package backup

Get desktop application:
View/edit binary Protocol Buffers messages

• rpc backup (BackupRequest, stream BackupResponse)

  brpb.proto:621

  message BackupRequest

  brpb.proto:334

  • uint64 cluster_id = 1

  • bytes start_key = 2

  • bytes end_key = 3

  • uint64 start_version = 4

  • uint64 end_version = 5

  • uint64 rate_limit = 7

    The I/O rate limit for backup request.

  • uint32 concurrency = 8

    The concurrency for executing the backup request in every tikv node.

  • optional StorageBackend storage_backend = 9

  • bool is_raw_kv = 10

    If raw kv mode is enabled, `start_version` and `end_version` will be ignored, and `cf` specifies which cf to backup.

  • string cf = 11

  • CompressionType compression_type = 12

    algorithm used for compress sst files

  • int32 compression_level = 13

    sst compression level, some algorithms support negative compression levels

  • optional CipherInfo cipher_info = 14

    The cipher_info is Used to encrypt sst

  • kvrpcpb.APIVersion dst_api_version = 15

    dst_api_version indicates the key-value encoding version used by the generated SST file. Accepted values: 1. "v1": the generated SST files are encoded with api-v1, can be restored to TiKV clusters whose api version is set to v1. 2. "v2": the generated SST files are encoded with api-v2, can be restored to TiKV clusters whose api version is set to v2.

  • repeated kvrpcpb.KeyRange sub_ranges = 16

    with checkpoint, some subintervals of the range have been backed up and recorded. only the remaining sub ranges of the range need to be backed up this time.

  • repeated SortedSubRanges sorted_sub_ranges_groups = 21

    only ranges in the same groups can be merged into one SST file. the ranges in each groups are sorted.

  • bool replica_read = 17

    replica_read indicates whether to use replica read for backup. If it is false, the backup will only read data from leader. If it is true, the backup will read data from both leader and follower.

  • BackupMode mode = 18

  • string unique_id = 19

    unique_id represents the handle of this backup. after we implement file-copy backup. we need generate some internal states during the whole backup precedure. this unique id is help to find the state effictively.

  • optional kvrpcpb.Context context = 20

  message BackupResponse

  brpb.proto:575

  • optional Error error = 1

  • bytes start_key = 2

  • bytes end_key = 3

  • repeated File files = 4

  • kvrpcpb.APIVersion api_version = 5

    API version implies the encode of the key and value.

• rpc CheckPendingAdminOp (CheckAdminRequest, stream CheckAdminResponse)

  brpb.proto:626

  CheckPendingAdminOp used for snapshot backup. before we start snapshot for a TiKV. we need stop all schedule first and make sure all in-flight schedule has finished. this rpc check all pending conf change for leader.

  message CheckAdminRequest

  brpb.proto:611

  (message has no fields)

  message CheckAdminResponse

  brpb.proto:614

  • optional Error error = 1

  • optional metapb.Region region = 2

  • bool has_pending_admin = 3

• rpc PrepareSnapshotBackup (stream PrepareSnapshotBackupRequest, stream PrepareSnapshotBackupResponse)

  brpb.proto:630

  PrepareSnapshotBackup is an advanced version of preparing snapshot backup. Check the defination of `PrepareSnapshotBackupRequest` for more.

  message PrepareSnapshotBackupRequest

  brpb.proto:35

  • PrepareSnapshotBackupRequestType ty = 1

    The type of the request.

  • repeated metapb.Region regions = 2

    The regions related to the request. If the request is "UpdateLease" or "Finish", it should be empty.

  • uint64 lease_in_seconds = 3

    The lease duration for "UpdateLease" request.

  message PrepareSnapshotBackupResponse

  brpb.proto:52

  • PrepareSnapshotBackupEventType ty = 1

    The type of the event.

  • optional metapb.Region region = 2

    The region bind to the type. For "WaitApplyDone".

  • optional errorpb.Error error = 3

    The error of the execution.

  • bool last_lease_is_valid = 4

    Whether the last lease is valid. For "UpdateLease" and "Finish".

• rpc prepare (PrepareRequest, PrepareResponse)

  brpb.proto:634

  prepare is used for file-copy backup. before we start the backup for a TiKV. we need invoke this function to generate the SST files map. or we get nothing to backup.

  message PrepareRequest

  brpb.proto:596

  • bool save_to_storage = 1

    whether save state to the storage.

  message PrepareResponse

  brpb.proto:601

  • optional Error error = 1

  • string unique_id = 2

    unique_id represents the unique handle of the whole backup predecure. if unique_id = 0 means prepare failed. if unique_id > 0 means prepare success and all states saved with this unique info.

  • uint64 collect_file_count = 3

  • uint64 collect_file_size = 4

• rpc cleanup (CleanupRequest, CleanupResponse)

  brpb.proto:638

  cleanup used for file-copy backup. after we finish the backup for a TiKV. we need clean some internel state. e.g. checkpoint, SST File maps

  message CleanupRequest

  brpb.proto:585

  • string unique_id = 1

    unique_id represents the unique handle of the whole backup predecure. it generated in prepare request and corrosponed to one specific backup.

  message CleanupResponse

  brpb.proto:591

  • optional Error error = 1

  • bool success = 2

ExternalStorage is a service for using a cloud backend from StorageBackend to store files. This can be used to backup and restore SST files.

• rpc restore (ExternalStorageRestoreRequest, ExternalStorageRestoreResponse)

  brpb.proto:645

  Restore to a file

  message ExternalStorageRestoreRequest

  brpb.proto:650

  • optional StorageBackend storage_backend = 1

  • string object_name = 2

  • string restore_name = 3

  • uint64 content_length = 4

  message ExternalStorageRestoreResponse

  brpb.proto:657

  (message has no fields)

• rpc save (ExternalStorageSaveRequest, ExternalStorageSaveResponse)

  brpb.proto:647

  Save a file to storage

  message ExternalStorageSaveRequest

  brpb.proto:660

  • optional StorageBackend storage_backend = 1

  • string object_name = 2

  • uint64 content_length = 3

  message ExternalStorageSaveResponse

  brpb.proto:666

  (message has no fields)

AzureBlobStorage storage backend saves files into azure blob storage.

Used in: StorageBackend

• string endpoint = 1

• string bucket = 2

  Alias: container

• string prefix = 3

  Notice: prefix starts without `/`, otherwise the first directory's name is empty.

• string storage_class = 4

  Alias: access_tier. See https://docs.microsoft.com/en-us/azure/storage/blobs/access-tiers-overview

• string account_name = 5

  if empty, try to read account_name from the node's environment variable $AZURE_STORAGE_ACCOUNT.

• string shared_key = 6

  There are 3 kinds of credentials, and the priority order is `SAS > Shared key > Azure AD (env) > Shared key (env)`. 1. Use shared key to access the azure blob If the node's environment variables($AZURE_CLIENT_ID, $AZURE_TENANT_ID, $AZURE_CLIENT_SECRET) exist, prefer to use token to access the azure blob. See https://learn.microsoft.com/en-us/rest/api/storageservices/authorize-with-shared-key Otherwise, if empty, try to read shared key from the node's environment variable $AZURE_STORAGE_KEY.

• string access_sig = 8

  2. Use Azure AD (Azure Active Directory) to access the azure blob See https://learn.microsoft.com/en-us/rest/api/storageservices/authorize-with-azure-active-directory The Azure AD would generate the token, which tasks some time. So it is not recommanded to generate the token in each request. // AzureActiveDirectory azure_ad = #; 3. Use SAS (shared access signature) See https://learn.microsoft.com/en-us/rest/api/storageservices/delegate-access-with-shared-access-signature

• string encryption_scope = 9

  Server Side Encryption, 2 types in total: 1. Specify an encryption scope for uploaded blobs. See https://learn.microsoft.com/en-us/azure/storage/blobs/encryption-scope-manage?tabs=powershell#upload-a-blob-with-an-encryption-scope

• optional AzureCustomerKey encryption_key = 10

  2. Provide an encryption key on a request to blob storage. See https://learn.microsoft.com/en-us/azure/storage/blobs/encryption-customer-provided-keys

The encryption algorithm must be AES256.

Used in: AzureBlobStorage

• string encryption_key = 1

  A Base64-encoded AES-256 encryption key value.

• string encryption_key_sha256 = 2

  The Base64-encoded SHA256 of the encryption key.

The message save the metadata of a backup.

• uint64 cluster_id = 1

  ID and version of backuped cluster.

• string cluster_version = 2

• string br_version = 11

  Save the version of BR running backup jobs.

• int32 version = 12

  The backupmeta scheme version.

• repeated File files = 4

  A set of files that compose a backup. Note: `files` is deprecated, as it bloats backupmeta. It is kept for compatibility, so new BR can restore older backups.

• optional MetaFile file_index = 13

  An index to files contains data files.

• uint64 start_version = 5

  A pair of timestamp specifies a time range of a backup. For full backup, the start_version equals to the end_version, it means point in time. For incremental backup, the time range is specified as (start_version, end_version].

• uint64 end_version = 6

• repeated Schema schemas = 7

  Table metadata describes database and table info. Note: `schemas` is deprecated, as it bloats backupmeta. It is kept for compatibility, so new BR can restore older backups.

• optional MetaFile schema_index = 14

  An index to files contains Schemas.

• bool is_raw_kv = 8

  If in raw kv mode, `start_versions`, `end_versions` and `schemas` will be ignored, and the backup data's range is represented by raw_ranges.

• repeated RawRange raw_ranges = 9

  Note: `raw_ranges` is deprecated, as it bloats backupmeta. It is kept for compatibility, so new BR can restore older backups.

• optional MetaFile raw_range_index = 15

  An index to files contains RawRanges.

• bytes ddls = 10

  In incremental backup, DDLs which are completed in (lastBackupTS, backupTS] will be stored here. Note: `raw_ranges` is deprecated, as it bloats backupmeta. It is kept for compatibility, so new BR can restore older backups.

• optional MetaFile ddl_indexes = 16

  An index to files contains DDLs.

• string backup_result = 17

  the backup result into `backupmeta` file

• kvrpcpb.APIVersion api_version = 18

  API version implies the encode of the key and value.

• repeated PlacementPolicy policies = 19

  the placement policy info in backup cluster. we assume the policy won't be too much for one cluster.

• string new_collations_enabled = 20

  new_collations_enabled specifies the config `new_collations_enabled_on_first_bootstrap` in tidb.

• bool is_txn_kv = 21

  If in txn kv mode, `schemas` will be ignored, the backup data's range is as same as normal backup.

• repeated PitrDBMap db_maps = 22

  maintain the id mapping from upstream cluster to downstream cluster.

• BackupMode mode = 23

• repeated BackupRange ranges = 24

  record the backup range and the correspond SST files when using file-copy backup.

• uint64 backup_size = 25

  record the size of the backup data files and meta files

BackupMpde represents the mode of this whole backup request to the cluster. and we need to store it in `backupmeta`.

Used in: BackupMeta, BackupRequest

• SCAN = 0

  scan and generate precise SST file of a backup range. then we don't need to do anything in future restore. Note: this mode will cost the CPU resource of TiKV.

• FILE = 1

  check and upload the coarse overlap SST files of a backup range. then we need to use a merge iterator to filter unexpected kv in future restore. Note: this mode will save the CPU resource of TiKV.

Used in: BackupMeta, MetaFile

• bytes start_key = 1

• bytes end_key = 2

• repeated File files = 3

Used in: CloudDynamic

• string endpoint = 1

• string region = 3

• string bucket = 4

• string prefix = 5

• string storage_class = 6

Used in: BackupRequest, StreamBackupTaskSecurityConfig, import_sstpb.ApplyRequest, import_sstpb.DownloadRequest

• encryptionpb.EncryptionMethod cipher_type = 1

• bytes cipher_key = 2

CloudDynamic allows testing new cloud providers and new fields without changing protobuf definitions

Used in: StorageBackend

• optional Bucket bucket = 1

• string provider_name = 2

  s3, gcs and azureBlobStorage are supported

• map<string, string> attrs = 3

Used in: Error

• uint64 current = 1

• uint64 request = 2

sst files or log files compression algorithm for log files, unknown means not use compression algorithm

Used in: BackupRequest, DataFileInfo, StreamBackupTaskInfo, import_sstpb.KVMeta

• UNKNOWN = 0

• LZ4 = 1

• SNAPPY = 2

• ZSTD = 3

DataFileGroup is the merged file info in log-backup

Used in: Metadata

• string path = 1

  Path of the file.

• repeated DataFileInfo data_files_info = 2

  Partitions of the file.

• uint64 min_ts = 3

  / Below are extra information of the file, for better filtering files. The min ts of the keys in the file.

• uint64 max_ts = 4

  The max ts of the keys in the file.

• uint64 min_resolved_ts = 5

  The resolved ts of the region when saving the file.

• uint64 length = 6

  The file length after compressed.

Used in: DataFileGroup, Metadata

• bytes sha256 = 1

  Checksum of the plaintext file, i.e., pre-compression, pre-encryption.

• string path = 2

  Path of the file.

• int64 number_of_entries = 3

• uint64 min_ts = 4

  / Below are extra information of the file, for better filtering files. The min ts of the keys in the file.

• uint64 max_ts = 5

  The max ts of the keys in the file.

• uint64 resolved_ts = 6

  The resolved ts of the region when saving the file.

• int64 region_id = 7

  The region of the file.

• bytes start_key = 8

  The key range of the file. Encoded and starts with 'z'(internal key).

• bytes end_key = 9

• string cf = 10

  The column family of the file.

• FileType type = 11

  The operation type of the file.

• bool is_meta = 12

  Whether the data file contains meta keys(m prefixed keys) only.

• int64 table_id = 13

  The table ID of the file contains, when `is_meta` is true, would be ignored.

• uint64 length = 14

  The file length.

• uint64 min_begin_ts_in_default_cf = 15

  The minimal begin ts in default cf if this file is write cf.

• uint64 range_offset = 16

  Offset of the partition. compatible with V1 and V2.

• uint64 range_length = 17

  The range length of the merged file, if it exists.

• CompressionType compression_type = 18

  The compression type for the file.

• uint64 crc64xor = 19

  The CRC64XOR of all contents of the file. In older versions, this might be empty.

• bytes region_start_key = 20

  The original region range the log file belongs to. In older versions, this might be empty. This may not be accurative: file may contain keys less than this.

• bytes region_end_key = 21

  The original region range the log file belongs to. In older versions, this might be empty. This may not be accurative: file may contain keys greater than this.

• repeated metapb.RegionEpoch region_epoch = 22

  The region epoch that the log file belongs to. In older versions, this might be empty. If a region get split or merged during observing, the file may contain multi epoches. This may not be complete: file may contain records from other versions. If there is exactly one epoch, `region_start_key` and `region_end_key` must match this epoch.

• optional encryptionpb.FileEncryptionInfo file_encryption_info = 23

  Encryption information of this data file, not set if plaintext.

An extended version of `SpansOfFile`, added more metadata for the execution of delayed deletion.

Used in: MetaEdit

• string path = 1

• repeated Span spans = 2

• uint64 whole_file_length = 3

  This is a hint for garbage collecting. So we can easily check whether a set of spans consists the whole file.

Used in: BackupResponse, CheckAdminResponse, CleanupResponse, PrepareResponse

• string msg = 1

• oneof detail

  • ClusterIDError cluster_id_error = 3

  • kvrpcpb.KeyError kv_error = 4

  • errorpb.Error region_error = 5

Used in: BackupMeta, BackupRange, BackupResponse, IngestedSSTs, LogFileSubcompaction, MetaFile

• string name = 1

• bytes sha256 = 2

• bytes start_key = 3

• bytes end_key = 4

• uint64 start_version = 5

• uint64 end_version = 6

• uint64 crc64xor = 7

  recorded the total statistics across multiple tables

• uint64 total_kvs = 8

• uint64 total_bytes = 9

• repeated TableMeta table_metas = 13

  recorded when the SST kvs are from different tables

• string cf = 10

• uint64 size = 11

• bytes cipher_iv = 12

  cipher_iv is used for AES cipher

Used in: DataFileInfo, LogFileSubcompactionMeta

• Delete = 0

• Put = 1

GCS storage backend saves files into google cloud storage.

Used in: StorageBackend

• string endpoint = 1

• string bucket = 2

• string prefix = 3

• string storage_class = 4

• string predefined_acl = 5

  If not empty, applies a predefined set of access controls. See https://cloud.google.com/storage/docs/access-control/lists#predefined-acl for valid values.

• string credentials_blob = 6

  Service Account Credentials JSON blob You can get one from https://console.cloud.google.com/apis/credentials, and copy the content, set it as string here.

HDFS storage backend saves file into HDFS compatible storages

Used in: StorageBackend

• string remote = 1

  a URL: hdfs:///some/path or hdfs://host:port/some/path

Used in: PitrDBMap, PitrTableMap

• int64 upstream_id = 1

• int64 downstream_id = 2

IngestedSSTs is created by a client that wants to put data to the cluster by `Ingest` APIs, these write cannot be directly recorded by log backup. The client should put a migration that contains the SSTs it uploaded.

• repeated RewrittenTableID rewritten_tables = 1

  The table IDs rewritten during restoring.

• repeated File files = 2

  The SST files restored.

• uint64 as_if_ts = 3

  Treating the whole batch of SSTs as an huge atomic write as this timestamp. That is, when user wants to PiTR to sometime after this, everything restored will present after the PiTR. Otherwise nothing will present.

• string files_prefix_hint = 4

  The hint of the common prefix of the files. Used for speed up garbage collecting.

• bool finished = 5

  Whether this batch have finished. If not, `as_if_ts` should be ignored and this should only be restored when there is a finished backup with the same `backup_uuid`.

• bytes backup_uuid = 6

  When checkpoint enabled, one restoration may be separated to many of `IngestedSST`s. `IngestedSST`s sharing the same UUID should be treated as one. That is: - The `as_if_ts` should be the largest one among all finished ingested SST batches. - When restoring, all batches sharing the same uuid should be restored or not.

Local storage backend saves files into local disk

Used in: StorageBackend

• string path = 1

Used in: Migration

• uint64 compaction_from_ts = 1

  The selector of compaction source. Files contains ANY record greater than this (and less than `until_ts`) will be selected. The source and artificats may contain records out of the `from_ts` and `until_ts` range. When restoring, please use `input_min_ts` to filter out not needed compactions instead.

• uint64 compaction_until_ts = 2

  The selector of compaction source. Files contains ANY record less than this (and not less than `from_ts`) will be selected. The source and artificats may contain records out of the `from_ts` and `until_ts` range. When restoring, please use `input_max_ts` to filter out not needed compactions instead.

• string name = 3

  The user provided name for the compaction.

• string comments = 4

  The statistics or other useful information for human.

• string artifacts = 5

  relative path to the generated subcompaction metas. ("artificates")

• uint64 artifacts_hash = 6

  the XOR of all artificate's hash.

• string generated_files = 7

  relative path to the subcompaction generated SST files. this is a hint for garbage-collecting, so we don't need to read the metas one by one. Those SSTs cannot be directly used, what you need is probably in `artificates`.

• uint64 input_min_ts = 8

  The smallest TS of record contained among all selected files of this compaction.

• uint64 input_max_ts = 9

  The largest TS of record contained among all selected files of this compaction.

A subcompaction from a compaction. Subcompactions are actions that collect then sort key values from one region.

Used in: LogFileSubcompactions

• optional LogFileSubcompactionMeta meta = 1

  The metadata of the subcompaction.

• repeated File sst_outputs = 2

  The generated SSTs.

• repeated RegionMetaHint region_meta_hints = 3

  The hint for the history of the region that contains the keys being compacted. This may be helpful for splitting, but it does not guarantee completeness and accuracy.

Used in: LogFileSubcompaction

• uint64 size = 1

  The whole size of this subcompaction. It should be the physical size of all input logs.

• uint64 region_id = 2

  The region that we perform the subcompaction.

• string cf = 3

• uint64 input_max_ts = 4

• uint64 input_min_ts = 5

• uint64 compact_from_ts = 6

  The whole compaction's configure.

• uint64 compact_until_ts = 7

• bytes min_key = 8

  The key range of the input logs. They are encoded user key. (i.e. mem-comparable, no 'z' prefix)

• bytes max_key = 9

  The largest key from the input.

• FileType ty = 10

  The type of operations in the file. (Write or Delete)

• int64 table_id = 11

• repeated SpansOfFile sources = 12

  The input log files of this.

Batched version of `LogFileSubcompaction`. So we can store many subcompactions to one file, to reduce the number of file.

• repeated LogFileSubcompaction subcompactions = 1

Used in: StreamBackupTaskSecurityConfig

• encryptionpb.EncryptionMethod encryption_type = 1

• repeated encryptionpb.MasterKey master_keys = 2

Used in: Migration

• string path = 1

  Path to the meta file.

• repeated string delete_physical_files = 2

  Delete the physical files (MetaFileGroup) in the meta file.

• repeated DeleteSpansOfFile delete_logical_files = 3

  Delete the logical files (MetaFileInfo) in the meta file. Note: Even the operation have been performed in the meta, this modification should be kept as long as the corresponding physical file not deleted. Or we may cannot know when to delete the physical file. Then the file will be leak until truncated.

• bool destruct_self = 4

  Whether to delete the metadata itself.

MetaFile describes a multi-level index of data used in backup.

Used in: BackupMeta

• repeated File meta_files = 1

  A set of files that contains a MetaFile. It is used as a multi-level index.

• repeated File data_files = 2

  A set of files that contains user data.

• repeated Schema schemas = 3

  A set of files that contains Schemas.

• repeated RawRange raw_ranges = 4

  A set of files that contains RawRanges.

• repeated BackupRange backup_ranges = 6

  A set of files that contains BackupRanges.

• repeated bytes ddls = 5

  A set of files that contains DDLs.

Used in: Metadata

• V1 = 0

• V2 = 1

• repeated DataFileInfo files = 1

  deprecated after(in) v6.3.0 TiKV cluster

• repeated DataFileGroup file_groups = 6

• int64 store_id = 2

• uint64 resolved_ts = 3

• uint64 max_ts = 4

• uint64 min_ts = 5

• MetaVersion meta_version = 7

• string creator = 5

• MigrationVersion version = 6

• repeated MetaEdit edit_meta = 1

• uint64 truncated_to = 2

• repeated LogFileCompaction compactions = 3

• repeated string destruct_prefix = 4

  Delete all files matches this prefix.

• repeated string ingested_sst_paths = 7

  The ingested SSTs. It is mutable during restoring, hence a reference.

Used in: Migration

• M0 = 0

  The initial version. Somehow "UNKNOWN".

• M1 = 1

  Added `creator` and `version`.

• M2 = 2

  Added `ingested_sst_paths`.

Noop storage backend saves files into void.

Used in: StorageBackend

(message has no fields)

Used in: BackupMeta

• string name = 1

• optional IDMap id_map = 2

• repeated PitrTableMap tables = 3

• bool filtered_out = 4

  whether this db has been filtered out

Used in: PitrDBMap

• string name = 1

• optional IDMap id_map = 2

• repeated IDMap partitions = 3

• bool filtered_out = 4

  whether this table has been filtered out

Used in: BackupMeta

• bytes info = 1

Used in: PrepareSnapshotBackupResponse

• WaitApplyDone = 0

  A region has finished wait apply.

• UpdateLeaseResult = 1

  A lease has been updated.

Used in: PrepareSnapshotBackupRequest

• UpdateLease = 0

  Update the lease of suspending some commands (Which may impact the disk snapshot backup) to be proposed. As long as the lease is kept, those commands may not be proposed. If the client has gone, the lease will be destroyed after its TTL.

• WaitApply = 1

  Wait a region apply to the last index.

• Finish = 2

  Hint before closing the stream. Server will go back to normal after this. This should also return a final "UpdateLeaseResult" to the client, which indices whether the last lease is valid.

Used in: BackupMeta, MetaFile

• bytes start_key = 1

• bytes end_key = 2

• string cf = 3

You may get region id from `LogFileSubcompactionMeta`.

Used in: LogFileSubcompaction

• bytes start_key = 1

  The left boundary of the region of this epoch. It is an byte-comparable encoded key without TS.

• bytes end_key = 2

  The right boundary of the region of the epoch.

• optional metapb.RegionEpoch region_epoch = 3

  The region epoch.

RewrittenTableID records a sort of modification over the SSTs during restoring. When "backing up" this "restored" SST, we want to backup the original SST before any rewrite. But in some cases, SSTs are not restored "as is", TiKV may rewrite or truncate its content by the client's request.

Used in: IngestedSSTs

• int64 ancestor_upstream = 1

  The table ID in the backup data.

• int64 upstream = 2

  The rewritten table ID during restoring.

S3 storage backend saves files into S3 compatible storages For non-aws providers, endpoint must be provided

Used in: StorageBackend

• string endpoint = 1

• string region = 2

• string bucket = 3

• string prefix = 4

• string storage_class = 5

• string sse = 6

  server side encryption

• string acl = 7

• string access_key = 8

• string secret_access_key = 9

• bool force_path_style = 10

• string sse_kms_key_id = 11

• string role_arn = 12

• string external_id = 13

• bool object_lock_enabled = 14

• string session_token = 15

• string provider = 16

• string profile = 17

Used in: BackupMeta, MetaFile

• bytes db = 1

• bytes table = 2

• uint64 crc64xor = 3

• uint64 total_kvs = 4

• uint64 total_bytes = 5

• uint32 tiflash_replicas = 6

• bytes stats = 7

  stats represents the dump stats for a analyzed table, which generate by DumpStatsToJSON https://github.com/pingcap/tidb/blob/e136429d8dc5d70f43cd3f94179b0b9f47595097/statistics/handle/dump.go#L116

• repeated StatsFileIndex stats_index = 8

  An index of files that contains stats data

Used in: BackupRequest

• repeated kvrpcpb.KeyRange sub_ranges = 1

Used in: DeleteSpansOfFile, SpansOfFile

• uint64 offset = 1

• uint64 length = 2

Used in: LogFileSubcompactionMeta

• string path = 1

• repeated Span spans = 2

Used in: StatsFile

• int64 physical_id = 1

  The physical id which the stats json data belongs to.

• bytes json_table = 2

  The marshal data of the part of json_table. Currently it is the partition table.

• repeated StatsBlock blocks = 1

  A set of files that contains json_tables.

Used in: Schema

• string name = 1

• bytes sha256 = 2

• uint64 size_enc = 3

  the size of the file content encrypted

• uint64 size_ori = 4

  the size of the original file content

• bytes cipher_iv = 5

  cipher_iv is used for AES cipher

• bytes inline_data = 6

  If the size of the stats is very small, save the data into the metafile

Used in: BackupRequest, ExternalStorageRestoreRequest, ExternalStorageSaveRequest, StreamBackupTaskInfo, import_sstpb.ApplyRequest, import_sstpb.DownloadRequest

• oneof backend

  • Noop noop = 1

  • Local local = 2

  • S3 s3 = 3

  • GCS gcs = 4

  • CloudDynamic cloud_dynamic = 5

  • HDFS hdfs = 6

  • AzureBlobStorage azure_blob_storage = 7

• uint64 happen_at = 1

  the unix epoch time (in millisecs) of the time the error reported.

• string error_code = 2

  the unified error code of the error.

• string error_message = 3

  the user-friendly error message.

• uint64 store_id = 4

  the store id of who issues the error.

• optional StorageBackend storage = 1

  The storage for backup, parsed by BR.

• uint64 start_ts = 2

  The time range for backing up.

• uint64 end_ts = 3

• string name = 4

  Misc meta data. The name of the task, also the ID of the task.

• repeated string table_filter = 5

  The table filter of the task. Only for displaying the task info.

• CompressionType compression_type = 6

  compression type

• optional StreamBackupTaskSecurityConfig security_config = 7

  security config for backup files

Used in: StreamBackupTaskInfo

• oneof encryption

  • CipherInfo plaintext_data_key = 1

    not recommended in production. configure KMS based or locally managed master key instead in TiKV StreamBackupConfig

  • MasterKeyConfig master_key_config = 2

The message saves the physical table checksum

Used in: File

• int64 physical_id = 1

  physical id is either table id or partition id

• uint64 crc64xor = 2

• uint64 total_kvs = 3

• uint64 total_bytes = 4