[
https://issues.apache.org/jira/browse/BEAM-2096?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15986795#comment-15986795
]
Daniel Halperin edited comment on BEAM-2096 at 4/27/17 3:20 PM:
----------------------------------------------------------------
This looks like an actual bug in the DataflowRunner exposed by actually running
tests. Woo!
Pablo, if this is something you know about please take a look -- but if not
something you can handle fairly quickly then please triage forward. Needs to
get fixed, but your work on Aggregators is a better use of your time.
[~bchambers] – this type of error worries me. Before first stable release, can
we make the DataflowMetrics code more resilient to evolution of metrics
representations? It seems very hard to not break the current code.
was (Author: [email protected]):
This looks like an actual bug in the DataflowRunner exposed by actually running
tests. Woo!
Pablo, if this is something you know about please take a look -- but if not
something we can triage fairly quickly then let's move on.
[~bchambers] – this type of error worries me. Before first stable release, can
we make the DataflowMetrics code more resilient to evolution of metrics
representations? It seems very hard to not break the current code.
> NullPointerException in DataflowMetrics
> ---------------------------------------
>
> Key: BEAM-2096
> URL: https://issues.apache.org/jira/browse/BEAM-2096
> Project: Beam
> Issue Type: Bug
> Components: runner-dataflow
> Reporter: Kenneth Knowles
> Assignee: Pablo Estrada
> Priority: Blocker
>
> This started failing some time yesterday it looks like:
> https://builds.apache.org/view/Beam/job/beam_PostCommit_Java_ValidatesRunner_Dataflow/2949/#showFailuresLink
> Possible culprit at
> https://github.com/apache/beam/commit/4faa8feba822db000b4b42636408245422ed324d#diff-fff6f69683ac0e26db7e3ad095d7cdf0
> since that is when it starting failing.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
