[FLINK-37460][state] Make state processor API checkpoint ID configurable

[gaborgsomogyi]

This is backport of #26285.

What is the purpose of the change

At the moment the state processor API is using hardcoded 0 as checkpoint ID.
In this PR I've made it configurable and savepoint modification keeps the original checkpoint ID.

Brief change log

Made state processor API checkpoint ID configurable and savepoint modification keeps the original ID.

Verifying this change

Changed automated tests.

Does this pull request potentially affect one of the following parts:

• Dependencies (does it add or upgrade a dependency): no
• The public API, i.e., is any changed class annotated with @Public(Evolving): no
• The serializers: no
• The runtime per-record code paths (performance sensitive): no
• Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Kubernetes/Yarn, ZooKeeper: no
• The S3 file system connector: no

Documentation

• Does this pull request introduce a new feature? no
• If yes, how is the feature documented? not applicable

[gaborgsomogyi]

This one is new!☝️The following error is shown by japicmp:

[ERROR] Failed to execute goal com.github.siom79.japicmp:japicmp-maven-plugin:0.17.1:cmp (default) on project flink-state-processor-api: There is at least one incompatibility: org.apache.flink.state.api.OperatorTransformation.bootstrapWith(org.apache.flink.api.java.DataSet,long):CLASS_GENERIC_TEMPLATE_CHANGED,org.apache.flink.state.api.OperatorTransformation.bootstrapWith(org.apache.flink.streaming.api.datastream.DataStream,long):CLASS_GENERIC_TEMPLATE_CHANGED -> [Help 1]

The problem is caused by siom79/japicmp#368.
The issue behind can be solved the following ways:

• We increase the jacmp version where the mentioned bug has been resolved
• We add class to the whitelist

[Zakelly]

Hi @gaborgsomogyi How about bump this on master branch and backport it here?

[gaborgsomogyi]

I'm fine with anything since this is ugly here. Separating the intentions is good in general so doing it.

[gaborgsomogyi]

https://issues.apache.org/jira/browse/FLINK-37480

[gaborgsomogyi]

When it's merged then we're going to have the possibility to remove this part.

[gaborgsomogyi]

It's no more in this change.

[flinkbot]

CI report:

• b3b13c5 Azure: SUCCESS

Bot commands

The @flinkbot bot supports the following commands:

• @flinkbot run azure re-run the last Azure build

[gaborgsomogyi]

cc @grzegorz-liter-getindata @Zakelly

[gaborgsomogyi]

@grzegorz-liter-getindata can you please check your side?

[grzegorz-liter-getindata]

@gaborgsomogyi yeah, give me couple of days thou please. Is there some date set for 1.20.2 release we need to met?

[gaborgsomogyi]

@grzegorz-liter-getindata any update? Related the 1.20.2 it's not yet planned but I'm pretty sure it's going to happen since 1.20 has LTS support.

[davidradl]

If this is a new configurable thing - it should have associated documentation to detail how, when and why this would be useful.

[gliter]

@gaborgsomogyi I finished similar test like before
Setup:
Cluster running from official dist 1.20.1
Job build with official 1.20.1 that generates random numbers and output to Kafka
State migration job build with 1.20-SNAPSHOT built from your branch

1. Start the job, let it run for a while
2. Stop with savepoint
3. Check if KafkaCommittable is present
   OK: KafkaCommittable{producerId=18, epoch=0, transactionalId=kafka-sink-0-18}
4. Change uid of stateful-function to stateful-function-tmp
5. Bootstrap stateful-function state using stateful-function-tmp as source and change value by 1000
6. Validate checkpointId is preserved after migration
   OK: checkpointId is preserved
7. Start job from savepoint
   OK: job continues from last checkpointId
8. state was updated and loaded correctly, outputs are now higher by 1000
9. stop the job with savepoint and check if KafkaCommitable was properly handled - that is, Commitable from checkpoint before migration was cleaned and there is only latest Commitable
   OK: Kafka commitable: KafkaCommittable{producerId=55, epoch=0, transactionalId=kafka-sink-0-54}

Everything seems to work in order. Great thanks for quick resolution 🙌 .