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.

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

table/tables: fix buildListColumnsPruner issue in the list partition (#32621) #32771

Closed

Conversation

ti-srebot
Copy link
Contributor

@ti-srebot ti-srebot commented Mar 3, 2022

cherry-pick #32621 to release-5.3
You can switch your code base to this Pull Request by using git-extras:

# In tidb repo:
git pr https://github.com/pingcap/tidb/pull/32771

After apply modifications, you can push your change to this PR via:

git push [email protected]:ti-srebot/tidb.git pr/32771:release-5.3-0e4084c02d69

What problem does this PR solve?

Issue Number: close #32416

Problem Summary:
This also affects the master. This problem can also occur in the following scenario:

  1. Set the new collation configuration item to false and run the SQL statement above (create list partition table).
  2. Restart the TiDB and execute the select statement.
    The above scenario is analyzed in detail:

In step 1

In the bootstrap phase, new collation is true
After bootstrap ends, new collation is false.

In step 2

  1. In the bootstrap phase, new collation is true. ForListColumnPruning builds the hashMap(generate the partition hash key, by calling EncodeKey) in TableFromMeta
  2. After bootstrap ends, new collation is false. The generated partition hash key(by calling EncodeKey) could not be found in the hashMap.

What is changed and how it works?

Instead of creating the hashMap in the TableFromMeta phase( In the bootstrap phase), we create the hashMap on a delayed basis (that is, when a partition needs to be located).
In this way, new_collations_enabled_on_first_bootstrap does not have inconsistency problems.

Check List

Tests

  • Unit test
  • Integration test
  • Manual test (add detailed scripts or steps below)
  1. Set the new_collations_enabled_on_first_bootstrap item to false and run the SQL statements as follows:
set @@session.tidb_enable_list_partition = ON;
set @@tidb_partition_prune_mode = 'dynamic';
drop table if exists github_events;
CREATE TABLE `github_events` (
`id` bigint(20) DEFAULT NULL,
`type` varchar(255) COLLATE utf8mb4_unicode_ci DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci
PARTITION BY LIST COLUMNS(`type`)
(PARTITION `push_event` VALUES IN ("PushEvent"),
PARTITION `watch_event` VALUES IN ("WatchEvent")
);
  1. Restart TiDB with new_collations_enabled_on_first_bootstrap = false and execute the following statements:
tidb> set @@session.tidb_enable_list_partition = ON;
Query OK, 0 rows affected (0.00 sec)

tidb> set @@tidb_partition_prune_mode = 'dynamic';
Query OK, 0 rows affected (0.00 sec)

tidb> desc select id,type from github_events  where  type = 'WatchEvent';
+-------------------------+----------+-----------+-----------------------+-------------------------------------------+
| id                      | estRows  | task      | access object         | operator info                             |
+-------------------------+----------+-----------+-----------------------+-------------------------------------------+
| TableReader_7           | 10.00    | root      | partition:watch_event | data:Selection_6                          |
| └─Selection_6           | 10.00    | cop[tikv] |                       | eq(test.github_events.type, "WatchEvent") |
|   └─TableFullScan_5     | 10000.00 | cop[tikv] | table:github_events   | keep order:false, stats:pseudo            |
+-------------------------+----------+-----------+-----------------------+-------------------------------------------+
3 rows in set (0.00 sec)

The above scenario is analyzed in detail:

  • 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

Fixed an issue where partition table pruning might not work after server restart (for list partition table)

@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
Copy link
Member

@ti-srebot: This cherry pick PR is for a release branch and has not yet been approved by release team.
Adding the do-not-merge/cherry-pick-not-approved label.

To merge this cherry pick, it must first be approved by the collaborators.

AFTER it has been approved by collaborators, please ping the release team in a comment to request a cherry pick review.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

@ti-srebot
Copy link
Contributor Author

/run-all-tests

@ti-chi-bot ti-chi-bot added do-not-merge/release-note-label-needed Indicates that a PR should not merge because it's missing one of the release note labels. release-note Denotes a PR that will be considered when it comes time to generate release notes. and removed do-not-merge/release-note-label-needed Indicates that a PR should not merge because it's missing one of the release note labels. labels Mar 3, 2022
@ti-srebot ti-srebot added sig/sql-infra SIG: SQL Infra size/M Denotes a PR that changes 30-99 lines, ignoring generated files. type/5.3-cherry-pick type/bugfix This PR fixes a bug. labels Mar 3, 2022
@ti-srebot ti-srebot added this to the v5.3.1 milestone Mar 3, 2022
@ti-srebot
Copy link
Contributor Author

@zimulala you're already a collaborator in bot's repo.

@zimulala
Copy link
Contributor

zimulala commented Mar 3, 2022

List partitions are experimental and this issue is not critical, so don't cherry pick

@zimulala zimulala closed this Mar 3, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
do-not-merge/cherry-pick-not-approved release-note Denotes a PR that will be considered when it comes time to generate release notes. sig/sql-infra SIG: SQL Infra size/M Denotes a PR that changes 30-99 lines, ignoring generated files. type/bugfix This PR fixes a bug. type/5.3-cherry-pick
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants