portability
> Fix For: 2.9.0
>
> Time Spent: 15h 50m
> Remaining Estimate: 0h
>
> Instead of having each language define a max timestamp themselves, define the
> max timestamps within proto to be shared across different languages.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
