[
https://issues.apache.org/jira/browse/FLINK-25360?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17463632#comment-17463632
]
刘方奇 commented on FLINK-25360:
-----------------------------
[~yunta] Hi, THX, in my case, we want to read the checkpoint, change the
maxParallelism, then store in the new checkpoint. And state-processor-api need
we to build a dataSet application which means we need to be aware of all the
state desc to write the UDF of the state-processor-api. And what we had is just
the checkpoint metadata.
Reply to your second words:
Certainly, flink store the state's Metadata which contains state name / value
serializer / key serializer/ namespace serializer, but it's not enough to build
a state desc. (PS: the state desc which i mentioned is the
org.apache.flink.api.common.state.StateDescriptor Object)
BTW, state-processor-api actually provide the power that we can handle the
state, but we should mock the stream operator which make us so hard to generate
the state and the checkpoint metadata.
> Add State Desc to CheckpointMetadata
> ------------------------------------
>
> Key: FLINK-25360
> URL: https://issues.apache.org/jira/browse/FLINK-25360
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Checkpointing, Runtime / State Backends
> Reporter: 刘方奇
> Priority: Major
> Attachments: image-2021-12-17-20-01-42-423.png
>
>
> Now we can't get the State Descriptor info in the checkpoint meta. Like the
> case if we use state-processor-api to load state then rewrite state, we can't
> flexible use the state.
> Maybe there are other cases we need the State Descriptor, so can we add this
> info?
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
