Merge pull request #16264 from taosdata/3.0

relrase: merge from 3.0 to main

Jenkinsfile2

@@ -43,6 +43,7 @@ def pre_test(){
         cd ${WKC}
         git reset --hard
         git clean -fxd
+        rm -rf examples/rust/
         git remote prune origin
         git fetch
     '''

cmake/cmake.install

-IF (EXISTS /var/lib/taos/dnode/dnodeCfg.json)
-  INSTALL(CODE "MESSAGE(\"The default data directory /var/lib/taos contains old data of tdengine 2.x, please clear it before installing!\")")
-ELSEIF (EXISTS C:/TDengine/data/dnode/dnodeCfg.json)
-  INSTALL(CODE "MESSAGE(\"The default data directory C:/TDengine/data contains old data of tdengine 2.x, please clear it before installing!\")")
-ELSEIF (TD_LINUX)
+IF (TD_LINUX)
   SET(TD_MAKE_INSTALL_SH "${TD_SOURCE_DIR}/packaging/tools/make_install.sh")
   INSTALL(CODE "MESSAGE(\"make install script: ${TD_MAKE_INSTALL_SH}\")")
   INSTALL(CODE "execute_process(COMMAND bash ${TD_MAKE_INSTALL_SH} ${TD_SOURCE_DIR} ${PROJECT_BINARY_DIR} Linux ${TD_VER_NUMBER})")
 ELSEIF (TD_WINDOWS)
-  SET(CMAKE_INSTALL_PREFIX C:/TDengine)
-
-  # INSTALL(DIRECTORY ${TD_SOURCE_DIR}/src/connector/go     DESTINATION connector)
-  # INSTALL(DIRECTORY ${TD_SOURCE_DIR}/src/connector/nodejs DESTINATION connector)
-  # INSTALL(DIRECTORY ${TD_SOURCE_DIR}/src/connector/python DESTINATION connector)
-  # INSTALL(DIRECTORY ${TD_SOURCE_DIR}/src/connector/C\# DESTINATION connector)
-  # INSTALL(DIRECTORY ${TD_SOURCE_DIR}/examples DESTINATION .)
-  INSTALL(CODE "IF (NOT EXISTS ${CMAKE_INSTALL_PREFIX}/cfg/taos.cfg)
-    execute_process(COMMAND ${CMAKE_COMMAND} -E copy ${TD_SOURCE_DIR}/packaging/cfg/taos.cfg ${CMAKE_INSTALL_PREFIX}/cfg/taos.cfg)
-  ENDIF ()")
-  INSTALL(FILES ${TD_SOURCE_DIR}/include/client/taos.h DESTINATION include)
-  INSTALL(FILES ${TD_SOURCE_DIR}/include/util/taoserror.h DESTINATION include)
-  INSTALL(FILES ${TD_SOURCE_DIR}/include/libs/function/taosudf.h DESTINATION include)
-  INSTALL(FILES ${LIBRARY_OUTPUT_PATH}/taos.lib DESTINATION driver)
-  INSTALL(FILES ${LIBRARY_OUTPUT_PATH}/taos_static.lib DESTINATION driver)
-  INSTALL(FILES ${LIBRARY_OUTPUT_PATH}/taos.dll DESTINATION driver)
-  INSTALL(FILES ${EXECUTABLE_OUTPUT_PATH}/taos.exe DESTINATION .)
-  INSTALL(FILES ${EXECUTABLE_OUTPUT_PATH}/taosd.exe DESTINATION .)
-  INSTALL(FILES ${EXECUTABLE_OUTPUT_PATH}/udfd.exe DESTINATION .)
-  IF (BUILD_TOOLS)
-    INSTALL(FILES ${EXECUTABLE_OUTPUT_PATH}/taosBenchmark.exe DESTINATION .)
-  ENDIF ()
-
-  IF (TD_MVN_INSTALLED)
-    INSTALL(FILES ${LIBRARY_OUTPUT_PATH}/taos-jdbcdriver-2.0.38-dist.jar DESTINATION connector/jdbc)
-  ENDIF ()
   SET(TD_MAKE_INSTALL_SH "${TD_SOURCE_DIR}/packaging/tools/make_install.bat")
   INSTALL(CODE "MESSAGE(\"make install script: ${TD_MAKE_INSTALL_SH}\")")
   INSTALL(CODE "execute_process(COMMAND ${TD_MAKE_INSTALL_SH} :needAdmin ${TD_SOURCE_DIR} ${PROJECT_BINARY_DIR} Windows ${TD_VER_NUMBER})")

cmake/rust-bindings_CMakeLists.txt.in

-
-# rust-bindings
-ExternalProject_Add(rust-bindings
-        GIT_REPOSITORY https://github.com/songtianyi/tdengine-rust-bindings.git
-        GIT_TAG 7ed7a97
-        SOURCE_DIR "${TD_SOURCE_DIR}/examples/rust"
-        BINARY_DIR "${TD_SOURCE_DIR}/examples/rust"
-        CONFIGURE_COMMAND "" 
-        BUILD_COMMAND ""  
-        INSTALL_COMMAND ""
-        TEST_COMMAND ""
-        )

cmake/taosadapter_CMakeLists.txt.in

@@ -2,7 +2,7 @@
 # taosadapter
 ExternalProject_Add(taosadapter
         GIT_REPOSITORY https://github.com/taosdata/taosadapter.git
-        GIT_TAG 3d21433
+        GIT_TAG abed566
         SOURCE_DIR "${TD_SOURCE_DIR}/tools/taosadapter"
         BINARY_DIR ""
         #BUILD_IN_SOURCE TRUE

contrib/CMakeLists.txt

@@ -105,11 +105,6 @@ if(${BUILD_WITH_SQLITE})
     cat("${TD_SUPPORT_DIR}/sqlite_CMakeLists.txt.in" ${CONTRIB_TMP_FILE})
 endif(${BUILD_WITH_SQLITE})
 
-# rust-bindings
-if(${RUST_BINDINGS})
-    cat("${TD_SUPPORT_DIR}/rust-bindings_CMakeLists.txt.in" ${CONTRIB_TMP_FILE})
-endif(${RUST_BINDINGS})
-
 # lucene
 if(${BUILD_WITH_LUCENE})
     cat("${TD_SUPPORT_DIR}/lucene_CMakeLists.txt.in" ${CONTRIB_TMP_FILE})
@@ -140,24 +135,6 @@ execute_process(COMMAND "${CMAKE_COMMAND}" -G "${CMAKE_GENERATOR}" .
         WORKING_DIRECTORY "${TD_CONTRIB_DIR}/deps-download")
 execute_process(COMMAND "${CMAKE_COMMAND}" --build .
         WORKING_DIRECTORY "${TD_CONTRIB_DIR}/deps-download")
-     
-# clear submodule
-execute_process(COMMAND git submodule deinit -f tools/taos-tools
-        WORKING_DIRECTORY "${TD_SOURCE_DIR}")
-execute_process(COMMAND git rm --cached tools/taos-tools
-        WORKING_DIRECTORY "${TD_SOURCE_DIR}")
-execute_process(COMMAND git submodule deinit -f tools/taosadapter
-        WORKING_DIRECTORY "${TD_SOURCE_DIR}")
-execute_process(COMMAND git rm --cached tools/taosadapter
-        WORKING_DIRECTORY "${TD_SOURCE_DIR}")
-execute_process(COMMAND git submodule deinit -f tools/taosws-rs
-        WORKING_DIRECTORY "${TD_SOURCE_DIR}")
-execute_process(COMMAND git rm --cached tools/taosws-rs
-        WORKING_DIRECTORY "${TD_SOURCE_DIR}")
-execute_process(COMMAND git submodule deinit -f examples/rust
-        WORKING_DIRECTORY "${TD_SOURCE_DIR}")
-execute_process(COMMAND git rm --cached examples/rust
-        WORKING_DIRECTORY "${TD_SOURCE_DIR}")
 
 # ================================================================================================
 # Build
@@ -354,9 +331,11 @@ endif(${BUILD_WITH_TRAFT})
 
 # LIBUV
 if(${BUILD_WITH_UV})
-    if (NOT ${CMAKE_SYSTEM_NAME} MATCHES "Windows")
-        MESSAGE("Windows need set no-sign-compare")
-        add_compile_options(-Wno-sign-compare)
+    if (TD_WINDOWS)
+        # There is no GetHostNameW function on win7.
+        file(READ "libuv/src/win/util.c" LIBUV_WIN_UTIL_CONTENT)
+        string(REPLACE "if (GetHostNameW(buf, UV_MAXHOSTNAMESIZE" "DWORD  nSize = UV_MAXHOSTNAMESIZE;\n  if (GetComputerNameW(buf, &nSize" LIBUV_WIN_UTIL_CONTENT "${LIBUV_WIN_UTIL_CONTENT}")
+        file(WRITE "libuv/src/win/util.c" "${LIBUV_WIN_UTIL_CONTENT}")
     endif ()
     add_subdirectory(libuv EXCLUDE_FROM_ALL)
 endif(${BUILD_WITH_UV})

docs/en/01-index.md

@@ -4,7 +4,8 @@ sidebar_label: Documentation Home
 slug: /
 ---
 
-TDengine is a [high-performance](https://tdengine.com/fast), [scalable](https://tdengine.com/scalable) time series database with [SQL support](https://tdengine.com/sql-support). This document is the TDengine user manual. It introduces the basic, as well as novel concepts, in TDengine, and also talks in detail about installation, features, SQL, APIs, operation, maintenance, kernel design and other topics. It’s written mainly for architects, developers and system administrators.
+
+TDengine is an open source, cloud native time-series database optimized for Internet of Things (IoT), Connected Cars, and Industrial IoT. It enables efficient, real-time data ingestion, processing, and monitoring of TB and even PB scale data per day, generated by billions of sensors and data collectors.  This document is the TDengine user manual. It introduces the basic, as well as novel concepts, in TDengine, and also talks in detail about installation, features, SQL, APIs, operation, maintenance, kernel design and other topics. It’s written mainly for architects, developers and system administrators.
 
 To get an overview of TDengine, such as a feature list, benchmarks, and competitive advantages, please browse through the [Introduction](./intro) section.
 

docs/en/02-intro/index.md

@@ -3,7 +3,7 @@ title: Introduction
 toc_max_heading_level: 2
 ---
 
-TDengine is a high-performance, scalable time-series database with SQL support. Its code, including its cluster feature is open source under GNU AGPL v3.0. Besides the database engine, it provides [caching](/develop/cache), [stream processing](/develop/continuous-query), [data subscription](/develop/subscribe)  and other functionalities to reduce the complexity and cost of development and operation.
+TDengine is an open source, high-performance, cloud native time-series database optimized for Internet of Things (IoT), Connected Cars, and Industrial IoT. Its code, including its cluster feature is open source under GNU AGPL v3.0. Besides the database engine, it provides [caching](/develop/cache), [stream processing](/develop/continuous-query), [data subscription](/develop/subscribe) and other functionalities to reduce the system complexity and cost of development and operation.
 
 This section introduces the major features, competitive advantages, typical use-cases and benchmarks to help you get a high level overview of TDengine.
 
@@ -31,25 +31,21 @@ For more details on features, please read through the entire documentation.
 
 ## Competitive Advantages
 
-Time-series data is structured, not transactional, and is rarely deleted or updated. TDengine makes full use of [these characteristics of time series data](https://tdengine.com/2019/07/09/86.html) to build its own innovative storage engine and computing engine to differentiate itself from other time series databases, with the following advantages.
+By making full use of [characteristics of time series data](https://tdengine.com/tsdb/characteristics-of-time-series-data/), TDengine differentiates itself from other time series databases, with the following advantages.
 
-- **[High Performance](https://tdengine.com/fast)**: With an innovatively designed and purpose-built storage engine, TDengine outperforms other time series databases in data ingestion and querying while significantly reducing storage costs and compute costs.
+- **High-Performance**: TDengine is the only time-series database to solve the high cardinality issue to support billions of data collection points while out performing other time-series databases for data ingestion, querying and data compression.
 
-- **[Scalable](https://tdengine.com/scalable)**: TDengine provides out-of-box scalability and high-availability through its native distributed design. Nodes can be added through simple configuration to achieve greater data processing power. In addition, this feature is open source.
+- **Simplified Solution**: Through built-in caching, stream processing and data subscription features, TDengine provides a simplified solution for time-series data processing. It reduces system design complexity and operation costs significantly.
 
-- **[SQL Support](https://tdengine.com/sql-support)**: TDengine uses SQL as the query language, thereby reducing learning and migration costs, while adding SQL extensions to better handle time-series. Keeping NoSQL developers in mind, TDengine also supports convenient and flexible, schemaless data ingestion.
+- **Cloud Native**: Through native distributed design, sharding and partitioning, separation of compute and storage, RAFT, support for kubernetes deployment and full observability, TDengine is a cloud native Time-Series Database and can be deployed on public, private or hybrid clouds.
 
-- **All in One**: TDengine has built-in caching, stream processing and data subscription functions. It is no longer necessary to integrate Kafka/Redis/HBase/Spark or other software in some scenarios. It makes the system architecture much simpler, cost-effective and easier to maintain.
+- **Ease of Use**: For administrators, TDengine significantly reduces the effort to deploy and maintain. For developers, it provides a simple interface, simplified solution and seamless integrations for third party tools. For data users, it gives easy data access. 
 
-- **Seamless Integration**: Without a single line of code, TDengine provide seamless, configurable integration with third-party tools such as Telegraf, Grafana, EMQX, Prometheus, StatsD, collectd, etc. More third-party tools are being integrated.
+- **Easy Data Analytics**: Through super tables, storage and compute separation, data partitioning by time interval, pre-computation and other means, TDengine makes it easy to explore, format, and get access to data in a highly efficient way. 
 
-- **Zero Management**: Installation and cluster setup can be done in seconds. Data partitioning and sharding are executed automatically. TDengine’s running status can be monitored via Grafana or other DevOps tools.
+- **Open Source**: TDengine’s core modules, including cluster feature, are all available under open source licenses. It has gathered 18.8k stars on GitHub. There is an active developer community, and over 139k running instances worldwide.
 
-- **Zero Learning Costs**: With SQL as the query language and support for ubiquitous tools like Python, Java, C/C++, Go, Rust, and Node.js connectors, and a REST API, there are zero learning costs.
-
-- **Interactive Console**: TDengine provides convenient console access to the database, through a CLI, to run ad hoc queries, maintain the database, or manage the cluster, without any programming.
-
-With TDengine, the total cost of ownership of your time-series data platform can be greatly reduced. 1: With its superior performance, the computing and storage resources are reduced significantly 2: With SQL support, it can be seamlessly integrated with many third party tools, and learning costs/migration costs are reduced significantly 3: With its simple architecture and zero management, the operation and maintenance costs are reduced. 
+With TDengine, the total cost of ownership of your time-series data platform can be greatly reduced. 1: With its superior performance, the computing and storage resources are reduced significantly；2: With SQL support, it can be seamlessly integrated with many third party tools, and learning costs/migration costs are reduced significantly；3: With its simplified solution and nearly zero management, the operation and maintenance costs are reduced significantly. 
 
 ## Technical Ecosystem
 This is how TDengine would be situated, in a typical time-series data processing platform:

docs/en/04-concept/index.md

@@ -2,7 +2,7 @@
 title: Concepts
 ---
 
-In order to explain the basic concepts and provide some sample code, the TDengine documentation smart meters as a typical time series use case. We assume the following: 1. Each smart meter collects three metrics i.e. current, voltage, and phase 2. There are multiple smart meters, and 3. Each meter has static attributes like location and group ID. Based on this, collected data will look similar to the following table:
+In order to explain the basic concepts and provide some sample code, the TDengine documentation smart meters as a typical time series use case. We assume the following: 1. Each smart meter collects three metrics i.e. current, voltage, and phase; 2. There are multiple smart meters; 3. Each meter has static attributes like location and group ID. Based on this, collected data will look similar to the following table:
 
 <div className="center-table">
 <table>

docs/en/07-develop/02-model/index.mdx

@@ -2,14 +2,16 @@
 title: Data Model
 ---
 
-The data model employed by TDengine is similar to that of a relational database. You have to create databases and tables. You must design the data model based on your own business and application requirements. You should design the STable (an abbreviation for super table) schema to fit your data. This chapter will explain the big picture without getting into syntactical details.
+The data model employed by TDengine is similar to that of a relational database. You have to create databases and tables. You must design the data model based on your own business and application requirements. You should design the [STable](/concept/#super-table-stable) (an abbreviation for super table) schema to fit your data. This chapter will explain the big picture without getting into syntactical details. 
+
+Note: before you read this chapter, please make sure you have already read through [Key Concepts](/concept/), since TDengine introduces new concepts like "one table for one [data collection point](/concept/#data-collection-point)" and "[super table](/concept/#super-table-stable)".
 
 ## Create Database
 
 The [characteristics of time-series data](https://www.taosdata.com/blog/2019/07/09/86.html) from different data collection points may be different. Characteristics include collection frequency, retention policy and others which determine how you create and configure the database. For e.g. days to keep, number of replicas, data block size, whether data updates are allowed and other configurable parameters would be determined by the characteristics of your data and your business requirements. For TDengine to operate with the best performance, we strongly recommend that you create and configure different databases for data with different characteristics. This allows you, for example, to set up different storage and retention policies. When creating a database, there are a lot of parameters that can be configured such as, the days to keep data, the number of replicas, the number of memory blocks, time precision, the minimum and maximum number of rows in each data block, whether compression is enabled, the time range of the data in single data file and so on. Below is an example of the SQL statement to create a database.
 
 ```sql
-CREATE DATABASE power KEEP 365 DURATION 10 BUFFER 16 VGROUPS 100 WAL 1;
+CREATE DATABASE power KEEP 365 DURATION 10 BUFFER 16 WAL_LEVEL 1;
 ```
 
 In the above SQL statement:

docs/en/07-develop/04-query-data/index.mdx

@@ -61,20 +61,20 @@ In summary, records across subtables can be aggregated by a simple query on thei
 In TDengine CLI `taos`, use the SQL below to get the average voltage of all the meters in California grouped by location.
 
 ```
-taos> SELECT AVG(voltage) FROM meters GROUP BY location;
-       avg(voltage)        |            location            |
-=============================================================
-             222.000000000 | California.LosAngeles                |
-             219.200000000 | California.SanFrancisco               |
-Query OK, 2 row(s) in set (0.002136s)
+taos> SELECT AVG(voltage), location FROM meters GROUP BY location;
+       avg(voltage)        |                             location                             |
+===============================================================================================
+             219.200000000 | California.SanFrancisco                                          |
+             221.666666667 | California.LosAngeles                                            |
+Query OK, 2 rows in database (0.005995s)
 ```
 
 ### Example 2
 
-In TDengine CLI `taos`, use the SQL below to get the number of rows and the maximum current in the past 24 hours from meters whose groupId is 2.
+In TDengine CLI `taos`, use the SQL below to get the number of rows and the maximum current from meters whose groupId is 2.
 
 ```
-taos> SELECT count(*), max(current) FROM meters where groupId = 2 and ts > now - 24h;
+taos> SELECT count(*), max(current) FROM meters where groupId = 2;
      count(*)  |    max(current)  |
 ==================================
             5 |             13.4 |
@@ -88,40 +88,41 @@ Join queries are only allowed between subtables of the same STable. In [Select](
 In IoT use cases, down sampling is widely used to aggregate data by time range. The `INTERVAL` keyword in TDengine can be used to simplify the query by time window. For example, the SQL statement below can be used to get the sum of current every 10 seconds from meters table d1001.
 
 ```
-taos> SELECT sum(current) FROM d1001 INTERVAL(10s);
-           ts            |       sum(current)        |
+taos> SELECT _wstart, sum(current) FROM d1001 INTERVAL(10s);
+         _wstart         |       sum(current)        |
 ======================================================
  2018-10-03 14:38:00.000 |              10.300000191 |
  2018-10-03 14:38:10.000 |              24.900000572 |
-Query OK, 2 row(s) in set (0.000883s)
+Query OK, 2 rows in database (0.003139s)
 ```
 
 Down sampling can also be used for STable. For example, the below SQL statement can be used to get the sum of current from all meters in California.
 
 ```
-taos> SELECT SUM(current) FROM meters where location like "California%" INTERVAL(1s);
-           ts            |       sum(current)        |
+taos> SELECT _wstart, SUM(current) FROM meters where location like "California%" INTERVAL(1s);
+         _wstart         |       sum(current)        |
 ======================================================
  2018-10-03 14:38:04.000 |              10.199999809 |
- 2018-10-03 14:38:05.000 |              32.900000572 |
+ 2018-10-03 14:38:05.000 |              23.699999809 |
  2018-10-03 14:38:06.000 |              11.500000000 |
  2018-10-03 14:38:15.000 |              12.600000381 |
- 2018-10-03 14:38:16.000 |              36.000000000 |
-Query OK, 5 row(s) in set (0.001538s)
+ 2018-10-03 14:38:16.000 |              34.400000572 |
+Query OK, 5 rows in database (0.007413s)
 ```
 
 Down sampling also supports time offset. For example, the below SQL statement can be used to get the sum of current from all meters but each time window must start at the boundary of 500 milliseconds.
 
 ```
-taos> SELECT SUM(current) FROM meters INTERVAL(1s, 500a);
-           ts            |       sum(current)        |
+taos> SELECT _wstart, SUM(current) FROM meters INTERVAL(1s, 500a);
+         _wstart         |       sum(current)        |
 ======================================================
- 2018-10-03 14:38:04.500 |              11.189999809 |
- 2018-10-03 14:38:05.500 |              31.900000572 |
- 2018-10-03 14:38:06.500 |              11.600000000 |
- 2018-10-03 14:38:15.500 |              12.300000381 |
- 2018-10-03 14:38:16.500 |              35.000000000 |
-Query OK, 5 row(s) in set (0.001521s)
+ 2018-10-03 14:38:03.500 |              10.199999809 |
+ 2018-10-03 14:38:04.500 |              10.300000191 |
+ 2018-10-03 14:38:05.500 |              13.399999619 |
+ 2018-10-03 14:38:06.500 |              11.500000000 |
+ 2018-10-03 14:38:14.500 |              12.600000381 |
+ 2018-10-03 14:38:16.500 |              34.400000572 |
+Query OK, 6 rows in database (0.005515s)
 ```
 
 In many use cases, it's hard to align the timestamp of the data collected by each collection point. However, a lot of algorithms like FFT require the data to be aligned with same time interval and application programs have to handle this by themselves. In TDengine, it's easy to achieve the alignment using down sampling.

docs/en/14-reference/03-connector/java.mdx

@@ -130,7 +130,7 @@ The configuration parameters in the URL are as follows:
 - charset: The character set used by the client, the default value is the system character set.
 - locale: Client locale, by default, use the system's current locale.
 - timezone: The time zone used by the client, the default value is the system's current time zone.
-- batchfetch: true: pulls result sets in batches when executing queries; false: pulls result sets row by row. The default value is: false. Enabling batch pulling and obtaining a batch of data can improve query performance when the query data volume is large.
+- batchfetch: true: pulls result sets in batches when executing queries; false: pulls result sets row by row. The default value is: true. Enabling batch pulling and obtaining a batch of data can improve query performance when the query data volume is large.
 - batchErrorIgnore:true: When executing statement executeBatch, if there is a SQL execution failure in the middle, the following SQL will continue to be executed. false: No more statements after the failed SQL are executed. The default value is: false.
 
 For more information about JDBC native connections, see [Video Tutorial](https://www.taosdata.com/blog/2020/11/11/1955.html).

docs/examples/c/tmq_example.c

+/*
+ * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
+ *
+ * This program is free software: you can use, redistribute, and/or modify
+ * it under the terms of the GNU Affero General Public License, version 3
+ * or later ("AGPL"), as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <assert.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+#include "taos.h"
+
+static int  running = 1;
+static char dbName[64] = "tmqdb";
+static char stbName[64] = "stb";
+static char topicName[64] = "topicname";
+
+static int32_t msg_process(TAOS_RES* msg) {
+  char    buf[1024];
+  int32_t rows = 0;
+
+  const char* topicName = tmq_get_topic_name(msg);
+  const char* dbName = tmq_get_db_name(msg);
+  int32_t     vgroupId = tmq_get_vgroup_id(msg);
+
+  printf("topic: %s\n", topicName);
+  printf("db: %s\n", dbName);
+  printf("vgroup id: %d\n", vgroupId);
+
+  while (1) {
+    TAOS_ROW row = taos_fetch_row(msg);
+    if (row == NULL) break;
+
+    TAOS_FIELD* fields = taos_fetch_fields(msg);
+    int32_t     numOfFields = taos_field_count(msg);
+    int32_t*    length = taos_fetch_lengths(msg);
+    int32_t     precision = taos_result_precision(msg);
+    rows++;
+    taos_print_row(buf, row, fields, numOfFields);
+    printf("row content: %s\n", buf);
+  }
+
+  return rows;
+}
+
+static int32_t init_env() {
+  TAOS* pConn = taos_connect("localhost", "root", "taosdata", NULL, 0);
+  if (pConn == NULL) {
+    return -1;
+  }
+
+  TAOS_RES* pRes;
+  // drop database if exists
+  printf("create database\n");
+  pRes = taos_query(pConn, "drop database if exists tmqdb");
+  if (taos_errno(pRes) != 0) {
+    printf("error in drop tmqdb, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  // create database
+  pRes = taos_query(pConn, "create database tmqdb");
+  if (taos_errno(pRes) != 0) {
+    printf("error in create tmqdb, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  // create super table
+  printf("create super table\n");
+  pRes = taos_query(
+      pConn, "create table tmqdb.stb (ts timestamp, c1 int, c2 float, c3 varchar(16)) tags(t1 int, t3 varchar(16))");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create super table stb, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  // create sub tables
+  printf("create sub tables\n");
+  pRes = taos_query(pConn, "create table tmqdb.ctb0 using tmqdb.stb tags(0, 'subtable0')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create super table ctb0, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "create table tmqdb.ctb1 using tmqdb.stb tags(1, 'subtable1')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create super table ctb1, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "create table tmqdb.ctb2 using tmqdb.stb tags(2, 'subtable2')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create super table ctb2, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "create table tmqdb.ctb3 using tmqdb.stb tags(3, 'subtable3')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create super table ctb3, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  // insert data
+  printf("insert data into sub tables\n");
+  pRes = taos_query(pConn, "insert into tmqdb.ctb0 values(now, 0, 0, 'a0')(now+1s, 0, 0, 'a00')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to insert into ctb0, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "insert into tmqdb.ctb1 values(now, 1, 1, 'a1')(now+1s, 11, 11, 'a11')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to insert into ctb0, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "insert into tmqdb.ctb2 values(now, 2, 2, 'a1')(now+1s, 22, 22, 'a22')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to insert into ctb0, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "insert into tmqdb.ctb3 values(now, 3, 3, 'a1')(now+1s, 33, 33, 'a33')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to insert into ctb0, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  taos_close(pConn);
+  return 0;
+}
+
+int32_t create_topic() {
+  printf("create topic\n");
+  TAOS_RES* pRes;
+  TAOS*     pConn = taos_connect("localhost", "root", "taosdata", NULL, 0);
+  if (pConn == NULL) {
+    return -1;
+  }
+
+  pRes = taos_query(pConn, "use tmqdb");
+  if (taos_errno(pRes) != 0) {
+    printf("error in use tmqdb, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "create topic topicname as select ts, c1, c2, c3, tbname from tmqdb.stb where c1 > 1");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create topic topicname, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  taos_close(pConn);
+  return 0;
+}
+
+void tmq_commit_cb_print(tmq_t* tmq, int32_t code, void* param) {
+  printf("tmq_commit_cb_print() code: %d, tmq: %p, param: %p\n", code, tmq, param);
+}
+
+tmq_t* build_consumer() {
+  tmq_conf_res_t code;
+  tmq_conf_t*    conf = tmq_conf_new();
+  code = tmq_conf_set(conf, "enable.auto.commit", "true");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "auto.commit.interval.ms", "1000");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "group.id", "cgrpName");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "client.id", "user defined name");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "td.connect.user", "root");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "td.connect.pass", "taosdata");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "auto.offset.reset", "earliest");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "experimental.snapshot.enable", "false");
+  if (TMQ_CONF_OK != code) return NULL;
+
+  tmq_conf_set_auto_commit_cb(conf, tmq_commit_cb_print, NULL);
+
+  tmq_t* tmq = tmq_consumer_new(conf, NULL, 0);
+  tmq_conf_destroy(conf);
+  return tmq;
+}
+
+tmq_list_t* build_topic_list() {
+  tmq_list_t* topicList = tmq_list_new();
+  int32_t     code = tmq_list_append(topicList, "topicname");
+  if (code) {
+    return NULL;
+  }
+  return topicList;
+}
+
+void basic_consume_loop(tmq_t* tmq) {
+  int32_t totalRows = 0;
+  int32_t msgCnt = 0;
+  int32_t timeout = 5000;
+  while (running) {
+    TAOS_RES* tmqmsg = tmq_consumer_poll(tmq, timeout);
+    if (tmqmsg) {
+      msgCnt++;
+      totalRows += msg_process(tmqmsg);
+      taos_free_result(tmqmsg);
+    } else {
+      break;
+    }
+  }
+
+  fprintf(stderr, "%d msg consumed, include %d rows\n", msgCnt, totalRows);
+}
+
+int main(int argc, char* argv[]) {
+  int32_t code;
+
+  if (init_env() < 0) {
+    return -1;
+  }
+
+  if (create_topic() < 0) {
+    return -1;
+  }
+
+  tmq_t* tmq = build_consumer();
+  if (NULL == tmq) {
+    fprintf(stderr, "%% build_consumer() fail!\n");
+    return -1;
+  }
+
+  tmq_list_t* topic_list = build_topic_list();
+  if (NULL == topic_list) {
+    return -1;
+  }
+
+  if ((code = tmq_subscribe(tmq, topic_list))) {
+    fprintf(stderr, "%% Failed to tmq_subscribe(): %s\n", tmq_err2str(code));
+  }
+  tmq_list_destroy(topic_list);
+
+  basic_consume_loop(tmq);
+
+  code = tmq_consumer_close(tmq);
+  if (code) {
+    fprintf(stderr, "%% Failed to close consumer: %s\n", tmq_err2str(code));
+  } else {
+    fprintf(stderr, "%% Consumer closed\n");
+  }
+
+  return 0;
+}

docs/examples/node/nativeexample/subscribe_demo.js

@@ -28,8 +28,7 @@ function runConsumer() {
         console.log(msg.topicPartition);
         console.log(msg.block);
         console.log(msg.fields)
-        // fixme(@xiaolei): commented temp, should be fixed.
-        //consumer.commit(msg);
+        consumer.commit(msg);
         console.log(`=======consumer ${i} done`)
     }
 

docs/examples/node/package.json

@@ -4,7 +4,7 @@
   "main": "index.js",
   "license": "MIT",
   "dependencies": {
-    "@tdengine/client": "^3.0.0",
+    "@tdengine/client": "^3.0.1",
     "@tdengine/rest": "^3.0.0"
   }
 }

docs/zh/05-get-started/03-package.md

@@ -7,12 +7,13 @@ import Tabs from "@theme/Tabs";
 import TabItem from "@theme/TabItem";
 import PkgListV3 from "/components/PkgListV3";
 
-TDengine 完整的软件包包括服务端（taosd）、用于与第三方系统对接并提供 RESTful 接口的 taosAdapter、应用驱动（taosc）、命令行程序 (CLI，taos) 和一些工具软件，目前服务端 taosd 和 taosAdapter 仅在 Linux 系统上安装和运行，后续将支持 Windows、macOS 等系统。应用驱动 taosc 与 TDengine CLI 可以在 Windows 或 Linux 上安装和运行。TDengine 除了提供多种语言的连接器之外，还通过 [taosAdapter](../../reference/taosadapter/) 提供 [RESTful 接口](../../reference/rest-api/)。
+您可以[用 Docker 立即体验](../../get-started/docker/) TDengine。如果您希望对 TDengine 贡献代码或对内部实现感兴趣，请参考我们的 [TDengine GitHub 主页](https://github.com/taosdata/TDengine) 下载源码构建和安装.
 
-为方便使用，标准的服务端安装包包含了 taos、taosd、taosAdapter、taosdump、taosBenchmark、TDinsight 安装脚本和示例代码；如果您只需要用到服务端程序和客户端连接的 C/C++ 语言支持，也可以仅下载 lite 版本的安装包。
+TDengine 完整的软件包包括服务端（taosd）、用于与第三方系统对接并提供 RESTful 接口的 taosAdapter、应用驱动（taosc）、命令行程序 (CLI，taos) 和一些工具软件。目前 taosAdapter 仅在 Linux 系统上安装和运行，后续将支持 Windows、macOS 等系统。TDengine 除了提供多种语言的连接器之外，还通过 [taosAdapter](../../reference/taosadapter/) 提供 [RESTful 接口](../../reference/rest-api/)。
 
-在 Linux 系统上，TDengine 开源版本提供 deb 和 rpm 格式安装包，用户可以根据自己的运行环境选择合适的安装包。其中 deb 支持 Debian/Ubuntu 及衍生系统，rpm 支持 CentOS/RHEL/SUSE 及衍生系统。同时我们也为企业用户提供 tar.gz 格式安装包，也支持通过 `apt-get` 工具从线上进行安装。TDengine 也提供 Windows x64 平台的安装包。您也可以[用 Docker 立即体验](../../get-started/docker/)。需要注意的是，rpm 和 deb 包不含 taosdump 和 TDinsight 安装脚本，这些工具需要通过安装 taosTool 包获得。如果您希望对 TDengine 贡献代码或对内部实现感兴趣，请参考我们的 [TDengine GitHub 主页](https://github.com/taosdata/TDengine) 下载源码构建和安装.
+为方便使用，标准的服务端安装包包含了 taosd、taosAdapter、taosc、taos、taosdump、taosBenchmark、TDinsight 安装脚本和示例代码；如果您只需要用到服务端程序和客户端连接的 C/C++ 语言支持，也可以仅下载 lite 版本的安装包。
 
+在 Linux 系统上，TDengine 开源版本提供 deb 和 rpm 格式安装包，用户可以根据自己的运行环境选择合适的安装包。其中 deb 支持 Debian/Ubuntu 及衍生系统，rpm 支持 CentOS/RHEL/SUSE 及衍生系统。同时我们也为企业用户提供 tar.gz 格式安装包，也支持通过 `apt-get` 工具从线上进行安装。需要注意的是，rpm 和 deb 包不含 taosdump 和 TDinsight 安装脚本，这些工具需要通过安装 taosTool 包获得。TDengine 也提供 Windows x64 平台的安装包。
 
 ## 安装
 
@@ -205,7 +206,7 @@ Query OK, 2 row(s) in set (0.003128s)
 
 ## 使用 taosBenchmark 体验写入速度
 
-启动 TDengine 的服务，在 Linux 终端执行 `taosBenchmark` （曾命名为 `taosdemo`）：
+启动 TDengine 的服务，在 Linux 或 windows 终端执行 `taosBenchmark` （曾命名为 `taosdemo`）：
 
 ```bash
 taosBenchmark

docs/zh/07-develop/02-model/index.mdx

@@ -11,10 +11,10 @@ TDengine 采用类关系型数据模型，需要建库、建表。因此对于
 不同类型的数据采集点往往具有不同的数据特征，包括数据采集频率的高低，数据保留时间的长短，副本的数目，数据块的大小，是否允许更新数据等等。为了在各种场景下 TDengine 都能最大效率的工作，TDengine 建议将不同数据特征的表创建在不同的库里，因为每个库可以配置不同的存储策略。创建一个库时，除 SQL 标准的选项外，还可以指定保留时长、副本数、缓存大小、时间精度、文件块里最大最小记录条数、是否压缩、一个数据文件覆盖的天数等多种参数。比如：
 
 ```sql
-CREATE DATABASE power KEEP 365 DURATION 10 BUFFER 16 VGROUPS 100 WAL 1;
+CREATE DATABASE power KEEP 365 DURATION 10 BUFFER 16 WAL_LEVEL 1;
 ```
 
-上述语句将创建一个名为 power 的库，这个库的数据将保留 365 天（超过 365 天将被自动删除），每 10 天一个数据文件，每个 VNODE 的写入内存池的大小为 16 MB，数据库的 VGROUPS 数量，对该数据库入会写 WAL 但不执行 FSYNC。详细的语法及参数请见 [数据库管理](/taos-sql/database) 章节。
+上述语句将创建一个名为 power 的库，这个库的数据将保留 365 天（超过 365 天将被自动删除），每 10 天一个数据文件，每个 VNODE 的写入内存池的大小为 16 MB，对该数据库入会写 WAL 但不执行 FSYNC。详细的语法及参数请见 [数据库管理](/taos-sql/database) 章节。
 
 创建库之后，需要使用 SQL 命令 `USE` 将当前库切换过来，例如：
 

docs/zh/07-develop/04-query-data/index.mdx

@@ -54,20 +54,20 @@ Query OK, 2 row(s) in set (0.001100s)
 在 TAOS Shell，查找加利福尼亚州所有智能电表采集的电压平均值，并按照 location 分组。
 
 ```
-taos> SELECT AVG(voltage) FROM meters GROUP BY location;
-       avg(voltage)        |            location            |
-=============================================================
-             222.000000000 | California.LosAngeles                |
-             219.200000000 | California.SanFrancisco               |
-Query OK, 2 row(s) in set (0.002136s)
+taos> SELECT AVG(voltage), location FROM meters GROUP BY location;
+       avg(voltage)        |                             location                             |
+===============================================================================================
+             219.200000000 | California.SanFrancisco                                          |
+             221.666666667 | California.LosAngeles                                            |
+Query OK, 2 rows in database (0.005995s)
 ```
 
 ### 示例二
 
-在 TAOS shell, 查找 groupId 为 2 的所有智能电表过去 24 小时的记录条数，电流的最大值。
+在 TAOS shell, 查找 groupId 为 2 的所有智能电表的记录条数，电流的最大值。
 
 ```
-taos> SELECT count(*), max(current) FROM meters where groupId = 2 and ts > now - 24h;
+taos> SELECT count(*), max(current) FROM meters where groupId = 2;
      cunt(*)  |    max(current)  |
 ==================================
             5 |             13.4 |
@@ -81,40 +81,41 @@ Query OK, 1 row(s) in set (0.002136s)
 物联网场景里，经常需要通过降采样（down sampling）将采集的数据按时间段进行聚合。TDengine 提供了一个简便的关键词 interval 让按照时间窗口的查询操作变得极为简单。比如，将智能电表 d1001 采集的电流值每 10 秒钟求和
 
 ```
-taos> SELECT sum(current) FROM d1001 INTERVAL(10s);
-           ts            |       sum(current)        |
+taos> SELECT _wstart, sum(current) FROM d1001 INTERVAL(10s);
+         _wstart         |       sum(current)        |
 ======================================================
  2018-10-03 14:38:00.000 |              10.300000191 |
  2018-10-03 14:38:10.000 |              24.900000572 |
-Query OK, 2 row(s) in set (0.000883s)
+Query OK, 2 rows in database (0.003139s)
 ```
 
 降采样操作也适用于超级表，比如：将加利福尼亚州所有智能电表采集的电流值每秒钟求和
 
 ```
-taos> SELECT SUM(current) FROM meters where location like "California%" INTERVAL(1s);
-           ts            |       sum(current)        |
+taos> SELECT _wstart, SUM(current) FROM meters where location like "California%" INTERVAL(1s);
+         _wstart         |       sum(current)        |
 ======================================================
  2018-10-03 14:38:04.000 |              10.199999809 |
- 2018-10-03 14:38:05.000 |              32.900000572 |
+ 2018-10-03 14:38:05.000 |              23.699999809 |
  2018-10-03 14:38:06.000 |              11.500000000 |
  2018-10-03 14:38:15.000 |              12.600000381 |
- 2018-10-03 14:38:16.000 |              36.000000000 |
-Query OK, 5 row(s) in set (0.001538s)
+ 2018-10-03 14:38:16.000 |              34.400000572 |
+Query OK, 5 rows in database (0.007413s)
 ```
 
 降采样操作也支持时间偏移，比如：将所有智能电表采集的电流值每秒钟求和，但要求每个时间窗口从 500 毫秒开始
 
 ```
-taos> SELECT SUM(current) FROM meters INTERVAL(1s, 500a);
-           ts            |       sum(current)        |
+taos> SELECT _wstart, SUM(current) FROM meters INTERVAL(1s, 500a);
+         _wstart         |       sum(current)        |
 ======================================================
- 2018-10-03 14:38:04.500 |              11.189999809 |
- 2018-10-03 14:38:05.500 |              31.900000572 |
- 2018-10-03 14:38:06.500 |              11.600000000 |
- 2018-10-03 14:38:15.500 |              12.300000381 |
- 2018-10-03 14:38:16.500 |              35.000000000 |
-Query OK, 5 row(s) in set (0.001521s)
+ 2018-10-03 14:38:03.500 |              10.199999809 |
+ 2018-10-03 14:38:04.500 |              10.300000191 |
+ 2018-10-03 14:38:05.500 |              13.399999619 |
+ 2018-10-03 14:38:06.500 |              11.500000000 |
+ 2018-10-03 14:38:14.500 |              12.600000381 |
+ 2018-10-03 14:38:16.500 |              34.400000572 |
+Query OK, 6 rows in database (0.005515s)
 ```
 
 物联网场景里，每个数据采集点采集数据的时间是难同步的，但很多分析算法(比如 FFT)需要把采集的数据严格按照时间等间隔的对齐，在很多系统里，需要应用自己写程序来处理，但使用 TDengine 的降采样操作就轻松解决。

docs/zh/07-develop/07-tmq.mdx

@@ -724,7 +724,6 @@ consumer.close();
 
 </TabItem>
 
-
 <TabItem value="Go" label="Go">
 
 ```go
@@ -769,6 +768,7 @@ consumer.Unsubscribe();
 // 关闭消费
 consumer.Close();
 ```
+
 </TabItem>
 
 </Tabs>
@@ -807,300 +807,9 @@ SHOW SUBSCRIPTIONS;
 以下是各语言的完整示例代码。
 
 <Tabs defaultValue="java" groupId="lang">
-<TabItem label="C" value="c">
-
-```c
-/*
- * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
- *
- * This program is free software: you can use, redistribute, and/or modify
- * it under the terms of the GNU Affero General Public License, version 3
- * or later ("AGPL"), as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.
- *
- * You should have received a copy of the GNU Affero General Public License
- * along with this program. If not, see <http://www.gnu.org/licenses/>.
- */
-
-#include <assert.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-#include <time.h>
-#include "taos.h"
-
-static int  running = 1;
-static char dbName[64] = "tmqdb";
-static char stbName[64] = "stb";
-static char topicName[64] = "topicname";
-
-static int32_t msg_process(TAOS_RES* msg) {
-  char    buf[1024];
-  int32_t rows = 0;
-
-  const char* topicName = tmq_get_topic_name(msg);
-  const char* dbName = tmq_get_db_name(msg);
-  int32_t     vgroupId = tmq_get_vgroup_id(msg);
-
-  printf("topic: %s\n", topicName);
-  printf("db: %s\n", dbName);
-  printf("vgroup id: %d\n", vgroupId);
-
-  while (1) {
-    TAOS_ROW row = taos_fetch_row(msg);
-    if (row == NULL) break;
-
-    TAOS_FIELD* fields = taos_fetch_fields(msg);
-    int32_t     numOfFields = taos_field_count(msg);
-    int32_t*    length = taos_fetch_lengths(msg);
-    int32_t     precision = taos_result_precision(msg);
-    const char* tbName = tmq_get_table_name(msg);
-    rows++;
-    taos_print_row(buf, row, fields, numOfFields);
-    printf("row content from %s: %s\n", (tbName != NULL ? tbName : "table null"), buf);
-  }
-
-  return rows;
-}
-
-static int32_t init_env() {
-  TAOS* pConn = taos_connect("localhost", "root", "taosdata", NULL, 0);
-  if (pConn == NULL) {
-    return -1;
-  }
-
-  TAOS_RES* pRes;
-  // drop database if exists
-  printf("create database\n");
-  pRes = taos_query(pConn, "drop database if exists tmqdb");
-  if (taos_errno(pRes) != 0) {
-    printf("error in drop tmqdb, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  // create database
-  pRes = taos_query(pConn, "create database tmqdb");
-  if (taos_errno(pRes) != 0) {
-    printf("error in create tmqdb, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  // create super table
-  printf("create super table\n");
-  pRes = taos_query(
-      pConn, "create table tmqdb.stb (ts timestamp, c1 int, c2 float, c3 varchar(16)) tags(t1 int, t3 varchar(16))");
-  if (taos_errno(pRes) != 0) {
-    printf("failed to create super table stb, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  // create sub tables
-  printf("create sub tables\n");
-  pRes = taos_query(pConn, "create table tmqdb.ctb0 using tmqdb.stb tags(0, 'subtable0')");
-  if (taos_errno(pRes) != 0) {
-    printf("failed to create super table ctb0, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  pRes = taos_query(pConn, "create table tmqdb.ctb1 using tmqdb.stb tags(1, 'subtable1')");
-  if (taos_errno(pRes) != 0) {
-    printf("failed to create super table ctb1, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  pRes = taos_query(pConn, "create table tmqdb.ctb2 using tmqdb.stb tags(2, 'subtable2')");
-  if (taos_errno(pRes) != 0) {
-    printf("failed to create super table ctb2, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  pRes = taos_query(pConn, "create table tmqdb.ctb3 using tmqdb.stb tags(3, 'subtable3')");
-  if (taos_errno(pRes) != 0) {
-    printf("failed to create super table ctb3, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  // insert data
-  printf("insert data into sub tables\n");
-  pRes = taos_query(pConn, "insert into tmqdb.ctb0 values(now, 0, 0, 'a0')(now+1s, 0, 0, 'a00')");
-  if (taos_errno(pRes) != 0) {
-    printf("failed to insert into ctb0, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  pRes = taos_query(pConn, "insert into tmqdb.ctb1 values(now, 1, 1, 'a1')(now+1s, 11, 11, 'a11')");
-  if (taos_errno(pRes) != 0) {
-    printf("failed to insert into ctb0, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  pRes = taos_query(pConn, "insert into tmqdb.ctb2 values(now, 2, 2, 'a1')(now+1s, 22, 22, 'a22')");
-  if (taos_errno(pRes) != 0) {
-    printf("failed to insert into ctb0, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  pRes = taos_query(pConn, "insert into tmqdb.ctb3 values(now, 3, 3, 'a1')(now+1s, 33, 33, 'a33')");
-  if (taos_errno(pRes) != 0) {
-    printf("failed to insert into ctb0, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  taos_close(pConn);
-  return 0;
-}
-
-int32_t create_topic() {
-  printf("create topic\n");
-  TAOS_RES* pRes;
-  TAOS*     pConn = taos_connect("localhost", "root", "taosdata", NULL, 0);
-  if (pConn == NULL) {
-    return -1;
-  }
-
-  pRes = taos_query(pConn, "use tmqdb");
-  if (taos_errno(pRes) != 0) {
-    printf("error in use tmqdb, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  pRes = taos_query(pConn, "create topic topicname as select ts, c1, c2, c3 from tmqdb.stb where c1 > 1");
-  if (taos_errno(pRes) != 0) {
-    printf("failed to create topic topicname, reason:%s\n", taos_errstr(pRes));
-    return -1;
-  }
-  taos_free_result(pRes);
-
-  taos_close(pConn);
-  return 0;
-}
-
-void tmq_commit_cb_print(tmq_t* tmq, int32_t code, void* param) {
-  printf("tmq_commit_cb_print() code: %d, tmq: %p, param: %p\n", code, tmq, param);
-}
-
-tmq_t* build_consumer() {
-  tmq_conf_res_t code;
-  tmq_conf_t*    conf = tmq_conf_new();
-  code = tmq_conf_set(conf, "enable.auto.commit", "true");
-  if (TMQ_CONF_OK != code) return NULL;
-  code = tmq_conf_set(conf, "auto.commit.interval.ms", "1000");
-  if (TMQ_CONF_OK != code) return NULL;
-  code = tmq_conf_set(conf, "group.id", "cgrpName");
-  if (TMQ_CONF_OK != code) return NULL;
-  code = tmq_conf_set(conf, "client.id", "user defined name");
-  if (TMQ_CONF_OK != code) return NULL;
-  code = tmq_conf_set(conf, "td.connect.user", "root");
-  if (TMQ_CONF_OK != code) return NULL;
-  code = tmq_conf_set(conf, "td.connect.pass", "taosdata");
-  if (TMQ_CONF_OK != code) return NULL;
-  code = tmq_conf_set(conf, "auto.offset.reset", "earliest");
-  if (TMQ_CONF_OK != code) return NULL;
-  code = tmq_conf_set(conf, "experimental.snapshot.enable", "true");
-  if (TMQ_CONF_OK != code) return NULL;
-  code = tmq_conf_set(conf, "msg.with.table.name", "true");
-  if (TMQ_CONF_OK != code) return NULL;
-
-  tmq_conf_set_auto_commit_cb(conf, tmq_commit_cb_print, NULL);
-
-  tmq_t* tmq = tmq_consumer_new(conf, NULL, 0);
-  tmq_conf_destroy(conf);
-  return tmq;
-}
-
-tmq_list_t* build_topic_list() {
-  tmq_list_t* topicList = tmq_list_new();
-  int32_t     code = tmq_list_append(topicList, "topicname");
-  if (code) {
-    return NULL;
-  }
-  return topicList;
-}
-
-void basic_consume_loop(tmq_t* tmq, tmq_list_t* topicList) {
-  int32_t code;
-
-  if ((code = tmq_subscribe(tmq, topicList))) {
-    fprintf(stderr, "%% Failed to tmq_subscribe(): %s\n", tmq_err2str(code));
-    return;
-  }
-
-  int32_t totalRows = 0;
-  int32_t msgCnt = 0;
-  int32_t timeout = 5000;
-  while (running) {
-    TAOS_RES* tmqmsg = tmq_consumer_poll(tmq, timeout);
-    if (tmqmsg) {
-      msgCnt++;
-      totalRows += msg_process(tmqmsg);
-      taos_free_result(tmqmsg);
-      /*} else {*/
-      /*break;*/
-    }
-  }
-
-  fprintf(stderr, "%d msg consumed, include %d rows\n", msgCnt, totalRows);
-}
-
-int main(int argc, char* argv[]) {
-  int32_t code;
-
-  if (init_env() < 0) {
-    return -1;
-  }
-
-  if (create_topic() < 0) {
-    return -1;
-  }
-
-  tmq_t* tmq = build_consumer();
-  if (NULL == tmq) {
-    fprintf(stderr, "%% build_consumer() fail!\n");
-    return -1;
-  }
-
-  tmq_list_t* topic_list = build_topic_list();
-  if (NULL == topic_list) {
-    return -1;
-  }
 
-  basic_consume_loop(tmq, topic_list);
-
-  code = tmq_unsubscribe(tmq);
-  if (code) {
-    fprintf(stderr, "%% Failed to unsubscribe: %s\n", tmq_err2str(code));
-  } else {
-    fprintf(stderr, "%% unsubscribe\n");
-  }
-
-  code = tmq_consumer_close(tmq);
-  if (code) {
-    fprintf(stderr, "%% Failed to close consumer: %s\n", tmq_err2str(code));
-  } else {
-    fprintf(stderr, "%% Consumer closed\n");
-  }
-
-  return 0;
-}
-
-```
-
-[查看源码](https://github.com/taosdata/TDengine/blob/develop/examples/c/tmq.c)
+<TabItem label="C" value="c">
+  <CDemo />
 </TabItem>
 
 <TabItem label="Java" value="java">
@@ -1116,22 +825,7 @@ int main(int argc, char* argv[]) {
 </TabItem>
 
 <TabItem label="Python" value="Python">
-
-```python
-import taos
-from taos.tmq import TaosConsumer
-
-import taos
-from taos.tmq import *
-consumer = TaosConsumer('topic_ctb_column', group_id='vg2')
-for msg in consumer:
-    for row in msg:
-        print(row)
-
-```
-
-[查看源码](https://github.com/taosdata/TDengine/blob/develop/docs/examples/python/tmq_example.py)
-
+    <Python />
 </TabItem>
 
 <TabItem label="Node.JS" value="Node.JS">

docs/zh/07-develop/_sub_c.mdx

 ```c
-{{#include docs/examples/c/subscribe_demo.c}}
-```
+{{#include docs/examples/c/tmq_example.c}}
+```
\ No newline at end of file

docs/zh/07-develop/_sub_python.mdx

 ```py
-{{#include docs/examples/python/subscribe_demo.py}}
-```
+{{#include docs/examples/python/tmq_example.py}}
+```
\ No newline at end of file

docs/zh/12-taos-sql/03-table.md

@@ -110,7 +110,7 @@ alter_table_option: {
 对普通表可以进行如下修改操作
 1. ADD COLUMN：添加列。
 2. DROP COLUMN：删除列。
-3. ODIFY COLUMN：修改列定义，如果数据列的类型是可变长类型，那么可以使用此指令修改其宽度，只能改大，不能改小。
+3. MODIFY COLUMN：修改列定义，如果数据列的类型是可变长类型，那么可以使用此指令修改其宽度，只能改大，不能改小。
 4. RENAME COLUMN：修改列名称。
 
 ### 增加列
@@ -195,4 +195,4 @@ SHOW CREATE TABLE tb_name;
 
 ```
 DESCRIBE [db_name.]tb_name;
-```
+```
\ No newline at end of file

docs/zh/14-reference/03-connector/java.mdx

@@ -131,7 +131,7 @@ url 中的配置参数如下：
 - charset：客户端使用的字符集，默认值为系统字符集。
 - locale：客户端语言环境，默认值系统当前 locale。
 - timezone：客户端使用的时区，默认值为系统当前时区。
-- batchfetch: true：在执行查询时批量拉取结果集；false：逐行拉取结果集。默认值为：false。开启批量拉取同时获取一批数据在查询数据量较大时批量拉取可以有效的提升查询性能。
+- batchfetch: true：在执行查询时批量拉取结果集；false：逐行拉取结果集。默认值为：true。开启批量拉取同时获取一批数据在查询数据量较大时批量拉取可以有效的提升查询性能。
 - batchErrorIgnore：true：在执行 Statement 的 executeBatch 时，如果中间有一条 SQL 执行失败将继续执行下面的 SQL。false：不再执行失败 SQL 后的任何语句。默认值为：false。
 
 JDBC 原生连接的使用请参见[视频教程](https://www.taosdata.com/blog/2020/11/11/1955.html)。

docs/zh/14-reference/15-taosKeeper.md → docs/zh/14-reference/14-taosKeeper.md

@@ -25,7 +25,7 @@ taosKeeper 安装方式：
 <!-- taosKeeper 需要在操作系统终端执行，该工具支持两种配置方式：[命令行参数](#命令行参数启动) 和 [配置文件](#配置文件启动)。命令行参数优先级高于配置文件参数。-->
 taosKeeper 需要在操作系统终端执行，该工具支持 [配置文件启动](#配置文件启动)。
 
-**在运行 taosKeeper 之前要确保 TDengine 集群与 taosAdapter 已经在正确运行。**
+**在运行 taosKeeper 之前要确保 TDengine 集群与 taosAdapter 已经在正确运行。** 并且 TDengine 已经开启监控服务，具体请参考：[TDengine 监控配置](../config/#监控相关)。
 
 <!--
 ### 命令行参数启动
@@ -93,7 +93,7 @@ taosKeeper 作为 TDengine 监控指标的导出工具，可以将 TDengine 产
 
 ```shell
 $ taos
-# 
+# 如上示例，使用 log 库作为监控日志存储位置
 > use log;
 > select * from cluster_info limit 1;
 ```

docs/zh/14-reference/14-taosx.md

----
-sidebar_label: taosX
-title: 使用 taosX 在集群间复制数据
----
\ No newline at end of file

docs/zh/17-operation/03-tolerance.md

@@ -26,5 +26,3 @@ TDengine 集群中的时序数据的副本数是与数据库关联的，一个
 TDengine 集群的节点数必须大于等于副本数，否则创建表时将报错。
 
 当 TDengine 集群中的节点部署在不同的物理机上，并设置多个副本数时，就实现了系统的高可靠性，无需再使用其他软件或工具。TDengine 企业版还可以将副本部署在不同机房，从而实现异地容灾。
-
-另外一种灾备方式是通过 `taosX` 将一个 TDengine 集群的数据同步复制到物理上位于不同数据中心的另一个 TDengine 集群。其详细使用方法请参考 [taosX 参考手册](../../reference/taosX)

docs/zh/20-third-party/01-grafana.mdx

@@ -193,7 +193,7 @@ docker run -d \
 
 如上图所示，在 Query 中选中 `TDengine` 数据源，在下方查询框可输入相应 SQL 进行查询，具体说明如下：
 
-- INPUT SQL：输入要查询的语句（该 SQL 语句的结果集应为两列多行），例如：`select avg(mem_system) from log.dn where ts >= $from and ts < $to interval($interval)` ，其中，from、to 和 interval 为 TDengine 插件的内置变量，表示从 Grafana 插件面板获取的查询范围和时间间隔。除了内置变量外，`也支持可以使用自定义模板变量`。
+- INPUT SQL：输入要查询的语句（该 SQL 语句的结果集应为两列多行），例如：`select _wstart, avg(mem_system) from log.dnodes_info where ts >= $from and ts < $to interval($interval)` ，其中，from、to 和 interval 为 TDengine 插件的内置变量，表示从 Grafana 插件面板获取的查询范围和时间间隔。除了内置变量外，`也支持可以使用自定义模板变量`。
 - ALIAS BY：可设置当前查询别名。
 - GENERATE SQL： 点击该按钮会自动替换相应变量，并生成最终执行的语句。
 
@@ -205,7 +205,11 @@ docker run -d \
 
 ### 导入 Dashboard
 
-在数据源配置页面，您可以为该数据源导入 TDinsight 面板，作为 TDengine 集群的监控可视化工具。该 Dashboard 已发布在 Grafana：[Dashboard 15167 - TDinsight](https://grafana.com/grafana/dashboards/15167)) 。其他安装方式和相关使用说明请见 [TDinsight 用户手册](/reference/tdinsight/)。
+在数据源配置页面，您可以为该数据源导入 TDinsight 面板，作为 TDengine 集群的监控可视化工具。如果 TDengine 服务端为 3.0 版本请选择 `TDinsight for 3.x` 导入。
+
+![TDengine Database Grafana plugine import dashboard](./import_dashboard.webp)
+
+其中适配 TDengine 2.* 的 Dashboard 已发布在 Grafana：[Dashboard 15167 - TDinsight](https://grafana.com/grafana/dashboards/15167)) 。其他安装方式和相关使用说明请见 [TDinsight 用户手册](/reference/tdinsight/)。
 
 使用 TDengine 作为数据源的其他面板，可以[在此搜索](https://grafana.com/grafana/dashboards/?dataSource=tdengine-datasource)。以下是一份不完全列表：
 

examples/c/stream_demo.c

@@ -98,10 +98,9 @@ int32_t create_stream() {
   /*const char* sql = "select min(k), max(k), sum(k) as sum_of_k from st1";*/
   /*const char* sql = "select sum(k) from tu1 interval(10m)";*/
   /*pRes = tmq_create_stream(pConn, "stream1", "out1", sql);*/
-  pRes =
-      taos_query(pConn,
-                 "create stream stream1 trigger max_delay 10s into outstb as select _wstart, sum(k) from st1 partition "
-                 "by tbname session(ts, 10s) ");
+  pRes = taos_query(pConn,
+                    "create stream stream1 trigger max_delay 10s watermark 10s into outstb as select _wstart start, "
+                    "count(k) from st1 partition by tbname interval(20s) ");
   if (taos_errno(pRes) != 0) {
     printf("failed to create stream stream1, reason:%s\n", taos_errstr(pRes));
     return -1;

examples/rust/.gitignore

+/target
+Cargo.lock

examples/rust/Cargo.toml

+[package]
+name = "rust"
+version = "0.1.0"
+edition = "2021"
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]
+taos = "*"
+
+[dev-dependencies]
+chrono = "0.4"
+itertools = "0.10.3"
+pretty_env_logger = "0.4.0"
+serde = { version = "1", features = ["derive"] }
+serde_json = "1"
+tokio = { version = "1", features = ["full"] }
+anyhow = "1"

examples/rust/examples/bind-tags.rs

+use anyhow::Result;
+use serde::Deserialize;
+use taos::*;
+
+#[tokio::main]
+async fn main() -> Result<()> {
+    let taos = TaosBuilder::from_dsn("taos://")?.build()?;
+    taos.exec_many([
+        "drop database if exists test",
+        "create database test keep 36500",
+        "use test",
+        "create table tb1 (ts timestamp, c1 bool, c2 tinyint, c3 smallint, c4 int, c5 bigint,
+            c6 tinyint unsigned, c7 smallint unsigned, c8 int unsigned, c9 bigint unsigned,
+            c10 float, c11 double, c12 varchar(100), c13 nchar(100)) tags(t1 varchar(100))",
+    ])
+    .await?;
+    let mut stmt = Stmt::init(&taos)?;
+    stmt.prepare(
+        "insert into ? using tb1 tags(?) values(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)",
+    )?;
+    stmt.set_tbname("d0")?;
+    stmt.set_tags(&[Value::VarChar("涛思".to_string())])?;
+
+    let params = vec![
+        ColumnView::from_millis_timestamp(vec![164000000000]),
+        ColumnView::from_bools(vec![true]),
+        ColumnView::from_tiny_ints(vec![i8::MAX]),
+        ColumnView::from_small_ints(vec![i16::MAX]),
+        ColumnView::from_ints(vec![i32::MAX]),
+        ColumnView::from_big_ints(vec![i64::MAX]),
+        ColumnView::from_unsigned_tiny_ints(vec![u8::MAX]),
+        ColumnView::from_unsigned_small_ints(vec![u16::MAX]),
+        ColumnView::from_unsigned_ints(vec![u32::MAX]),
+        ColumnView::from_unsigned_big_ints(vec![u64::MAX]),
+        ColumnView::from_floats(vec![f32::MAX]),
+        ColumnView::from_doubles(vec![f64::MAX]),
+        ColumnView::from_varchar(vec!["ABC"]),
+        ColumnView::from_nchar(vec!["涛思数据"]),
+    ];
+    let rows = stmt.bind(&params)?.add_batch()?.execute()?;
+    assert_eq!(rows, 1);
+
+    #[derive(Debug, Deserialize)]
+    #[allow(dead_code)]
+    struct Row {
+        ts: String,
+        c1: bool,
+        c2: i8,
+        c3: i16,
+        c4: i32,
+        c5: i64,
+        c6: u8,
+        c7: u16,
+        c8: u32,
+        c9: u64,
+        c10: Option<f32>,
+        c11: f64,
+        c12: String,
+        c13: String,
+        t1: serde_json::Value,
+    }
+
+    let rows: Vec<Row> = taos
+        .query("select * from tb1")
+        .await?
+        .deserialize()
+        .try_collect()
+        .await?;
+    let row = &rows[0];
+    dbg!(&row);
+    assert_eq!(row.c5, i64::MAX);
+    assert_eq!(row.c8, u32::MAX);
+    assert_eq!(row.c9, u64::MAX);
+    assert_eq!(row.c10.unwrap(), f32::MAX);
+    // assert_eq!(row.c11, f64::MAX);
+    assert_eq!(row.c12, "ABC");
+    assert_eq!(row.c13, "涛思数据");
+
+    Ok(())
+}

examples/rust/examples/bind.rs

+use anyhow::Result;
+use serde::Deserialize;
+use taos::*;
+
+#[tokio::main]
+async fn main() -> Result<()> {
+    let taos = TaosBuilder::from_dsn("taos://")?.build()?;
+    taos.exec_many([
+        "drop database if exists test_bindable",
+        "create database test_bindable keep 36500",
+        "use test_bindable",
+        "create table tb1 (ts timestamp, c1 bool, c2 tinyint, c3 smallint, c4 int, c5 bigint,
+            c6 tinyint unsigned, c7 smallint unsigned, c8 int unsigned, c9 bigint unsigned,
+            c10 float, c11 double, c12 varchar(100), c13 nchar(100))",
+    ])
+    .await?;
+    let mut stmt = Stmt::init(&taos)?;
+    stmt.prepare("insert into tb1 values(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)")?;
+    let params = vec![
+        ColumnView::from_millis_timestamp(vec![0]),
+        ColumnView::from_bools(vec![true]),
+        ColumnView::from_tiny_ints(vec![i8::MAX]),
+        ColumnView::from_small_ints(vec![i16::MAX]),
+        ColumnView::from_ints(vec![i32::MAX]),
+        ColumnView::from_big_ints(vec![i64::MAX]),
+        ColumnView::from_unsigned_tiny_ints(vec![u8::MAX]),
+        ColumnView::from_unsigned_small_ints(vec![u16::MAX]),
+        ColumnView::from_unsigned_ints(vec![u32::MAX]),
+        ColumnView::from_unsigned_big_ints(vec![u64::MAX]),
+        ColumnView::from_floats(vec![f32::MAX]),
+        ColumnView::from_doubles(vec![f64::MAX]),
+        ColumnView::from_varchar(vec!["ABC"]),
+        ColumnView::from_nchar(vec!["涛思数据"]),
+    ];
+    let rows = stmt.bind(&params)?.add_batch()?.execute()?;
+    assert_eq!(rows, 1);
+
+    #[derive(Debug, Deserialize)]
+    #[allow(dead_code)]
+    struct Row {
+        ts: String,
+        c1: bool,
+        c2: i8,
+        c3: i16,
+        c4: i32,
+        c5: i64,
+        c6: u8,
+        c7: u16,
+        c8: u32,
+        c9: u64,
+        c10: Option<f32>,
+        c11: f64,
+        c12: String,
+        c13: String,
+    }
+
+    let rows: Vec<Row> = taos
+        .query("select * from tb1")
+        .await?
+        .deserialize()
+        .try_collect()
+        .await?;
+    let row = &rows[0];
+    dbg!(&row);
+    assert_eq!(row.c5, i64::MAX);
+    assert_eq!(row.c8, u32::MAX);
+    assert_eq!(row.c9, u64::MAX);
+    assert_eq!(row.c10.unwrap(), f32::MAX);
+    // assert_eq!(row.c11, f64::MAX);
+    assert_eq!(row.c12, "ABC");
+    assert_eq!(row.c13, "涛思数据");
+
+    Ok(())
+}

examples/rust/examples/query.rs

+use std::time::Duration;
+
+use chrono::{DateTime, Local};
+use taos::*;
+
+#[tokio::main]
+async fn main() -> anyhow::Result<()> {
+    let dsn = "taos://";
+
+    let opts = PoolBuilder::new()
+        .max_size(5000) // max connections
+        .max_lifetime(Some(Duration::from_secs(60 * 60))) // lifetime of each connection
+        .min_idle(Some(1000)) // minimal idle connections
+        .connection_timeout(Duration::from_secs(2));
+
+    let pool = TaosBuilder::from_dsn(dsn)?.with_pool_builder(opts)?;
+
+    let taos = pool.get()?;
+
+    let db = "query";
+
+    // prepare database
+    taos.exec_many([
+        format!("DROP DATABASE IF EXISTS `{db}`"),
+        format!("CREATE DATABASE `{db}`"),
+        format!("USE `{db}`"),
+    ])
+    .await?;
+
+    let inserted = taos.exec_many([
+        // create super table
+        "CREATE TABLE `meters` (`ts` TIMESTAMP, `current` FLOAT, `voltage` INT, `phase` FLOAT) TAGS (`groupid` INT, `location` BINARY(16))",
+        // create child table
+        "CREATE TABLE `d0` USING `meters` TAGS(0, 'Los Angles')",
+        // insert into child table
+        "INSERT INTO `d0` values(now - 10s, 10, 116, 0.32)",
+        // insert with NULL values
+        "INSERT INTO `d0` values(now - 8s, NULL, NULL, NULL)",
+        // insert and automatically create table with tags if not exists
+        "INSERT INTO `d1` USING `meters` TAGS(1, 'San Francisco') values(now - 9s, 10.1, 119, 0.33)",
+        // insert many records in a single sql
+        "INSERT INTO `d1` values (now-8s, 10, 120, 0.33) (now - 6s, 10, 119, 0.34) (now - 4s, 11.2, 118, 0.322)",
+    ]).await?;
+
+    assert_eq!(inserted, 6);
+    loop {
+        let count: usize = taos
+            .query_one("select count(*) from `meters`")
+            .await?
+            .unwrap_or_default();
+
+        if count >= 6 {
+            break;
+        } else {
+            println!("waiting for data");
+        }
+    }
+
+    let mut result = taos.query("select tbname, * from `meters`").await?;
+
+    for field in result.fields() {
+        println!("got field: {}", field.name());
+    }
+
+    // Query option 1, use rows stream.
+    let mut rows = result.rows();
+    let mut nrows = 0;
+    while let Some(row) = rows.try_next().await? {
+        for (col, (name, value)) in row.enumerate() {
+            println!(
+                "[{}] got value in col {} (named `{:>8}`): {}",
+                nrows, col, name, value
+            );
+        }
+        nrows += 1;
+    }
+
+    // Query options 2, use deserialization with serde.
+    #[derive(Debug, serde::Deserialize)]
+    #[allow(dead_code)]
+    struct Record {
+        tbname: String,
+        // deserialize timestamp to chrono::DateTime<Local>
+        ts: DateTime<Local>,
+        // float to f32
+        current: Option<f32>,
+        // int to i32
+        voltage: Option<i32>,
+        phase: Option<f32>,
+        groupid: i32,
+        // binary/varchar to String
+        location: String,
+    }
+
+    let records: Vec<Record> = taos
+        .query("select tbname, * from `meters`")
+        .await?
+        .deserialize()
+        .try_collect()
+        .await?;
+
+    dbg!(result.summary());
+    assert_eq!(records.len(), 6);
+    dbg!(records);
+    Ok(())
+}

examples/rust/examples/subscribe.rs

+use std::time::Duration;
+
+use chrono::{DateTime, Local};
+use taos::*;
+
+// Query options 2, use deserialization with serde.
+#[derive(Debug, serde::Deserialize)]
+#[allow(dead_code)]
+struct Record {
+    // deserialize timestamp to chrono::DateTime<Local>
+    ts: DateTime<Local>,
+    // float to f32
+    current: Option<f32>,
+    // int to i32
+    voltage: Option<i32>,
+    phase: Option<f32>,
+}
+
+async fn prepare(taos: Taos) -> anyhow::Result<()> {
+    let inserted = taos.exec_many([
+        // create child table
+        "CREATE TABLE `d0` USING `meters` TAGS(0, 'Los Angles')",
+        // insert into child table
+        "INSERT INTO `d0` values(now - 10s, 10, 116, 0.32)",
+        // insert with NULL values
+        "INSERT INTO `d0` values(now - 8s, NULL, NULL, NULL)",
+        // insert and automatically create table with tags if not exists
+        "INSERT INTO `d1` USING `meters` TAGS(1, 'San Francisco') values(now - 9s, 10.1, 119, 0.33)",
+        // insert many records in a single sql
+        "INSERT INTO `d1` values (now-8s, 10, 120, 0.33) (now - 6s, 10, 119, 0.34) (now - 4s, 11.2, 118, 0.322)",
+    ]).await?;
+    assert_eq!(inserted, 6);
+    Ok(())
+}
+
+#[tokio::main]
+async fn main() -> anyhow::Result<()> {
+    // std::env::set_var("RUST_LOG", "debug");
+    pretty_env_logger::init();
+    let dsn = "taos://localhost:6030";
+    let builder = TaosBuilder::from_dsn(dsn)?;
+
+    let taos = builder.build()?;
+    let db = "tmq";
+
+    // prepare database
+    taos.exec_many([
+        "DROP TOPIC IF EXISTS tmq_meters".to_string(),
+        format!("DROP DATABASE IF EXISTS `{db}`"),
+        format!("CREATE DATABASE `{db}`"),
+        format!("USE `{db}`"),
+        // create super table
+        "CREATE TABLE `meters` (`ts` TIMESTAMP, `current` FLOAT, `voltage` INT, `phase` FLOAT) TAGS (`groupid` INT, `location` BINARY(16))".to_string(),
+        // create topic for subscription
+        format!("CREATE TOPIC tmq_meters with META AS DATABASE {db}")
+    ])
+    .await?;
+
+    let task = tokio::spawn(prepare(taos));
+
+    tokio::time::sleep(Duration::from_secs(1)).await;
+
+    // subscribe
+    let tmq = TmqBuilder::from_dsn("taos://localhost:6030/?group.id=test")?;
+
+    let mut consumer = tmq.build()?;
+    consumer.subscribe(["tmq_meters"]).await?;
+
+    {
+        let mut stream = consumer.stream();
+
+        while let Some((offset, message)) = stream.try_next().await? {
+            // get information from offset
+
+            // the topic
+            let topic = offset.topic();
+            // the vgroup id, like partition id in kafka.
+            let vgroup_id = offset.vgroup_id();
+            println!("* in vgroup id {vgroup_id} of topic {topic}\n");
+
+            if let Some(data) = message.into_data() {
+                while let Some(block) = data.fetch_raw_block().await? {
+                    // one block for one table, get table name if needed
+                    let name = block.table_name();
+                    let records: Vec<Record> = block.deserialize().try_collect()?;
+                    println!(
+                        "** table: {}, got {} records: {:#?}\n",
+                        name.unwrap(),
+                        records.len(),
+                        records
+                    );
+                }
+            }
+            consumer.commit(offset).await?;
+        }
+    }
+
+    consumer.unsubscribe().await;
+
+    task.await??;
+
+    Ok(())
+}

examples/rust/src/main.rs

+fn main() {
+    println!("Hello, world!");
+}

include/common/tcommon.h

@@ -60,6 +60,7 @@ enum {
   STREAM_INPUT__DATA_RETRIEVE,
   STREAM_INPUT__GET_RES,
   STREAM_INPUT__CHECKPOINT,
+  STREAM_INPUT__DESTROY,
 };
 
 typedef enum EStreamType {

include/common/tdataformat.h

@@ -38,22 +38,18 @@ typedef struct STagVal       STagVal;
 typedef struct STag          STag;
 
 // bitmap
-#define N1(n)        ((1 << (n)) - 1)
-#define BIT1_SIZE(n) (((n)-1) / 8 + 1)
-#define BIT2_SIZE(n) (((n)-1) / 4 + 1)
-#define SET_BIT1(p, i, v)                            \
-  do {                                               \
-    (p)[(i) / 8] &= N1((i) % 8);                     \
-    (p)[(i) / 8] |= (((uint8_t)(v)) << (((i) % 8))); \
-  } while (0)
-
-#define GET_BIT1(p, i) (((p)[(i) / 8] >> ((i) % 8)) & ((uint8_t)1))
-#define SET_BIT2(p, i, v)                                \
-  do {                                                   \
-    p[(i) / 4] &= N1((i) % 4 * 2);                       \
-    (p)[(i) / 4] |= (((uint8_t)(v)) << (((i) % 4) * 2)); \
-  } while (0)
-#define GET_BIT2(p, i) (((p)[(i) / 4] >> (((i) % 4) * 2)) & ((uint8_t)3))
+const static uint8_t BIT2_MAP[4][4] = {{0b00000000, 0b00000001, 0b00000010, 0},
+                                       {0b00000000, 0b00000100, 0b00001000, 2},
+                                       {0b00000000, 0b00010000, 0b00100000, 4},
+                                       {0b00000000, 0b01000000, 0b10000000, 6}};
+
+#define N1(n)             ((((uint8_t)1) << (n)) - 1)
+#define BIT1_SIZE(n)      ((((n)-1) >> 3) + 1)
+#define BIT2_SIZE(n)      ((((n)-1) >> 2) + 1)
+#define SET_BIT1(p, i, v) ((p)[(i) >> 3] = (p)[(i) >> 3] & N1((i)&7) | (((uint8_t)(v)) << ((i)&7)))
+#define GET_BIT1(p, i)    (((p)[(i) >> 3] >> ((i)&7)) & ((uint8_t)1))
+#define SET_BIT2(p, i, v) ((p)[(i) >> 2] = (p)[(i) >> 2] & N1(BIT2_MAP[(i)&3][3]) | BIT2_MAP[(i)&3][(v)])
+#define GET_BIT2(p, i)    (((p)[(i) >> 2] >> BIT2_MAP[(i)&3][3]) & ((uint8_t)3))
 
 // STSchema
 int32_t tTSchemaCreate(int32_t sver, SSchema *pSchema, int32_t nCols, STSchema **ppTSchema);
@@ -171,7 +167,7 @@ struct SColVal {
 
 #pragma pack(push, 1)
 struct STagVal {
-//  char colName[TSDB_COL_NAME_LEN]; // only used for tmq_get_meta
+  //  char colName[TSDB_COL_NAME_LEN]; // only used for tmq_get_meta
   union {
     int16_t cid;
     char   *pKey;

include/common/tglobal.h

@@ -139,7 +139,6 @@ int32_t taosInitCfg(const char *cfgDir, const char **envCmd, const char *envFile
                     bool tsc);
 void    taosCleanupCfg();
 void    taosCfgDynamicOptions(const char *option, const char *value);
-void    taosAddDataDir(int32_t index, char *v1, int32_t level, int32_t primary);
 
 struct SConfig *taosGetCfg();
 

include/common/tmsg.h

@@ -2555,10 +2555,14 @@ typedef struct {
   char    topic[TSDB_TOPIC_FNAME_LEN];
   int64_t ntbUid;
   SArray* colIdList;  // SArray<int16_t>
-} SCheckAlterInfo;
+} STqCheckInfo;
 
-int32_t tEncodeSCheckAlterInfo(SEncoder* pEncoder, const SCheckAlterInfo* pInfo);
-int32_t tDecodeSCheckAlterInfo(SDecoder* pDecoder, SCheckAlterInfo* pInfo);
+int32_t tEncodeSTqCheckInfo(SEncoder* pEncoder, const STqCheckInfo* pInfo);
+int32_t tDecodeSTqCheckInfo(SDecoder* pDecoder, STqCheckInfo* pInfo);
+
+typedef struct {
+  char topic[TSDB_TOPIC_FNAME_LEN];
+} STqDelCheckInfoReq;
 
 typedef struct {
   int32_t vgId;
@@ -2659,6 +2663,10 @@ typedef struct {
   SEpSet  epSet;
 } SVgEpSet;
 
+typedef struct {
+  int32_t padding;
+} SRSmaExecMsg;
+
 typedef struct {
   int64_t suid;
   int8_t  level;

include/common/tmsgdef.h

@@ -188,7 +188,8 @@ enum {
   TD_DEF_MSG_TYPE(TDMT_VND_MQ_VG_CHANGE, "vnode-mq-vg-change", SMqRebVgReq, SMqRebVgRsp)
   TD_DEF_MSG_TYPE(TDMT_VND_MQ_VG_DELETE, "vnode-mq-vg-delete", SMqVDeleteReq, SMqVDeleteRsp)
   TD_DEF_MSG_TYPE(TDMT_VND_MQ_COMMIT_OFFSET, "vnode-commit-offset", STqOffset, STqOffset)
-  TD_DEF_MSG_TYPE(TDMT_VND_CHECK_ALTER_INFO, "vnode-alter-check-info", NULL, NULL)
+  TD_DEF_MSG_TYPE(TDMT_VND_ADD_CHECK_INFO, "vnode-add-check-info", NULL, NULL)
+  TD_DEF_MSG_TYPE(TDMT_VND_DELETE_CHECK_INFO, "vnode-delete-check-info", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_VND_CREATE_TOPIC, "vnode-create-topic", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_VND_ALTER_TOPIC, "vnode-alter-topic", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_VND_DROP_TOPIC, "vnode-drop-topic", NULL, NULL)
@@ -201,6 +202,7 @@ enum {
   TD_DEF_MSG_TYPE(TDMT_VND_DROP_SMA, "vnode-drop-sma", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_VND_SUBMIT_RSMA, "vnode-submit-rsma", SSubmitReq, SSubmitRsp)
   TD_DEF_MSG_TYPE(TDMT_VND_FETCH_RSMA, "vnode-fetch-rsma", SRSmaFetchMsg, NULL)
+  TD_DEF_MSG_TYPE(TDMT_VND_EXEC_RSMA, "vnode-exec-rsma", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_VND_DELETE, "delete-data", SVDeleteReq, SVDeleteRsp)
   TD_DEF_MSG_TYPE(TDMT_VND_BATCH_DEL, "batch-delete", SBatchDeleteReq, NULL)
   TD_DEF_MSG_TYPE(TDMT_VND_ALTER_CONFIG, "alter-config", NULL, NULL)

include/libs/nodes/nodes.h

@@ -105,7 +105,7 @@ typedef enum ENodeType {
   QUERY_NODE_COLUMN_REF,
 
   // Statement nodes are used in parser and planner module.
-  QUERY_NODE_SET_OPERATOR,
+  QUERY_NODE_SET_OPERATOR = 100,
   QUERY_NODE_SELECT_STMT,
   QUERY_NODE_VNODE_MODIF_STMT,
   QUERY_NODE_CREATE_DATABASE_STMT,
@@ -198,7 +198,7 @@ typedef enum ENodeType {
   QUERY_NODE_QUERY,
 
   // logic plan node
-  QUERY_NODE_LOGIC_PLAN_SCAN,
+  QUERY_NODE_LOGIC_PLAN_SCAN = 1000,
   QUERY_NODE_LOGIC_PLAN_JOIN,
   QUERY_NODE_LOGIC_PLAN_AGG,
   QUERY_NODE_LOGIC_PLAN_PROJECT,
@@ -215,7 +215,7 @@ typedef enum ENodeType {
   QUERY_NODE_LOGIC_PLAN,
 
   // physical plan node
-  QUERY_NODE_PHYSICAL_PLAN_TAG_SCAN,
+  QUERY_NODE_PHYSICAL_PLAN_TAG_SCAN = 1100,
   QUERY_NODE_PHYSICAL_PLAN_TABLE_SCAN,
   QUERY_NODE_PHYSICAL_PLAN_TABLE_SEQ_SCAN,
   QUERY_NODE_PHYSICAL_PLAN_TABLE_MERGE_SCAN,

include/libs/nodes/querynodes.h

@@ -428,6 +428,9 @@ void    nodesValueNodeToVariant(const SValueNode* pNode, SVariant* pVal);
 char*   nodesGetFillModeString(EFillMode mode);
 int32_t nodesMergeConds(SNode** pDst, SNodeList** pSrc);
 
+const char* operatorTypeStr(EOperatorType type);
+const char* logicConditionTypeStr(ELogicConditionType type);
+
 #ifdef __cplusplus
 }
 #endif

include/libs/stream/tstream.h

@@ -53,6 +53,7 @@ enum {
   TASK_SCHED_STATUS__WAITING,
   TASK_SCHED_STATUS__ACTIVE,
   TASK_SCHED_STATUS__FAILED,
+  TASK_SCHED_STATUS__DROPPING,
 };
 
 enum {
@@ -127,6 +128,10 @@ typedef struct {
   int8_t type;
 } SStreamCheckpoint;
 
+typedef struct {
+  int8_t type;
+} SStreamTaskDestroy;
+
 typedef struct {
   int8_t       type;
   SSDataBlock* pBlock;
@@ -211,7 +216,6 @@ typedef struct {
   void*     vnode;
   FTbSink*  tbSinkFunc;
   STSchema* pTSchema;
-  SHashObj* pHash;  // groupId to tbuid
 } STaskSinkTb;
 
 typedef void FSmaSink(void* vnode, int64_t smaId, const SArray* data);
@@ -515,7 +519,7 @@ SStreamMeta* streamMetaOpen(const char* path, void* ahandle, FTaskExpand expandF
 void         streamMetaClose(SStreamMeta* streamMeta);
 
 int32_t      streamMetaAddTask(SStreamMeta* pMeta, SStreamTask* pTask);
-int32_t      streamMetaAddSerializedTask(SStreamMeta* pMeta, char* msg, int32_t msgLen);
+int32_t      streamMetaAddSerializedTask(SStreamMeta* pMeta, int64_t startVer, char* msg, int32_t msgLen);
 int32_t      streamMetaRemoveTask(SStreamMeta* pMeta, int32_t taskId);
 SStreamTask* streamMetaGetTask(SStreamMeta* pMeta, int32_t taskId);
 

include/libs/sync/sync.h

@@ -26,11 +26,15 @@ extern "C" {
 
 extern bool gRaftDetailLog;
 
-#define SYNC_RESP_TTL_MS       10000000
-#define SYNC_SPEED_UP_HB_TIMER 400
-#define SYNC_SPEED_UP_AFTER_MS (1000 * 20)
-#define SYNC_SLOW_DOWN_RANGE   100
-#define SYNC_MAX_READ_RANGE    10
+#define SYNC_RESP_TTL_MS             10000000
+#define SYNC_SPEED_UP_HB_TIMER       400
+#define SYNC_SPEED_UP_AFTER_MS       (1000 * 20)
+#define SYNC_SLOW_DOWN_RANGE         100
+#define SYNC_MAX_READ_RANGE          2
+#define SYNC_MAX_PROGRESS_WAIT_MS    4000
+#define SYNC_MAX_START_TIME_RANGE_MS (1000 * 20)
+#define SYNC_MAX_RECV_TIME_RANGE_MS  1200
+#define SYNC_ADD_QUORUM_COUNT        3
 
 #define SYNC_MAX_BATCH_SIZE 1
 #define SYNC_INDEX_BEGIN    0

include/libs/sync/syncTools.h

@@ -423,6 +423,7 @@ typedef struct SyncAppendEntriesReply {
   SyncTerm  privateTerm;
   bool      success;
   SyncIndex matchIndex;
+  int64_t   startTime;
 } SyncAppendEntriesReply;
 
 SyncAppendEntriesReply* syncAppendEntriesReplyBuild(int32_t vgId);

include/util/taoserror.h

@@ -291,6 +291,7 @@ int32_t* taosGetErrno();
 #define TSDB_CODE_MND_STREAM_NOT_EXIST          TAOS_DEF_ERROR_CODE(0, 0x03F1)
 #define TSDB_CODE_MND_INVALID_STREAM_OPTION     TAOS_DEF_ERROR_CODE(0, 0x03F2)
 #define TSDB_CODE_MND_STREAM_MUST_BE_DELETED    TAOS_DEF_ERROR_CODE(0, 0x03F3)
+#define TSDB_CODE_MND_STREAM_TASK_DROPPED       TAOS_DEF_ERROR_CODE(0, 0x03F4)
 
 // mnode-sma
 #define TSDB_CODE_MND_SMA_ALREADY_EXIST         TAOS_DEF_ERROR_CODE(0, 0x0480)
@@ -614,6 +615,7 @@ int32_t* taosGetErrno();
 #define TSDB_CODE_RSMA_REMOVE_EXISTS             TAOS_DEF_ERROR_CODE(0, 0x3154)
 #define TSDB_CODE_RSMA_FETCH_MSG_MSSED_UP        TAOS_DEF_ERROR_CODE(0, 0x3155)
 #define TSDB_CODE_RSMA_EMPTY_INFO                TAOS_DEF_ERROR_CODE(0, 0x3156)
+#define TSDB_CODE_RSMA_INVALID_SCHEMA            TAOS_DEF_ERROR_CODE(0, 0x3157)
 
 //index
 #define TSDB_CODE_INDEX_REBUILDING               TAOS_DEF_ERROR_CODE(0, 0x3200)

include/util/tdef.h

@@ -132,15 +132,14 @@ typedef enum EOperatorType {
   OP_TYPE_DIV,
   OP_TYPE_REM,
   // unary arithmetic operator
-  OP_TYPE_MINUS,
-  OP_TYPE_ASSIGN,
+  OP_TYPE_MINUS = 20,
 
   // bitwise operator
-  OP_TYPE_BIT_AND,
+  OP_TYPE_BIT_AND = 30,
   OP_TYPE_BIT_OR,
 
   // binary comparison operator
-  OP_TYPE_GREATER_THAN,
+  OP_TYPE_GREATER_THAN = 40,
   OP_TYPE_GREATER_EQUAL,
   OP_TYPE_LOWER_THAN,
   OP_TYPE_LOWER_EQUAL,
@@ -153,7 +152,7 @@ typedef enum EOperatorType {
   OP_TYPE_MATCH,
   OP_TYPE_NMATCH,
   // unary comparison operator
-  OP_TYPE_IS_NULL,
+  OP_TYPE_IS_NULL = 100,
   OP_TYPE_IS_NOT_NULL,
   OP_TYPE_IS_TRUE,
   OP_TYPE_IS_FALSE,
@@ -163,8 +162,11 @@ typedef enum EOperatorType {
   OP_TYPE_IS_NOT_UNKNOWN,
 
   // json operator
-  OP_TYPE_JSON_GET_VALUE,
-  OP_TYPE_JSON_CONTAINS
+  OP_TYPE_JSON_GET_VALUE = 150,
+  OP_TYPE_JSON_CONTAINS,
+
+  // internal operator
+  OP_TYPE_ASSIGN = 200
 } EOperatorType;
 
 #define OP_TYPE_CALC_MAX OP_TYPE_BIT_OR

packaging/tools/make_install.bat

 @echo off
 goto %1
 :needAdmin
+
+if exist C:\\TDengine\\data\\dnode\\dnodeCfg.json (
+  echo The default data directory C:/TDengine/data contains old data of tdengine 2.x, please clear it before installing!
+)
+set source_dir=%2
+set source_dir=%source_dir:/=\\%
+set binary_dir=%3
+set binary_dir=%binary_dir:/=\\%
+set osType=%4
+set verNumber=%5
+set tagert_dir=C:\\TDengine
+
+if not exist %tagert_dir% (
+    mkdir %tagert_dir%
+)
+if not exist %tagert_dir%\\cfg (
+    mkdir %tagert_dir%\\cfg
+)
+if not exist %tagert_dir%\\include (
+    mkdir %tagert_dir%\\include
+)
+if not exist %tagert_dir%\\driver (
+    mkdir %tagert_dir%\\driver
+)
+if not exist C:\\TDengine\\cfg\\taos.cfg (
+    copy %source_dir%\\packaging\\cfg\\taos.cfg %tagert_dir%\\cfg\\taos.cfg > nul
+)
+
+if exist %binary_dir%\\test\\cfg\\taosadapter.toml (
+    if not exist %tagert_dir%\\cfg\\taosadapter.toml (
+        copy %binary_dir%\\test\\cfg\\taosadapter.toml %tagert_dir%\\cfg\\taosadapter.toml > nul
+    )
+)
+
+copy %source_dir%\\include\\client\\taos.h %tagert_dir%\\include > nul
+copy %source_dir%\\include\\util\\taoserror.h %tagert_dir%\\include > nul
+copy %source_dir%\\include\\libs\\function\\taosudf.h %tagert_dir%\\include > nul
+copy %binary_dir%\\build\\lib\\taos.lib %tagert_dir%\\driver > nul
+copy %binary_dir%\\build\\lib\\taos_static.lib %tagert_dir%\\driver > nul
+copy %binary_dir%\\build\\lib\\taos.dll %tagert_dir%\\driver > nul
+copy %binary_dir%\\build\\bin\\taos.exe %tagert_dir% > nul
+copy %binary_dir%\\build\\bin\\taosd.exe %tagert_dir% > nul
+copy %binary_dir%\\build\\bin\\udfd.exe %tagert_dir% > nul
+if exist %binary_dir%\\build\\bin\\taosBenchmark.exe (
+    copy %binary_dir%\\build\\bin\\taosBenchmark.exe %tagert_dir% > nul
+)
+if exist %binary_dir%\\build\\bin\\taosadapter.exe (
+    copy %binary_dir%\\build\\bin\\taosadapter.exe %tagert_dir% > nul
+)
+
 mshta vbscript:createobject("shell.application").shellexecute("%~s0",":hasAdmin","","runas",1)(window.close)&& echo To start/stop TDengine with administrator privileges: sc start/stop taosd &goto :eof
 :hasAdmin
-cp -f C:\\TDengine\\driver\\taos.dll C:\\Windows\\System32
+copy /y C:\\TDengine\\driver\\taos.dll C:\\Windows\\System32 > nul
 sc query "taosd" >nul || sc create "taosd" binPath= "C:\\TDengine\\taosd.exe --win_service" start= DEMAND

packaging/tools/make_install.sh

@@ -664,7 +664,9 @@ function install_TDengine() {
 ## ==============================Main program starts from here============================
 echo source directory: $1
 echo binary directory: $2
-if [ "$osType" != "Darwin" ]; then
+if [ -x ${data_dir}/dnode/dnodeCfg.json ]; then
+  echo -e "\033[44;31;5mThe default data directory ${data_dir} contains old data of tdengine 2.x, please clear it before installing!\033[0m"
+elif [ "$osType" != "Darwin" ]; then
   if [ -x ${bin_dir}/${clientName} ]; then
     update_TDengine
   else

source/client/test/clientTests.cpp

@@ -123,7 +123,7 @@ void createNewTable(TAOS* pConn, int32_t index) {
   }
   taos_free_result(pRes);
 
-  for(int32_t i = 0; i < 100000; i += 20) {
+  for(int32_t i = 0; i < 3280; i += 20) {
     char sql[1024] = {0};
     sprintf(sql,
             "insert into tu%d values(now+%da, %d)(now+%da, %d)(now+%da, %d)(now+%da, %d)"
@@ -679,30 +679,28 @@ TEST(testCase, projection_query_tables) {
   TAOS_RES* pRes = taos_query(pConn, "use abc1");
   taos_free_result(pRes);
 
-  pRes = taos_query(pConn, "explain verbose true select _wstart,count(*),a from st1 partition by a interval(1s)");
-  printResult(pRes);
-//  pRes = taos_query(pConn, "create stable st1 (ts timestamp, k int) tags(a int)");
-//  if (taos_errno(pRes) != 0) {
-//    printf("failed to create table tu, reason:%s\n", taos_errstr(pRes));
-//  }
-//  taos_free_result(pRes);
-//
-//  pRes = taos_query(pConn, "create stable st2 (ts timestamp, k int) tags(a int)");
-//  if (taos_errno(pRes) != 0) {
-//    printf("failed to create table tu, reason:%s\n", taos_errstr(pRes));
-//  }
-//  taos_free_result(pRes);
-//
-//  pRes = taos_query(pConn, "create table tu using st1 tags(1)");
-//  if (taos_errno(pRes) != 0) {
-//    printf("failed to create table tu, reason:%s\n", taos_errstr(pRes));
-//  }
-//  taos_free_result(pRes);
-//
-//  for(int32_t i = 0; i < 1; ++i) {
-//    printf("create table :%d\n", i);
-//    createNewTable(pConn, i);
-//  }
+  pRes = taos_query(pConn, "create stable st1 (ts timestamp, k int) tags(a int)");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create table tu, reason:%s\n", taos_errstr(pRes));
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "create stable st2 (ts timestamp, k int) tags(a int)");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create table tu, reason:%s\n", taos_errstr(pRes));
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "create table tu using st1 tags(1)");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create table tu, reason:%s\n", taos_errstr(pRes));
+  }
+  taos_free_result(pRes);
+
+  for(int32_t i = 0; i < 2; ++i) {
+    printf("create table :%d\n", i);
+    createNewTable(pConn, i);
+  }
 //
 //  pRes = taos_query(pConn, "select * from tu");
 //  if (taos_errno(pRes) != 0) {

source/common/src/systable.c

@@ -88,7 +88,7 @@ static const SSysDbTableSchema userDBSchema[] = {
     {.name = "comp", .bytes = 1, .type = TSDB_DATA_TYPE_TINYINT},
     {.name = "precision", .bytes = 2 + VARSTR_HEADER_SIZE, .type = TSDB_DATA_TYPE_VARCHAR},
     {.name = "status", .bytes = 10 + VARSTR_HEADER_SIZE, .type = TSDB_DATA_TYPE_VARCHAR},
-    {.name = "retention", .bytes = 60 + VARSTR_HEADER_SIZE, .type = TSDB_DATA_TYPE_VARCHAR},
+    {.name = "retentions", .bytes = 60 + VARSTR_HEADER_SIZE, .type = TSDB_DATA_TYPE_VARCHAR},
     {.name = "single_stable", .bytes = 1, .type = TSDB_DATA_TYPE_BOOL},
     {.name = "cachemodel", .bytes = TSDB_CACHE_MODEL_STR_LEN + VARSTR_HEADER_SIZE, .type = TSDB_DATA_TYPE_VARCHAR},
     {.name = "cachesize", .bytes = 4, .type = TSDB_DATA_TYPE_INT},

source/common/src/tdatablock.c

@@ -1343,12 +1343,14 @@ SSDataBlock* createDataBlock() {
   SSDataBlock* pBlock = taosMemoryCalloc(1, sizeof(SSDataBlock));
   if (pBlock == NULL) {
     terrno = TSDB_CODE_OUT_OF_MEMORY;
+    return NULL;
   }
 
   pBlock->pDataBlock = taosArrayInit(4, sizeof(SColumnInfoData));
   if (pBlock->pDataBlock == NULL) {
     terrno = TSDB_CODE_OUT_OF_MEMORY;
     taosMemoryFree(pBlock);
+    return NULL;
   }
 
   return pBlock;
@@ -1423,6 +1425,7 @@ size_t blockDataGetCapacityInRow(const SSDataBlock* pBlock, size_t pageSize) {
 }
 
 void colDataDestroy(SColumnInfoData* pColData) {
+  if(!pColData) return;
   if (IS_VAR_DATA_TYPE(pColData->info.type)) {
     taosMemoryFreeClear(pColData->varmeta.offset);
   } else {

source/common/src/tglobal.c

@@ -75,7 +75,7 @@ int32_t  tsMonitorMaxLogs = 100;
 bool     tsMonitorComp = false;
 
 // telem
-bool     tsEnableTelem = false;
+bool     tsEnableTelem = true;
 int32_t  tsTelemInterval = 86400;
 char     tsTelemServer[TSDB_FQDN_LEN] = "telemetry.taosdata.com";
 uint16_t tsTelemPort = 80;
@@ -166,7 +166,22 @@ int32_t tsTtlPushInterval = 86400;
 int32_t tsGrantHBInterval = 60;
 
 #ifndef _STORAGE
-int32_t taosSetTfsCfg(SConfig *pCfg) { return 0; }
+int32_t taosSetTfsCfg(SConfig *pCfg) {
+  SConfigItem *pItem = cfgGetItem(pCfg, "dataDir");
+  memset(tsDataDir, 0, PATH_MAX);
+
+  int32_t size = taosArrayGetSize(pItem->array);
+  tsDiskCfgNum = 1;
+  tstrncpy(tsDiskCfg[0].dir, pItem->str, TSDB_FILENAME_LEN);
+  tsDiskCfg[0].level = 0;
+  tsDiskCfg[0].primary = 1;
+  tstrncpy(tsDataDir, pItem->str, PATH_MAX);
+  if (taosMulMkDir(tsDataDir) != 0) {
+    uError("failed to create dataDir:%s", tsDataDir);
+    return -1;
+  }
+  return 0;
+}
 #else
 int32_t taosSetTfsCfg(SConfig *pCfg);
 #endif

source/common/src/tmsg.c

@@ -4262,7 +4262,6 @@ int32_t tDeserializeSServerStatusRsp(void *buf, int32_t bufLen, SServerStatusRsp
   tDecoderClear(&decoder);
   return 0;
 }
-
 int32_t tEncodeSMqOffset(SEncoder *encoder, const SMqOffset *pOffset) {
   if (tEncodeI32(encoder, pOffset->vgId) < 0) return -1;
   if (tEncodeI64(encoder, pOffset->offset) < 0) return -1;
@@ -4300,7 +4299,6 @@ int32_t tDecodeSMqCMCommitOffsetReq(SDecoder *decoder, SMqCMCommitOffsetReq *pRe
   tEndDecode(decoder);
   return 0;
 }
-
 int32_t tSerializeSExplainRsp(void *buf, int32_t bufLen, SExplainRsp *pRsp) {
   SEncoder encoder = {0};
   tEncoderInit(&encoder, buf, bufLen);
@@ -5590,7 +5588,6 @@ int32_t tDecodeSTqOffsetVal(SDecoder *pDecoder, STqOffsetVal *pOffsetVal) {
   return 0;
 }
 
-#if 1
 int32_t tFormatOffset(char *buf, int32_t maxLen, const STqOffsetVal *pVal) {
   if (pVal->type == TMQ_OFFSET__RESET_NONE) {
     snprintf(buf, maxLen, "offset(reset to none)");
@@ -5609,7 +5606,6 @@ int32_t tFormatOffset(char *buf, int32_t maxLen, const STqOffsetVal *pVal) {
   }
   return 0;
 }
-#endif
 
 bool tOffsetEqual(const STqOffsetVal *pLeft, const STqOffsetVal *pRight) {
   if (pLeft->type == pRight->type) {
@@ -5643,7 +5639,7 @@ int32_t tDecodeSTqOffset(SDecoder *pDecoder, STqOffset *pOffset) {
   return 0;
 }
 
-int32_t tEncodeSCheckAlterInfo(SEncoder *pEncoder, const SCheckAlterInfo *pInfo) {
+int32_t tEncodeSTqCheckInfo(SEncoder *pEncoder, const STqCheckInfo *pInfo) {
   if (tEncodeCStr(pEncoder, pInfo->topic) < 0) return -1;
   if (tEncodeI64(pEncoder, pInfo->ntbUid) < 0) return -1;
   int32_t sz = taosArrayGetSize(pInfo->colIdList);
@@ -5655,7 +5651,7 @@ int32_t tEncodeSCheckAlterInfo(SEncoder *pEncoder, const SCheckAlterInfo *pInfo)
   return pEncoder->pos;
 }
 
-int32_t tDecodeSCheckAlterInfo(SDecoder *pDecoder, SCheckAlterInfo *pInfo) {
+int32_t tDecodeSTqCheckInfo(SDecoder *pDecoder, STqCheckInfo *pInfo) {
   if (tDecodeCStrTo(pDecoder, pInfo->topic) < 0) return -1;
   if (tDecodeI64(pDecoder, &pInfo->ntbUid) < 0) return -1;
   int32_t sz;

source/common/src/ttypes.c

@@ -392,10 +392,10 @@ tDataTypeDescriptor tDataTypes[TSDB_DATA_TYPE_MAX] = {
      getStatics_i64},
     {TSDB_DATA_TYPE_FLOAT, 5, FLOAT_BYTES, "FLOAT", 0, 0, tsCompressFloat, tsDecompressFloat, getStatics_f},
     {TSDB_DATA_TYPE_DOUBLE, 6, DOUBLE_BYTES, "DOUBLE", 0, 0, tsCompressDouble, tsDecompressDouble, getStatics_d},
-    {TSDB_DATA_TYPE_VARCHAR, 6, 0, "VARCHAR", 0, 0, tsCompressString, tsDecompressString, getStatics_bin},
+    {TSDB_DATA_TYPE_VARCHAR, 6, 1, "VARCHAR", 0, 0, tsCompressString, tsDecompressString, getStatics_bin},
     {TSDB_DATA_TYPE_TIMESTAMP, 9, LONG_BYTES, "TIMESTAMP", INT64_MIN, INT64_MAX, tsCompressTimestamp,
      tsDecompressTimestamp, getStatics_i64},
-    {TSDB_DATA_TYPE_NCHAR, 5, 8, "NCHAR", 0, 0, tsCompressString, tsDecompressString, getStatics_nchr},
+    {TSDB_DATA_TYPE_NCHAR, 5, 1, "NCHAR", 0, 0, tsCompressString, tsDecompressString, getStatics_nchr},
     {TSDB_DATA_TYPE_UTINYINT, 16, CHAR_BYTES, "TINYINT UNSIGNED", 0, UINT8_MAX, tsCompressTinyint, tsDecompressTinyint,
      getStatics_u8},
     {TSDB_DATA_TYPE_USMALLINT, 17, SHORT_BYTES, "SMALLINT UNSIGNED", 0, UINT16_MAX, tsCompressSmallint,

source/dnode/mgmt/mgmt_mnode/src/mmHandle.c

@@ -225,7 +225,8 @@ SArray *mmGetMsgHandles() {
   if (dmSetMgmtHandle(pArray, TDMT_VND_DROP_SMA_RSP, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_VND_MQ_VG_CHANGE_RSP, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_VND_MQ_VG_DELETE_RSP, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
-  if (dmSetMgmtHandle(pArray, TDMT_VND_CHECK_ALTER_INFO_RSP, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_VND_ADD_CHECK_INFO_RSP, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_VND_DELETE_CHECK_INFO_RSP, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_SCH_DROP_TASK, mmPutMsgToFetchQueue, 1) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_DEPLOY_RSP, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_DROP_RSP, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;

source/dnode/mgmt/mgmt_vnode/src/vmHandle.c

@@ -338,6 +338,7 @@ SArray *vmGetMsgHandles() {
   if (dmSetMgmtHandle(pArray, TDMT_SCH_MERGE_QUERY, vmPutMsgToQueryQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_SCH_QUERY_CONTINUE, vmPutMsgToQueryQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_VND_FETCH_RSMA, vmPutMsgToQueryQueue, 0) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_VND_EXEC_RSMA, vmPutMsgToQueryQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_SCH_FETCH, vmPutMsgToFetchQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_SCH_MERGE_FETCH, vmPutMsgToFetchQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_VND_ALTER_TABLE, vmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
@@ -361,7 +362,8 @@ SArray *vmGetMsgHandles() {
   if (dmSetMgmtHandle(pArray, TDMT_VND_MQ_VG_CHANGE, vmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_VND_MQ_VG_DELETE, vmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_VND_MQ_COMMIT_OFFSET, vmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
-  if (dmSetMgmtHandle(pArray, TDMT_VND_CHECK_ALTER_INFO, vmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_VND_ADD_CHECK_INFO, vmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_VND_DELETE_CHECK_INFO, vmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_VND_CONSUME, vmPutMsgToFetchQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_VND_DELETE, vmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_VND_BATCH_DEL, vmPutMsgToWriteQueue, 0) == NULL) goto _OVER;

source/dnode/mnode/impl/inc/mndDef.h

@@ -636,6 +636,7 @@ typedef struct {
 
 int32_t tEncodeSStreamObj(SEncoder* pEncoder, const SStreamObj* pObj);
 int32_t tDecodeSStreamObj(SDecoder* pDecoder, SStreamObj* pObj);
+void    tFreeStreamObj(SStreamObj* pObj);
 
 typedef struct {
   char    streamName[TSDB_STREAM_FNAME_LEN];

source/dnode/mnode/impl/inc/mndStb.h

@@ -34,6 +34,7 @@ int32_t mndCheckCreateStbReq(SMCreateStbReq *pCreate);
 SDbObj *mndAcquireDbByStb(SMnode *pMnode, const char *stbName);
 int32_t mndBuildStbFromReq(SMnode *pMnode, SStbObj *pDst, SMCreateStbReq *pCreate, SDbObj *pDb);
 int32_t mndAddStbToTrans(SMnode *pMnode, STrans *pTrans, SDbObj *pDb, SStbObj *pStb);
+void    mndFreeStb(SStbObj *pStb);
 
 void mndExtractDbNameFromStbFullName(const char *stbFullName, char *dst);
 void mndExtractTbNameFromStbFullName(const char *stbFullName, char *dst, int32_t dstSize);

source/dnode/mnode/impl/src/mndDef.c

@@ -116,6 +116,25 @@ int32_t tDecodeSStreamObj(SDecoder *pDecoder, SStreamObj *pObj) {
   return 0;
 }
 
+void tFreeStreamObj(SStreamObj *pStream) {
+  taosMemoryFree(pStream->sql);
+  taosMemoryFree(pStream->ast);
+  taosMemoryFree(pStream->physicalPlan);
+  if (pStream->outputSchema.nCols) taosMemoryFree(pStream->outputSchema.pSchema);
+
+  int32_t sz = taosArrayGetSize(pStream->tasks);
+  for (int32_t i = 0; i < sz; i++) {
+    SArray *pLevel = taosArrayGetP(pStream->tasks, i);
+    int32_t taskSz = taosArrayGetSize(pLevel);
+    for (int32_t j = 0; j < taskSz; j++) {
+      SStreamTask *pTask = taosArrayGetP(pLevel, j);
+      tFreeSStreamTask(pTask);
+    }
+    taosArrayDestroy(pLevel);
+  }
+  taosArrayDestroy(pStream->tasks);
+}
+
 SMqVgEp *tCloneSMqVgEp(const SMqVgEp *pVgEp) {
   SMqVgEp *pVgEpNew = taosMemoryMalloc(sizeof(SMqVgEp));
   if (pVgEpNew == NULL) return NULL;

source/dnode/mnode/impl/src/mndOffset.c

@@ -15,10 +15,10 @@
 
 #define _DEFAULT_SOURCE
 #include "mndOffset.h"
-#include "mndPrivilege.h"
 #include "mndDb.h"
 #include "mndDnode.h"
 #include "mndMnode.h"
+#include "mndPrivilege.h"
 #include "mndShow.h"
 #include "mndStb.h"
 #include "mndTopic.h"
@@ -305,7 +305,7 @@ int32_t mndDropOffsetByDB(SMnode *pMnode, STrans *pTrans, SDbObj *pDb) {
     sdbRelease(pSdb, pOffset);
   }
 
-   return code;
+  return code;
 }
 
 int32_t mndDropOffsetByTopic(SMnode *pMnode, STrans *pTrans, const char *topic) {

source/dnode/mnode/impl/src/mndScheduler.c

@@ -424,6 +424,8 @@ int32_t mndScheduleStream(SMnode* pMnode, SStreamObj* pStream) {
       }
       mndAddTaskToTaskSet(taskSourceLevel, pTask);
 
+      pTask->triggerParam = 0;
+
       // source
       pTask->taskLevel = TASK_LEVEL__SOURCE;
 

source/dnode/mnode/impl/src/mndSma.c

@@ -489,7 +489,7 @@ static int32_t mndCreateSma(SMnode *pMnode, SRpcMsg *pReq, SMCreateSmaReq *pCrea
   smaObj.uid = mndGenerateUid(pCreate->name, TSDB_TABLE_FNAME_LEN);
   ASSERT(smaObj.uid != 0);
   char resultTbName[TSDB_TABLE_FNAME_LEN + 16] = {0};
-  snprintf(resultTbName, TSDB_TABLE_FNAME_LEN + 16, "%s_td_tsma_rst_tb",pCreate->name);
+  snprintf(resultTbName, TSDB_TABLE_FNAME_LEN + 16, "%s_td_tsma_rst_tb", pCreate->name);
   memcpy(smaObj.dstTbName, resultTbName, TSDB_TABLE_FNAME_LEN);
   smaObj.dstTbUid = mndGenerateUid(smaObj.dstTbName, TSDB_TABLE_FNAME_LEN);
   smaObj.stbUid = pStb->uid;
@@ -530,7 +530,7 @@ static int32_t mndCreateSma(SMnode *pMnode, SRpcMsg *pReq, SMCreateSmaReq *pCrea
   streamObj.sourceDbUid = pDb->uid;
   streamObj.targetDbUid = pDb->uid;
   streamObj.version = 1;
-  streamObj.sql = pCreate->sql;
+  streamObj.sql = strdup(pCreate->sql);
   streamObj.smaId = smaObj.uid;
   streamObj.watermark = pCreate->watermark;
   streamObj.trigger = STREAM_TRIGGER_WINDOW_CLOSE;
@@ -585,6 +585,7 @@ static int32_t mndCreateSma(SMnode *pMnode, SRpcMsg *pReq, SMCreateSmaReq *pCrea
     return -1;
   }
   if (pAst != NULL) nodesDestroyNode(pAst);
+  nodesDestroyNode((SNode *)pPlan);
 
   int32_t code = -1;
   STrans *pTrans = mndTransCreate(pMnode, TRN_POLICY_RETRY, TRN_CONFLICT_DB, pReq);
@@ -609,6 +610,7 @@ static int32_t mndCreateSma(SMnode *pMnode, SRpcMsg *pReq, SMCreateSmaReq *pCrea
   code = 0;
 
 _OVER:
+  tFreeStreamObj(&streamObj);
   mndDestroySmaObj(&smaObj);
   mndTransDrop(pTrans);
   return code;

source/dnode/mnode/impl/src/mndStb.c

@@ -266,6 +266,15 @@ _OVER:
   return pRow;
 }
 
+void mndFreeStb(SStbObj *pStb) {
+  taosArrayDestroy(pStb->pFuncs);
+  taosMemoryFreeClear(pStb->pColumns);
+  taosMemoryFreeClear(pStb->pTags);
+  taosMemoryFreeClear(pStb->comment);
+  taosMemoryFreeClear(pStb->pAst1);
+  taosMemoryFreeClear(pStb->pAst2);
+}
+
 static int32_t mndStbActionInsert(SSdb *pSdb, SStbObj *pStb) {
   mTrace("stb:%s, perform insert action, row:%p", pStb->name, pStb);
   return 0;
@@ -273,12 +282,7 @@ static int32_t mndStbActionInsert(SSdb *pSdb, SStbObj *pStb) {
 
 static int32_t mndStbActionDelete(SSdb *pSdb, SStbObj *pStb) {
   mTrace("stb:%s, perform delete action, row:%p", pStb->name, pStb);
-  taosArrayDestroy(pStb->pFuncs);
-  taosMemoryFreeClear(pStb->pColumns);
-  taosMemoryFreeClear(pStb->pTags);
-  taosMemoryFreeClear(pStb->comment);
-  taosMemoryFreeClear(pStb->pAst1);
-  taosMemoryFreeClear(pStb->pAst2);
+  mndFreeStb(pStb);
   return 0;
 }
 
@@ -1145,7 +1149,7 @@ static int32_t mndAddSuperTableTag(const SStbObj *pOld, SStbObj *pNew, SArray *p
   return 0;
 }
 
-int32_t mndCheckColAndTagModifiable(SMnode *pMnode, const char *stbname, int64_t suid, col_id_t colId) {
+static int32_t mndCheckAlterColForTopic(SMnode *pMnode, const char *stbFullName, int64_t suid, col_id_t colId) {
   SSdb *pSdb = pMnode->pSdb;
   void *pIter = NULL;
   while (1) {
@@ -1154,7 +1158,7 @@ int32_t mndCheckColAndTagModifiable(SMnode *pMnode, const char *stbname, int64_t
     if (pIter == NULL) break;
 
     mDebug("topic:%s, check tag and column modifiable, stb:%s suid:%" PRId64 " colId:%d, subType:%d sql:%s",
-           pTopic->name, stbname, suid, colId, pTopic->subType, pTopic->sql);
+           pTopic->name, stbFullName, suid, colId, pTopic->subType, pTopic->sql);
     if (pTopic->subType != TOPIC_SUB_TYPE__COLUMN) {
       sdbRelease(pSdb, pTopic);
       continue;
@@ -1192,20 +1196,66 @@ int32_t mndCheckColAndTagModifiable(SMnode *pMnode, const char *stbname, int64_t
     sdbRelease(pSdb, pTopic);
     nodesDestroyNode(pAst);
   }
+  return 0;
+}
 
+static int32_t mndCheckAlterColForStream(SMnode *pMnode, const char *stbFullName, int64_t suid, col_id_t colId) {
+  SSdb *pSdb = pMnode->pSdb;
+  void *pIter = NULL;
+  while (1) {
+    SStreamObj *pStream = NULL;
+    pIter = sdbFetch(pSdb, SDB_STREAM, pIter, (void **)&pStream);
+    if (pIter == NULL) break;
+
+    SNode *pAst = NULL;
+    if (nodesStringToNode(pStream->ast, &pAst) != 0) {
+      ASSERT(0);
+      return -1;
+    }
+
+    SNodeList *pNodeList = NULL;
+    nodesCollectColumns((SSelectStmt *)pAst, SQL_CLAUSE_FROM, NULL, COLLECT_COL_TYPE_ALL, &pNodeList);
+    SNode *pNode = NULL;
+    FOREACH(pNode, pNodeList) {
+      SColumnNode *pCol = (SColumnNode *)pNode;
+
+      if (pCol->tableId != suid) {
+        mDebug("stream:%s, check colId:%d passed", pStream->name, pCol->colId);
+        goto NEXT;
+      }
+      if (pCol->colId > 0 && pCol->colId == colId) {
+        sdbRelease(pSdb, pStream);
+        nodesDestroyNode(pAst);
+        terrno = TSDB_CODE_MND_STREAM_MUST_BE_DELETED;
+        mError("stream:%s, check colId:%d conflicted", pStream->name, pCol->colId);
+        return -1;
+      }
+      mDebug("stream:%s, check colId:%d passed", pStream->name, pCol->colId);
+    }
+
+  NEXT:
+    sdbRelease(pSdb, pStream);
+    nodesDestroyNode(pAst);
+  }
+  return 0;
+}
+
+static int32_t mndCheckAlterColForTSma(SMnode *pMnode, const char *stbFullName, int64_t suid, col_id_t colId) {
+  SSdb *pSdb = pMnode->pSdb;
+  void *pIter = NULL;
   while (1) {
     SSmaObj *pSma = NULL;
     pIter = sdbFetch(pSdb, SDB_SMA, pIter, (void **)&pSma);
     if (pIter == NULL) break;
 
-    mDebug("tsma:%s, check tag and column modifiable, stb:%s suid:%" PRId64 " colId:%d, sql:%s", pSma->name, stbname,
-           suid, colId, pSma->sql);
+    mDebug("tsma:%s, check tag and column modifiable, stb:%s suid:%" PRId64 " colId:%d, sql:%s", pSma->name,
+           stbFullName, suid, colId, pSma->sql);
 
     SNode *pAst = NULL;
     if (nodesStringToNode(pSma->ast, &pAst) != 0) {
       terrno = TSDB_CODE_SDB_INVALID_DATA_CONTENT;
       mError("tsma:%s, check tag and column modifiable, stb:%s suid:%" PRId64 " colId:%d failed since parse AST err",
-             pSma->name, stbname, suid, colId);
+             pSma->name, stbFullName, suid, colId);
       return -1;
     }
 
@@ -1218,7 +1268,7 @@ int32_t mndCheckColAndTagModifiable(SMnode *pMnode, const char *stbname, int64_t
 
       if ((pCol->tableId != suid) && (pSma->stbUid != suid)) {
         mDebug("tsma:%s, check colId:%d passed", pSma->name, pCol->colId);
-        goto NEXT2;
+        goto NEXT;
       }
       if ((pCol->colId) > 0 && (pCol->colId == colId)) {
         sdbRelease(pSdb, pSma);
@@ -1230,11 +1280,24 @@ int32_t mndCheckColAndTagModifiable(SMnode *pMnode, const char *stbname, int64_t
       mDebug("tsma:%s, check colId:%d passed", pSma->name, pCol->colId);
     }
 
-  NEXT2:
+  NEXT:
     sdbRelease(pSdb, pSma);
     nodesDestroyNode(pAst);
   }
+  return 0;
+}
 
+int32_t mndCheckColAndTagModifiable(SMnode *pMnode, const char *stbFullName, int64_t suid, col_id_t colId) {
+  if (mndCheckAlterColForTopic(pMnode, stbFullName, suid, colId) < 0) {
+    return -1;
+  }
+  if (mndCheckAlterColForStream(pMnode, stbFullName, suid, colId) < 0) {
+    return -1;
+  }
+
+  if (mndCheckAlterColForTSma(pMnode, stbFullName, suid, colId) < 0) {
+    return -1;
+  }
   return 0;
 }
 
@@ -1930,6 +1993,98 @@ _OVER:
   return code;
 }
 
+static int32_t mndCheckDropStbForTopic(SMnode *pMnode, const char *stbFullName, int64_t suid) {
+  SSdb *pSdb = pMnode->pSdb;
+  void *pIter = NULL;
+  while (1) {
+    SMqTopicObj *pTopic = NULL;
+    pIter = sdbFetch(pSdb, SDB_TOPIC, pIter, (void **)&pTopic);
+    if (pIter == NULL) break;
+
+    if (pTopic->subType == TOPIC_SUB_TYPE__TABLE) {
+      if (pTopic->stbUid == suid) {
+        sdbRelease(pSdb, pTopic);
+        return -1;
+      }
+    }
+
+    if (pTopic->subType != TOPIC_SUB_TYPE__COLUMN) {
+      sdbRelease(pSdb, pTopic);
+      continue;
+    }
+
+    SNode *pAst = NULL;
+    if (nodesStringToNode(pTopic->ast, &pAst) != 0) {
+      ASSERT(0);
+      return -1;
+    }
+
+    SNodeList *pNodeList = NULL;
+    nodesCollectColumns((SSelectStmt *)pAst, SQL_CLAUSE_FROM, NULL, COLLECT_COL_TYPE_ALL, &pNodeList);
+    SNode *pNode = NULL;
+    FOREACH(pNode, pNodeList) {
+      SColumnNode *pCol = (SColumnNode *)pNode;
+
+      if (pCol->tableId == suid) {
+        sdbRelease(pSdb, pTopic);
+        nodesDestroyNode(pAst);
+        return -1;
+      } else {
+        goto NEXT;
+      }
+    }
+  NEXT:
+    sdbRelease(pSdb, pTopic);
+    nodesDestroyNode(pAst);
+  }
+  return 0;
+}
+
+static int32_t mndCheckDropStbForStream(SMnode *pMnode, const char *stbFullName, int64_t suid) {
+  SSdb *pSdb = pMnode->pSdb;
+  void *pIter = NULL;
+  while (1) {
+    SStreamObj *pStream = NULL;
+    pIter = sdbFetch(pSdb, SDB_STREAM, pIter, (void **)&pStream);
+    if (pIter == NULL) break;
+
+    if (pStream->smaId != 0) {
+      sdbRelease(pSdb, pStream);
+      continue;
+    }
+
+    if (pStream->targetStbUid == suid) {
+      sdbRelease(pSdb, pStream);
+      return -1;
+    }
+
+    SNode *pAst = NULL;
+    if (nodesStringToNode(pStream->ast, &pAst) != 0) {
+      ASSERT(0);
+      return -1;
+    }
+
+    SNodeList *pNodeList = NULL;
+    nodesCollectColumns((SSelectStmt *)pAst, SQL_CLAUSE_FROM, NULL, COLLECT_COL_TYPE_ALL, &pNodeList);
+    SNode *pNode = NULL;
+    FOREACH(pNode, pNodeList) {
+      SColumnNode *pCol = (SColumnNode *)pNode;
+
+      if (pCol->tableId == suid) {
+        sdbRelease(pSdb, pStream);
+        nodesDestroyNode(pAst);
+        return -1;
+      } else {
+        goto NEXT;
+      }
+    }
+  NEXT:
+    sdbRelease(pSdb, pStream);
+    nodesDestroyNode(pAst);
+  }
+  return 0;
+}
+
 static int32_t mndProcessDropStbReq(SRpcMsg *pReq) {
   SMnode      *pMnode = pReq->info.node;
   int32_t      code = -1;
@@ -1971,6 +2126,16 @@ static int32_t mndProcessDropStbReq(SRpcMsg *pReq) {
     goto _OVER;
   }
 
+  if (mndCheckDropStbForTopic(pMnode, dropReq.name, pStb->uid) < 0) {
+    terrno = TSDB_CODE_MND_TOPIC_MUST_BE_DELETED;
+    goto _OVER;
+  }
+
+  if (mndCheckDropStbForStream(pMnode, dropReq.name, pStb->uid) < 0) {
+    terrno = TSDB_CODE_MND_STREAM_MUST_BE_DELETED;
+    goto _OVER;
+  }
+
   code = mndDropStb(pMnode, pReq, pDb, pStb);
   if (code == 0) code = TSDB_CODE_ACTION_IN_PROGRESS;
 

source/dnode/mnode/impl/src/mndStream.c

@@ -167,6 +167,9 @@ static int32_t mndStreamActionInsert(SSdb *pSdb, SStreamObj *pStream) {
 
 static int32_t mndStreamActionDelete(SSdb *pSdb, SStreamObj *pStream) {
   mTrace("stream:%s, perform delete action", pStream->name);
+  taosWLockLatch(&pStream->lock);
+  tFreeStreamObj(pStream);
+  taosWUnLockLatch(&pStream->lock);
   return 0;
 }
 
@@ -493,10 +496,17 @@ static int32_t mndCreateStbForStream(SMnode *pMnode, STrans *pTrans, const SStre
 
   stbObj.uid = pStream->targetStbUid;
 
-  if (mndAddStbToTrans(pMnode, pTrans, pDb, &stbObj) < 0) goto _OVER;
+  if (mndAddStbToTrans(pMnode, pTrans, pDb, &stbObj) < 0) {
+    mndFreeStb(&stbObj);
+    goto _OVER;
+  }
+
+  tFreeSMCreateStbReq(&createReq);
+  mndFreeStb(&stbObj);
 
   return 0;
 _OVER:
+  tFreeSMCreateStbReq(&createReq);
   mndReleaseStb(pMnode, pStb);
   mndReleaseDb(pMnode, pDb);
   return -1;
@@ -715,6 +725,7 @@ _OVER:
   mndReleaseDb(pMnode, pDb);
 
   tFreeSCMCreateStreamReq(&createStreamReq);
+  tFreeStreamObj(&streamObj);
   return code;
 }
 

source/dnode/mnode/impl/src/mndSubscribe.c

@@ -356,31 +356,44 @@ static int32_t mndDoRebalance(SMnode *pMnode, const SMqRebInputObj *pInput, SMqR
       taosArrayPush(pConsumerEp->vgs, &pRebVg->pVgEp);
       pRebVg->newConsumerId = pConsumerEp->consumerId;
       taosArrayPush(pOutput->rebVgs, pRebVg);
-      mInfo("mq rebalance: add vgId:%d to consumer:%" PRId64 ",(second scan)", pRebVg->pVgEp->vgId,
+      mInfo("mq rebalance: add vgId:%d to consumer:%" PRId64 " (second scan) (not enough)", pRebVg->pVgEp->vgId,
             pConsumerEp->consumerId);
     }
   }
 
+  ASSERT(pIter == NULL);
   // 7. handle unassigned vg
   if (taosHashGetSize(pOutput->pSub->consumerHash) != 0) {
     // if has consumer, assign all left vg
     while (1) {
+      SMqConsumerEp *pConsumerEp = NULL;
       pRemovedIter = taosHashIterate(pHash, pRemovedIter);
-      if (pRemovedIter == NULL) break;
-      pIter = taosHashIterate(pOutput->pSub->consumerHash, pIter);
-      ASSERT(pIter);
+      if (pRemovedIter == NULL) {
+        if (pIter != NULL) {
+          taosHashCancelIterate(pOutput->pSub->consumerHash, pIter);
+          pIter = NULL;
+        }
+        break;
+      }
+      while (1) {
+        pIter = taosHashIterate(pOutput->pSub->consumerHash, pIter);
+        ASSERT(pIter);
+        pConsumerEp = (SMqConsumerEp *)pIter;
+        ASSERT(pConsumerEp->consumerId > 0);
+        if (taosArrayGetSize(pConsumerEp->vgs) == minVgCnt) {
+          break;
+        }
+      }
       pRebVg = (SMqRebOutputVg *)pRemovedIter;
-      SMqConsumerEp *pConsumerEp = (SMqConsumerEp *)pIter;
-      ASSERT(pConsumerEp->consumerId > 0);
       taosArrayPush(pConsumerEp->vgs, &pRebVg->pVgEp);
       pRebVg->newConsumerId = pConsumerEp->consumerId;
       if (pRebVg->newConsumerId == pRebVg->oldConsumerId) {
-        mInfo("mq rebalance: skip vg %d for same consumer:%" PRId64 ",(second scan)", pRebVg->pVgEp->vgId,
+        mInfo("mq rebalance: skip vg %d for same consumer:%" PRId64 " (second scan)", pRebVg->pVgEp->vgId,
               pConsumerEp->consumerId);
         continue;
       }
       taosArrayPush(pOutput->rebVgs, pRebVg);
-      mInfo("mq rebalance: add vgId:%d to consumer:%" PRId64 ",(second scan)", pRebVg->pVgEp->vgId,
+      mInfo("mq rebalance: add vgId:%d to consumer:%" PRId64 " (second scan) (unassigned)", pRebVg->pVgEp->vgId,
             pConsumerEp->consumerId);
     }
   } else {
@@ -571,7 +584,7 @@ static int32_t mndProcessRebalanceReq(SRpcMsg *pMsg) {
       SMqTopicObj *pTopic = mndAcquireTopic(pMnode, topic);
       /*ASSERT(pTopic);*/
       if (pTopic == NULL) {
-        mError("rebalance %s failed since topic %s was dropped, abort", pRebInfo->key, topic);
+        mError("mq rebalance %s failed since topic %s not exist, abort", pRebInfo->key, topic);
         continue;
       }
       taosRLockLatch(&pTopic->lock);
@@ -601,7 +614,7 @@ static int32_t mndProcessRebalanceReq(SRpcMsg *pMsg) {
 
     // TODO replace assert with error check
     if (mndPersistRebResult(pMnode, pMsg, &rebOutput) < 0) {
-      mError("persist rebalance output error, possibly vnode splitted or dropped");
+      mError("mq rebalance persist rebalance output error, possibly vnode splitted or dropped");
     }
     taosArrayDestroy(pRebInfo->lostConsumers);
     taosArrayDestroy(pRebInfo->newConsumers);

source/dnode/mnode/impl/src/mndTopic.c

@@ -57,7 +57,8 @@ int32_t mndInitTopic(SMnode *pMnode) {
   mndSetMsgHandle(pMnode, TDMT_MND_CREATE_TOPIC, mndProcessCreateTopicReq);
   mndSetMsgHandle(pMnode, TDMT_MND_DROP_TOPIC, mndProcessDropTopicReq);
   mndSetMsgHandle(pMnode, TDMT_VND_DROP_TOPIC_RSP, mndTransProcessRsp);
-  mndSetMsgHandle(pMnode, TDMT_VND_CHECK_ALTER_INFO_RSP, mndTransProcessRsp);
+  mndSetMsgHandle(pMnode, TDMT_VND_ADD_CHECK_INFO_RSP, mndTransProcessRsp);
+  mndSetMsgHandle(pMnode, TDMT_VND_DELETE_CHECK_INFO_RSP, mndTransProcessRsp);
 
   mndAddShowRetrieveHandle(pMnode, TSDB_MGMT_TABLE_TOPICS, mndRetrieveTopic);
   mndAddShowFreeIterHandle(pMnode, TSDB_MGMT_TABLE_TOPICS, mndCancelGetNextTopic);
@@ -450,7 +451,7 @@ static int32_t mndCreateTopic(SMnode *pMnode, SRpcMsg *pReq, SCMCreateTopicReq *
   sdbSetRawStatus(pCommitRaw, SDB_STATUS_READY);
 
   if (topicObj.ntbUid != 0) {
-    SCheckAlterInfo info;
+    STqCheckInfo info;
     memcpy(info.topic, topicObj.name, TSDB_TOPIC_FNAME_LEN);
     info.ntbUid = topicObj.ntbUid;
     info.colIdList = topicObj.ntbColIds;
@@ -470,7 +471,7 @@ static int32_t mndCreateTopic(SMnode *pMnode, SRpcMsg *pReq, SCMCreateTopicReq *
       // encoder check alter info
       int32_t len;
       int32_t code;
-      tEncodeSize(tEncodeSCheckAlterInfo, &info, len, code);
+      tEncodeSize(tEncodeSTqCheckInfo, &info, len, code);
       if (code < 0) {
         sdbRelease(pSdb, pVgroup);
         mndTransDrop(pTrans);
@@ -481,7 +482,7 @@ static int32_t mndCreateTopic(SMnode *pMnode, SRpcMsg *pReq, SCMCreateTopicReq *
       void    *abuf = POINTER_SHIFT(buf, sizeof(SMsgHead));
       SEncoder encoder;
       tEncoderInit(&encoder, abuf, len);
-      if (tEncodeSCheckAlterInfo(&encoder, &info) < 0) {
+      if (tEncodeSTqCheckInfo(&encoder, &info) < 0) {
         sdbRelease(pSdb, pVgroup);
         mndTransDrop(pTrans);
         return -1;
@@ -493,7 +494,7 @@ static int32_t mndCreateTopic(SMnode *pMnode, SRpcMsg *pReq, SCMCreateTopicReq *
       action.epSet = mndGetVgroupEpset(pMnode, pVgroup);
       action.pCont = buf;
       action.contLen = sizeof(SMsgHead) + len;
-      action.msgType = TDMT_VND_CHECK_ALTER_INFO;
+      action.msgType = TDMT_VND_ADD_CHECK_INFO;
       if (mndTransAppendRedoAction(pTrans, &action) != 0) {
         taosMemoryFree(buf);
         sdbRelease(pSdb, pVgroup);
@@ -659,12 +660,14 @@ static int32_t mndProcessDropTopicReq(SRpcMsg *pReq) {
 
   mDebug("trans:%d, used to drop topic:%s", pTrans->id, pTopic->name);
 
+#if 0
   if (mndDropOffsetByTopic(pMnode, pTrans, dropReq.name) < 0) {
     ASSERT(0);
     mndTransDrop(pTrans);
     mndReleaseTopic(pMnode, pTopic);
     return -1;
   }
+#endif
 
   // TODO check if rebalancing
   if (mndDropSubByTopic(pMnode, pTrans, dropReq.name) < 0) {
@@ -675,6 +678,37 @@ static int32_t mndProcessDropTopicReq(SRpcMsg *pReq) {
     return -1;
   }
 
+  if (pTopic->ntbUid != 0) {
+    // broadcast to all vnode
+    void   *pIter = NULL;
+    SVgObj *pVgroup = NULL;
+    while (1) {
+      pIter = sdbFetch(pSdb, SDB_VGROUP, pIter, (void **)&pVgroup);
+      if (pIter == NULL) break;
+      if (!mndVgroupInDb(pVgroup, pTopic->dbUid)) {
+        sdbRelease(pSdb, pVgroup);
+        continue;
+      }
+
+      void *buf = taosMemoryCalloc(1, sizeof(SMsgHead) + TSDB_TOPIC_FNAME_LEN);
+      void *abuf = POINTER_SHIFT(buf, sizeof(SMsgHead));
+      ((SMsgHead *)buf)->vgId = htonl(pVgroup->vgId);
+      memcpy(abuf, pTopic->name, TSDB_TOPIC_FNAME_LEN);
+
+      STransAction action = {0};
+      action.epSet = mndGetVgroupEpset(pMnode, pVgroup);
+      action.pCont = buf;
+      action.contLen = sizeof(SMsgHead) + TSDB_TOPIC_FNAME_LEN;
+      action.msgType = TDMT_VND_DELETE_CHECK_INFO;
+      if (mndTransAppendRedoAction(pTrans, &action) != 0) {
+        taosMemoryFree(buf);
+        sdbRelease(pSdb, pVgroup);
+        mndTransDrop(pTrans);
+        return -1;
+      }
+    }
+  }
+
   int32_t code = mndDropTopic(pMnode, pTrans, pReq, pTopic);
   mndReleaseTopic(pMnode, pTopic);
 

source/dnode/mnode/impl/src/mndVgroup.c

@@ -509,6 +509,7 @@ int32_t mndAllocSmaVgroup(SMnode *pMnode, SDbObj *pDb, SVgObj *pVgroup) {
   pVgroup->replica = 1;
 
   if (mndGetAvailableDnode(pMnode, pDb, pVgroup, pArray) != 0) return -1;
+  taosArrayDestroy(pArray);
 
   mInfo("db:%s, sma vgId:%d is alloced", pDb->name, pVgroup->vgId);
   return 0;
@@ -1862,4 +1863,4 @@ _OVER:
 #endif
 }
 
-bool mndVgroupInDb(SVgObj *pVgroup, int64_t dbUid) { return !pVgroup->isTsma && pVgroup->dbUid == dbUid; }
+bool mndVgroupInDb(SVgObj *pVgroup, int64_t dbUid) { return !pVgroup->isTsma && pVgroup->dbUid == dbUid; }
\ No newline at end of file

source/dnode/vnode/CMakeLists.txt

@@ -24,6 +24,7 @@ target_sources(
     "src/meta/metaCommit.c"
     "src/meta/metaEntry.c"
     "src/meta/metaSnapshot.c"
+    "src/meta/metaCache.c"
 
     # sma
     "src/sma/smaEnv.c"

source/dnode/vnode/inc/vnode.h

@@ -63,6 +63,7 @@ void    vnodeGetInfo(SVnode *pVnode, const char **dbname, int32_t *vgId);
 int32_t vnodeProcessCreateTSma(SVnode *pVnode, void *pCont, uint32_t contLen);
 int32_t vnodeGetAllTableList(SVnode *pVnode, uint64_t uid, SArray *list);
 int32_t vnodeGetCtbIdList(SVnode *pVnode, int64_t suid, SArray *list);
+int32_t vnodeGetStbIdList(SVnode *pVnode, int64_t suid, SArray* list);
 void   *vnodeGetIdx(SVnode *pVnode);
 void   *vnodeGetIvtIdx(SVnode *pVnode);
 
@@ -91,9 +92,11 @@ typedef struct SMetaEntry  SMetaEntry;
 void        metaReaderInit(SMetaReader *pReader, SMeta *pMeta, int32_t flags);
 void        metaReaderClear(SMetaReader *pReader);
 int32_t     metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid);
+int32_t     metaGetTableTags(SMeta *pMeta, uint64_t suid, SArray *uidList, SHashObj *tags);
 int32_t     metaReadNext(SMetaReader *pReader);
-const void *metaGetTableTagVal(SMetaEntry *pEntry, int16_t type, STagVal *tagVal);
+const void *metaGetTableTagVal(void *tag, int16_t type, STagVal *tagVal);
 int         metaGetTableNameByUid(void *meta, uint64_t uid, char *tbName);
+bool        metaIsTableExist(SMeta  *pMeta, tb_uid_t uid);
 
 typedef struct SMetaFltParam {
   tb_uid_t suid;

source/dnode/vnode/src/inc/meta.h

@@ -23,8 +23,9 @@
 extern "C" {
 #endif
 
-typedef struct SMetaIdx SMetaIdx;
-typedef struct SMetaDB  SMetaDB;
+typedef struct SMetaIdx   SMetaIdx;
+typedef struct SMetaDB    SMetaDB;
+typedef struct SMetaCache SMetaCache;
 
 // metaDebug ==================
 // clang-format off
@@ -60,6 +61,12 @@ static FORCE_INLINE tb_uid_t metaGenerateUid(SMeta* pMeta) { return tGenIdPI64()
 // metaTable ==================
 int metaHandleEntry(SMeta* pMeta, const SMetaEntry* pME);
 
+// metaCache ==================
+int32_t metaCacheOpen(SMeta* pMeta);
+void    metaCacheClose(SMeta* pMeta);
+int32_t metaCacheUpsert(SMeta* pMeta, SMetaInfo* pInfo);
+int32_t metaCacheDrop(SMeta* pMeta, int64_t uid);
+
 struct SMeta {
   TdThreadRwlock lock;
 
@@ -84,6 +91,8 @@ struct SMeta {
   TTB* pStreamDb;
 
   SMetaIdx* pIdx;
+
+  SMetaCache* pCache;
 };
 
 typedef struct {
@@ -92,6 +101,12 @@ typedef struct {
 } STbDbKey;
 
 #pragma pack(push, 1)
+typedef struct {
+  tb_uid_t suid;
+  int64_t  version;
+  int32_t  skmVer;
+} SUidIdxVal;
+
 typedef struct {
   tb_uid_t uid;
   int32_t  sver;

source/dnode/vnode/src/inc/sma.h

@@ -57,9 +57,10 @@ typedef struct {
   void   *tmrHandle;  // shared by all fetch tasks
 } SSmaMgmt;
 
-#define SMA_ENV_LOCK(env) (&(env)->lock)
-#define SMA_ENV_TYPE(env) ((env)->type)
-#define SMA_ENV_STAT(env) ((env)->pStat)
+#define SMA_ENV_LOCK(env)  (&(env)->lock)
+#define SMA_ENV_TYPE(env)  ((env)->type)
+#define SMA_ENV_STAT(env)  ((env)->pStat)
+#define SMA_RSMA_STAT(sma) ((SRSmaStat *)SMA_ENV_STAT((SSmaEnv *)(sma)->pRSmaEnv))
 
 struct STSmaStat {
   int8_t    state;  // ETsdbSmaStat
@@ -86,15 +87,17 @@ struct SQTaskFWriter {
 };
 
 struct SRSmaStat {
-  SSma     *pSma;
-  int64_t   commitAppliedVer;  // vnode applied version for async commit
-  int64_t   refId;             // shared by fetch tasks
-  SRWLatch  lock;              // r/w lock for rsma fs(e.g. qtaskinfo)
-  int8_t    triggerStat;       // shared by fetch tasks
-  int8_t    commitStat;        // 0 not in committing, 1 in committing
-  SArray   *aTaskFile;         // qTaskFiles committed recently(for recovery/snapshot r/w)
-  SHashObj *rsmaInfoHash;      // key: stbUid, value: SRSmaInfo;
-  SHashObj *iRsmaInfoHash;     // key: stbUid, value: SRSmaInfo; immutable rsmaInfoHash
+  SSma            *pSma;
+  int64_t          commitAppliedVer;  // vnode applied version for async commit
+  int64_t          refId;             // shared by fetch tasks
+  volatile int64_t qBufSize;          // queue buffer size
+  SRWLatch         lock;              // r/w lock for rsma fs(e.g. qtaskinfo)
+  int8_t           triggerStat;       // shared by fetch tasks
+  int8_t           commitStat;        // 0 not in committing, 1 in committing
+  int8_t           execStat;          // 0 not in exec , 1 in exec
+  SArray          *aTaskFile;         // qTaskFiles committed recently(for recovery/snapshot r/w)
+  SHashObj        *infoHash;          // key: suid, value: SRSmaInfo
+  SHashObj        *fetchHash;         // key: suid, value: L1 or L2 or L1|L2
 };
 
 struct SSmaStat {
@@ -105,34 +108,40 @@ struct SSmaStat {
   T_REF_DECLARE()
 };
 
-#define SMA_TSMA_STAT(s)      (&(s)->tsmaStat)
-#define SMA_RSMA_STAT(s)      (&(s)->rsmaStat)
-#define RSMA_INFO_HASH(r)     ((r)->rsmaInfoHash)
-#define RSMA_IMU_INFO_HASH(r) ((r)->iRsmaInfoHash)
-#define RSMA_TRIGGER_STAT(r)  (&(r)->triggerStat)
-#define RSMA_COMMIT_STAT(r)   (&(r)->commitStat)
-#define RSMA_REF_ID(r)        ((r)->refId)
-#define RSMA_FS_LOCK(r)       (&(r)->lock)
+#define SMA_STAT_TSMA(s)     (&(s)->tsmaStat)
+#define SMA_STAT_RSMA(s)     (&(s)->rsmaStat)
+#define RSMA_INFO_HASH(r)    ((r)->infoHash)
+#define RSMA_FETCH_HASH(r)   ((r)->fetchHash)
+#define RSMA_TRIGGER_STAT(r) (&(r)->triggerStat)
+#define RSMA_COMMIT_STAT(r)  (&(r)->commitStat)
+#define RSMA_REF_ID(r)       ((r)->refId)
+#define RSMA_FS_LOCK(r)      (&(r)->lock)
 
 struct SRSmaInfoItem {
-  int8_t  level;
-  int8_t  triggerStat;
-  int32_t maxDelay;
-  tmr_h   tmrId;
+  int8_t   level;
+  int8_t   triggerStat;
+  uint16_t interval;  // second
+  int32_t  maxDelay;
+  tmr_h    tmrId;
 };
 
 struct SRSmaInfo {
   STSchema *pTSchema;
   int64_t   suid;
   int64_t   refId;  // refId of SRSmaStat
-  int8_t    delFlag;
+  uint64_t  delFlag : 1;
+  uint64_t  lastReceived : 63;  // second
   T_REF_DECLARE()
   SRSmaInfoItem items[TSDB_RETENTION_L2];
   void         *taskInfo[TSDB_RETENTION_L2];   // qTaskInfo_t
-  void         *iTaskInfo[TSDB_RETENTION_L2];  // immutable
+  STaosQueue   *queue;                         // buffer queue of SubmitReq
+  STaosQall    *qall;                          // buffer qall of SubmitReq
+  void         *iTaskInfo[TSDB_RETENTION_L2];  // immutable qTaskInfo_t
+  STaosQueue   *iQueue;                        // immutable buffer queue of SubmitReq
+  STaosQall    *iQall;                         // immutable buffer qall of SubmitReq
 };
 
-#define RSMA_INFO_HEAD_LEN     32
+#define RSMA_INFO_HEAD_LEN     offsetof(SRSmaInfo, items)
 #define RSMA_INFO_IS_DEL(r)    ((r)->delFlag == 1)
 #define RSMA_INFO_SET_DEL(r)   ((r)->delFlag = 1)
 #define RSMA_INFO_QTASK(r, i)  ((r)->taskInfo[i])
@@ -161,6 +170,12 @@ enum {
   RSMA_RESTORE_SYNC = 2,
 };
 
+typedef enum {
+  RSMA_EXEC_OVERFLOW = 1,  // triggered by queue buf overflow
+  RSMA_EXEC_TIMEOUT = 2,   // triggered by timer
+  RSMA_EXEC_COMMIT = 3,    // triggered by commit
+} ERsmaExecType;
+
 void  tdDestroySmaEnv(SSmaEnv *pSmaEnv);
 void *tdFreeSmaEnv(SSmaEnv *pSmaEnv);
 
@@ -228,12 +243,13 @@ static FORCE_INLINE void tdSmaStatSetDropped(STSmaStat *pTStat) {
 
 void           tdRSmaQTaskInfoGetFileName(int32_t vid, int64_t version, char *outputName);
 void           tdRSmaQTaskInfoGetFullName(int32_t vid, int64_t version, const char *path, char *outputName);
-int32_t        tdCloneRSmaInfo(SSma *pSma, SRSmaInfo **pDest, SRSmaInfo *pSrc);
+int32_t        tdCloneRSmaInfo(SSma *pSma, SRSmaInfo *pInfo);
 void           tdFreeQTaskInfo(qTaskInfo_t *taskHandle, int32_t vgId, int32_t level);
 static int32_t tdDestroySmaState(SSmaStat *pSmaStat, int8_t smaType);
 void          *tdFreeSmaState(SSmaStat *pSmaStat, int8_t smaType);
 void          *tdFreeRSmaInfo(SSma *pSma, SRSmaInfo *pInfo, bool isDeepFree);
 int32_t        tdRSmaPersistExecImpl(SRSmaStat *pRSmaStat, SHashObj *pInfoHash);
+int32_t        tdRSmaProcessExecImpl(SSma *pSma, ERsmaExecType type);
 
 int32_t tdProcessRSmaCreateImpl(SSma *pSma, SRSmaParam *param, int64_t suid, const char *tbName);
 int32_t tdProcessRSmaRestoreImpl(SSma *pSma, int8_t type, int64_t qtaskFileVer);

source/dnode/vnode/src/inc/tq.h

@@ -117,16 +117,15 @@ typedef struct {
 struct STQ {
   SVnode*   pVnode;
   char*     path;
-  SHashObj* pushMgr;     // consumerId -> STqHandle*
-  SHashObj* handles;     // subKey -> STqHandle
-  SHashObj* pAlterInfo;  // topic -> SAlterCheckInfo
+  SHashObj* pPushMgr;    // consumerId -> STqHandle*
+  SHashObj* pHandle;     // subKey -> STqHandle
+  SHashObj* pCheckInfo;  // topic -> SAlterCheckInfo
 
   STqOffsetStore* pOffsetStore;
 
-  TDB* pMetaStore;
+  TDB* pMetaDB;
   TTB* pExecStore;
-
-  TTB* pAlterInfoStore;
+  TTB* pCheckStore;
 
   SStreamMeta* pStreamMeta;
 };
@@ -155,6 +154,9 @@ int32_t tqMetaClose(STQ* pTq);
 int32_t tqMetaSaveHandle(STQ* pTq, const char* key, const STqHandle* pHandle);
 int32_t tqMetaDeleteHandle(STQ* pTq, const char* key);
 int32_t tqMetaRestoreHandle(STQ* pTq);
+int32_t tqMetaSaveCheckInfo(STQ* pTq, const char* key, const void* value, int32_t vLen);
+int32_t tqMetaDeleteCheckInfo(STQ* pTq, const char* key);
+int32_t tqMetaRestoreCheckInfo(STQ* pTq);
 
 typedef struct {
   int32_t size;

source/dnode/vnode/src/inc/tsdb.h

@@ -45,7 +45,7 @@ typedef struct SBlockIdx     SBlockIdx;
 typedef struct SBlock        SBlock;
 typedef struct SBlockL       SBlockL;
 typedef struct SColData      SColData;
-typedef struct SBlockDataHdr SBlockDataHdr;
+typedef struct SDiskDataHdr  SDiskDataHdr;
 typedef struct SBlockData    SBlockData;
 typedef struct SDelFile      SDelFile;
 typedef struct SHeadFile     SHeadFile;
@@ -61,7 +61,11 @@ typedef struct SRowIter      SRowIter;
 typedef struct STsdbFS       STsdbFS;
 typedef struct SRowMerger    SRowMerger;
 typedef struct STsdbReadSnap STsdbReadSnap;
+typedef struct SBlockInfo    SBlockInfo;
+typedef struct SSmaInfo      SSmaInfo;
+typedef struct SBlockCol     SBlockCol;
 
+#define TSDB_FILE_DLMT     ((uint32_t)0xF00AFA0F)
 #define TSDB_MAX_SUBBLOCKS 8
 #define TSDB_FHDR_SIZE     512
 
@@ -113,10 +117,14 @@ int32_t tPutBlock(uint8_t *p, void *ph);
 int32_t tGetBlock(uint8_t *p, void *ph);
 int32_t tBlockCmprFn(const void *p1, const void *p2);
 bool    tBlockHasSma(SBlock *pBlock);
+// SBlockL
+int32_t tPutBlockL(uint8_t *p, void *ph);
+int32_t tGetBlockL(uint8_t *p, void *ph);
 // SBlockIdx
 int32_t tPutBlockIdx(uint8_t *p, void *ph);
 int32_t tGetBlockIdx(uint8_t *p, void *ph);
 int32_t tCmprBlockIdx(void const *lhs, void const *rhs);
+int32_t tCmprBlockL(void const *lhs, void const *rhs);
 // SColdata
 void    tColDataInit(SColData *pColData, int16_t cid, int8_t type, int8_t smaOn);
 void    tColDataReset(SColData *pColData);
@@ -131,20 +139,25 @@ int32_t tGetColData(uint8_t *p, SColData *pColData);
 #define tBlockDataLastRow(PBLOCKDATA)  tsdbRowFromBlockData(PBLOCKDATA, (PBLOCKDATA)->nRow - 1)
 #define tBlockDataFirstKey(PBLOCKDATA) TSDBROW_KEY(&tBlockDataFirstRow(PBLOCKDATA))
 #define tBlockDataLastKey(PBLOCKDATA)  TSDBROW_KEY(&tBlockDataLastRow(PBLOCKDATA))
-int32_t   tBlockDataInit(SBlockData *pBlockData);
+
+int32_t   tBlockDataCreate(SBlockData *pBlockData);
+void      tBlockDataDestroy(SBlockData *pBlockData, int8_t deepClear);
+int32_t   tBlockDataInit(SBlockData *pBlockData, int64_t suid, int64_t uid, STSchema *pTSchema);
+int32_t   tBlockDataInitEx(SBlockData *pBlockData, SBlockData *pBlockDataFrom);
 void      tBlockDataReset(SBlockData *pBlockData);
-int32_t   tBlockDataSetSchema(SBlockData *pBlockData, STSchema *pTSchema);
-int32_t   tBlockDataCorrectSchema(SBlockData *pBlockData, SBlockData *pBlockDataFrom);
-void      tBlockDataClearData(SBlockData *pBlockData);
-void      tBlockDataClear(SBlockData *pBlockData, int8_t deepClear);
-int32_t   tBlockDataAddColData(SBlockData *pBlockData, int32_t iColData, SColData **ppColData);
-int32_t   tBlockDataAppendRow(SBlockData *pBlockData, TSDBROW *pRow, STSchema *pTSchema);
-int32_t   tBlockDataMerge(SBlockData *pBlockData1, SBlockData *pBlockData2, SBlockData *pBlockData);
-int32_t   tBlockDataCopy(SBlockData *pBlockDataSrc, SBlockData *pBlockDataDest);
+int32_t   tBlockDataAppendRow(SBlockData *pBlockData, TSDBROW *pRow, STSchema *pTSchema, int64_t uid);
+void      tBlockDataClear(SBlockData *pBlockData);
 SColData *tBlockDataGetColDataByIdx(SBlockData *pBlockData, int32_t idx);
 void      tBlockDataGetColData(SBlockData *pBlockData, int16_t cid, SColData **ppColData);
-int32_t   tPutBlockData(uint8_t *p, SBlockData *pBlockData);
-int32_t   tGetBlockData(uint8_t *p, SBlockData *pBlockData);
+int32_t   tBlockDataCopy(SBlockData *pBlockDataSrc, SBlockData *pBlockDataDest);
+int32_t   tBlockDataMerge(SBlockData *pBlockData1, SBlockData *pBlockData2, SBlockData *pBlockData);
+int32_t   tBlockDataAddColData(SBlockData *pBlockData, int32_t iColData, SColData **ppColData);
+int32_t   tCmprBlockData(SBlockData *pBlockData, int8_t cmprAlg, uint8_t **ppOut, int32_t *szOut, uint8_t *aBuf[],
+                         int32_t aBufN[]);
+int32_t   tDecmprBlockData(uint8_t *pIn, int32_t szIn, SBlockData *pBlockData, uint8_t *aBuf[]);
+// SDiskDataHdr
+int32_t tPutDiskDataHdr(uint8_t *p, void *ph);
+int32_t tGetDiskDataHdr(uint8_t *p, void *ph);
 // SDelIdx
 int32_t tPutDelIdx(uint8_t *p, void *ph);
 int32_t tGetDelIdx(uint8_t *p, void *ph);
@@ -168,13 +181,25 @@ void    tsdbFidKeyRange(int32_t fid, int32_t minutes, int8_t precision, TSKEY *m
 int32_t tsdbFidLevel(int32_t fid, STsdbKeepCfg *pKeepCfg, int64_t now);
 int32_t tsdbBuildDeleteSkyline(SArray *aDelData, int32_t sidx, int32_t eidx, SArray *aSkyline);
 void    tsdbCalcColDataSMA(SColData *pColData, SColumnDataAgg *pColAgg);
+int32_t tPutColumnDataAgg(uint8_t *p, SColumnDataAgg *pColAgg);
+int32_t tGetColumnDataAgg(uint8_t *p, SColumnDataAgg *pColAgg);
+int32_t tsdbCmprData(uint8_t *pIn, int32_t szIn, int8_t type, int8_t cmprAlg, uint8_t **ppOut, int32_t nOut,
+                     int32_t *szOut, uint8_t **ppBuf);
+int32_t tsdbDecmprData(uint8_t *pIn, int32_t szIn, int8_t type, int8_t cmprAlg, uint8_t **ppOut, int32_t szOut,
+                       uint8_t **ppBuf);
+int32_t tsdbCmprColData(SColData *pColData, int8_t cmprAlg, SBlockCol *pBlockCol, uint8_t **ppOut, int32_t nOut,
+                        uint8_t **ppBuf);
+int32_t tsdbDecmprColData(uint8_t *pIn, SBlockCol *pBlockCol, int8_t cmprAlg, int32_t nVal, SColData *pColData,
+                          uint8_t **ppBuf);
+int32_t tsdbReadAndCheck(TdFilePtr pFD, int64_t offset, uint8_t **ppOut, int32_t size, int8_t toCheck);
 // tsdbMemTable ==============================================================================================
 // SMemTable
-int32_t tsdbMemTableCreate(STsdb *pTsdb, SMemTable **ppMemTable);
-void    tsdbMemTableDestroy(SMemTable *pMemTable);
-void    tsdbGetTbDataFromMemTable(SMemTable *pMemTable, tb_uid_t suid, tb_uid_t uid, STbData **ppTbData);
-void    tsdbRefMemTable(SMemTable *pMemTable);
-void    tsdbUnrefMemTable(SMemTable *pMemTable);
+int32_t  tsdbMemTableCreate(STsdb *pTsdb, SMemTable **ppMemTable);
+void     tsdbMemTableDestroy(SMemTable *pMemTable);
+STbData *tsdbGetTbDataFromMemTable(SMemTable *pMemTable, tb_uid_t suid, tb_uid_t uid);
+void     tsdbRefMemTable(SMemTable *pMemTable);
+void     tsdbUnrefMemTable(SMemTable *pMemTable);
+SArray  *tsdbMemTableGetTbDataArray(SMemTable *pMemTable);
 // STbDataIter
 int32_t  tsdbTbDataIterCreate(STbData *pTbData, TSDBKEY *pFrom, int8_t backward, STbDataIter **ppIter);
 void    *tsdbTbDataIterDestroy(STbDataIter *pIter);
@@ -223,33 +248,33 @@ int32_t tsdbFSUpsertDelFile(STsdbFS *pFS, SDelFile *pDelFile);
 int32_t tsdbDataFWriterOpen(SDataFWriter **ppWriter, STsdb *pTsdb, SDFileSet *pSet);
 int32_t tsdbDataFWriterClose(SDataFWriter **ppWriter, int8_t sync);
 int32_t tsdbUpdateDFileSetHeader(SDataFWriter *pWriter);
-int32_t tsdbWriteBlockIdx(SDataFWriter *pWriter, SArray *aBlockIdx, uint8_t **ppBuf);
-int32_t tsdbWriteBlock(SDataFWriter *pWriter, SMapData *pMapData, uint8_t **ppBuf, SBlockIdx *pBlockIdx);
-int32_t tsdbWriteBlockData(SDataFWriter *pWriter, SBlockData *pBlockData, uint8_t **ppBuf1, uint8_t **ppBuf2,
-                           SBlockIdx *pBlockIdx, SBlock *pBlock, int8_t cmprAlg);
+int32_t tsdbWriteBlockIdx(SDataFWriter *pWriter, SArray *aBlockIdx);
+int32_t tsdbWriteBlock(SDataFWriter *pWriter, SMapData *pMapData, SBlockIdx *pBlockIdx);
+int32_t tsdbWriteBlockL(SDataFWriter *pWriter, SArray *aBlockL);
+int32_t tsdbWriteBlockData(SDataFWriter *pWriter, SBlockData *pBlockData, SBlockInfo *pBlkInfo, SSmaInfo *pSmaInfo,
+                           int8_t cmprAlg, int8_t toLast);
 
 int32_t tsdbDFileSetCopy(STsdb *pTsdb, SDFileSet *pSetFrom, SDFileSet *pSetTo);
 // SDataFReader
 int32_t tsdbDataFReaderOpen(SDataFReader **ppReader, STsdb *pTsdb, SDFileSet *pSet);
 int32_t tsdbDataFReaderClose(SDataFReader **ppReader);
-int32_t tsdbReadBlockIdx(SDataFReader *pReader, SArray *aBlockIdx, uint8_t **ppBuf);
-int32_t tsdbReadBlock(SDataFReader *pReader, SBlockIdx *pBlockIdx, SMapData *pMapData, uint8_t **ppBuf);
-int32_t tsdbReadColData(SDataFReader *pReader, SBlockIdx *pBlockIdx, SBlock *pBlock, int16_t *aColId, int32_t nCol,
-                        SBlockData *pBlockData, uint8_t **ppBuf1, uint8_t **ppBuf2);
-int32_t tsdbReadBlockData(SDataFReader *pReader, SBlockIdx *pBlockIdx, SBlock *pBlock, SBlockData *pBlockData,
-                          uint8_t **ppBuf1, uint8_t **ppBuf2);
-int32_t tsdbReadBlockSma(SDataFReader *pReader, SBlock *pBlock, SArray *aColumnDataAgg, uint8_t **ppBuf);
+int32_t tsdbReadBlockIdx(SDataFReader *pReader, SArray *aBlockIdx);
+int32_t tsdbReadBlock(SDataFReader *pReader, SBlockIdx *pBlockIdx, SMapData *pMapData);
+int32_t tsdbReadBlockL(SDataFReader *pReader, SArray *aBlockL);
+int32_t tsdbReadBlockSma(SDataFReader *pReader, SBlock *pBlock, SArray *aColumnDataAgg);
+int32_t tsdbReadDataBlock(SDataFReader *pReader, SBlock *pBlock, SBlockData *pBlockData);
+int32_t tsdbReadLastBlock(SDataFReader *pReader, SBlockL *pBlockL, SBlockData *pBlockData);
 // SDelFWriter
 int32_t tsdbDelFWriterOpen(SDelFWriter **ppWriter, SDelFile *pFile, STsdb *pTsdb);
 int32_t tsdbDelFWriterClose(SDelFWriter **ppWriter, int8_t sync);
-int32_t tsdbWriteDelData(SDelFWriter *pWriter, SArray *aDelData, uint8_t **ppBuf, SDelIdx *pDelIdx);
-int32_t tsdbWriteDelIdx(SDelFWriter *pWriter, SArray *aDelIdx, uint8_t **ppBuf);
+int32_t tsdbWriteDelData(SDelFWriter *pWriter, SArray *aDelData, SDelIdx *pDelIdx);
+int32_t tsdbWriteDelIdx(SDelFWriter *pWriter, SArray *aDelIdx);
 int32_t tsdbUpdateDelFileHdr(SDelFWriter *pWriter);
 // SDelFReader
-int32_t tsdbDelFReaderOpen(SDelFReader **ppReader, SDelFile *pFile, STsdb *pTsdb, uint8_t **ppBuf);
+int32_t tsdbDelFReaderOpen(SDelFReader **ppReader, SDelFile *pFile, STsdb *pTsdb);
 int32_t tsdbDelFReaderClose(SDelFReader **ppReader);
-int32_t tsdbReadDelData(SDelFReader *pReader, SDelIdx *pDelIdx, SArray *aDelData, uint8_t **ppBuf);
-int32_t tsdbReadDelIdx(SDelFReader *pReader, SArray *aDelIdx, uint8_t **ppBuf);
+int32_t tsdbReadDelData(SDelFReader *pReader, SDelIdx *pDelIdx, SArray *aDelData);
+int32_t tsdbReadDelIdx(SDelFReader *pReader, SArray *aDelIdx);
 // tsdbRead.c ==============================================================================================
 int32_t tsdbTakeReadSnap(STsdb *pTsdb, STsdbReadSnap **ppSnap);
 void    tsdbUntakeReadSnap(STsdb *pTsdb, STsdbReadSnap *pSnap);
@@ -260,7 +285,7 @@ void    tsdbUntakeReadSnap(STsdb *pTsdb, STsdbReadSnap *pSnap);
 
 // tsdbCache
 int32_t tsdbOpenCache(STsdb *pTsdb);
-void    tsdbCloseCache(SLRUCache *pCache);
+void    tsdbCloseCache(STsdb *pTsdb);
 int32_t tsdbCacheInsertLast(SLRUCache *pCache, tb_uid_t uid, STSRow *row, STsdb *pTsdb);
 int32_t tsdbCacheInsertLastrow(SLRUCache *pCache, STsdb *pTsdb, tb_uid_t uid, STSRow *row, bool dup);
 int32_t tsdbCacheGetLastH(SLRUCache *pCache, tb_uid_t uid, STsdb *pTsdb, LRUHandle **h);
@@ -277,13 +302,6 @@ size_t tsdbCacheGetCapacity(SVnode *pVnode);
 int32_t tsdbCacheLastArray2Row(SArray *pLastArray, STSRow **ppRow, STSchema *pSchema);
 
 // structs =======================
-typedef struct {
-  int   minFid;
-  int   midFid;
-  int   maxFid;
-  TSKEY minKey;
-} SRtn;
-
 struct STsdbFS {
   SDelFile *pDelFile;
   SArray   *aDFileSet;  // SArray<SDFileSet>
@@ -298,6 +316,7 @@ struct STsdb {
   SMemTable     *imem;
   STsdbFS        fs;
   SLRUCache     *lruCache;
+  TdThreadMutex  lruMutex;
 };
 
 struct TSDBKEY {
@@ -311,30 +330,23 @@ struct SMemSkipListNode {
   SMemSkipListNode *forwards[0];
 };
 typedef struct SMemSkipList {
-  uint32_t          seed;
   int64_t           size;
+  uint32_t          seed;
   int8_t            maxLevel;
   int8_t            level;
   SMemSkipListNode *pHead;
   SMemSkipListNode *pTail;
 } SMemSkipList;
 
-struct SDelDataInfo {
-  tb_uid_t suid;
-  tb_uid_t uid;
-};
-
 struct STbData {
   tb_uid_t     suid;
   tb_uid_t     uid;
   TSKEY        minKey;
   TSKEY        maxKey;
-  int64_t      minVersion;
-  int64_t      maxVersion;
-  int32_t      maxSkmVer;
   SDelData    *pHead;
   SDelData    *pTail;
   SMemSkipList sl;
+  STbData     *next;
 };
 
 struct SMemTable {
@@ -344,11 +356,13 @@ struct SMemTable {
   volatile int32_t nRef;
   TSKEY            minKey;
   TSKEY            maxKey;
-  int64_t          minVersion;
-  int64_t          maxVersion;
   int64_t          nRow;
   int64_t          nDel;
-  SArray          *aTbData;  // SArray<STbData*>
+  struct {
+    int32_t   nTbData;
+    int32_t   nBucket;
+    STbData **aBucket;
+  };
 };
 
 struct TSDBROW {
@@ -379,63 +393,51 @@ struct SMapData {
   uint8_t *pData;
 };
 
-typedef struct {
+struct SBlockCol {
   int16_t cid;
   int8_t  type;
   int8_t  smaOn;
-  int8_t  flag;  // HAS_NONE|HAS_NULL|HAS_VALUE
-  int32_t offset;
-  int32_t szBitmap;  // bitmap size
-  int32_t szOffset;  // size of offset, only for variant-length data type
-  int32_t szValue;   // compressed column value size
+  int8_t  flag;      // HAS_NONE|HAS_NULL|HAS_VALUE
   int32_t szOrigin;  // original column value size (only save for variant data type)
-} SBlockCol;
-
-typedef struct {
-  int32_t nRow;
-  int8_t  cmprAlg;
-  int64_t offset;      // block data offset
-  int32_t szBlockCol;  // SBlockCol size
-  int32_t szVersion;   // VERSION size
-  int32_t szTSKEY;     // TSKEY size
-  int32_t szBlock;     // total block size
-  int64_t sOffset;     // sma offset
-  int32_t nSma;        // sma size
-} SSubBlock;
+  int32_t szBitmap;  // bitmap size, 0 only for flag == HAS_VAL
+  int32_t szOffset;  // offset size, 0 only for non-variant-length type
+  int32_t szValue;   // value size, 0 when flag == (HAS_NULL | HAS_NONE)
+  int32_t offset;
+};
+
+struct SBlockInfo {
+  int64_t offset;  // block data offset
+  int32_t szBlock;
+  int32_t szKey;
+};
+
+struct SSmaInfo {
+  int64_t offset;
+  int32_t size;
+};
 
 struct SBlock {
-  TSDBKEY   minKey;
-  TSDBKEY   maxKey;
-  int64_t   minVersion;
-  int64_t   maxVersion;
-  int32_t   nRow;
-  int8_t    last;
-  int8_t    hasDup;
-  int8_t    nSubBlock;
-  SSubBlock aSubBlock[TSDB_MAX_SUBBLOCKS];
+  TSDBKEY    minKey;
+  TSDBKEY    maxKey;
+  int64_t    minVer;
+  int64_t    maxVer;
+  int32_t    nRow;
+  int8_t     hasDup;
+  int8_t     nSubBlock;
+  SBlockInfo aSubBlock[TSDB_MAX_SUBBLOCKS];
+  SSmaInfo   smaInfo;
 };
 
 struct SBlockL {
-  struct {
-    int64_t uid;
-    int64_t version;
-    TSKEY   ts;
-  } minKey;
-  struct {
-    int64_t uid;
-    int64_t version;
-    TSKEY   ts;
-  } maxKey;
-  int64_t minVer;
-  int64_t maxVer;
-  int32_t nRow;
-  int8_t  cmprAlg;
-  int64_t offset;
-  int32_t szBlock;
-  int32_t szBlockCol;
-  int32_t szUid;
-  int32_t szVer;
-  int32_t szTSKEY;
+  int64_t    suid;
+  int64_t    minUid;
+  int64_t    maxUid;
+  TSKEY      minKey;
+  TSKEY      maxKey;
+  int64_t    minVer;
+  int64_t    maxVer;
+  int32_t    nRow;
+  SBlockInfo bInfo;
 };
 
 struct SColData {
@@ -450,10 +452,17 @@ struct SColData {
   uint8_t *pData;
 };
 
+// (SBlockData){.suid = 0, .uid = 0}: block data not initialized
+// (SBlockData){.suid = suid, .uid = uid}: block data for ONE child table int .data file
+// (SBlockData){.suid = suid, .uid = 0}: block data for N child tables int .last file
+// (SBlockData){.suid = 0, .uid = uid}: block data for 1 normal table int .last/.data file
 struct SBlockData {
-  int32_t  nRow;
-  int64_t *aVersion;
-  TSKEY   *aTSKEY;
+  int64_t  suid;      // 0 means normal table block data, otherwise child table block data
+  int64_t  uid;       // 0 means block data in .last file, otherwise in .data file
+  int32_t  nRow;      // number of rows
+  int64_t *aUid;      // uids of each row, only exist in block data in .last file (uid == 0)
+  int64_t *aVersion;  // versions of each row
+  TSKEY   *aTSKEY;    // timestamp of each row
   SArray  *aIdx;      // SArray<int32_t>
   SArray  *aColData;  // SArray<SColData>
 };
@@ -491,13 +500,18 @@ struct SDelIdx {
   int64_t  size;
 };
 
-#pragma pack(push, 1)
-struct SBlockDataHdr {
+struct SDiskDataHdr {
   uint32_t delimiter;
+  uint32_t fmtVer;
   int64_t  suid;
   int64_t  uid;
+  int32_t  szUid;
+  int32_t  szVer;
+  int32_t  szKey;
+  int32_t  szBlkCol;
+  int32_t  nRow;
+  int8_t   cmprAlg;
 };
-#pragma pack(pop)
 
 struct SDelFile {
   volatile int32_t nRef;
@@ -527,6 +541,7 @@ struct SLastFile {
 
   int64_t commitID;
   int64_t size;
+  int64_t offset;
 };
 
 struct SSmaFile {
@@ -561,6 +576,8 @@ struct SDelFWriter {
   STsdb    *pTsdb;
   SDelFile  fDel;
   TdFilePtr pWriteH;
+
+  uint8_t *aBuf[1];
 };
 
 struct SDataFWriter {
@@ -576,6 +593,8 @@ struct SDataFWriter {
   SDataFile fData;
   SLastFile fLast;
   SSmaFile  fSma;
+
+  uint8_t *aBuf[4];
 };
 
 struct STsdbReadSnap {

source/dnode/vnode/src/inc/vnd.h

@@ -65,6 +65,7 @@ struct SVBufPool {
   SVBufPool*       next;
   SVnode*          pVnode;
   volatile int32_t nRef;
+  TdThreadSpinlock lock;
   int64_t          size;
   uint8_t*         ptr;
   SVBufPoolNode*   pTail;

source/dnode/vnode/src/inc/vnodeInt.h

@@ -130,6 +130,14 @@ int             metaTtlSmaller(SMeta* pMeta, uint64_t time, SArray* uidList);
 int32_t metaCreateTSma(SMeta* pMeta, int64_t version, SSmaCfg* pCfg);
 int32_t metaDropTSma(SMeta* pMeta, int64_t indexUid);
 
+typedef struct SMetaInfo {
+  int64_t uid;
+  int64_t suid;
+  int64_t version;
+  int32_t skmVer;
+} SMetaInfo;
+int32_t metaGetInfo(SMeta* pMeta, int64_t uid, SMetaInfo* pInfo);
+
 // tsdb
 int         tsdbOpen(SVnode* pVnode, STsdb** ppTsdb, const char* dir, STsdbKeepCfg* pKeepCfg);
 int         tsdbClose(STsdb** pTsdb);
@@ -155,13 +163,16 @@ int     tqPushMsg(STQ*, void* msg, int32_t msgLen, tmsg_t msgType, int64_t ver);
 int     tqCommit(STQ*);
 int32_t tqUpdateTbUidList(STQ* pTq, const SArray* tbUidList, bool isAdd);
 int32_t tqCheckColModifiable(STQ* pTq, int64_t tbUid, int32_t colId);
-int32_t tqProcessCheckAlterInfoReq(STQ* pTq, char* msg, int32_t msgLen);
-int32_t tqProcessVgChangeReq(STQ* pTq, char* msg, int32_t msgLen);
-int32_t tqProcessVgDeleteReq(STQ* pTq, char* msg, int32_t msgLen);
-int32_t tqProcessOffsetCommitReq(STQ* pTq, char* msg, int32_t msgLen, int64_t ver);
+// tq-mq
+int32_t tqProcessAddCheckInfoReq(STQ* pTq, int64_t version, char* msg, int32_t msgLen);
+int32_t tqProcessDelCheckInfoReq(STQ* pTq, int64_t version, char* msg, int32_t msgLen);
+int32_t tqProcessVgChangeReq(STQ* pTq, int64_t version, char* msg, int32_t msgLen);
+int32_t tqProcessVgDeleteReq(STQ* pTq, int64_t version, char* msg, int32_t msgLen);
+int32_t tqProcessOffsetCommitReq(STQ* pTq, int64_t version, char* msg, int32_t msgLen);
 int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg);
-int32_t tqProcessTaskDeployReq(STQ* pTq, char* msg, int32_t msgLen);
-int32_t tqProcessTaskDropReq(STQ* pTq, char* msg, int32_t msgLen);
+// tq-stream
+int32_t tqProcessTaskDeployReq(STQ* pTq, int64_t version, char* msg, int32_t msgLen);
+int32_t tqProcessTaskDropReq(STQ* pTq, int64_t version, char* msg, int32_t msgLen);
 int32_t tqProcessStreamTrigger(STQ* pTq, SSubmitReq* data, int64_t ver);
 int32_t tqProcessTaskRunReq(STQ* pTq, SRpcMsg* pMsg);
 int32_t tqProcessTaskDispatchReq(STQ* pTq, SRpcMsg* pMsg, bool exec);
@@ -188,6 +199,7 @@ int32_t smaAsyncCommit(SSma* pSma);
 int32_t smaAsyncPostCommit(SSma* pSma);
 int32_t smaDoRetention(SSma* pSma, int64_t now);
 int32_t smaProcessFetch(SSma* pSma, void* pMsg);
+int32_t smaProcessExec(SSma* pSma, void* pMsg);
 
 int32_t tdProcessTSmaCreate(SSma* pSma, int64_t version, const char* msg);
 int32_t tdProcessTSmaInsert(SSma* pSma, int64_t indexUid, const char* msg);
@@ -357,6 +369,7 @@ struct SSma {
 void smaHandleRes(void* pVnode, int64_t smaId, const SArray* data);
 
 enum {
+  SNAP_DATA_CFG = 0,
   SNAP_DATA_META = 1,
   SNAP_DATA_TSDB = 2,
   SNAP_DATA_DEL = 3,

source/dnode/vnode/src/meta/metaCache.c

+/*
+ * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
+ *
+ * This program is free software: you can use, redistribute, and/or modify
+ * it under the terms of the GNU Affero General Public License, version 3
+ * or later ("AGPL"), as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+#include "meta.h"
+
+#define META_CACHE_BASE_BUCKET 1024
+
+// (uid , suid) : child table
+// (uid,     0) : normal table
+// (suid, suid) : super table
+typedef struct SMetaCacheEntry SMetaCacheEntry;
+struct SMetaCacheEntry {
+  SMetaCacheEntry* next;
+  SMetaInfo        info;
+};
+
+struct SMetaCache {
+  int32_t           nEntry;
+  int32_t           nBucket;
+  SMetaCacheEntry** aBucket;
+};
+
+int32_t metaCacheOpen(SMeta* pMeta) {
+  int32_t     code = 0;
+  SMetaCache* pCache = NULL;
+
+  pCache = (SMetaCache*)taosMemoryMalloc(sizeof(SMetaCache));
+  if (pCache == NULL) {
+    code = TSDB_CODE_OUT_OF_MEMORY;
+    goto _err;
+  }
+
+  pCache->nEntry = 0;
+  pCache->nBucket = META_CACHE_BASE_BUCKET;
+  pCache->aBucket = (SMetaCacheEntry**)taosMemoryCalloc(pCache->nBucket, sizeof(SMetaCacheEntry*));
+  if (pCache->aBucket == NULL) {
+    code = TSDB_CODE_OUT_OF_MEMORY;
+    taosMemoryFree(pCache);
+    goto _err;
+  }
+
+  pMeta->pCache = pCache;
+
+_exit:
+  return code;
+
+_err:
+  metaError("vgId:%d meta open cache failed since %s", TD_VID(pMeta->pVnode), tstrerror(code));
+  return code;
+}
+
+void metaCacheClose(SMeta* pMeta) {
+  if (pMeta->pCache) {
+    for (int32_t iBucket = 0; iBucket < pMeta->pCache->nBucket; iBucket++) {
+      SMetaCacheEntry* pEntry = pMeta->pCache->aBucket[iBucket];
+      while (pEntry) {
+        SMetaCacheEntry* tEntry = pEntry->next;
+        taosMemoryFree(pEntry);
+        pEntry = tEntry;
+      }
+    }
+    taosMemoryFree(pMeta->pCache->aBucket);
+    taosMemoryFree(pMeta->pCache);
+    pMeta->pCache = NULL;
+  }
+}
+
+static int32_t metaRehashCache(SMetaCache* pCache, int8_t expand) {
+  int32_t code = 0;
+  int32_t nBucket;
+
+  if (expand) {
+    nBucket = pCache->nBucket * 2;
+  } else {
+    nBucket = pCache->nBucket / 2;
+  }
+
+  SMetaCacheEntry** aBucket = (SMetaCacheEntry**)taosMemoryCalloc(nBucket, sizeof(SMetaCacheEntry*));
+  if (aBucket == NULL) {
+    code = TSDB_CODE_OUT_OF_MEMORY;
+    goto _exit;
+  }
+
+  // rehash
+  for (int32_t iBucket = 0; iBucket < pCache->nBucket; iBucket++) {
+    SMetaCacheEntry* pEntry = pCache->aBucket[iBucket];
+
+    while (pEntry) {
+      SMetaCacheEntry* pTEntry = pEntry->next;
+
+      pEntry->next = aBucket[TABS(pEntry->info.uid) % nBucket];
+      aBucket[TABS(pEntry->info.uid) % nBucket] = pEntry;
+
+      pEntry = pTEntry;
+    }
+  }
+
+  // final set
+  taosMemoryFree(pCache->aBucket);
+  pCache->nBucket = nBucket;
+  pCache->aBucket = aBucket;
+
+_exit:
+  return code;
+}
+
+int32_t metaCacheUpsert(SMeta* pMeta, SMetaInfo* pInfo) {
+  int32_t code = 0;
+
+  // ASSERT(metaIsWLocked(pMeta));
+
+  // search
+  SMetaCache*       pCache = pMeta->pCache;
+  int32_t           iBucket = TABS(pInfo->uid) % pCache->nBucket;
+  SMetaCacheEntry** ppEntry = &pCache->aBucket[iBucket];
+  while (*ppEntry && (*ppEntry)->info.uid != pInfo->uid) {
+    ppEntry = &(*ppEntry)->next;
+  }
+
+  if (*ppEntry) {  // update
+    ASSERT(pInfo->suid == (*ppEntry)->info.suid);
+    if (pInfo->version > (*ppEntry)->info.version) {
+      (*ppEntry)->info.version = pInfo->version;
+      (*ppEntry)->info.skmVer = pInfo->skmVer;
+    }
+  } else {  // insert
+    if (pCache->nEntry >= pCache->nBucket) {
+      code = metaRehashCache(pCache, 1);
+      if (code) goto _exit;
+
+      iBucket = TABS(pInfo->uid) % pCache->nBucket;
+    }
+
+    SMetaCacheEntry* pEntryNew = (SMetaCacheEntry*)taosMemoryMalloc(sizeof(*pEntryNew));
+    if (pEntryNew == NULL) {
+      code = TSDB_CODE_OUT_OF_MEMORY;
+      goto _exit;
+    }
+
+    pEntryNew->info = *pInfo;
+    pEntryNew->next = pCache->aBucket[iBucket];
+    pCache->aBucket[iBucket] = pEntryNew;
+    pCache->nEntry++;
+  }
+
+_exit:
+  return code;
+}
+
+int32_t metaCacheDrop(SMeta* pMeta, int64_t uid) {
+  int32_t code = 0;
+
+  SMetaCache*       pCache = pMeta->pCache;
+  int32_t           iBucket = TABS(uid) % pCache->nBucket;
+  SMetaCacheEntry** ppEntry = &pCache->aBucket[iBucket];
+  while (*ppEntry && (*ppEntry)->info.uid != uid) {
+    ppEntry = &(*ppEntry)->next;
+  }
+
+  SMetaCacheEntry* pEntry = *ppEntry;
+  if (pEntry) {
+    *ppEntry = pEntry->next;
+    taosMemoryFree(pEntry);
+    pCache->nEntry--;
+    if (pCache->nEntry < pCache->nBucket / 4 && pCache->nBucket > META_CACHE_BASE_BUCKET) {
+      code = metaRehashCache(pCache, 0);
+      if (code) goto _exit;
+    }
+  } else {
+    code = TSDB_CODE_NOT_FOUND;
+  }
+
+_exit:
+  return code;
+}
+
+int32_t metaCacheGet(SMeta* pMeta, int64_t uid, SMetaInfo* pInfo) {
+  int32_t code = 0;
+
+  SMetaCache*      pCache = pMeta->pCache;
+  int32_t          iBucket = TABS(uid) % pCache->nBucket;
+  SMetaCacheEntry* pEntry = pCache->aBucket[iBucket];
+
+  while (pEntry && pEntry->info.uid != uid) {
+    pEntry = pEntry->next;
+  }
+
+  if (pEntry) {
+    *pInfo = pEntry->info;
+  } else {
+    code = TSDB_CODE_NOT_FOUND;
+  }
+
+  return code;
+}

source/dnode/vnode/src/meta/metaOpen.c

@@ -73,7 +73,7 @@ int metaOpen(SVnode *pVnode, SMeta **ppMeta) {
   }
 
   // open pUidIdx
-  ret = tdbTbOpen("uid.idx", sizeof(tb_uid_t), sizeof(int64_t), uidIdxKeyCmpr, pMeta->pEnv, &pMeta->pUidIdx);
+  ret = tdbTbOpen("uid.idx", sizeof(tb_uid_t), sizeof(SUidIdxVal), uidIdxKeyCmpr, pMeta->pEnv, &pMeta->pUidIdx);
   if (ret < 0) {
     metaError("vgId:%d, failed to open meta uid idx since %s", TD_VID(pVnode), tstrerror(terrno));
     goto _err;
@@ -87,7 +87,7 @@ int metaOpen(SVnode *pVnode, SMeta **ppMeta) {
   }
 
   // open pCtbIdx
-  ret = tdbTbOpen("ctb.idx", sizeof(SCtbIdxKey), 0, ctbIdxKeyCmpr, pMeta->pEnv, &pMeta->pCtbIdx);
+  ret = tdbTbOpen("ctb.idx", sizeof(SCtbIdxKey), -1, ctbIdxKeyCmpr, pMeta->pEnv, &pMeta->pCtbIdx);
   if (ret < 0) {
     metaError("vgId:%d, failed to open meta child table index since %s", TD_VID(pVnode), tstrerror(terrno));
     goto _err;
@@ -143,6 +143,13 @@ int metaOpen(SVnode *pVnode, SMeta **ppMeta) {
     goto _err;
   }
 
+  int32_t code = metaCacheOpen(pMeta);
+  if (code) {
+    terrno = code;
+    metaError("vgId:%d, failed to open meta cache since %s", TD_VID(pVnode), tstrerror(terrno));
+    goto _err;
+  }
+
   metaDebug("vgId:%d, meta is opened", TD_VID(pVnode));
 
   *ppMeta = pMeta;
@@ -169,6 +176,7 @@ _err:
 
 int metaClose(SMeta *pMeta) {
   if (pMeta) {
+    if (pMeta->pCache) metaCacheClose(pMeta);
     if (pMeta->pIdx) metaCloseIdx(pMeta);
     if (pMeta->pStreamDb) tdbTbClose(pMeta->pStreamDb);
     if (pMeta->pSmaIdx) tdbTbClose(pMeta->pSmaIdx);

source/dnode/vnode/src/meta/metaQuery.c

@@ -53,6 +53,89 @@ _err:
   return -1;
 }
 
+// int metaGetTableEntryByUidTest(void* meta, SArray *uidList) {
+//
+//   SArray* readerList = taosArrayInit(taosArrayGetSize(uidList), sizeof(SMetaReader));
+//   SArray* uidVersion = taosArrayInit(taosArrayGetSize(uidList), sizeof(STbDbKey));
+//   SMeta  *pMeta = meta;
+//   int64_t version;
+//   SHashObj *uHash = taosHashInit(32, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), false, HASH_NO_LOCK);
+//
+//   int64_t stt1 = taosGetTimestampUs();
+//   for(int i = 0; i < taosArrayGetSize(uidList); i++) {
+//     void* ppVal = NULL;
+//     int vlen = 0;
+//     uint64_t *  uid = taosArrayGet(uidList, i);
+//     // query uid.idx
+//     if (tdbTbGet(pMeta->pUidIdx, uid, sizeof(*uid), &ppVal, &vlen) < 0) {
+//       continue;
+//     }
+//     version = *(int64_t *)ppVal;
+//
+//     STbDbKey tbDbKey = {.version = version, .uid = *uid};
+//     taosArrayPush(uidVersion, &tbDbKey);
+//     taosHashPut(uHash, uid, sizeof(int64_t), ppVal, sizeof(int64_t));
+//   }
+//   int64_t stt2 = taosGetTimestampUs();
+//   qDebug("metaGetTableEntryByUidTest1 rows:%d, cost:%ld us", taosArrayGetSize(uidList), stt2-stt1);
+//
+//   TBC        *pCur = NULL;
+//   tdbTbcOpen(pMeta->pTbDb, &pCur, NULL);
+//   tdbTbcMoveToFirst(pCur);
+//   void *pKey = NULL;
+//   int   kLen = 0;
+//
+//   while(1){
+//     SMetaReader pReader = {0};
+//     int32_t ret = tdbTbcNext(pCur, &pKey, &kLen, &pReader.pBuf, &pReader.szBuf);
+//     if (ret < 0) break;
+//     STbDbKey *tmp = (STbDbKey*)pKey;
+//     int64_t *ver = (int64_t*)taosHashGet(uHash, &tmp->uid, sizeof(int64_t));
+//     if(ver == NULL || *ver != tmp->version) continue;
+//     taosArrayPush(readerList, &pReader);
+//   }
+//   tdbTbcClose(pCur);
+//
+//   taosArrayClear(readerList);
+//   int64_t stt3 = taosGetTimestampUs();
+//   qDebug("metaGetTableEntryByUidTest2 rows:%d, cost:%ld us", taosArrayGetSize(uidList), stt3-stt2);
+//   for(int i = 0; i < taosArrayGetSize(uidVersion); i++) {
+//     SMetaReader pReader = {0};
+//
+//     STbDbKey *tbDbKey = taosArrayGet(uidVersion, i);
+//     // query table.db
+//     if (tdbTbGet(pMeta->pTbDb, tbDbKey, sizeof(STbDbKey), &pReader.pBuf, &pReader.szBuf) < 0) {
+//       continue;
+//     }
+//     taosArrayPush(readerList, &pReader);
+//   }
+//   int64_t stt4 = taosGetTimestampUs();
+//   qDebug("metaGetTableEntryByUidTest3 rows:%d, cost:%ld us", taosArrayGetSize(uidList), stt4-stt3);
+//
+//   for(int i = 0; i < taosArrayGetSize(readerList); i++){
+//     SMetaReader* pReader  = taosArrayGet(readerList, i);
+//     metaReaderInit(pReader, meta, 0);
+//     // decode the entry
+//     tDecoderInit(&pReader->coder, pReader->pBuf, pReader->szBuf);
+//
+//     if (metaDecodeEntry(&pReader->coder, &pReader->me) < 0) {
+//     }
+//     metaReaderClear(pReader);
+//   }
+//   int64_t stt5 = taosGetTimestampUs();
+//   qDebug("metaGetTableEntryByUidTest4 rows:%d, cost:%ld us", taosArrayGetSize(readerList), stt5-stt4);
+//   return 0;
+// }
+
+bool metaIsTableExist(SMeta *pMeta, tb_uid_t uid) {
+  // query uid.idx
+  if (tdbTbGet(pMeta->pUidIdx, &uid, sizeof(uid), NULL, NULL) < 0) {
+    return false;
+  }
+
+  return true;
+}
+
 int metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid) {
   SMeta  *pMeta = pReader->pMeta;
   int64_t version;
@@ -63,7 +146,7 @@ int metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid) {
     return -1;
   }
 
-  version = *(int64_t *)pReader->pBuf;
+  version = ((SUidIdxVal *)pReader->pBuf)[0].version;
   return metaGetTableEntryByVersion(pReader, version, uid);
 }
 
@@ -160,7 +243,7 @@ int metaTbCursorNext(SMTbCursor *pTbCur) {
 
     tDecoderClear(&pTbCur->mr.coder);
 
-    metaGetTableEntryByVersion(&pTbCur->mr, *(int64_t *)pTbCur->pVal, *(tb_uid_t *)pTbCur->pKey);
+    metaGetTableEntryByVersion(&pTbCur->mr, ((SUidIdxVal *)pTbCur->pVal)[0].version, *(tb_uid_t *)pTbCur->pKey);
     if (pTbCur->mr.me.type == TSDB_SUPER_TABLE) {
       continue;
     }
@@ -185,7 +268,7 @@ _query:
     goto _err;
   }
 
-  version = *(int64_t *)pData;
+  version = ((SUidIdxVal *)pData)[0].version;
 
   tdbTbGet(pMeta->pTbDb, &(STbDbKey){.uid = uid, .version = version}, sizeof(STbDbKey), &pData, &nData);
   SMetaEntry me = {0};
@@ -429,18 +512,65 @@ STSchema *metaGetTbTSchema(SMeta *pMeta, tb_uid_t uid, int32_t sver) {
 }
 
 int32_t metaGetTbTSchemaEx(SMeta *pMeta, tb_uid_t suid, tb_uid_t uid, int32_t sver, STSchema **ppTSchema) {
-  int32_t   code = 0;
-  STSchema *pTSchema = NULL;
-  SSkmDbKey skmDbKey = {.uid = suid ? suid : uid, .sver = sver};
+  int32_t code = 0;
+
   void     *pData = NULL;
   int       nData = 0;
+  SSkmDbKey skmDbKey;
+  if (sver <= 0) {
+    SMetaInfo info;
+    if (metaGetInfo(pMeta, suid ? suid : uid, &info) == 0) {
+      sver = info.skmVer;
+    } else {
+      TBC *pSkmDbC = NULL;
+      int  c;
+
+      skmDbKey.uid = suid ? suid : uid;
+      skmDbKey.sver = INT32_MAX;
+
+      tdbTbcOpen(pMeta->pSkmDb, &pSkmDbC, NULL);
+      metaRLock(pMeta);
+
+      if (tdbTbcMoveTo(pSkmDbC, &skmDbKey, sizeof(skmDbKey), &c) < 0) {
+        metaULock(pMeta);
+        tdbTbcClose(pSkmDbC);
+        code = TSDB_CODE_NOT_FOUND;
+        goto _exit;
+      }
+
+      ASSERT(c);
+
+      if (c < 0) {
+        tdbTbcMoveToPrev(pSkmDbC);
+      }
+
+      const void *pKey = NULL;
+      int32_t     nKey = 0;
+      tdbTbcGet(pSkmDbC, &pKey, &nKey, NULL, NULL);
+
+      if (((SSkmDbKey *)pKey)->uid != skmDbKey.uid) {
+        metaULock(pMeta);
+        tdbTbcClose(pSkmDbC);
+        code = TSDB_CODE_NOT_FOUND;
+        goto _exit;
+      }
 
-  // query
+      sver = ((SSkmDbKey *)pKey)->sver;
+
+      metaULock(pMeta);
+      tdbTbcClose(pSkmDbC);
+    }
+  }
+
+  ASSERT(sver > 0);
+
+  skmDbKey.uid = suid ? suid : uid;
+  skmDbKey.sver = sver;
   metaRLock(pMeta);
-  if (tdbTbGet(pMeta->pSkmDb, &skmDbKey, sizeof(skmDbKey), &pData, &nData) < 0) {
-    code = TSDB_CODE_NOT_FOUND;
+  if (tdbTbGet(pMeta->pSkmDb, &skmDbKey, sizeof(SSkmDbKey), &pData, &nData) < 0) {
     metaULock(pMeta);
-    goto _err;
+    code = TSDB_CODE_NOT_FOUND;
+    goto _exit;
   }
   metaULock(pMeta);
 
@@ -462,15 +592,13 @@ int32_t metaGetTbTSchemaEx(SMeta *pMeta, tb_uid_t suid, tb_uid_t uid, int32_t sv
     SSchema *pSchema = pSchemaWrapper->pSchema + i;
     tdAddColToSchema(&sb, pSchema->type, pSchema->flags, pSchema->colId, pSchema->bytes);
   }
-  pTSchema = tdGetSchemaFromBuilder(&sb);
+  STSchema *pTSchema = tdGetSchemaFromBuilder(&sb);
   tdDestroyTSchemaBuilder(&sb);
 
   *ppTSchema = pTSchema;
   taosMemoryFree(pSchemaWrapper->pSchema);
-  return code;
 
-_err:
-  *ppTSchema = NULL;
+_exit:
   return code;
 }
 
@@ -749,9 +877,8 @@ SArray *metaGetSmaTbUids(SMeta *pMeta) {
 
 #endif
 
-const void *metaGetTableTagVal(SMetaEntry *pEntry, int16_t type, STagVal *val) {
-  ASSERT(pEntry->type == TSDB_CHILD_TABLE);
-  STag *tag = (STag *)pEntry->ctbEntry.pTags;
+const void *metaGetTableTagVal(void *pTag, int16_t type, STagVal *val) {
+  STag *tag = (STag *)pTag;
   if (type == TSDB_DATA_TYPE_JSON) {
     return tag;
   }
@@ -853,6 +980,9 @@ int32_t metaFilterTableIds(SMeta *pMeta, SMetaFltParam *param, SArray *pUids) {
         break;
       }
     }
+    if (p->suid != pKey->suid) {
+      break;
+    }
     first = false;
     if (p != NULL) {
       int32_t cmp = (*param->filterFunc)(p->data, pKey->data, pKey->type);
@@ -888,3 +1018,75 @@ END:
 
   return ret;
 }
+
+int32_t metaGetTableTags(SMeta *pMeta, uint64_t suid, SArray *uidList, SHashObj *tags) {
+  SMCtbCursor *pCur = metaOpenCtbCursor(pMeta, suid);
+
+  SHashObj *uHash = NULL;
+  size_t    len = taosArrayGetSize(uidList);  // len > 0 means there already have uids
+  if (len > 0) {
+    uHash = taosHashInit(32, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), false, HASH_NO_LOCK);
+    for (int i = 0; i < len; i++) {
+      int64_t *uid = taosArrayGet(uidList, i);
+      taosHashPut(uHash, uid, sizeof(int64_t), &i, sizeof(i));
+    }
+  }
+  while (1) {
+    tb_uid_t id = metaCtbCursorNext(pCur);
+    if (id == 0) {
+      break;
+    }
+
+    if (len > 0 && taosHashGet(uHash, &id, sizeof(int64_t)) == NULL) {
+      continue;
+    } else if (len == 0) {
+      taosArrayPush(uidList, &id);
+    }
+
+    taosHashPut(tags, &id, sizeof(int64_t), pCur->pVal, pCur->vLen);
+  }
+
+  taosHashCleanup(uHash);
+  metaCloseCtbCursor(pCur);
+  return TSDB_CODE_SUCCESS;
+}
+
+int32_t metaCacheGet(SMeta *pMeta, int64_t uid, SMetaInfo *pInfo);
+
+int32_t metaGetInfo(SMeta *pMeta, int64_t uid, SMetaInfo *pInfo) {
+  int32_t code = 0;
+  void   *pData = NULL;
+  int     nData = 0;
+
+  metaRLock(pMeta);
+
+  // search cache
+  if (metaCacheGet(pMeta, uid, pInfo) == 0) {
+    metaULock(pMeta);
+    goto _exit;
+  }
+
+  // search TDB
+  if (tdbTbGet(pMeta->pUidIdx, &uid, sizeof(uid), &pData, &nData) < 0) {
+    // not found
+    metaULock(pMeta);
+    code = TSDB_CODE_NOT_FOUND;
+    goto _exit;
+  }
+
+  metaULock(pMeta);
+
+  pInfo->uid = uid;
+  pInfo->suid = ((SUidIdxVal *)pData)->suid;
+  pInfo->version = ((SUidIdxVal *)pData)->version;
+  pInfo->skmVer = ((SUidIdxVal *)pData)->skmVer;
+
+  // upsert the cache
+  metaWLock(pMeta);
+  metaCacheUpsert(pMeta, pInfo);
+  metaULock(pMeta);
+
+_exit:
+  tdbFree(pData);
+  return code;
+}

source/dnode/vnode/src/meta/metaSma.c

@@ -28,9 +28,9 @@ int32_t metaCreateTSma(SMeta *pMeta, int64_t version, SSmaCfg *pCfg) {
   int         vLen = 0;
   const void *pKey = NULL;
   const void *pVal = NULL;
-  void *      pBuf = NULL;
+  void       *pBuf = NULL;
   int32_t     szBuf = 0;
-  void *      p = NULL;
+  void       *p = NULL;
   SMetaReader mr = {0};
 
   // validate req
@@ -83,8 +83,8 @@ int32_t metaDropTSma(SMeta *pMeta, int64_t indexUid) {
 
 static int metaSaveSmaToDB(SMeta *pMeta, const SMetaEntry *pME) {
   STbDbKey tbDbKey;
-  void *   pKey = NULL;
-  void *   pVal = NULL;
+  void    *pKey = NULL;
+  void    *pVal = NULL;
   int      kLen = 0;
   int      vLen = 0;
   SEncoder coder = {0};
@@ -130,7 +130,8 @@ _err:
 }
 
 static int metaUpdateUidIdx(SMeta *pMeta, const SMetaEntry *pME) {
-  return tdbTbInsert(pMeta->pUidIdx, &pME->uid, sizeof(tb_uid_t), &pME->version, sizeof(int64_t), &pMeta->txn);
+  SUidIdxVal uidIdxVal = {.suid = pME->smaEntry.tsma->indexUid, .version = pME->version, .skmVer = 0};
+  return tdbTbInsert(pMeta->pUidIdx, &pME->uid, sizeof(tb_uid_t), &uidIdxVal, sizeof(uidIdxVal), &pMeta->txn);
 }
 
 static int metaUpdateNameIdx(SMeta *pMeta, const SMetaEntry *pME) {

source/dnode/vnode/src/meta/metaTable.c

@@ -27,6 +27,23 @@ static int metaUpdateSuidIdx(SMeta *pMeta, const SMetaEntry *pME);
 static int metaUpdateTagIdx(SMeta *pMeta, const SMetaEntry *pCtbEntry);
 static int metaDropTableByUid(SMeta *pMeta, tb_uid_t uid, int *type);
 
+static void metaGetEntryInfo(const SMetaEntry *pEntry, SMetaInfo *pInfo) {
+  pInfo->uid = pEntry->uid;
+  pInfo->version = pEntry->version;
+  if (pEntry->type == TSDB_SUPER_TABLE) {
+    pInfo->suid = pEntry->uid;
+    pInfo->skmVer = pEntry->stbEntry.schemaRow.version;
+  } else if (pEntry->type == TSDB_CHILD_TABLE) {
+    pInfo->suid = pEntry->ctbEntry.suid;
+    pInfo->skmVer = 0;
+  } else if (pEntry->type == TSDB_NORMAL_TABLE) {
+    pInfo->suid = 0;
+    pInfo->skmVer = pEntry->ntbEntry.schemaRow.version;
+  } else {
+    ASSERT(0);
+  }
+}
+
 static int metaUpdateMetaRsp(tb_uid_t uid, char *tbName, SSchemaWrapper *pSchema, STableMetaRsp *pMetaRsp) {
   pMetaRsp->pSchemas = taosMemoryMalloc(pSchema->nCols * sizeof(SSchema));
   if (NULL == pMetaRsp->pSchemas) {
@@ -171,22 +188,22 @@ int metaCreateSTable(SMeta *pMeta, int64_t version, SVCreateStbReq *pReq) {
   void       *pBuf = NULL;
   int32_t     szBuf = 0;
   void       *p = NULL;
-  SMetaReader mr = {0};
 
   // validate req
-  metaReaderInit(&mr, pMeta, 0);
-  if (metaGetTableEntryByName(&mr, pReq->name) == 0) {
-// TODO: just for pass case
-#if 0
-    terrno = TSDB_CODE_TDB_STB_ALREADY_EXIST;
-    metaReaderClear(&mr);
-    return -1;
-#else
-    metaReaderClear(&mr);
-    return 0;
-#endif
+  void *pData = NULL;
+  int   nData = 0;
+  if (tdbTbGet(pMeta->pNameIdx, pReq->name, strlen(pReq->name) + 1, &pData, &nData) == 0) {
+    tb_uid_t uid = *(tb_uid_t *)pData;
+    tdbFree(pData);
+    SMetaInfo info;
+    metaGetInfo(pMeta, uid, &info);
+    if (info.uid == info.suid) {
+      return 0;
+    } else {
+      terrno = TSDB_CODE_TDB_TABLE_ALREADY_EXIST;
+      return -1;
+    }
   }
-  metaReaderClear(&mr);
 
   // set structs
   me.version = version;
@@ -265,8 +282,8 @@ int metaDropSTable(SMeta *pMeta, int64_t verison, SVDropStbReq *pReq, SArray *tb
   // drop super table
 _drop_super_table:
   tdbTbGet(pMeta->pUidIdx, &pReq->suid, sizeof(tb_uid_t), &pData, &nData);
-  tdbTbDelete(pMeta->pTbDb, &(STbDbKey){.version = *(int64_t *)pData, .uid = pReq->suid}, sizeof(STbDbKey),
-              &pMeta->txn);
+  tdbTbDelete(pMeta->pTbDb, &(STbDbKey){.version = ((SUidIdxVal *)pData)[0].version, .uid = pReq->suid},
+              sizeof(STbDbKey), &pMeta->txn);
   tdbTbDelete(pMeta->pNameIdx, pReq->name, strlen(pReq->name) + 1, &pMeta->txn);
   tdbTbDelete(pMeta->pUidIdx, &pReq->suid, sizeof(tb_uid_t), &pMeta->txn);
   tdbTbDelete(pMeta->pSuidIdx, &pReq->suid, sizeof(tb_uid_t), &pMeta->txn);
@@ -309,7 +326,7 @@ int metaAlterSTable(SMeta *pMeta, int64_t version, SVCreateStbReq *pReq) {
     return -1;
   }
 
-  oversion = *(int64_t *)pData;
+  oversion = ((SUidIdxVal *)pData)[0].version;
 
   tdbTbcOpen(pMeta->pTbDb, &pTbDbc, &pMeta->txn);
   ret = tdbTbcMoveTo(pTbDbc, &((STbDbKey){.uid = pReq->suid, .version = oversion}), sizeof(STbDbKey), &c);
@@ -336,15 +353,11 @@ int metaAlterSTable(SMeta *pMeta, int64_t version, SVCreateStbReq *pReq) {
     metaSaveToSkmDb(pMeta, &nStbEntry);
   }
 
-  // if (oStbEntry.stbEntry.schemaTag.sver != pReq->schemaTag.sver) {
-  //   // change tag schema
-  // }
-
   // update table.db
   metaSaveToTbDb(pMeta, &nStbEntry);
 
   // update uid index
-  tdbTbcUpsert(pUidIdxc, &pReq->suid, sizeof(tb_uid_t), &version, sizeof(version), 0);
+  metaUpdateUidIdx(pMeta, &nStbEntry);
 
   if (oStbEntry.pBuf) taosMemoryFree(oStbEntry.pBuf);
   metaULock(pMeta);
@@ -503,7 +516,7 @@ static int metaDropTableByUid(SMeta *pMeta, tb_uid_t uid, int *type) {
   SDecoder   dc = {0};
 
   rc = tdbTbGet(pMeta->pUidIdx, &uid, sizeof(uid), &pData, &nData);
-  int64_t version = *(int64_t *)pData;
+  int64_t version = ((SUidIdxVal *)pData)[0].version;
 
   tdbTbGet(pMeta->pTbDb, &(STbDbKey){.version = version, .uid = uid}, sizeof(STbDbKey), &pData, &nData);
 
@@ -517,7 +530,7 @@ static int metaDropTableByUid(SMeta *pMeta, tb_uid_t uid, int *type) {
     int   tLen = 0;
 
     if (tdbTbGet(pMeta->pUidIdx, &e.ctbEntry.suid, sizeof(tb_uid_t), &tData, &tLen) == 0) {
-      version = *(int64_t *)tData;
+      version = ((SUidIdxVal *)tData)[0].version;
       STbDbKey tbDbKey = {.uid = e.ctbEntry.suid, .version = version};
       if (tdbTbGet(pMeta->pTbDb, &tbDbKey, sizeof(tbDbKey), &tData, &tLen) == 0) {
         SDecoder   tdc = {0};
@@ -556,6 +569,8 @@ static int metaDropTableByUid(SMeta *pMeta, tb_uid_t uid, int *type) {
     --pMeta->pVnode->config.vndStats.numOfSTables;
   }
 
+  metaCacheDrop(pMeta, uid);
+
   tDecoderClear(&dc);
   tdbFree(pData);
 
@@ -594,7 +609,7 @@ static int metaAlterTableColumn(SMeta *pMeta, int64_t version, SVAlterTbReq *pAl
   ASSERT(c == 0);
 
   tdbTbcGet(pUidIdxc, NULL, NULL, &pData, &nData);
-  oversion = *(int64_t *)pData;
+  oversion = ((SUidIdxVal *)pData)[0].version;
 
   // search table.db
   TBC *pTbDbc = NULL;
@@ -708,7 +723,7 @@ static int metaAlterTableColumn(SMeta *pMeta, int64_t version, SVAlterTbReq *pAl
   // save to table db
   metaSaveToTbDb(pMeta, &entry);
 
-  tdbTbcUpsert(pUidIdxc, &entry.uid, sizeof(tb_uid_t), &version, sizeof(version), 0);
+  metaUpdateUidIdx(pMeta, &entry);
 
   metaSaveToSkmDb(pMeta, &entry);
 
@@ -764,7 +779,7 @@ static int metaUpdateTableTagVal(SMeta *pMeta, int64_t version, SVAlterTbReq *pA
   ASSERT(c == 0);
 
   tdbTbcGet(pUidIdxc, NULL, NULL, &pData, &nData);
-  oversion = *(int64_t *)pData;
+  oversion = ((SUidIdxVal *)pData)[0].version;
 
   // search table.db
   TBC     *pTbDbc = NULL;
@@ -784,8 +799,8 @@ static int metaUpdateTableTagVal(SMeta *pMeta, int64_t version, SVAlterTbReq *pA
 
   /* get stbEntry*/
   tdbTbGet(pMeta->pUidIdx, &ctbEntry.ctbEntry.suid, sizeof(tb_uid_t), &pVal, &nVal);
-  tdbTbGet(pMeta->pTbDb, &((STbDbKey){.uid = ctbEntry.ctbEntry.suid, .version = *(int64_t *)pVal}), sizeof(STbDbKey),
-           (void **)&stbEntry.pBuf, &nVal);
+  tdbTbGet(pMeta->pTbDb, &((STbDbKey){.uid = ctbEntry.ctbEntry.suid, .version = ((SUidIdxVal *)pVal)[0].version}),
+           sizeof(STbDbKey), (void **)&stbEntry.pBuf, &nVal);
   tdbFree(pVal);
   tDecoderInit(&dc2, stbEntry.pBuf, nVal);
   metaDecodeEntry(&dc2, &stbEntry);
@@ -859,12 +874,16 @@ static int metaUpdateTableTagVal(SMeta *pMeta, int64_t version, SVAlterTbReq *pA
   metaSaveToTbDb(pMeta, &ctbEntry);
 
   // save to uid.idx
-  tdbTbUpsert(pMeta->pUidIdx, &ctbEntry.uid, sizeof(tb_uid_t), &version, sizeof(version), &pMeta->txn);
+  metaUpdateUidIdx(pMeta, &ctbEntry);
 
   if (iCol == 0) {
     metaUpdateTagIdx(pMeta, &ctbEntry);
   }
 
+  SCtbIdxKey ctbIdxKey = {.suid = ctbEntry.ctbEntry.suid, .uid = uid};
+  tdbTbUpsert(pMeta->pCtbIdx, &ctbIdxKey, sizeof(ctbIdxKey), ctbEntry.ctbEntry.pTags,
+              ((STag *)(ctbEntry.ctbEntry.pTags))->len, &pMeta->txn);
+
   tDecoderClear(&dc1);
   tDecoderClear(&dc2);
   if (ctbEntry.ctbEntry.pTags) taosMemoryFree((void *)ctbEntry.ctbEntry.pTags);
@@ -914,7 +933,7 @@ static int metaUpdateTableOptions(SMeta *pMeta, int64_t version, SVAlterTbReq *p
   ASSERT(c == 0);
 
   tdbTbcGet(pUidIdxc, NULL, NULL, &pData, &nData);
-  oversion = *(int64_t *)pData;
+  oversion = ((SUidIdxVal *)pData)[0].version;
 
   // search table.db
   TBC *pTbDbc = NULL;
@@ -959,7 +978,7 @@ static int metaUpdateTableOptions(SMeta *pMeta, int64_t version, SVAlterTbReq *p
 
   // save to table db
   metaSaveToTbDb(pMeta, &entry);
-  tdbTbcUpsert(pUidIdxc, &entry.uid, sizeof(tb_uid_t), &version, sizeof(version), 0);
+  metaUpdateUidIdx(pMeta, &entry);
   metaULock(pMeta);
 
   tdbTbcClose(pTbDbc);
@@ -1042,7 +1061,14 @@ _err:
 }
 
 static int metaUpdateUidIdx(SMeta *pMeta, const SMetaEntry *pME) {
-  return tdbTbInsert(pMeta->pUidIdx, &pME->uid, sizeof(tb_uid_t), &pME->version, sizeof(int64_t), &pMeta->txn);
+  // upsert cache
+  SMetaInfo info;
+  metaGetEntryInfo(pME, &info);
+  metaCacheUpsert(pMeta, &info);
+
+  SUidIdxVal uidIdxVal = {.suid = info.suid, .version = info.version, .skmVer = info.skmVer};
+
+  return tdbTbUpsert(pMeta->pUidIdx, &pME->uid, sizeof(tb_uid_t), &uidIdxVal, sizeof(uidIdxVal), &pMeta->txn);
 }
 
 static int metaUpdateSuidIdx(SMeta *pMeta, const SMetaEntry *pME) {
@@ -1062,7 +1088,9 @@ static int metaUpdateTtlIdx(SMeta *pMeta, const SMetaEntry *pME) {
 
 static int metaUpdateCtbIdx(SMeta *pMeta, const SMetaEntry *pME) {
   SCtbIdxKey ctbIdxKey = {.suid = pME->ctbEntry.suid, .uid = pME->uid};
-  return tdbTbInsert(pMeta->pCtbIdx, &ctbIdxKey, sizeof(ctbIdxKey), NULL, 0, &pMeta->txn);
+
+  return tdbTbInsert(pMeta->pCtbIdx, &ctbIdxKey, sizeof(ctbIdxKey), pME->ctbEntry.pTags,
+                     ((STag *)(pME->ctbEntry.pTags))->len, &pMeta->txn);
 }
 
 int metaCreateTagIdxKey(tb_uid_t suid, int32_t cid, const void *pTagData, int32_t nTagData, int8_t type, tb_uid_t uid,
@@ -1118,7 +1146,7 @@ static int metaUpdateTagIdx(SMeta *pMeta, const SMetaEntry *pCtbEntry) {
     return -1;
   }
   tbDbKey.uid = pCtbEntry->ctbEntry.suid;
-  tbDbKey.version = *(int64_t *)pData;
+  tbDbKey.version = ((SUidIdxVal *)pData)[0].version;
   tdbTbGet(pMeta->pTbDb, &tbDbKey, sizeof(tbDbKey), &pData, &nData);
 
   tDecoderInit(&dc, pData, nData);

source/dnode/vnode/src/sma/smaCommit.c

@@ -83,8 +83,7 @@ int32_t smaBegin(SSma *pSma) {
     return TSDB_CODE_SUCCESS;
   }
 
-  SSmaStat  *pStat = SMA_ENV_STAT(pSmaEnv);
-  SRSmaStat *pRSmaStat = SMA_RSMA_STAT(pStat);
+  SRSmaStat *pRSmaStat = (SRSmaStat *)SMA_ENV_STAT(pSmaEnv);
 
   int8_t rsmaTriggerStat =
       atomic_val_compare_exchange_8(RSMA_TRIGGER_STAT(pRSmaStat), TASK_TRIGGER_STAT_PAUSED, TASK_TRIGGER_STAT_ACTIVE);
@@ -123,7 +122,7 @@ static int32_t tdProcessRSmaSyncPreCommitImpl(SSma *pSma) {
   }
 
   SSmaStat  *pStat = SMA_ENV_STAT(pSmaEnv);
-  SRSmaStat *pRSmaStat = SMA_RSMA_STAT(pStat);
+  SRSmaStat *pRSmaStat = SMA_STAT_RSMA(pStat);
 
   // step 1: set rsma stat paused
   atomic_store_8(RSMA_TRIGGER_STAT(pRSmaStat), TASK_TRIGGER_STAT_PAUSED);
@@ -289,8 +288,7 @@ static int32_t tdProcessRSmaSyncPostCommitImpl(SSma *pSma) {
     return TSDB_CODE_SUCCESS;
   }
 
-  SSmaEnv   *pSmaEnv = SMA_RSMA_ENV(pSma);
-  SRSmaStat *pRSmaStat = SMA_RSMA_STAT(SMA_ENV_STAT(pSmaEnv));
+  SRSmaStat *pRSmaStat = SMA_RSMA_STAT(pSma);
 
   // cleanup outdated qtaskinfo files
   tdCleanupQTaskInfoFiles(pSma, pRSmaStat);
@@ -299,10 +297,9 @@ static int32_t tdProcessRSmaSyncPostCommitImpl(SSma *pSma) {
 }
 
 /**
- * @brief Rsma async commit implementation
+ * @brief Rsma async commit implementation(only do some necessary light weighted task)
  *  1) set rsma stat TASK_TRIGGER_STAT_PAUSED
  *  2) Wait all running fetch task finish to fetch and put submitMsg into level 2/3 wQueue(blocking level 1 write)
- *  3)
  *
  * @param pSma
  * @return int32_t
@@ -314,7 +311,7 @@ static int32_t tdProcessRSmaAsyncPreCommitImpl(SSma *pSma) {
   }
 
   SSmaStat  *pStat = SMA_ENV_STAT(pEnv);
-  SRSmaStat *pRSmaStat = SMA_RSMA_STAT(pStat);
+  SRSmaStat *pRSmaStat = SMA_STAT_RSMA(pStat);
 
   // step 1: set rsma stat
   atomic_store_8(RSMA_TRIGGER_STAT(pRSmaStat), TASK_TRIGGER_STAT_PAUSED);
@@ -336,28 +333,55 @@ static int32_t tdProcessRSmaAsyncPreCommitImpl(SSma *pSma) {
     }
   }
 
-  // step 3:  swap rsmaInfoHash and iRsmaInfoHash
+  /**
+   * @brief step 3: consume the SubmitReq in buffer
+   *  1) This is high cost task and should not put in asyncPreCommit originally.
+   *  2) But, if put in asyncCommit, would trigger taskInfo cloning frequently.
+   */
+  nLoops = 0;
+  smaInfo("vgId:%d, start to wait for rsma qtask free, TID:%p", SMA_VID(pSma), (void *)taosGetSelfPthreadId());
+
+  int8_t old;
+  while (1) {
+    old = atomic_val_compare_exchange_8(&pRSmaStat->execStat, 0, 1);
+    if (old == 0) break;
+    if (++nLoops > 1000) {
+      sched_yield();
+      nLoops = 0;
+      smaDebug("vgId:%d, wait for rsma qtask free, TID:%p", SMA_VID(pSma), (void *)taosGetSelfPthreadId());
+    }
+  }
+
+  smaInfo("vgId:%d, end to wait for rsma qtask free, TID:%p", SMA_VID(pSma), (void *)taosGetSelfPthreadId());
+
+  if (tdRSmaProcessExecImpl(pSma, RSMA_EXEC_COMMIT) < 0) {
+    atomic_store_8(&pRSmaStat->execStat, 0);
+    return TSDB_CODE_FAILED;
+  }
+
+  // step 4:  swap queue/qall and iQueue/iQall
   // lock
   taosWLockLatch(SMA_ENV_LOCK(pEnv));
 
   ASSERT(RSMA_INFO_HASH(pRSmaStat));
-  ASSERT(!RSMA_IMU_INFO_HASH(pRSmaStat));
 
-  RSMA_IMU_INFO_HASH(pRSmaStat) = RSMA_INFO_HASH(pRSmaStat);
-  RSMA_INFO_HASH(pRSmaStat) =
-      taosHashInit(RSMA_TASK_INFO_HASH_SLOT, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), true, HASH_ENTRY_LOCK);
+  void *pIter = taosHashIterate(RSMA_INFO_HASH(pRSmaStat), NULL);
 
-  if (!RSMA_INFO_HASH(pRSmaStat)) {
-    // unlock
-    taosWUnLockLatch(SMA_ENV_LOCK(pEnv));
-    smaError("vgId:%d, rsma async commit failed since %s", SMA_VID(pSma), terrstr());
-    return TSDB_CODE_FAILED;
+  while (pIter) {
+    SRSmaInfo *pInfo = *(SRSmaInfo **)pIter;
+    TSWAP(pInfo->iQall, pInfo->qall);
+    TSWAP(pInfo->iQueue, pInfo->queue);
+    TSWAP(pInfo->iTaskInfo[0], pInfo->taskInfo[0]);
+    TSWAP(pInfo->iTaskInfo[1], pInfo->taskInfo[1]);
+    pIter = taosHashIterate(RSMA_INFO_HASH(pRSmaStat), pIter);
   }
 
+  atomic_store_64(&pRSmaStat->qBufSize, 0);
+  atomic_store_8(&pRSmaStat->execStat, 0);
   // unlock
   taosWUnLockLatch(SMA_ENV_LOCK(pEnv));
 
-  // step 4: others
+  // step 5: others
   pRSmaStat->commitAppliedVer = pSma->pVnode->state.applied;
 
   return TSDB_CODE_SUCCESS;
@@ -375,17 +399,18 @@ static int32_t tdProcessRSmaAsyncCommitImpl(SSma *pSma) {
     return TSDB_CODE_SUCCESS;
   }
 
-  SSmaStat  *pStat = SMA_ENV_STAT(pSmaEnv);
-  SRSmaStat *pRSmaStat = SMA_RSMA_STAT(pStat);
+  SRSmaStat *pRSmaStat = (SRSmaStat *)SMA_ENV_STAT(pSmaEnv);
 
-  // perform persist task for qTaskInfo
-  tdRSmaPersistExecImpl(pRSmaStat, RSMA_IMU_INFO_HASH(pRSmaStat));
+  // perform persist task for qTaskInfo operator
+  if (tdRSmaPersistExecImpl(pRSmaStat, RSMA_INFO_HASH(pRSmaStat)) < 0) {
+    return TSDB_CODE_FAILED;
+  }
 
   return TSDB_CODE_SUCCESS;
 }
 
 /**
- * @brief Migrate rsmaInfo from iRsmaInfo to rsmaInfo if rsmaInfoHash not empty.
+ * @brief Migrate rsmaInfo from iRsmaInfo to rsmaInfo if rsma infoHash not empty.
  *
  * @param pSma
  * @return int32_t
@@ -396,65 +421,66 @@ static int32_t tdProcessRSmaAsyncPostCommitImpl(SSma *pSma) {
     return TSDB_CODE_SUCCESS;
   }
 
-  SSmaStat  *pStat = SMA_ENV_STAT(pEnv);
-  SRSmaStat *pRSmaStat = SMA_RSMA_STAT(pStat);
+  SRSmaStat *pRSmaStat = (SRSmaStat *)SMA_ENV_STAT(pEnv);
+  SArray    *rsmaDeleted = NULL;
 
-  // step 1: merge rsmaInfoHash and iRsmaInfoHash
+  // step 1: merge qTaskInfo and iQTaskInfo
   // lock
   taosWLockLatch(SMA_ENV_LOCK(pEnv));
-#if 0
-  if (taosHashGetSize(RSMA_INFO_HASH(pRSmaStat)) <= 0) {
-    // just switch the hash pointer if rsmaInfoHash is empty
-    if (taosHashGetSize(RSMA_IMU_INFO_HASH(pRSmaStat)) > 0) {
-      SHashObj *infoHash = RSMA_INFO_HASH(pRSmaStat);
-      RSMA_INFO_HASH(pRSmaStat) = RSMA_IMU_INFO_HASH(pRSmaStat);
-      RSMA_IMU_INFO_HASH(pRSmaStat) = infoHash;
-    }
-  } else {
-#endif
-#if 1
-  void *pIter = taosHashIterate(RSMA_IMU_INFO_HASH(pRSmaStat), NULL);
+
+  void *pIter = taosHashIterate(RSMA_INFO_HASH(pRSmaStat), NULL);
   while (pIter) {
-    tb_uid_t *pSuid = (tb_uid_t *)taosHashGetKey(pIter, NULL);
-
-    if (!taosHashGet(RSMA_INFO_HASH(pRSmaStat), pSuid, sizeof(tb_uid_t))) {
-      SRSmaInfo *pRSmaInfo = *(SRSmaInfo **)pIter;
-      if (RSMA_INFO_IS_DEL(pRSmaInfo)) {
-        int32_t refVal = T_REF_VAL_GET(pRSmaInfo);
-        if (refVal == 0) {
-          tdFreeRSmaInfo(pSma, pRSmaInfo, true);
-          smaDebug(
-              "vgId:%d, rsma async post commit, free rsma info since already deleted and ref is 0 for "
-              "table:%" PRIi64,
-              SMA_VID(pSma), *pSuid);
-        } else {
-          smaDebug(
-              "vgId:%d, rsma async post commit, not free rsma info since ref is %d although already deleted for "
-              "table:%" PRIi64,
-              SMA_VID(pSma), refVal, *pSuid);
+    tb_uid_t  *pSuid = (tb_uid_t *)taosHashGetKey(pIter, NULL);
+    SRSmaInfo *pRSmaInfo = *(SRSmaInfo **)pIter;
+    if (RSMA_INFO_IS_DEL(pRSmaInfo)) {
+      int32_t refVal = T_REF_VAL_GET(pRSmaInfo);
+      if (refVal == 0) {
+        if (!rsmaDeleted) {
+          if ((rsmaDeleted = taosArrayInit(1, sizeof(tb_uid_t)))) {
+            taosArrayPush(rsmaDeleted, pSuid);
+          }
         }
+      } else {
+        smaDebug(
+            "vgId:%d, rsma async post commit, not free rsma info since ref is %d although already deleted for "
+            "table:%" PRIi64,
+            SMA_VID(pSma), refVal, *pSuid);
+      }
 
-        pIter = taosHashIterate(RSMA_IMU_INFO_HASH(pRSmaStat), pIter);
-        continue;
+      pIter = taosHashIterate(RSMA_INFO_HASH(pRSmaStat), pIter);
+      continue;
+    }
+
+    if (pRSmaInfo->taskInfo[0]) {
+      if (pRSmaInfo->iTaskInfo[0]) {
+        SRSmaInfo *pRSmaInfo = *(SRSmaInfo **)pRSmaInfo->iTaskInfo[0];
+        tdFreeRSmaInfo(pSma, pRSmaInfo, true);
+        pRSmaInfo->iTaskInfo[0] = NULL;
       }
-      taosHashPut(RSMA_INFO_HASH(pRSmaStat), pSuid, sizeof(tb_uid_t), pIter, sizeof(pIter));
-      smaDebug("vgId:%d, rsma async post commit, migrated from iRsmaInfoHash for table:%" PRIi64, SMA_VID(pSma),
-               *pSuid);
     } else {
-      // free the resources
-      SRSmaInfo *pRSmaInfo = *(SRSmaInfo **)pIter;
-      tdFreeRSmaInfo(pSma, pRSmaInfo, false);
-      smaDebug("vgId:%d, rsma async post commit, free rsma info since already COW for table:%" PRIi64, SMA_VID(pSma),
-               *pSuid);
+      TSWAP(pRSmaInfo->taskInfo[0], pRSmaInfo->iTaskInfo[0]);
     }
 
-    pIter = taosHashIterate(RSMA_IMU_INFO_HASH(pRSmaStat), pIter);
+    taosHashPut(RSMA_INFO_HASH(pRSmaStat), pSuid, sizeof(tb_uid_t), pIter, sizeof(pIter));
+    smaDebug("vgId:%d, rsma async post commit, migrated from iRsmaInfoHash for table:%" PRIi64, SMA_VID(pSma), *pSuid);
+
+    pIter = taosHashIterate(RSMA_INFO_HASH(pRSmaStat), pIter);
   }
-#endif
-  // }
 
-  taosHashCleanup(RSMA_IMU_INFO_HASH(pRSmaStat));
-  RSMA_IMU_INFO_HASH(pRSmaStat) = NULL;
+  for (int32_t i = 0; i < taosArrayGetSize(rsmaDeleted); ++i) {
+    tb_uid_t *pSuid = taosArrayGet(rsmaDeleted, i);
+    void     *pRSmaInfo = taosHashGet(RSMA_INFO_HASH(pRSmaStat), pSuid, sizeof(tb_uid_t));
+    if ((pRSmaInfo = *(SRSmaInfo **)pRSmaInfo)) {
+      tdFreeRSmaInfo(pSma, pRSmaInfo, true);
+      smaDebug(
+          "vgId:%d, rsma async post commit, free rsma info since already deleted and ref is 0 for "
+          "table:%" PRIi64,
+          SMA_VID(pSma), *pSuid);
+    }
+    taosHashRemove(RSMA_INFO_HASH(pRSmaStat), pSuid, sizeof(tb_uid_t));
+  }
+  taosArrayDestroy(rsmaDeleted);
+  // TODO: remove suid in files?
 
   // unlock
   taosWUnLockLatch(SMA_ENV_LOCK(pEnv));

source/dnode/vnode/src/sma/smaEnv.c

@@ -171,7 +171,7 @@ int32_t tdUnRefSmaStat(SSma *pSma, SSmaStat *pStat) {
 
 int32_t tdRefRSmaInfo(SSma *pSma, SRSmaInfo *pRSmaInfo) {
   if (!pRSmaInfo) return 0;
-  
+
   int ref = T_REF_INC(pRSmaInfo);
   smaDebug("vgId:%d, ref rsma info:%p, val:%d", SMA_VID(pSma), pRSmaInfo, ref);
   return 0;
@@ -228,7 +228,12 @@ static int32_t tdInitSmaStat(SSmaStat **pSmaStat, int8_t smaType, const SSma *pS
       RSMA_INFO_HASH(pRSmaStat) = taosHashInit(
           RSMA_TASK_INFO_HASH_SLOT, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), true, HASH_ENTRY_LOCK);
       if (!RSMA_INFO_HASH(pRSmaStat)) {
-        taosMemoryFreeClear(*pSmaStat);
+        return TSDB_CODE_FAILED;
+      }
+
+      RSMA_FETCH_HASH(pRSmaStat) = taosHashInit(
+          RSMA_TASK_INFO_HASH_SLOT, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), true, HASH_ENTRY_LOCK);
+      if (!RSMA_FETCH_HASH(pRSmaStat)) {
         return TSDB_CODE_FAILED;
       }
     } else if (smaType == TSDB_SMA_TYPE_TIME_RANGE) {
@@ -264,8 +269,6 @@ static void tdDestroyRSmaStat(void *pRSmaStat) {
     atomic_store_8(RSMA_TRIGGER_STAT(pStat), TASK_TRIGGER_STAT_CANCELLED);
 
     // step 2: destroy the rsma info and associated fetch tasks
-    // TODO: use taosHashSetFreeFp when taosHashSetFreeFp is ready.
-#if 1
     if (taosHashGetSize(RSMA_INFO_HASH(pStat)) > 0) {
       void *infoHash = taosHashIterate(RSMA_INFO_HASH(pStat), NULL);
       while (infoHash) {
@@ -274,10 +277,12 @@ static void tdDestroyRSmaStat(void *pRSmaStat) {
         infoHash = taosHashIterate(RSMA_INFO_HASH(pStat), infoHash);
       }
     }
-#endif
     taosHashCleanup(RSMA_INFO_HASH(pStat));
 
-    // step 3: wait all triggered fetch tasks finished
+    // step 3: destroy the rsma fetch hash
+    taosHashCleanup(RSMA_FETCH_HASH(pStat));
+
+    // step 4: wait all triggered fetch tasks finished
     int32_t nLoops = 0;
     while (1) {
       if (T_REF_VAL_GET((SSmaStat *)pStat) == 0) {
@@ -293,7 +298,7 @@ static void tdDestroyRSmaStat(void *pRSmaStat) {
       }
     }
 
-    // step 4: free pStat
+    // step 5: free pStat
     taosMemoryFreeClear(pStat);
   }
 }
@@ -318,9 +323,9 @@ void *tdFreeSmaState(SSmaStat *pSmaStat, int8_t smaType) {
 int32_t tdDestroySmaState(SSmaStat *pSmaStat, int8_t smaType) {
   if (pSmaStat) {
     if (smaType == TSDB_SMA_TYPE_TIME_RANGE) {
-      tdDestroyTSmaStat(SMA_TSMA_STAT(pSmaStat));
+      tdDestroyTSmaStat(SMA_STAT_TSMA(pSmaStat));
     } else if (smaType == TSDB_SMA_TYPE_ROLLUP) {
-      SRSmaStat *pRSmaStat = SMA_RSMA_STAT(pSmaStat);
+      SRSmaStat *pRSmaStat = &pSmaStat->rsmaStat;
       int32_t    vid = SMA_VID(pRSmaStat->pSma);
       int64_t    refId = RSMA_REF_ID(pRSmaStat);
       if (taosRemoveRef(smaMgmt.rsetId, RSMA_REF_ID(pRSmaStat)) < 0) {

source/dnode/vnode/src/sma/smaSnapshot.c

@@ -139,7 +139,6 @@ static int32_t rsmaSnapReadQTaskInfo(SRsmaSnapReader* pReader, uint8_t** ppBuf)
 
   smaInfo("vgId:%d, vnode snapshot rsma read qtaskinfo, size:%" PRIi64, SMA_VID(pSma), size);
 
-
   SSnapDataHdr* pHdr = (SSnapDataHdr*)(*ppBuf);
   pHdr->type = SNAP_DATA_QTASK;
   pHdr->size = size;
@@ -279,7 +278,8 @@ int32_t rsmaSnapWriterOpen(SSma* pSma, int64_t sver, int64_t ever, SRsmaSnapWrit
   TdFilePtr qTaskF = taosCreateFile(qTaskInfoFullName, TD_FILE_CREATE | TD_FILE_WRITE | TD_FILE_TRUNC);
   if (!qTaskF) {
     code = TAOS_SYSTEM_ERROR(errno);
-    smaError("vgId:%d, rsma snapshot writer open %s failed since %s", TD_VID(pSma->pVnode), qTaskInfoFullName, tstrerror(code));
+    smaError("vgId:%d, rsma snapshot writer open %s failed since %s", TD_VID(pSma->pVnode), qTaskInfoFullName,
+             tstrerror(code));
     goto _err;
   }
   qWriter->pWriteH = qTaskF;
@@ -309,7 +309,7 @@ int32_t rsmaSnapWriterClose(SRsmaSnapWriter** ppWriter, int8_t rollback) {
   if (rollback) {
     // TODO: rsma1/rsma2
     // qtaskinfo
-    if(pWriter->pQTaskFWriter) {
+    if (pWriter->pQTaskFWriter) {
       taosRemoveFile(pWriter->pQTaskFWriter->fname);
     }
   } else {

source/dnode/vnode/src/sma/smaTimeRange.c

@@ -175,7 +175,7 @@ int32_t tdProcessTSmaInsertImpl(SSma *pSma, int64_t indexUid, const char *msg) {
   }
 
   tdRefSmaStat(pSma, pStat);
-  pTsmaStat = SMA_TSMA_STAT(pStat);
+  pTsmaStat = SMA_STAT_TSMA(pStat);
 
   if (!pTsmaStat->pTSma) {
     STSma *pTSma = metaGetSmaInfoByIndex(SMA_META(pSma), indexUid);

source/dnode/vnode/src/sma/smaUtil.c

@@ -350,49 +350,45 @@ _err:
 }
 
 /**
- * @brief pTSchema is shared
+ * @brief Clone qTaskInfo of SRSmaInfo
  *
  * @param pSma
- * @param pDest
- * @param pSrc
+ * @param pInfo
  * @return int32_t
  */
-int32_t tdCloneRSmaInfo(SSma *pSma, SRSmaInfo **pDest, SRSmaInfo *pSrc) {
-  SVnode     *pVnode = pSma->pVnode;
+int32_t tdCloneRSmaInfo(SSma *pSma, SRSmaInfo *pInfo) {
   SRSmaParam *param = NULL;
-  if (!pSrc) {
-    *pDest = NULL;
+  if (!pInfo) {
     return TSDB_CODE_SUCCESS;
   }
 
   SMetaReader mr = {0};
   metaReaderInit(&mr, SMA_META(pSma), 0);
-  smaDebug("vgId:%d, rsma clone, suid is %" PRIi64, TD_VID(pVnode), pSrc->suid);
-  if (metaGetTableEntryByUid(&mr, pSrc->suid) < 0) {
-    smaError("vgId:%d, rsma clone, failed to get table meta for %" PRIi64 " since %s", TD_VID(pVnode), pSrc->suid,
+  smaDebug("vgId:%d, rsma clone qTaskInfo for suid:%" PRIi64, SMA_VID(pSma), pInfo->suid);
+  if (metaGetTableEntryByUid(&mr, pInfo->suid) < 0) {
+    smaError("vgId:%d, rsma clone, failed to get table meta for %" PRIi64 " since %s", SMA_VID(pSma), pInfo->suid,
              terrstr());
     goto _err;
   }
   ASSERT(mr.me.type == TSDB_SUPER_TABLE);
-  ASSERT(mr.me.uid == pSrc->suid);
+  ASSERT(mr.me.uid == pInfo->suid);
   if (TABLE_IS_ROLLUP(mr.me.flags)) {
     param = &mr.me.stbEntry.rsmaParam;
     for (int32_t i = 0; i < TSDB_RETENTION_L2; ++i) {
-      if (tdCloneQTaskInfo(pSma, pSrc->iTaskInfo[i], pSrc->taskInfo[i], param, pSrc->suid, i) < 0) {
+      if (tdCloneQTaskInfo(pSma, pInfo->taskInfo[i], pInfo->iTaskInfo[i], param, pInfo->suid, i) < 0) {
         goto _err;
       }
     }
-    smaDebug("vgId:%d, rsma clone env success for %" PRIi64, TD_VID(pVnode), pSrc->suid);
+    smaDebug("vgId:%d, rsma clone env success for %" PRIi64, SMA_VID(pSma), pInfo->suid);
+  } else {
+    terrno = TSDB_CODE_RSMA_INVALID_SCHEMA;
+    goto _err;
   }
 
   metaReaderClear(&mr);
-
-  *pDest = pSrc;  // pointer copy
-
   return TSDB_CODE_SUCCESS;
 _err:
-  *pDest = NULL;
   metaReaderClear(&mr);
-  smaError("vgId:%d, rsma clone env failed for %" PRIi64 " since %s", TD_VID(pVnode), pSrc->suid, terrstr());
+  smaError("vgId:%d, rsma clone env failed for %" PRIi64 " since %s", SMA_VID(pSma), pInfo->suid, terrstr());
   return TSDB_CODE_FAILED;
 }
\ No newline at end of file

source/dnode/vnode/src/tsdb/tsdbOpen.c

@@ -86,7 +86,7 @@ int tsdbClose(STsdb **pTsdb) {
   if (*pTsdb) {
     taosThreadRwlockDestroy(&(*pTsdb)->rwLock);
     tsdbFSClose(*pTsdb);
-    tsdbCloseCache((*pTsdb)->lruCache);
+    tsdbCloseCache(*pTsdb);
     taosMemoryFreeClear(*pTsdb);
   }
   return 0;