From 75b13020225d0df9e3f93348191249762ae4d7a5 Mon Sep 17 00:00:00 2001 From: Leto_b Date: Tue, 27 Jan 2026 10:39:00 +0800 Subject: [PATCH] add pipe permission control specifications --- .../Tree/User-Manual/Data-Sync_apache.md | 23 ++++++++++++++--- .../Tree/User-Manual/Data-Sync_timecho.md | 19 ++++++++++++-- .../latest/User-Manual/Data-Sync_apache.md | 18 ++++++++++++- .../latest/User-Manual/Data-Sync_timecho.md | 18 ++++++++++++- .../Tree/User-Manual/Data-Sync_apache.md | 24 +++++++++++++++--- .../Tree/User-Manual/Data-Sync_timecho.md | 25 ++++++++++++++++--- .../latest/User-Manual/Data-Sync_apache.md | 20 ++++++++++++++- .../latest/User-Manual/Data-Sync_timecho.md | 20 ++++++++++++++- 8 files changed, 150 insertions(+), 17 deletions(-) diff --git a/src/UserGuide/Master/Tree/User-Manual/Data-Sync_apache.md b/src/UserGuide/Master/Tree/User-Manual/Data-Sync_apache.md index 212c3be0a..59de59159 100644 --- a/src/UserGuide/Master/Tree/User-Manual/Data-Sync_apache.md +++ b/src/UserGuide/Master/Tree/User-Manual/Data-Sync_apache.md @@ -79,7 +79,7 @@ By declaratively configuring the specific content of the three parts through SQL ### 1.2 Functional limitations and instructions -The schema and auth synchronization functions have the following limitations: +1. The schema and auth synchronization functions have the following limitations: - When using schema synchronization, it is required that the consensus protocol for `Schema region` and `ConfigNode` must be the default ratis protocol. This means that the `iotdb-system.properties` configuration file should contain the settings `config_node_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus` and `schema_region_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus`. If these are not specified, the default ratis protocol is used. @@ -89,6 +89,22 @@ The schema and auth synchronization functions have the following limitations: - During data synchronization tasks, please avoid performing any deletion operations to prevent inconsistent states between the two ends. +2. Pipe Permission Control Specifications + +- When creating a pipe, a username and password can be specified for the extraction/write‑back plugins. If the password is incorrect, creation is prohibited. If not specified, the current user is used for synchronization by default. + +- During data/metadata synchronization, filtering is first performed according to the path pattern (pattern/path) configured in the Pipe, followed by authentication based on the user’s read permissions: + - If the permission scope is greater than or equal to the write path: full synchronization. + - If the permission scope has no intersection with the write path: no synchronization. + - If the permission scope is smaller than the write path or overlaps partially: synchronize only the intersecting part. + +- When encountering data for which the user lacks permission: + - If the sender’s skipIf=no‑privileges, the unauthorized data is skipped. + - If skipIf is left empty (unconfigured), the task reports an error (Error 803). + - Note: This skipIf configuration is independent of the receiver’s skipIf setting (which defaults to empty). + +- Data under root.__system and root.__audit will not be synchronized. + ## 2. Usage Instructions Data synchronization tasks have three states: RUNNING, STOPPED, and DROPPED. The task state transitions are shown in the following diagram: @@ -514,7 +530,7 @@ pipe_all_sinks_rate_limit_bytes_per_second=-1 | password | Password for the username. | String | No | root | | batch.enable | Enables batch mode for log transmission to improve throughput and reduce IOPS. | Boolean: true, false | No | true | | batch.max-delay-seconds | Maximum delay (in seconds) for batch transmission. | Integer | No | 1 | -| batch.max-delay-ms | Maximum delay (in ms) for batch transmission. (Available since v2.0.5) | Integer | No | 1 | +| batch.max-delay-ms | Maximum delay (in ms) for batch transmission. (Available since v2.0.5) | Integer | No | 1 | | batch.size-bytes | Maximum batch size (in bytes) for batch transmission. | Long | No | 16*1024*1024 | | compressor | The selected RPC compression algorithm. Multiple algorithms can be configured and will be adopted in sequence for each request. | String: snappy / gzip / lz4 / zstd / lzma2 | No | "" | | compressor.zstd.level | When the selected RPC compression algorithm is zstd, this parameter can be used to additionally configure the compression level of the zstd algorithm. | Int: [-131072, 22] | No | 3 | @@ -524,7 +540,6 @@ pipe_all_sinks_rate_limit_bytes_per_second=-1 #### iotdb-thrift-ssl-sink - | **Parameter** | **Description** | Value Range | Required | Default Value | |:----------------------------|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------| :------------ | | sink | iotdb-thrift-ssl-sink | String: iotdb-thrift-ssl-sink | Yes | - | @@ -533,7 +548,7 @@ pipe_all_sinks_rate_limit_bytes_per_second=-1 | password | Password for the username. | String | No | root | | batch.enable | Enables batch mode for log transmission to improve throughput and reduce IOPS. | Boolean: true, false | No | true | | batch.max-delay-seconds | Maximum delay (in seconds) for batch transmission. | Integer | No | 1 | -| batch.max-delay-ms | Maximum delay (in ms) for batch transmission. (Available since v2.0.5) | Integer | No | 1 | +| batch.max-delay-ms | Maximum delay (in ms) for batch transmission. (Available since v2.0.5) | Integer | No | 1 | | batch.size-bytes | Maximum batch size (in bytes) for batch transmission. | Long | No | 16*1024*1024 | | compressor | The selected RPC compression algorithm. Multiple algorithms can be configured and will be adopted in sequence for each request. | String: snappy / gzip / lz4 / zstd / lzma2 | No | "" | | compressor.zstd.level | When the selected RPC compression algorithm is zstd, this parameter can be used to additionally configure the compression level of the zstd algorithm. | Int: [-131072, 22] | No | 3 | diff --git a/src/UserGuide/Master/Tree/User-Manual/Data-Sync_timecho.md b/src/UserGuide/Master/Tree/User-Manual/Data-Sync_timecho.md index ae0813daa..b17d26d7c 100644 --- a/src/UserGuide/Master/Tree/User-Manual/Data-Sync_timecho.md +++ b/src/UserGuide/Master/Tree/User-Manual/Data-Sync_timecho.md @@ -79,7 +79,7 @@ By declaratively configuring the specific content of the three parts through SQL ### 1.2 Functional limitations and instructions -The schema and auth synchronization functions have the following limitations: +1. The schema and auth synchronization functions have the following limitations: - When using schema synchronization, it is required that the consensus protocol for `Schema region` and `ConfigNode` must be the default ratis protocol. This means that the `iotdb-system.properties` configuration file should contain the settings `config_node_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus` and `schema_region_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus`. If these are not specified, the default ratis protocol is used. @@ -91,6 +91,22 @@ The schema and auth synchronization functions have the following limitations: - During data synchronization tasks, please avoid performing any deletion operations to prevent inconsistent states between the two ends. +2. Pipe Permission Control Specifications + +- When creating a pipe, a username and password can be specified for the extraction/write‑back plugins. If the password is incorrect, creation is prohibited. If not specified, the current user is used for synchronization by default. + +- During data/metadata synchronization, filtering is first performed according to the path pattern (pattern/path) configured in the Pipe, followed by authentication based on the user’s read permissions: + - If the permission scope is greater than or equal to the write path: full synchronization. + - If the permission scope has no intersection with the write path: no synchronization. + - If the permission scope is smaller than the write path or overlaps partially: synchronize only the intersecting part. + +- When encountering data for which the user lacks permission: + - If the sender’s skipIf=no‑privileges, the unauthorized data is skipped. + - If skipIf is left empty (unconfigured), the task reports an error (Error 803). + - Note: This skipIf configuration is independent of the receiver’s skipIf setting (which defaults to empty). + +- Data under root.__system and root.__audit will not be synchronized. + ## 2. Usage Instructions Data synchronization tasks have three states: RUNNING, STOPPED, and DROPPED. The task state transitions are shown in the following diagram: @@ -607,7 +623,6 @@ pipe_all_sinks_rate_limit_bytes_per_second=-1 | load-tsfile-strategy | When synchronizing file data, ​​whether the receiver waits for the local load tsfile operation to complete before responding to the sender​​:
​​sync​​: Wait for the local load tsfile operation to complete before returning the response.
​​async​​: Do not wait for the local load tsfile operation to complete; return the response immediately. | String: sync / async | No | sync | | format | The payload formats for data transmission include the following options:
- hybrid: The format depends on what is passed from the processor (either tsfile or tablet), and the sink performs no conversion.
- tsfile: Data is forcibly converted to tsfile format before transmission. This is suitable for scenarios like data file backup.
- tablet: Data is forcibly converted to tsfile format before transmission. This is useful for data synchronization when the sender and receiver have incompatible data types (to minimize errors). | String: hybrid / tsfile / tablet | No | hybrid | - #### iotdb-air-gap-sink | key | value | value Range | required or not | Default Value | diff --git a/src/UserGuide/latest/User-Manual/Data-Sync_apache.md b/src/UserGuide/latest/User-Manual/Data-Sync_apache.md index 66f196575..59de59159 100644 --- a/src/UserGuide/latest/User-Manual/Data-Sync_apache.md +++ b/src/UserGuide/latest/User-Manual/Data-Sync_apache.md @@ -79,7 +79,7 @@ By declaratively configuring the specific content of the three parts through SQL ### 1.2 Functional limitations and instructions -The schema and auth synchronization functions have the following limitations: +1. The schema and auth synchronization functions have the following limitations: - When using schema synchronization, it is required that the consensus protocol for `Schema region` and `ConfigNode` must be the default ratis protocol. This means that the `iotdb-system.properties` configuration file should contain the settings `config_node_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus` and `schema_region_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus`. If these are not specified, the default ratis protocol is used. @@ -89,6 +89,22 @@ The schema and auth synchronization functions have the following limitations: - During data synchronization tasks, please avoid performing any deletion operations to prevent inconsistent states between the two ends. +2. Pipe Permission Control Specifications + +- When creating a pipe, a username and password can be specified for the extraction/write‑back plugins. If the password is incorrect, creation is prohibited. If not specified, the current user is used for synchronization by default. + +- During data/metadata synchronization, filtering is first performed according to the path pattern (pattern/path) configured in the Pipe, followed by authentication based on the user’s read permissions: + - If the permission scope is greater than or equal to the write path: full synchronization. + - If the permission scope has no intersection with the write path: no synchronization. + - If the permission scope is smaller than the write path or overlaps partially: synchronize only the intersecting part. + +- When encountering data for which the user lacks permission: + - If the sender’s skipIf=no‑privileges, the unauthorized data is skipped. + - If skipIf is left empty (unconfigured), the task reports an error (Error 803). + - Note: This skipIf configuration is independent of the receiver’s skipIf setting (which defaults to empty). + +- Data under root.__system and root.__audit will not be synchronized. + ## 2. Usage Instructions Data synchronization tasks have three states: RUNNING, STOPPED, and DROPPED. The task state transitions are shown in the following diagram: diff --git a/src/UserGuide/latest/User-Manual/Data-Sync_timecho.md b/src/UserGuide/latest/User-Manual/Data-Sync_timecho.md index 83282a7e1..b17d26d7c 100644 --- a/src/UserGuide/latest/User-Manual/Data-Sync_timecho.md +++ b/src/UserGuide/latest/User-Manual/Data-Sync_timecho.md @@ -79,7 +79,7 @@ By declaratively configuring the specific content of the three parts through SQL ### 1.2 Functional limitations and instructions -The schema and auth synchronization functions have the following limitations: +1. The schema and auth synchronization functions have the following limitations: - When using schema synchronization, it is required that the consensus protocol for `Schema region` and `ConfigNode` must be the default ratis protocol. This means that the `iotdb-system.properties` configuration file should contain the settings `config_node_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus` and `schema_region_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus`. If these are not specified, the default ratis protocol is used. @@ -91,6 +91,22 @@ The schema and auth synchronization functions have the following limitations: - During data synchronization tasks, please avoid performing any deletion operations to prevent inconsistent states between the two ends. +2. Pipe Permission Control Specifications + +- When creating a pipe, a username and password can be specified for the extraction/write‑back plugins. If the password is incorrect, creation is prohibited. If not specified, the current user is used for synchronization by default. + +- During data/metadata synchronization, filtering is first performed according to the path pattern (pattern/path) configured in the Pipe, followed by authentication based on the user’s read permissions: + - If the permission scope is greater than or equal to the write path: full synchronization. + - If the permission scope has no intersection with the write path: no synchronization. + - If the permission scope is smaller than the write path or overlaps partially: synchronize only the intersecting part. + +- When encountering data for which the user lacks permission: + - If the sender’s skipIf=no‑privileges, the unauthorized data is skipped. + - If skipIf is left empty (unconfigured), the task reports an error (Error 803). + - Note: This skipIf configuration is independent of the receiver’s skipIf setting (which defaults to empty). + +- Data under root.__system and root.__audit will not be synchronized. + ## 2. Usage Instructions Data synchronization tasks have three states: RUNNING, STOPPED, and DROPPED. The task state transitions are shown in the following diagram: diff --git a/src/zh/UserGuide/Master/Tree/User-Manual/Data-Sync_apache.md b/src/zh/UserGuide/Master/Tree/User-Manual/Data-Sync_apache.md index 97100bb68..8ebaccab0 100644 --- a/src/zh/UserGuide/Master/Tree/User-Manual/Data-Sync_apache.md +++ b/src/zh/UserGuide/Master/Tree/User-Manual/Data-Sync_apache.md @@ -78,7 +78,7 @@ ### 1.2 功能限制及说明 -元数据（schema）、权限（auth）同步功能存在如下限制： +1. 元数据（schema）、权限（auth）同步功能存在如下限制： - 使用元数据同步时，要求`Schema region`、`ConfigNode` 的共识协议必须为默认的 ratis 协议，即`iotdb-system.properties`配置文件中是否包含`config_node_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus`、`schema_region_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus`，不包含即为默认值ratis 协议。 @@ -88,6 +88,24 @@ - 在进行数据同步任务时，请避免执行任何删除操作，防止两端状态不一致。 +2. Pipe 权限控制规范如下： + +- 创建 pipe 时，可以对抽取/写回插件指定用户名和密码。密码错误则禁止创建，未指定时默认使用当前用户进行同步。 + +- 数据/元数据同步时，先根据 Pipe 配置的路径模式(pattern/path)筛选，再基于用户读取权限进行鉴权 + + - 权限范围≥写入路径：完整同步 + + - 权限范围与写入路径无交集：不同步 + + - 权限范围<写入路径或存在交集：同步交集部分 + +- 遇到无权限数据时，若发送端 skipIf=no-privileges，则跳过无权限数据；若 skipIf 配置为空，任务报错（803错误） + + - 注意：此 skipIf 配置与接收端的 skipIf（默认为空）相互独立 + +- 对于 root.__system, root.__audit 均不会同步 + ## 2. 使用说明 数据同步任务有三种状态：RUNNING、STOPPED 和 DROPPED。任务状态转换如下图所示： @@ -509,7 +527,7 @@ pipe_all_sinks_rate_limit_bytes_per_second=-1 | password | 连接接收端使用的用户名对应的密码，同步要求该用户具备相应的操作权限 | String | 选填 | root | | batch.enable | 是否开启日志攒批发送模式，用于提高传输吞吐，降低 IOPS | Boolean: true, false | 选填 | true | | batch.max-delay-seconds | 在开启日志攒批发送模式时生效，表示一批数据在发送前的最长等待时间（单位：s） | Integer | 选填 | 1 | -| batch.max-delay-ms | 在开启日志攒批发送模式时生效，表示一批数据在发送前的最长等待时间（单位：ms）(V2.0.5及以后版本支持） | Integer | 选填 | 1 | +| batch.max-delay-ms | 在开启日志攒批发送模式时生效，表示一批数据在发送前的最长等待时间（单位：ms）(V2.0.5及以后版本支持） | Integer | 选填 | 1 | | batch.size-bytes | 在开启日志攒批发送模式时生效，表示一批数据最大的攒批大小（单位：byte） | Long | 选填 | 16*1024*1024 | | compressor | 所选取的 rpc 压缩算法，可配置多个，对每个请求顺序采用 | String: snappy / gzip / lz4 / zstd / lzma2 | 选填 | "" | | compressor.zstd.level | 所选取的 rpc 压缩算法为 zstd 时，可使用该参数额外配置 zstd 算法的压缩等级 | Int: [-131072, 22] | 选填 | 3 | @@ -527,7 +545,7 @@ pipe_all_sinks_rate_limit_bytes_per_second=-1 | password | 连接接收端使用的用户名对应的密码，同步要求该用户具备相应的操作权限 | String | 选填 | root | | batch.enable | 是否开启日志攒批发送模式，用于提高传输吞吐，降低 IOPS | Boolean: true, false | 选填 | true | | batch.max-delay-seconds | 在开启日志攒批发送模式时生效，表示一批数据在发送前的最长等待时间（单位：s） | Integer | 选填 | 1 | -| batch.max-delay-ms | 在开启日志攒批发送模式时生效，表示一批数据在发送前的最长等待时间（单位：ms）(V2.0.5及以后版本支持） | Integer | 选填 | 1 | +| batch.max-delay-ms | 在开启日志攒批发送模式时生效，表示一批数据在发送前的最长等待时间（单位：ms）(V2.0.5及以后版本支持） | Integer | 选填 | 1 | | batch.size-bytes | 在开启日志攒批发送模式时生效，表示一批数据最大的攒批大小（单位：byte） | Long | 选填 | 16*1024*1024 | | compressor | 所选取的 rpc 压缩算法，可配置多个，对每个请求顺序采用 | String: snappy / gzip / lz4 / zstd / lzma2 | 选填 | "" | | compressor.zstd.level | 所选取的 rpc 压缩算法为 zstd 时，可使用该参数额外配置 zstd 算法的压缩等级 | Int: [-131072, 22] | 选填 | 3 | diff --git a/src/zh/UserGuide/Master/Tree/User-Manual/Data-Sync_timecho.md b/src/zh/UserGuide/Master/Tree/User-Manual/Data-Sync_timecho.md index cf6547635..edc35641b 100644 --- a/src/zh/UserGuide/Master/Tree/User-Manual/Data-Sync_timecho.md +++ b/src/zh/UserGuide/Master/Tree/User-Manual/Data-Sync_timecho.md @@ -78,7 +78,7 @@ ### 1.2 功能限制及说明 -元数据（schema）、权限（auth）同步功能存在如下限制： +1. 元数据（schema）、权限（auth）同步功能存在如下限制： - 使用元数据同步时，要求`Schema region`、`ConfigNode` 的共识协议必须为默认的 ratis 协议，即`iotdb-system.properties`配置文件中是否包含`config_node_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus`、`schema_region_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus`，不包含即为默认值ratis 协议。 @@ -90,6 +90,24 @@ - 在进行数据同步任务时，请避免执行任何删除操作，防止两端状态不一致。 +2. Pipe 权限控制规范如下： + +- 创建 pipe 时，可以对抽取/写回插件指定用户名和密码。密码错误则禁止创建，未指定时默认使用当前用户进行同步。 + +- 数据/元数据同步时，先根据 Pipe 配置的路径模式(pattern/path)筛选，再基于用户读取权限进行鉴权 + + - 权限范围≥写入路径：完整同步 + + - 权限范围与写入路径无交集：不同步 + + - 权限范围<写入路径或存在交集：同步交集部分 + +- 遇到无权限数据时，若发送端 skipIf=no-privileges，则跳过无权限数据；若 skipIf 配置为空，任务报错（803错误） + + - 注意：此 skipIf 配置与接收端的 skipIf（默认为空）相互独立 + +- 对于 root.__system, root.__audit 均不会同步 + ## 2. 使用说明 数据同步任务有三种状态：RUNNING、STOPPED 和 DROPPED。任务状态转换如下图所示： @@ -594,7 +612,7 @@ pipe_all_sinks_rate_limit_bytes_per_second=-1 | password | 连接接收端使用的用户名对应的密码，同步要求该用户具备相应的操作权限 | String | 选填 | TimechoDB@2021, V2.0.6.x 之前为root | | batch.enable | 是否开启日志攒批发送模式，用于提高传输吞吐，降低 IOPS | Boolean: true, false | 选填 | true | | batch.max-delay-seconds | 在开启日志攒批发送模式时生效，表示一批数据在发送前的最长等待时间（单位：s） | Integer | 选填 | 1 | -| batch.max-delay-ms | 在开启日志攒批发送模式时生效，表示一批数据在发送前的最长等待时间（单位：ms）(V2.0.5及以后版本支持） | Integer | 选填 | 1 | +| batch.max-delay-ms | 在开启日志攒批发送模式时生效，表示一批数据在发送前的最长等待时间（单位：ms）(V2.0.5及以后版本支持） | Integer | 选填 | 1 | | batch.size-bytes | 在开启日志攒批发送模式时生效，表示一批数据最大的攒批大小（单位：byte） | Long | 选填 | 16*1024*1024 | | compressor | 所选取的 rpc 压缩算法，可配置多个，对每个请求顺序采用 | String: snappy / gzip / lz4 / zstd / lzma2 | 选填 | "" | | compressor.zstd.level | 所选取的 rpc 压缩算法为 zstd 时，可使用该参数额外配置 zstd 算法的压缩等级 | Int: [-131072, 22] | 选填 | 3 | @@ -602,7 +620,6 @@ pipe_all_sinks_rate_limit_bytes_per_second=-1 | load-tsfile-strategy | 文件同步数据时，接收端请求返回发送端前，是否等待接收端本地的 load tsfile 执行结果返回。
sync：等待本地的 load tsfile 执行结果返回；
async：不等待本地的 load tsfile 执行结果返回。 | String: sync / async | 选填 | sync | | format | 数据传输的payload格式， 可选项包括：
- hybrid： 取决于 processor 传递过来的格式（tsfile或tablet），sink不做任何转换。
- tsfile：强制转换成tsfile发送，可用于数据文件备份等场景。
- tablet：强制转换成tsfile发送，可用于发送端/接收端数据类型不完全兼容时的数据同步（以减少报错）。 | String: hybrid / tsfile / tablet | 选填 | hybrid | - #### iotdb-air-gap-sink | key | value | value 取值范围 | 是否必填 | 默认取值 | @@ -627,7 +644,7 @@ pipe_all_sinks_rate_limit_bytes_per_second=-1 | password | 连接接收端使用的用户名对应的密码，同步要求该用户具备相应的操作权限 | String | 选填 | TimechoDB@2021, V2.0.6.x 之前为root | | batch.enable | 是否开启日志攒批发送模式，用于提高传输吞吐，降低 IOPS | Boolean: true, false | 选填 | true | | batch.max-delay-seconds | 在开启日志攒批发送模式时生效，表示一批数据在发送前的最长等待时间（单位：s） | Integer | 选填 | 1 | -| batch.max-delay-ms | 在开启日志攒批发送模式时生效，表示一批数据在发送前的最长等待时间（单位：ms）(V2.0.5及以后版本支持） | Integer | 选填 | 1 | +| batch.max-delay-ms | 在开启日志攒批发送模式时生效，表示一批数据在发送前的最长等待时间（单位：ms）(V2.0.5及以后版本支持） | Integer | 选填 | 1 | | batch.size-bytes | 在开启日志攒批发送模式时生效，表示一批数据最大的攒批大小（单位：byte） | Long | 选填 | 16*1024*1024 | | compressor | 所选取的 rpc 压缩算法，可配置多个，对每个请求顺序采用 | String: snappy / gzip / lz4 / zstd / lzma2 | 选填 | "" | | compressor.zstd.level | 所选取的 rpc 压缩算法为 zstd 时，可使用该参数额外配置 zstd 算法的压缩等级 | Int: [-131072, 22] | 选填 | 3 | diff --git a/src/zh/UserGuide/latest/User-Manual/Data-Sync_apache.md b/src/zh/UserGuide/latest/User-Manual/Data-Sync_apache.md index e029b66b7..8ebaccab0 100644 --- a/src/zh/UserGuide/latest/User-Manual/Data-Sync_apache.md +++ b/src/zh/UserGuide/latest/User-Manual/Data-Sync_apache.md @@ -78,7 +78,7 @@ ### 1.2 功能限制及说明 -元数据（schema）、权限（auth）同步功能存在如下限制： +1. 元数据（schema）、权限（auth）同步功能存在如下限制： - 使用元数据同步时，要求`Schema region`、`ConfigNode` 的共识协议必须为默认的 ratis 协议，即`iotdb-system.properties`配置文件中是否包含`config_node_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus`、`schema_region_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus`，不包含即为默认值ratis 协议。 @@ -88,6 +88,24 @@ - 在进行数据同步任务时，请避免执行任何删除操作，防止两端状态不一致。 +2. Pipe 权限控制规范如下： + +- 创建 pipe 时，可以对抽取/写回插件指定用户名和密码。密码错误则禁止创建，未指定时默认使用当前用户进行同步。 + +- 数据/元数据同步时，先根据 Pipe 配置的路径模式(pattern/path)筛选，再基于用户读取权限进行鉴权 + + - 权限范围≥写入路径：完整同步 + + - 权限范围与写入路径无交集：不同步 + + - 权限范围<写入路径或存在交集：同步交集部分 + +- 遇到无权限数据时，若发送端 skipIf=no-privileges，则跳过无权限数据；若 skipIf 配置为空，任务报错（803错误） + + - 注意：此 skipIf 配置与接收端的 skipIf（默认为空）相互独立 + +- 对于 root.__system, root.__audit 均不会同步 + ## 2. 使用说明 数据同步任务有三种状态：RUNNING、STOPPED 和 DROPPED。任务状态转换如下图所示： diff --git a/src/zh/UserGuide/latest/User-Manual/Data-Sync_timecho.md b/src/zh/UserGuide/latest/User-Manual/Data-Sync_timecho.md index c0a869aa5..edc35641b 100644 --- a/src/zh/UserGuide/latest/User-Manual/Data-Sync_timecho.md +++ b/src/zh/UserGuide/latest/User-Manual/Data-Sync_timecho.md @@ -78,7 +78,7 @@ ### 1.2 功能限制及说明 -元数据（schema）、权限（auth）同步功能存在如下限制： +1. 元数据（schema）、权限（auth）同步功能存在如下限制： - 使用元数据同步时，要求`Schema region`、`ConfigNode` 的共识协议必须为默认的 ratis 协议，即`iotdb-system.properties`配置文件中是否包含`config_node_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus`、`schema_region_consensus_protocol_class=org.apache.iotdb.consensus.ratis.RatisConsensus`，不包含即为默认值ratis 协议。 @@ -90,6 +90,24 @@ - 在进行数据同步任务时，请避免执行任何删除操作，防止两端状态不一致。 +2. Pipe 权限控制规范如下： + +- 创建 pipe 时，可以对抽取/写回插件指定用户名和密码。密码错误则禁止创建，未指定时默认使用当前用户进行同步。 + +- 数据/元数据同步时，先根据 Pipe 配置的路径模式(pattern/path)筛选，再基于用户读取权限进行鉴权 + + - 权限范围≥写入路径：完整同步 + + - 权限范围与写入路径无交集：不同步 + + - 权限范围<写入路径或存在交集：同步交集部分 + +- 遇到无权限数据时，若发送端 skipIf=no-privileges，则跳过无权限数据；若 skipIf 配置为空，任务报错（803错误） + + - 注意：此 skipIf 配置与接收端的 skipIf（默认为空）相互独立 + +- 对于 root.__system, root.__audit 均不会同步 + ## 2. 使用说明 数据同步任务有三种状态：RUNNING、STOPPED 和 DROPPED。任务状态转换如下图所示：