sink(ticdc): check virtual columns in column dispatcher (#12254) #12324
Conversation
Signed-off-by: ti-chi-bot <[email protected]>
ti-chi-bot
added
do-not-merge/hold
|
This cherry pick PR is for a release branch and has not yet been approved by triage owners. To merge this cherry pick:
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-sigs/prow repository. |
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
@wk989898 This PR has conflicts, I have hold it. |
|
@ti-chi-bot: ## If you want to know how to resolve it, please read the guide in TiDB Dev Guide. 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
bot
added
size/XXL
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Summary of Changes
Hello @ti-chi-bot, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!
This pull request addresses a critical issue where using virtual generated columns in the column dispatcher could lead to panics. It introduces robust validation to prevent such configurations and expands the dispatching capabilities by allowing events to be routed based on specific columns, ensuring data integrity and system stability.
Highlights
- Panic Prevention: Prevented panics in the column dispatcher by explicitly checking for and rejecting virtual generated columns, returning an ErrDispatcherFailed error if such columns are specified.
- New Dispatcher Capability: Introduced a new ColumnsDispatcher for message queue sinks, allowing event dispatching based on user-defined columns.
- Table Information Enhancement: Enhanced the TableInfo model with new methods for retrieving column offsets by name and improved index information access, crucial for robust column-based dispatching.
Using Gemini Code Assist
The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.
Invoking Gemini
You can request assistance from Gemini at any point in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.
| Feature | Command | Description |
|---|---|---|
| Code Review | /gemini review |
Performs a code review for the current pull request in its current state. |
| Pull Request Summary | /gemini summary |
Provides a summary of the current pull request in its current state. |
| Comment | @gemini-code-assist | Responds in comments when explicitly tagged, both in issue comments and review comments. |
| Help | /gemini help |
Displays a list of available commands. |
Customization
To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.
Limitations & Feedback
Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.
You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.
Footnotes
-
Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩
There was a problem hiding this comment.
Code Review
This pull request introduces a check for virtual columns in the column dispatcher to prevent a panic. While the core logic for identifying virtual columns is sound, the pull request appears to be an incomplete or conflicted cherry-pick. There are multiple merge conflicts across several files that will prevent the code from compiling. Additionally, there's a critical interface mismatch with the new ColumnsDispatcher and a performance issue in the OffsetsByNames function. These issues need to be addressed before this PR can be merged.
| <<<<<<< HEAD | ||
| "github.com/pingcap/tidb/parser/model" | ||
| "github.com/pingcap/tidb/parser/mysql" | ||
| "github.com/pingcap/tidb/parser/types" | ||
| "github.com/pingcap/tidb/table/tables" | ||
| "github.com/pingcap/tidb/util/rowcodec" | ||
| ======= | ||
| "github.com/pingcap/log" | ||
| "github.com/pingcap/tidb/pkg/meta/model" | ||
| "github.com/pingcap/tidb/pkg/parser/mysql" | ||
| "github.com/pingcap/tidb/pkg/parser/types" | ||
| "github.com/pingcap/tidb/pkg/table/tables" | ||
| "github.com/pingcap/tidb/pkg/util/rowcodec" | ||
| "github.com/pingcap/tiflow/pkg/errors" | ||
| "go.uber.org/zap" | ||
| >>>>>>> 35ff0a22ec (sink(ticdc): check virtual columns in column dispatcher (#12254)) |
There was a problem hiding this comment.
This file contains merge conflict markers (here and at lines 341-425) which will prevent the code from compiling. Please resolve them.
| <<<<<<< HEAD | ||
| pmodel "github.com/pingcap/tidb/parser/model" | ||
| timodel "github.com/pingcap/tidb/parser/model" | ||
| "github.com/pingcap/tidb/parser/mysql" | ||
| parser_types "github.com/pingcap/tidb/parser/types" | ||
| ======= | ||
| timodel "github.com/pingcap/tidb/pkg/meta/model" | ||
| pmodel "github.com/pingcap/tidb/pkg/parser/ast" | ||
| "github.com/pingcap/tidb/pkg/parser/charset" | ||
| "github.com/pingcap/tidb/pkg/parser/mysql" | ||
| parser_types "github.com/pingcap/tidb/pkg/parser/types" | ||
| "github.com/pingcap/tiflow/pkg/errors" | ||
| >>>>>>> 35ff0a22ec (sink(ticdc): check virtual columns in column dispatcher (#12254)) |
There was a problem hiding this comment.
This file contains merge conflict markers (here and at lines 284-529) which will prevent the code from compiling. Please resolve them.
| func (r *ColumnsDispatcher) DispatchRowChangedEvent(row *model.RowChangedEvent, partitionNum int32) (int32, string, error) { | ||
| r.lock.Lock() | ||
| defer r.lock.Unlock() | ||
| r.hasher.Reset() | ||
|
|
||
| r.hasher.Write([]byte(row.TableInfo.GetSchemaName()), []byte(row.TableInfo.GetTableName())) | ||
|
|
||
| dispatchCols := row.Columns | ||
| if len(dispatchCols) == 0 { | ||
| dispatchCols = row.PreColumns | ||
| } | ||
|
|
||
| offsets, err := row.TableInfo.OffsetsByNames(r.Columns) | ||
| if err != nil { | ||
| log.Error("dispatch event failed", zap.Error(err)) | ||
| return 0, "", err | ||
| } | ||
|
|
||
| for idx := 0; idx < len(r.Columns); idx++ { | ||
| col := dispatchCols[offsets[idx]] | ||
| if col == nil { | ||
| continue | ||
| } | ||
| colInfo := row.TableInfo.ForceGetColumnInfo(col.ColumnID) | ||
| r.hasher.Write([]byte(colInfo.Name.O), []byte(model.ColumnValueString(col.Value))) | ||
| } | ||
|
|
||
| sum32 := r.hasher.Sum32() | ||
| return int32(sum32 % uint32(partitionNum)), strconv.FormatInt(int64(sum32), 10), nil | ||
| } |
There was a problem hiding this comment.
The signature of DispatchRowChangedEvent (int32, string, error) does not match the partition.Dispatcher interface, which, based on the provided context files, expects a signature of (row *model.RowChangedEvent, partitionNum int32) int32. This will cause a compilation error as ColumnsDispatcher will not correctly implement the interface. The Dispatcher interface and all its implementations should be updated consistently to handle error propagation and the additional string return value, or this implementation must be changed to match the existing interface.
| <<<<<<< HEAD:cdc/sink/mq/dispatcher/event_router.go | ||
| // GetDLLDispatchRuleByProtocol returns the DDL | ||
| // distribution rule according to the protocol. | ||
| func (s *EventRouter) GetDLLDispatchRuleByProtocol( | ||
| protocol config.Protocol, | ||
| ) DDLDispatchRule { | ||
| if protocol == config.ProtocolCanal || protocol == config.ProtocolCanalJSON { | ||
| return PartitionZero | ||
| ======= | ||
| // GetPartitionDispatcher returns the partition dispatcher for a specific table. | ||
| func (s *EventRouter) GetPartitionDispatcher(schema, table string) partition.Dispatcher { | ||
| _, partitionDispatcher := s.matchDispatcher(schema, table) | ||
| return partitionDispatcher | ||
| } | ||
|
|
||
| // VerifyTables return error if any one table route rule is invalid. | ||
| func (s *EventRouter) VerifyTables(infos []*model.TableInfo) error { | ||
| for _, table := range infos { | ||
| _, partitionDispatcher := s.matchDispatcher(table.TableName.Schema, table.TableName.Table) | ||
| switch v := partitionDispatcher.(type) { | ||
| case *partition.IndexValueDispatcher: | ||
| if v.IndexName != "" { | ||
| index := table.GetIndex(v.IndexName) | ||
| if index == nil { | ||
| return cerror.ErrDispatcherFailed.GenWithStack( | ||
| "index not found when verify the table, table: %v, index: %s", table.TableName, v.IndexName) | ||
| } | ||
| // only allow the unique index to be set. | ||
| // For the non-unique index, if any column belongs to the index is updated, | ||
| // the event is not split, it may cause incorrect data consumption. | ||
| if !index.Unique { | ||
| return cerror.ErrDispatcherFailed.GenWithStack( | ||
| "index is not unique when verify the table, table: %v, index: %s", table.TableName, v.IndexName) | ||
| } | ||
| } | ||
| case *partition.ColumnsDispatcher: | ||
| _, err := table.OffsetsByNames(v.Columns) | ||
| if err != nil { | ||
| return err | ||
| } | ||
| default: | ||
| } | ||
| >>>>>>> 35ff0a22ec (sink(ticdc): check virtual columns in column dispatcher (#12254)):cdc/sink/dmlsink/mq/dispatcher/event_router.go |
There was a problem hiding this comment.
This file contains merge conflict markers, which will prevent the code from compiling. Please resolve them.
| func (ti *TableInfo) OffsetsByNames(names []string) ([]int, error) { | ||
| // todo: optimize it | ||
| columnOffsets := make(map[string]int) | ||
| virtualGeneratedColumn := make(map[string]struct{}) | ||
| for _, col := range ti.Columns { | ||
| if col != nil { | ||
| if IsColCDCVisible(col) { | ||
| columnOffsets[col.Name.L] = ti.columnsOffset[col.ID] | ||
| } else { | ||
| virtualGeneratedColumn[col.Name.L] = struct{}{} | ||
| } | ||
| } | ||
| } | ||
|
|
||
| result := make([]int, 0, len(names)) | ||
| for _, col := range names { | ||
| name := strings.ToLower(col) | ||
| if _, ok := virtualGeneratedColumn[name]; ok { | ||
| return nil, errors.ErrDispatcherFailed.GenWithStack( | ||
| "found virtual generated columns when dispatch event, table: %v, columns: %v column: %v", ti.GetTableName(), names, name) | ||
| } | ||
| offset, ok := columnOffsets[name] | ||
| if !ok { | ||
| return nil, errors.ErrDispatcherFailed.GenWithStack( | ||
| "columns not found when dispatch event, table: %v, columns: %v, column: %v", ti.GetTableName(), names, name) | ||
| } | ||
| result = append(result, offset) | ||
| } | ||
|
|
||
| return result, nil | ||
| } |
There was a problem hiding this comment.
The // todo: optimize it comment is valid. This function is called on the hot path for event dispatching, and it currently rebuilds columnOffsets and virtualGeneratedColumn maps on every call. This is inefficient.
For better performance, these maps should be pre-computed and cached in the TableInfo struct when it's initialized in WrapTableInfo. This would avoid repeated computations and significantly improve performance, especially for tables with many columns.
|
@ti-chi-bot: The following tests failed, say
Full PR test history. Your PR dashboard. 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-sigs/prow repository. I understand the commands that are listed here. |
2 participants
This is an automated cherry-pick of #12254
What problem does this PR solve?
Issue Number: close #12241
What is changed and how it works?
Since column values don't include virtual generated columns, we also filter the virtual generated columns in the column dispatcher.
The column dispatcher now explicitly checks for and prevents the use of virtual generated columns, returning an ErrDispatcherFailed error if such columns are specified. This prevents potential panics when configuring the dispatcher.
Check List
Tests
Questions
Will it cause performance regression or break compatibility?
Do you need to update user documentation, design documentation or monitoring documentation?
Release note