-
Notifications
You must be signed in to change notification settings - Fork 409
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.
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
fix bug that mpp_fail test cost too much time / Fix test clustered index test cases #1591
Conversation
Will this PR fix the problem that Related to #1586 (comment) |
/run-all-tests |
/run-all-tests |
Mark that this PR also fixes the test cases of clustered index. pingcap/tidb#23270 |
cherry pick to release-5.0 in PR #1600 |
Signed-off-by: ti-srebot <[email protected]> Co-authored-by: xufei <[email protected]>
Yes, but this pr still have a small chances that TiFlash server will crash when meet some mpp errors, #1577 fix it. |
What problem does this PR solve?
Issue Number: close #xxx
Problem Summary:
After pingcap/tidb#23056, fullstack test
mpp_fail
takes too much time(more than 10 minutes), which is expected.What is changed and how it works?
Proposal: xxx
What's Changed:
How it Works:
The root cause is after pingcap/tidb#23056, TiDB will send
kill
command to kill all the mpp tasks for a specified query, and currently implementation of cancelling a mpp task will callwriteErrToAllTunnel
to write error message to other mpp tasks, since all the mpp tasks are canceled at the same time,writeErrToAllTunnel
might hangs for a little time(write error message until tunnel is timeout), so the test execution time is much more than expected. In this pr, when canceling a mpp task, it just callcloseAllTunnel
to close the tunnel, which only write error message to other mpp task if the tunnel is already connected.Related changes
pingcap/docs
/pingcap/docs-cn
:Check List
Tests
Side effects
Release note