Skip to content

Conversation

@rkhachatryan
Copy link
Contributor

@rkhachatryan rkhachatryan commented Dec 28, 2020

What is the purpose of the change

In scenarios with multiple inputs (e.g. co-group; not union) one input may receive a
checkpoint barrier while the second input is still restoring state. This (previous)
state is currently not included into the snapshot, which therefore will be incomplete.

Brief change log

  • Disable alignment timeout by default in UnalignedCheckpointITCase
  • Decline checkpoint if state of the input gate was not yet consumed

Verifying this change

  • Added unit test SingleInputGateTest#testCheckpointsDeclinedUnlessStateConsumed.
  • Existing UnalignedCheckpointITCase un-ignored - without the change it fails once per ~10 runs

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: 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
Copy link
Collaborator

flinkbot commented Dec 28, 2020

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 c020556 (Fri May 28 06:58:28 UTC 2021)

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.

Details
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
Copy link
Collaborator

flinkbot commented Dec 28, 2020

CI report:

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

Copy link
Contributor

@pnowojski pnowojski left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM % one comment and compilation failure

@rkhachatryan rkhachatryan force-pushed the f20654 branch 3 times, most recently from 89cc24d to ad643ba Compare December 28, 2020 22:50
@rkhachatryan rkhachatryan force-pushed the f20654 branch 2 times, most recently from 72bcb26 to eda1fc4 Compare December 29, 2020 09:31
…kpointITCase

Most of the bugs in UC are revealed with higher back-pressure which is not created with alignment timeout.
This change disables it by default and adds a new test (p=20) with the timeout enabled.
In StreamTaskNetworkInput.prepareSnapshot, internal channelIndex
is inconsistent with CheckpointedInputGate channelIndex.

This change fixes data loss in UnalignedCheckpointITCase.cogroup.
…el state is consumed

In scenarios with multiple inputs (e.g. co-group; not union) one input may receive a
checkpoint barrier while the second input is still restoring state. This (previous)
state is currently not included into the snapshot, which therefore will be incomplete.
… snapshotState

This reduces the number of failures in UnalignedCheckpointITCase.union to ~2/100.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants