[UPDATE] Release 1.13 feature freeze

[dwysakowicz]

Hi all,

The agreed date of a feature freeze is due in about a month. Therefore we thought it would be a good time to give an update of the current progress.

From the information we gathered there are currently no known obstacles or foreseeable delays. We are still aiming for the end of March as the date for a feature freeze.

Some update on the progress of particular issues. Most of the issues listed in the https://cwiki.apache.org/confluence/x/jAhRCg are well on track. We will list only the issues that we feel might not end up in the release or end up there partially:

• Tolerate temporarily suspended ZooKeeper connections (FLINK-10052)
• Support shipping local/remote files for yarn/k8s integrations (FLINK-20681 / FLINK-20811 / FLINK-20867)
• Consistent Flink SQL time function behavior (FLIP-162)

These are long lasting efforts and are planned to be delivered based on best effort, therefore they might not be fully functional in 1.13:

• Generalized incremental checkpoints
  
• Incremental snapshots for heap-based state backend

We have 4 release blockers as of today:

Checkpointing:

• https://issues.apache.org/jira/browse/FLINK-21453 BoundedOneInput.endInput is NOT called when doing stop with savepoint WITH drain

Dependencies:

• https://issues.apache.org/jira/browse/FLINK-21152 Bump flink-shaded to 13.0

Coordination:

• https://issues.apache.org/jira/browse/FLINK-21030 Broken job restart for job with disjoint graph

Connectors:

• https://issues.apache.org/jira/browse/FLINK-20188 Add Documentation for new File Source

We have 100 test instabilities across all versions out of which 37 occurred on master (1.13) branch.

Let me know if we made a mistake somewhere or something is inaccurate. Moreover everyone is welcome to give more thorough update if you feel like so.

Your release managers

Guowei & Dawid