[FLINK-38558] Require copy when use heap timer

[Zakelly]

What is the purpose of the change

The StateBackend#isSafeToReuseKVState is exposed to SQL layer to indicate whether it can perform object reuse. Currently it does not consider the heap timer when using RocksDB/ForSt, which may cause checkpoint broken as the FLINK-38558 describes. This PR fixes the implementation.

Brief change log

• Change StateBackend#isSafeToReuseKVState to consider heap timer factory.

Verifying this change

This change is a trivial rework and changes the corresponding test of StateBackendTestBase.

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

[flinkbot]

CI report:

• 3811919 Azure: FAILURE

Bot commands

The @flinkbot bot supports the following commands:

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

[fsk119]

LGTM

[davidradl]

As this is set in the constructor - it would be better to do the instanceof there and then set an instance field flag. At the moment we do the instanceof every call of this method.

[Zakelly]

If you read the code that invokes this method, you may find it only invoked once when the subtask start. So don't worry about the performance. I'd prefer keep this as it's more readable.