StephanEwen opened a new pull request #11202: [FLINK-16247][checkpoints] 
Various cleanups in checkpointing code
URL: https://github.com/apache/flink/pull/11202
 
 
   ## Prerequisite
   
   *This PR is based on #11179*
   
   ## What is the purpose of the change
   
   **This is part of the preparation for implementing Fault Tolerance for 
Operator Coordinators for the new Source Interface (FLIP-27).**
   
   The PR contains various cleanups and simplifications to the checkpoint 
metadata handling code.
   
   ## Brief change log
   
     - [FLINK-16247][checkpoints] Rename 'checkpoints.savepoint.' package and 
classes to 'checkpoint.metadata.'
   
       These classes describe the persistence of checkpoint metadata. 
Originally, savepoints where the only snapshots with persistent metadata, while 
checkpoints had the metadata only in memory or in ZooKeeper (Java Serialized). 
   
       Nowadays, checkpoints and savepoints both persist metadata in the same 
way, and hence the package is not actually related to savepoints any more.
   
     - [FLINK-16259][checkpoints] Drop interface/impl separation for checkpoint 
Metadata and rename MetadataV2 to CheckpointMetadata
   
       The interface 'Metadata' (nee Savepoint) and its different 
implementations was misleading. There is only ever one real implementation: the 
latest version. All deserializers need to go to that version, to keep the 
complex versioning logic out of other parts of the code.
   
       Further more, this is a simple "data holder" class, simply storing 
collections, there is no behavior to be abstracted behind an interface.
   
       Even if there were different versions (as in an earlier version of the 
code, where the SavepointV1 class for tests was in the main scope), the 
interface was a union of all accessors to all different collections from 
different versions (with all accessors but the ones for the latest version 
deprecated).
   
       As a result, the 'Metadata' interface was misleading and suggesting 
something that is not
   there, namely that multiple versions of the metadata are handled in various 
places in the code, and not just during deserialization.
   
   ## Verifying this change
   
   This change is pure refactoring without any changes to functionality.
   
   ## 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/Mesos, 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**
   

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

Reply via email to