[FLINK-18990][task] Read channel state sequentially

[rkhachatryan]

What is the purpose of the change

Read channel state sequentially to reduce the number of seeks.

Verifying this change

• Added SequentialChannelStateReaderImplTest and ChannelStateChunkReaderTest for the new code.
• Added RecoveredInputChannelTest to check that conversion is not permitted untill fully recovered and state consumed
• some tests are eliminated because: 1) state in StreamTask is read unconditionally; 2) recovered channel can't be converted without being fully consumed anymore

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): (yes / 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: yes
• The S3 file system connector: no

Documentation

• Does this pull request introduce a new feature? no
• If yes, how is the feature documented? no

[flinkbot]

Thanks a lot for your contribution to the Apache Flink project. I'm the @flinkbot. I help the community
to review your pull request. We will use this comment to track the progress of the review.

Automated Checks

Last check on commit 3c79716 (Tue Sep 08 09:04:30 UTC 2020)

Warnings:

• No documentation files were touched! Remember to keep the Flink docs up to date!

_{Mention the bot in a comment to re-run the automated checks.}

Review Progress

• ❓ 1. The [description] looks good.
• ❓ 2. There is [consensus] that the contribution should go into to Flink.
• ❓ 3. Needs [attention] from.
• ❓ 4. The change fits into the overall [architecture].
• ❓ 5. Overall code [quality] is good.

Please see the Pull Request Review Guide for a full explanation of the review process.

The Bot is tracking the review progress through labels. Labels are applied according to the order of the review items. For consensus, approval by a Flink committer of PMC member is required

Bot commands

The @flinkbot bot supports the following commands:

• @flinkbot approve description to approve one or more aspects (aspects: description, consensus, architecture and quality)
• @flinkbot approve all to approve all aspects
• @flinkbot approve-until architecture to approve everything until architecture
• @flinkbot attention @username1 [@username2 ..] to require somebody's attention
• @flinkbot disapprove architecture to remove an approval you gave earlier

[flinkbot]

CI report:

• 3041c48 UNKNOWN
• c0b8f40 Azure: SUCCESS

Bot commands

The @flinkbot bot supports the following commands:

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

[AHeise]

Very intriguing solution. I left a couple of comments but nothing deep. While reviewing my main issue was to understand the relationship between Sequential and old ChannelStateReader (become only clear after last commit). Last commits looks a bit incomplete.
Note that I still struggle a bit on the concepts of StateHandle and delegate. I dropped a question at the main commit.

[rkhachatryan]

Thanks for reviewing @AHeise and @pnowojski
I've addressed your feedback, please take a look.

[AHeise]

Addressed most of my comments, a few open discussions.

I'm thinking that the last 3 commits (remove tests, methods, class of channel state reader) should be squashed.

[rkhachatryan]

@flinkbot run azure

[AHeise]

LGTM % squash last 3 commits. I'm still waiting for @pnowojski to look at the changed test.

[rkhachatryan]

Thanks for reviewing @AHeise and @pnowojski .
I've updated StreamTaskMultipleInputSelectiveReadingTest.testInputStarvation and squashed the last three commits.

[rkhachatryan]

Build failure unrelated: FLINK-19427
Private build passed.

[rkhachatryan]

I rebased the PR and resolved conflicts with FLINK-19027 by moving the assignment of exclusive buffers for RecoveredInputChannel into it's requestBufferBlocking method.

I also replaced RemoteInputChannel.assignExclusiveSegments() with InputChannel.setup() make it clear that RecoveredInputChannel doesn't need setup and avoid if/casts (e4946c0).

@pnowojski, @AHeise please take a look.