This closes #12683.

The current timestamp format in JSON format is not SQL standard which uses RFC-3339. This commit changes the default behavior to parse/generate timestamp using SQL standard. Besides, it introduces an option "json.timestamp-format.standard" to have the ability to fallback to ISO standard. 

This closes #12661

This closes #12691

 [FLINK-18238][checkpoint] Broadcast CancelCheckpointMarker while executing checkpoint aborted by coordinator RPC

In the case of aborting checkpoint RPC from CheckpointCoordinator, it will prevent executing the
respective checkpoint which was already triggered before. But we also need to broadcast the
CancelCheckpointMarker before exiting the execution , otherwise the downstream side would
probably wait for barrier alignment until deadlock.
    
This closes #12664.

This reverts commit 8ca388ca.

Revert "[FLINK-17800][roksdb] Support customized RocksDB write/read options and use RocksDBResourceContainer to get them"

This reverts commit f1250625.

[FLINK-18330][python][legal] Update the NOTICE file of flink-python module adding beam-runners-core-java and beam-vendor-bytebuddy

This closes #12692.

This closes #12313

This closes #12311

Revert "[FLINK-18072][hbase] Fix HBaseLookupFunction can not work with new internal data structure RowData"

This reverts commit b2711c5d because it broke master

This closes #12594

[FLINK-18303][filesystem][hive] Fix Filesystem connector doesn't flush part files after rolling interval

This commit introduces option 'sink.rolling-policy.check-interval' (default 1min) to control the frequency to check part file rollover.

This closes #12657

This closes #12632

This closes #12663

Before, it could happen that the new messages written in the test were
not written to the new partition. Now we use explicit keys of which we
know that they will hash to the second partition.

You can verify this by changing "key" to "keya". The test will then fail
deterministically.

This was broken because the behaviour of the Kafka/ZooKeeper command
line tools on Kafka 2.4.1 is slightly different:

zookeeper_shell.sh does not print debug output to stderr as it did
before. We change queryBrokerStatus() to instead consume stdout and
check that we get valid information for the broker.

The output of kafka-topics.sh now has a space between "PartitionCount:"
and the partition count. Before we had "PartitionCount:2", now it's
"PartitionCount: 2". We fix this by making the regex more lenient.

This also splits the waiting on ZooKeeper/Kafka into two loops to better
see which one we're blocking on.

This closes #12643

This closes #12654.