package coprocessor

Get desktop application:
View/edit binary Protocol Buffers messages

Used as request type in: tikvpb.Tikv.BatchCoprocessor

• optional kvrpcpb.Context context = 1

• int64 tp = 2

• bytes data = 3

• repeated RegionInfo regions = 4

• uint64 start_ts = 5

• int64 schema_ver = 6

  Any schema-ful storage to validate schema correctness if necessary.

• repeated TableRegions table_regions = 7

  Used for partition table scan

• string log_id = 8

• uint64 connection_id = 9

  This is the session id between a client and tidb

• string connection_alias = 10

  This is the session alias between a client and tidb

• repeated TableShardInfos table_shard_infos = 11

  Shard infos for FTS index, used by TiFlash reading TiCI.

Used as response type in: tikvpb.Tikv.BatchCoprocessor

• bytes data = 1

• string other_error = 2

• optional kvrpcpb.ExecDetails exec_details = 3

• repeated metapb.Region retry_regions = 4

• repeated ShardInfo retry_shards = 5

Used as request type in: tikvpb.Tikv.DelegateCoprocessor

• optional kvrpcpb.Context context = 1

• uint64 start_ts = 2

• repeated KeyRange ranges = 3

• uint64 mem_table_sequence = 4

  Used for avoid redundant mem-table copying. If the sequence is the same, tikv-server will not return the mem-table.

• uint64 snapshot_sequence = 5

  Used for avoid redundant snapshot copying. If the sequence is the same, tikv-server will not return the snapshot.

Used as response type in: tikvpb.Tikv.DelegateCoprocessor

• bytes mem_table_data = 1

• bytes snapshot = 2

• optional errorpb.Error region_error = 3

• optional kvrpcpb.LockInfo locked = 4

• string other_error = 5

• uint64 mem_table_sequence = 6

  Used for avoid redundant mem-table copying.

[start, end)

Used in: DelegateRequest, RegionInfo, Request, Response, ShardInfo, StoreBatchTask, VersionedKeyRange

• bytes start = 1

• bytes end = 2

Used in: BatchRequest, TableRegions, disaggregated.EstablishDisaggTaskRequest, mpp.DispatchTaskRequest

• uint64 region_id = 1

• optional metapb.RegionEpoch region_epoch = 2

• repeated KeyRange ranges = 3

Used as request type in: tikvpb.Tikv.Coprocessor, tikvpb.Tikv.CoprocessorStream, tikvpb.VersionedKv.VersionedCoprocessor

Used as field type in: tikvpb.BatchCommandsRequest.Request

• optional kvrpcpb.Context context = 1

• int64 tp = 2

• bytes data = 3

• uint64 start_ts = 7

• repeated KeyRange ranges = 4

• bool is_cache_enabled = 5

  If cache is enabled, TiKV returns cache hit instead of data if its last version matches this `cache_if_match_version`.

• uint64 cache_if_match_version = 6

• int64 schema_ver = 8

  Any schema-ful storage to validate schema correctness if necessary.

• bool is_trace_enabled = 9

• uint64 paging_size = 10

  paging_size is 0 when it's disabled, otherwise, it should be a positive number.

• repeated StoreBatchTask tasks = 11

  tasks stores the batched coprocessor tasks sent to the same tikv store.

• uint64 connection_id = 12

  This is the session id between a client and tidb

• string connection_alias = 13

  This is the session alias between a client and tidb

• repeated TableShardInfos table_shard_infos = 14

  Shard infos for FTS index, used by TiFlash reading TiCI.

• repeated VersionedKeyRange versioned_ranges = 15

  Versioned point ranges for TiCI lookup. When `versioned_ranges` is non-empty, all `versioned_ranges[i].range` must be point range.

• uint64 max_keys_read = 16

  max_keys_read is 0 when disabled, otherwise limits the number of storage engine keys scanned per coprocessor task. It is a per-task hard ceiling applied uniformly to any coprocessor request, used to bound worst-case scan amplification regardless of how the request is paginated.

• uint64 paging_size_bytes = 17

  paging_size_bytes is 0 when disabled, otherwise it should be a positive number. When set, within a paged coprocessor request, TiKV stops scanning the current page once accumulated scanned bytes reach this limit and returns the page boundary so the next page can resume from there. Unlike max_keys_read, which is a per-task hard ceiling on keys scanned, paging_size_bytes is an independent rate-control dimension intended for Resource Control: it bounds the byte volume of a single page so that RU pre-charging in PD's resource controller can be performed at byte granularity, complementing the existing row-count-based paging (paging_size). The two fields target different scenarios and may be set independently.

• bool allow_batch_task_data_merge = 18

  Signals that the client can handle results of the batched tasks in `tasks` being merged into `Response.data` instead of each task returning its result in its own `StoreBatchTaskResponse.data`. Merging lets the store combine mergeable payloads, such as per-region ANALYZE sampling statistics, into one result, reducing the response size and the client's decode and merge work. See the batched ANALYZE design for the motivating use case: https://github.com/pingcap/tidb/pull/70355. For every merged task, the store still adds a `StoreBatchTaskResponse` with `data_merged_into_response` set. The store may return some or all task results separately even when this field is set, for example when a task fails, when its result is not mergeable, or when the store predates this field, so the client must handle both merged and per-task results. This field only negotiates the response shape. It does not affect task scheduling and may be set independently of `execute_batch_tasks_serially`.

• bool execute_batch_tasks_serially = 19

  When true, the store executes the primary task and all tasks in `tasks` one at a time, so at most one task of this request is running at any moment. Task execution order is not guaranteed. Clients that already bound their per-store concurrency by the number of in-flight requests, such as batched ANALYZE, set this so that batching more tasks into one request does not multiply that concurrency. The trade-off is a longer-lived request. This field only affects task scheduling and may be set independently of `allow_batch_task_data_merge`.

Used as response type in: tikvpb.Tikv.Coprocessor, tikvpb.Tikv.CoprocessorStream, tikvpb.VersionedKv.VersionedCoprocessor

Used as field type in: tikvpb.BatchCommandsResponse.Response

• bytes data = 1

• optional errorpb.Error region_error = 2

• optional kvrpcpb.LockInfo locked = 3

• string other_error = 4

• optional KeyRange range = 5

• optional kvrpcpb.ExecDetails exec_details = 6

  This field is always filled for compatibility consideration. However newer TiDB should respect `exec_details_v2` field instead.

• optional kvrpcpb.ExecDetailsV2 exec_details_v2 = 11

  This field is provided in later versions, containing more detailed information.

• bool is_cache_hit = 7

• uint64 cache_last_version = 8

• bool can_be_cached = 9

• uint64 latest_buckets_version = 12

  Contains the latest buckets version of the region. Clients should query PD to update buckets in cache if its is stale.

• repeated StoreBatchTaskResponse batch_responses = 13

  StoreBatchTaskResponse is the collection of batch task responses.

Used in: BatchResponse, TableShardInfos, TiCIEstimateCountRequest

• uint64 shard_id = 1

• uint64 shard_epoch = 2

• repeated KeyRange ranges = 3

Used in: Request

• uint64 region_id = 1

• optional metapb.RegionEpoch region_epoch = 2

• optional metapb.Peer peer = 3

• repeated KeyRange ranges = 4

• uint64 task_id = 5

• repeated VersionedKeyRange versioned_ranges = 6

  Versioned point ranges for TiCI lookup. When `versioned_ranges` is non-empty, all `versioned_ranges[i].range` must be point range.

• uint64 buckets_version = 7

  Bucket metadata version used to validate this child task.

Used in: Response

• bytes data = 1

• optional errorpb.Error region_error = 2

• optional kvrpcpb.LockInfo locked = 3

• string other_error = 4

• uint64 task_id = 5

• optional kvrpcpb.ExecDetailsV2 exec_details_v2 = 6

• bool data_merged_into_response = 7

  Indicates that this task's result was merged into the enclosing `Response.data`, so this message's `data` is empty. The store sets this field only when the client enables `Request.allow_batch_task_data_merge`, and only for a task that succeeded: a failed or non-mergeable task is returned as a normal per-task response with this field unset. Merging is atomic. If the store fails while merging, it returns no partial merged data and no response with this field set, so the client can safely retry the entire batch without losing or double-counting any task's result. This message still identifies the merged task by `task_id` and carries its execution details.

Used in: BatchRequest, disaggregated.EstablishDisaggTaskRequest, mpp.DispatchTaskRequest

• int64 physical_table_id = 1

• repeated RegionInfo regions = 2

Used in: BatchRequest, Request, mpp.DispatchTaskRequest

• string executor_id = 1

  The executor ID is used to identify the tici executor.

• repeated ShardInfo shard_infos = 2

  The shard_infos contains the shard information for each tici executor.

Used as request type in: tikvpb.Tikv.GetEstimateTiCICount

• optional kvrpcpb.Context context = 1

• uint64 start_ts = 2

• int64 table_id = 3

• int64 index_id = 4

• bytes fts_query_info = 5

• string time_zone_name = 6

• int64 time_zone_offset = 7

• repeated ShardInfo shard_infos = 8

Used as response type in: tikvpb.Tikv.GetEstimateTiCICount

• uint64 est_count = 1

• string other_error = 2

KeyRange with an attached read_ts (version). It is used by TiCI versioned lookup. Callers must ensure `range` is a point range.

Used in: Request, StoreBatchTask

• optional KeyRange range = 1

• uint64 read_ts = 2