[ 
https://issues.apache.org/jira/browse/BEAM-9423?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17074678#comment-17074678
 ] 

Maximilian Michels commented on BEAM-9423:
------------------------------------------

If you don't make structural changes to the pipeline it is as easy as 
restarting from a savepoint.

There is limited support for modifying the pipeline. It is essentially as 
tricky as with Flink applications. You should assign stable names to all your 
transforms such that the state can be remapped. It can help to use the 
{{allowNonRestoredState}} option which allows you to drop the state of removed 
operators. If your state format changes, you need to write a migration pipeline 
which transforms the operator state.

> Re-Add the stop button to the Flink web interface for pipelines
> ---------------------------------------------------------------
>
>                 Key: BEAM-9423
>                 URL: https://issues.apache.org/jira/browse/BEAM-9423
>             Project: Beam
>          Issue Type: Bug
>          Components: runner-flink
>    Affects Versions: 2.19.0, 2.20.0
>            Reporter: Tobias Kaymak
>            Assignee: Maximilian Michels
>            Priority: Major
>
> The Flink 1.9/1.10 web interface is missing the Stop button for a streaming 
> pipeline, which is pretty useful to drain a streaming pipeline.
> For example to be able to cleanly shutdown a pipeline when a feature is going 
> to be released tomorrow, and the pipeline processing code should be updated 
> before, without the overhead of re-reading the complete past. Taking a 
> snapshot uses the current representation of the Beam code as a Flink job, but 
> what if one wants to update the Beam code and thus the pipeline code without 
> the need to reprocess the whole history? Moreover, a stop button is very 
> useful when a new Flink version is going to be rolled out: then one can drain 
> all pipelines, rollout a new Flink version and start them at the point where 
> they left off with their last committed offset in Kafka.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to