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

ddl, util/codec: fix add index OOM and prevent panic in logging (#29925) #30124

Merged
merged 1 commit into from
Nov 24, 2021

Conversation

ti-srebot
Copy link
Contributor

@ti-srebot ti-srebot commented Nov 24, 2021

cherry-pick #29925 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/30124

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

git push git@github.com:ti-srebot/tidb.git pr/30124:release-5.3-d04a1285a54c

What problem does this PR solve?

Issue Number: close #27687

Reproducible step:

drop table if exists t;
create table t (a bigint primary key);
insert into t values (-9078130948872077313);
alter table t add index idx(a);

Problem Summary:

This problem may occur in the following conditions:

  • The table contains an clustered integer primary key.
    • If it is unsigned, at least one large primary key(> MaxInt64) is explicitly inserted.
    • If it is signed, at least one small negative number is explicitly inserted.
  • An adding index job is running.

There is a bug in the current implementation of NewCommonHandle():

tidb/kv/key.go

Lines 237 to 262 in 37bfd57

// NewCommonHandle creates a CommonHandle from a encoded bytes which is encoded by code.EncodeKey.
func NewCommonHandle(encoded []byte) (*CommonHandle, error) {
ch := &CommonHandle{encoded: encoded}
if len(encoded) < 9 {
padded := make([]byte, 9)
copy(padded, encoded)
ch.encoded = padded
}
remain := encoded
endOff := uint16(0)
for len(remain) > 0 {
if remain[0] == 0 {
// padded data
break
}
var err error
var col []byte
col, remain, err = codec.CutOne(remain)
if err != nil {
return nil, err
}
endOff += uint16(len(col))
ch.colEndOffsets = append(ch.colEndOffsets, endOff)
}
return ch, nil
}

If the argument bytes have a prefix like 0x8203, CutOne() cuts nothing so that the remain is unchanged in the loop. This dead loop may cause TiDB OOM quickly because it endlessly append() colEndOffsets.

What is changed and how it works?

  1. The function peek() in the decoding process does not handle the invalid key. It is possible to run into a dead loop and makes TiDB OOM.
  2. The function tryDecodeHandleToString() is only used for logging, it should be wrapped with recover() so that the critical operation is not interrupted.

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

Fix an issue that adding index panics by chance.

Signed-off-by: ti-srebot <ti-srebot@pingcap.com>
@ti-chi-bot
Copy link
Member

ti-chi-bot commented Nov 24, 2021

[REVIEW NOTIFICATION]

This pull request has been approved by:

  • wjhuang2016
  • xiongjiwei

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-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 Nov 24, 2021
@ti-srebot ti-srebot added affects-5.0 This bug affects 5.0.x versions. affects-5.1 This bug affects 5.1.x versions. affects-5.2 This bug affects 5.2.x versions. affects-5.3 This bug affects 5.3.x versions. size/S Denotes a PR that changes 10-29 lines, ignoring generated files. type/5.3-cherry-pick labels Nov 24, 2021
@ti-srebot
Copy link
Contributor Author

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

@hawkingrei
Copy link
Member

/lgtm

@ti-chi-bot
Copy link
Member

@hawkingrei: Thanks for your review. The bot only counts approvals from reviewers and higher roles in list, but you're still welcome to leave your comments.

In response to this:

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 ti-community-infra/tichi repository.

@ti-chi-bot ti-chi-bot added the status/LGT1 Indicates that a PR has LGTM 1. label Nov 24, 2021
@ti-chi-bot ti-chi-bot added status/LGT2 Indicates that a PR has LGTM 2. and removed status/LGT1 Indicates that a PR has LGTM 1. labels Nov 24, 2021
@zhouqiang-cl zhouqiang-cl added the cherry-pick-approved Cherry pick PR approved by release team. label Nov 24, 2021
@zhouqiang-cl
Copy link
Contributor

/merge

@ti-chi-bot
Copy link
Member

This pull request has been accepted and is ready to merge.

Commit hash: b44fccd

@ti-chi-bot ti-chi-bot added the status/can-merge Indicates a PR has been approved by a committer. label Nov 24, 2021
@ti-chi-bot ti-chi-bot merged commit 4a1b2e9 into pingcap:release-5.3 Nov 24, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
affects-5.0 This bug affects 5.0.x versions. affects-5.1 This bug affects 5.1.x versions. affects-5.2 This bug affects 5.2.x versions. affects-5.3 This bug affects 5.3.x versions. cherry-pick-approved Cherry pick PR approved by release team. release-note Denotes a PR that will be considered when it comes time to generate release notes. size/S Denotes a PR that changes 10-29 lines, ignoring generated files. status/can-merge Indicates a PR has been approved by a committer. status/LGT2 Indicates that a PR has LGTM 2. type/5.3-cherry-pick
Projects
None yet
Development

Successfully merging this pull request may close these issues.

7 participants