[
https://issues.apache.org/jira/browse/FLINK-19011?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17181981#comment-17181981
]
Jiayi Liao commented on FLINK-19011:
------------------------------------
[~sewen] Mmm... probably you're right because we also find other problems
caused by Union State like full gc on JobManager during failover. But I think
the Union state is still useful in some cases. For example we store the
watermark as a value state so that the job can recover with a correct
watermark.
I'd vote dropping this if we can find a better replacement of Union State.
> Parallelize the restore operation in OperatorStateBackend
> ----------------------------------------------------------
>
> Key: FLINK-19011
> URL: https://issues.apache.org/jira/browse/FLINK-19011
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Checkpointing
> Affects Versions: 1.11.1
> Reporter: Jiayi Liao
> Priority: Major
>
> To restore the states, union state needs to read state handles produced by
> all operators. And currently during the restore operation, Flink iterates the
> state handles one by one, which could last tens of minutes if the magnitude
> of state handles exceeds ten thousand.
> To accelerate the process, I propose to parallelize the random reads on HDFS
> and deserialization. We can create a runnable for each state handle and let
> it return the metadata and deserialized data, which can be aggregated in main
> thread.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
