[
https://issues.apache.org/jira/browse/BEAM-9505?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17071109#comment-17071109
]
Niel Markwick commented on BEAM-9505:
-------------------------------------
This issue came was reported from StackOverflow. To reproduce it you would
have to run a streaming pipeline in dataflow. The SO bug was attempting to
write to Spanner (and in the SpannerIO, I have some code in StartBundle to
sanity-check the state, which reports the error.).
[https://stackoverflow.com/questions/60658135/spannerio-java-lang-illegalstateexception-sorter-should-be-null-here]
I would have thought that a trivial DoFn in a pipeline taking pub/sub input
would be able to reproduce it (but have not created that pipeline myself yet)
> DoFn.StartBundle called multiple times when streaming
> -----------------------------------------------------
>
> Key: BEAM-9505
> URL: https://issues.apache.org/jira/browse/BEAM-9505
> Project: Beam
> Issue Type: Bug
> Components: runner-dataflow
> Affects Versions: 2.18.0, 2.19.0
> Reporter: Niel Markwick
> Assignee: Boyuan Zhang
> Priority: Major
>
> When using DataflowRunner in streaming mode. DoFn.StartBundle is called
> multiple times for the same bundle.
>
> This does not occur with DirectRunner.
> This breaks DoFn's which require per-bundle setup and teardown procedures.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
