[
https://issues.apache.org/jira/browse/BEAM-2259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16006727#comment-16006727
]
ASF GitHub Bot commented on BEAM-2259:
--------------------------------------
GitHub user tgroh opened a pull request:
https://github.com/apache/beam/pull/3080
[BEAM-2259] Do not use infinite lateness in Reshuffle
Be sure to do all of the following to help us incorporate your contribution
quickly and easily:
- [ ] Make sure the PR title is formatted like:
`[BEAM-<Jira issue #>] Description of pull request`
- [ ] Make sure tests pass via `mvn clean verify`.
- [ ] Replace `<Jira issue #>` in the title with the actual Jira issue
number, if there is one.
- [ ] If this contribution is large, please file an Apache
[Individual Contributor License
Agreement](https://www.apache.org/licenses/icla.pdf).
---
Elements will be dropped only if the window they are assigned to is
expired, which can generally be performed arbitrarily.
Using infinite allowed lateness can cause an IllegalStateException if
the ReduceFnRunner observes certain combinations of timers and elements,
and attempts to buffer an element, setting a watermark hold for window
expiry (which is past the end of time).
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/tgroh/beam reshuffle_exception
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/beam/pull/3080.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #3080
----
----
> Reshuffle may set watermark holds past the end of time
> ------------------------------------------------------
>
> Key: BEAM-2259
> URL: https://issues.apache.org/jira/browse/BEAM-2259
> Project: Beam
> Issue Type: Bug
> Components: sdk-java-core
> Reporter: Thomas Groh
> Assignee: Thomas Groh
> Fix For: 2.0.0
>
>
> When it does so, the ReduceFnRunner throws an exception that the watermark
> hold is past the end of time
> This affects all non-Dataflow runners.
> (With additional checks)
> Caused by: java.lang.IllegalArgumentException: Assigned
> TimestampedValueInSingleWindow{value=KV{key: gameexamples:test_tgroh.l
> eaderboard_team shard: 33,
> org.apache.beam.sdk.io.gcp.bigquery.TableRowInfo@31436a52},
> timestamp=2017-05-11T15:29:59.999Z, window=[20
> 17-05-11T15:15:00.000Z..2017-05-11T15:30:00.000Z),
> pane=PaneInfo{isFirst=true, timing=EARLY, index=0}} to a window
> [2017-05-11T15:15:
> 00.000Z..2017-05-11T15:30:00.000Z) with fn
> org.apache.beam.sdk.util.IdentityWindowFn@58543f9f. Expiry is past end of
> time. Windowing
> strategy
> WindowingStrategy{windowFn=org.apache.beam.sdk.util.IdentityWindowFn@58543f9f,
> allowedLateness=PT9223372036854.775S, trigger
> =ReshuffleTrigger(), accumulationMode=DISCARDING_FIRED_PANES,
> timestampCombiner=EARLIEST}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
