Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

fix a bug that ExchangeReceiver can't be canceled #4436

Closed
wants to merge 6 commits into from
Closed

fix a bug that ExchangeReceiver can't be canceled #4436

wants to merge 6 commits into from

Conversation

bestwoody
Copy link
Contributor

@bestwoody bestwoody commented Mar 25, 2022 •
edited
Loading

Signed-off-by: bestwoody [email protected]

What problem does this PR solve?

Issue Number: close #4432

Problem Summary:

What is changed and how it works?

fix a bug that ExchangeReceiver can't be canceled

Check List

Tests

  • Unit test
  • Integration test
  • Manual test (add detailed scripts or steps below)
  • No code

Side effects

  • Performance regression: Consumes more CPU
  • Performance regression: Consumes more Memory
  • Breaking backward compatibility

Documentation

  • Affects user behaviors
  • Contains syntax changes
  • Contains variable changes
  • Contains experimental features
  • Changes MySQL compatibility

Release note

None

@bestwoody
close receiver early
c14a5e9 
Signed-off-by: bestwoody <[email protected]>
@ti-chi-bot
Copy link
Member

[REVIEW NOTIFICATION]

This pull request has not been approved.

To complete the pull request process, please ask the reviewers in the list to review by filling /cc @reviewer in the comment.
After your PR has acquired the required number of LGTMs, you can assign this pull request to the committer in the list by filling /assign @committer in the comment to help you merge this pull request.

The full list of commands accepted by this bot can be found here.

Reviewer can indicate their review by submitting an approval review.
Reviewer can cancel approval by submitting a request changes review.

@ti-chi-bot ti-chi-bot added do-not-merge/needs-linked-issue release-note-none Denotes a PR that doesn't merit a release note. size/M Denotes a PR that changes 30-99 lines, ignoring generated files. labels Mar 25, 2022
@fuzhe1989
Copy link
Contributor

add a failpoint in preprocess? more precisely, before executeQuery.cpp:L426

windtalker
windtalker reviewed Mar 25, 2022
View reviewed changes
dbms/src/Flash/Mpp/ExchangeReceiver.cpp
template <typename RPCContext>
void ExchangeReceiverBase<RPCContext>::close()
{
setState(ExchangeReceiverState::CLOSED);
Copy link
Contributor

Choose a reason for hiding this comment

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

Make sure close and cancel only work if the state is normal?

Copy link
Contributor Author

@bestwoody bestwoody Mar 25, 2022 •
edited
Loading

Choose a reason for hiding this comment

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

I think it's safe to call it multiple times since they are idempotent.

Copy link
Contributor

Choose a reason for hiding this comment

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

But it's weird that you can change the status from CLOSE to CANCELLED, and CANCELLED to CLOSE

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Another annoying point is that ExchangeReceiver also has another ExchangeReceiverState::ERROR state. So any suggestion about state checking?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

But it's weird that you can change the status from CLOSE to CANCELLED, and CANCELLED to CLOSE

The current version of TiFlash may also suffer from the same problem. How about refining it in future?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

check added.

windtalker
windtalker approved these changes Mar 25, 2022
View reviewed changes
Copy link
Contributor

@windtalker windtalker left a comment

Choose a reason for hiding this comment

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

LGTM

fuzhe1989
fuzhe1989 reviewed Mar 25, 2022
View reviewed changes
dbms/src/Flash/Mpp/ExchangeReceiver.cpp
Comment on lines +334 to +337
if (cur_state == ExchangeReceiverState::CANCELED || cur_state == ExchangeReceiverState::CLOSED)
{
return;
}
Copy link
Contributor

Choose a reason for hiding this comment

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

this check should be moved into setState and under protection of mu.

fzhedu
fzhedu reviewed Mar 25, 2022
View reviewed changes
dbms/src/DataStreams/TiRemoteBlockInputStream.h
@@ -241,6 +241,7 @@ class TiRemoteBlockInputStream : public IProfilingBlockInputStream
virtual void readSuffix() override
{
LOG_FMT_DEBUG(log, "finish read {} rows from remote", total_rows);
remote_reader->close();
Copy link
Contributor

Choose a reason for hiding this comment

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

how about local-reader?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

can you explain more? I didn't get it.

Copy link
Contributor

Choose a reason for hiding this comment

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

does local-reader not need to close?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

where islocal-reader?

fzhedu
fzhedu reviewed Mar 25, 2022
View reviewed changes
dbms/src/Flash/Mpp/ExchangeReceiver.cpp
Comment on lines +332 to +340
{
auto cur_state = getState();
if (cur_state == ExchangeReceiverState::CANCELED || cur_state == ExchangeReceiverState::CLOSED)
{
return;
}
setState(ExchangeReceiverState::CLOSED);
msg_channel.finish();
}
Copy link
Contributor

Choose a reason for hiding this comment

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

it seems that those lines are almost similar to L321-327

Copy link
Contributor Author

@bestwoody bestwoody Mar 25, 2022 •
edited
Loading

Choose a reason for hiding this comment

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

refined in another pr, since my local repo is broken, see another PR #4441

fzhedu
fzhedu reviewed Mar 25, 2022
View reviewed changes
dbms/src/Flash/Mpp/MPPTask.cpp
Comment on lines +355 to +356
if (dag_context)
dag_context->cancelAllExchangeReceiver();
Copy link
Contributor

Choose a reason for hiding this comment

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

why before L357? is it better to write errors to the turnnel first?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Refering to chat history in our OOT group, xufei has explained the reason. Receiver needs close first

@bestwoody bestwoody closed this Mar 25, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@fzhedu fzhedu fzhedu left review comments

@fuzhe1989 fuzhe1989 fuzhe1989 left review comments

@windtalker windtalker windtalker approved these changes

Assignees
No one assigned
Labels
needs-cherry-pick-release-6.0 Type: Need cherry pick to release-6.0 release-note-none Denotes a PR that doesn't merit a release note. size/M Denotes a PR that changes 30-99 lines, ignoring generated files.
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

The mvcc snapshot can't be released for a long time
5 participants
@bestwoody @ti-chi-bot @fuzhe1989 @windtalker @fzhedu