[
https://issues.apache.org/jira/browse/FLINK-33316?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17778510#comment-17778510
]
Rui Fan edited comment on FLINK-33316 at 10/26/23 1:31 PM:
-----------------------------------------------------------
The change is subtle, so I push this commit directly.
Merged 1.19: a2681f6a85aaad21179f91e03a91b4a05158841e and
0388b760fc66975c70f797ad07f2e073738a7171
Merged 1.17: 024fa4776d0246a283a70743f1ce3c04981daeb9
Merged 1.18: 0dd3b4ce9f0b9f193926445bf9c1f8579fa86161
was (Author: fanrui):
The change is subtle, so I push this commit directly.
Merged 1.17: 024fa4776d0246a283a70743f1ce3c04981daeb9
Merged 1.18: 0dd3b4ce9f0b9f193926445bf9c1f8579fa86161
> Avoid unnecessary heavy getStreamOperatorFactory
> ------------------------------------------------
>
> Key: FLINK-33316
> URL: https://issues.apache.org/jira/browse/FLINK-33316
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / Configuration
> Affects Versions: 1.17.0, 1.18.0
> Reporter: Rui Fan
> Assignee: Rui Fan
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.17.2, 1.19.0, 1.18.1
>
>
> See FLINK-33315 for details.
> This Jira focus on avoid unnecessary heavy getStreamOperatorFactory, it can
> optimize the memory and cpu cost of Replica_2 in FLINK-33315.
> Solution: We can store the serializedUdfClassName at StreamConfig, and using
> the getStreamOperatorFactoryClassName instead of the heavy
> getStreamOperatorFactory in OperatorChain#getOperatorRecordsOutCounter.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
