Merge branch '3.0' into test/jcy

README-CN.md

@@ -107,7 +107,7 @@ sudo yum config-manager --set-enabled Powertools
 ### macOS
 
 ```
-sudo brew install argp-standalone pkgconfig
+brew install argp-standalone pkgconfig
 ```
 
 ### 设置 golang 开发环境
@@ -276,7 +276,7 @@ sudo make install
 安装成功后，可以在应用程序中双击 TDengine 图标启动服务，或者在终端中启动 TDengine 服务：
 
 ```bash
-launchctl start taosd
+launchctl start com.tdengine.taosd
 ```
 
 用户可以使用 TDengine CLI 来连接 TDengine 服务，在终端中，输入：

README.md

@@ -108,7 +108,7 @@ sudo yum config-manager --set-enabled powertools
 ### macOS
 
 ```
-sudo brew install argp-standalone pkgconfig
+brew install argp-standalone pkgconfig
 ```
 
 ### Setup golang environment
@@ -280,7 +280,7 @@ Installing from source code will also configure service management for TDengine.
 To start the service after installation, double-click the /applications/TDengine to start the program, or in a terminal, use:
 
 ```bash
-launchctl start taosd
+launchctl start com.tdengine.taosd
 ```
 
 Then users can use the TDengine CLI to connect the TDengine server. In a terminal, use:

cmake/cmake.platform

@@ -45,10 +45,19 @@ IF (${CMAKE_SYSTEM_NAME} MATCHES "Linux" OR ${CMAKE_SYSTEM_NAME} MATCHES "Darwin
         ADD_DEFINITIONS("-DDARWIN -Wno-tautological-pointer-compare")
 
         MESSAGE("Current system processor is ${CMAKE_SYSTEM_PROCESSOR}.")
-        IF (${CMAKE_SYSTEM_PROCESSOR} MATCHES "arm64" OR ${CMAKE_SYSTEM_PROCESSOR} MATCHES "x86_64")
-            MESSAGE("Current system arch is 64")
+        IF (${CMAKE_SYSTEM_PROCESSOR} MATCHES "arm64")
+            MESSAGE("Current system arch is arm64")
             SET(TD_DARWIN_64 TRUE)
+            SET(TD_DARWIN_ARM64 TRUE)
             ADD_DEFINITIONS("-D_TD_DARWIN_64")
+            ADD_DEFINITIONS("-D_TD_DARWIN_ARM64")
+        ENDIF ()
+        IF (${CMAKE_SYSTEM_PROCESSOR} MATCHES "x86_64")
+            MESSAGE("Current system arch is x86_64")
+            SET(TD_DARWIN_64 TRUE)
+            SET(TD_DARWIN_X64 TRUE)
+            ADD_DEFINITIONS("-D_TD_DARWIN_64")
+            ADD_DEFINITIONS("-D_TD_DARWIN_X64")
         ENDIF ()
 
         ADD_DEFINITIONS("-DHAVE_UNISTD_H")

cmake/taostools_CMakeLists.txt.in

@@ -2,7 +2,7 @@
 # taos-tools
 ExternalProject_Add(taos-tools
         GIT_REPOSITORY https://github.com/taosdata/taos-tools.git
-        GIT_TAG d58230c
+        GIT_TAG 2849aa4
         SOURCE_DIR "${TD_SOURCE_DIR}/tools/taos-tools"
         BINARY_DIR ""
         #BUILD_IN_SOURCE TRUE

contrib/CMakeLists.txt

@@ -270,7 +270,7 @@ if(${JEMALLOC_ENABLED})
         PREFIX        "jemalloc"
         SOURCE_DIR    ${CMAKE_CURRENT_SOURCE_DIR}/jemalloc
         BUILD_IN_SOURCE     1
-        CONFIGURE_COMMAND   ./autogen.sh COMMAND ./configure --prefix=${CMAKE_BINARY_DIR}/build/ --disable-initial-exec-tls
+        CONFIGURE_COMMAND   ./autogen.sh COMMAND ./configure --prefix=${CMAKE_BINARY_DIR}/build/ --disable-initial-exec-tls --with-malloc-conf='background_thread:true,metadata_thp:auto'
         BUILD_COMMAND       ${MAKE}
     )
     INCLUDE_DIRECTORIES(${CMAKE_BINARY_DIR}/build/include)

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

@@ -189,13 +189,13 @@ After the installation is complete, run `C:\TDengine\taosd.exe` to start TDengin
 
 <TabItem label="macOS" value="macos">
 
-After the installation is complete, double-click the /applications/TDengine to start the program, or run `launchctl start taosd` to start TDengine Server.
+After the installation is complete, double-click the /applications/TDengine to start the program, or run `launchctl start com.tdengine.taosd` to start TDengine Server.
 
 The following `launchctl` commands can help you manage TDengine service:
 
-- Start TDengine Server: `launchctl start taosd`
+- Start TDengine Server: `launchctl start com.tdengine.taosd`
 
-- Stop TDengine Server: `launchctl stop taosd`
+- Stop TDengine Server: `launchctl stop com.tdengine.taosd`
 
 - Check TDengine Server status: `launchctl list | grep taosd`
 

docs/en/12-taos-sql/28-recovery.md

@@ -16,10 +16,10 @@ You can use the SHOW CONNECTIONS statement to find the conn_id.
 ## Terminate a Query
 
 ```sql
-SHOW QUERY query_id;
+KILL QUERY kill_id;
 ```
 
-You can use the SHOW QUERIES statement to find the query_id.
+You can use the SHOW QUERIES statement to find the kill_id.
 
 ## Terminate a Transaction
 

docs/en/14-reference/05-taosbenchmark.md

@@ -231,7 +231,7 @@ The parameters related to database creation are configured in `dbinfo` in the js
 
 - **name**: specify the name of the database.
 
-- **drop**: indicate whether to delete the database before inserting. The default is true.
+- **drop**: indicate whether to delete the database before inserting. The value can be 'yes' or 'no'. No means do not drop. The default is to drop.
 
 #### Stream processing related configuration parameters
 

docs/en/14-reference/12-config/index.md

@@ -675,7 +675,7 @@ To prevent system resource from being exhausted by multiple concurrent streams,
 | Meaning       | Whether to generate core file when server crashes                                                                                                                       |
 | Value Range   | 0: false, 1: true                                                                                                                                                       |
 | Default Value | 1                                                                                                                                                                       |
-| Note          | The core file is generated under root directory `systemctl/launchctl start taosd` is used to start, or under the working directory if `taosd` is started directly on Linux/macOS Shell. |
+| Note          | The core file is generated under root directory `systemctl start taosd`/`launchctl start com.tdengine.taosd` is used to start, or under the working directory if `taosd` is started directly on Linux/macOS Shell. |
 
 ### udf
 

docs/en/21-tdinternal/01-arch.md

@@ -181,7 +181,7 @@ To make full use of the characteristics of time-series data, TDengine splits the
 
 For time-series data, there is generally a retention policy, which is determined by the system configuration parameter `keep`. Data files exceeding this set number of days will be automatically deleted by the system to free up storage space.
 
-Given `duration` and `keep` parameters, the total number of data files in a vnode is: keep/duration. The total number of data files should not be too large or too small. 10 to 100 is appropriate. Based on this principle, reasonable `duration` can be set. In the current version, parameter `keep` can be modified, but parameter `duration` cannot be modified once it is set.
+Given `duration` and `keep` parameters, the total number of data files in a vnode is: round up of (keep/duration+1). The total number of data files should not be too large or too small. 10 to 100 is appropriate. Based on this principle, reasonable `duration` can be set. In the current version, parameter `keep` can be modified, but parameter `duration` cannot be modified once it is set.
 
 In each data file, the data of a table is stored in blocks. A table can have one or more data file blocks. In a file block, data is stored in columns, occupying a continuous storage space, thus greatly improving the reading speed. The size of file block is determined by the system parameter `maxRows` (the maximum number of records per block), and the default value is 4096. This value should not be too large or too small. If it is too large, data location for queries will take a longer time. If it is too small, the index of data block is too large, and the compression efficiency will be low with slower reading speed.
 

docs/examples/python/tmq_example.py

 import taos
-from taos.tmq import TaosConsumer
-consumer = TaosConsumer('topic_ctb_column', group_id='vg2')
-for msg in consumer:
-    for row in msg:
-        print(row)
+from taos.tmq import *
+
+conn = taos.connect()
+
+print("init")
+conn.execute("drop database if exists py_tmq")
+conn.execute("create database if not exists py_tmq vgroups 2")
+conn.select_db("py_tmq")
+conn.execute(
+    "create stable if not exists stb1 (ts timestamp, c1 int, c2 float, c3 binary(10)) tags(t1 int)"
+)
+conn.execute("create table if not exists tb1 using stb1 tags(1)")
+conn.execute("create table if not exists tb2 using stb1 tags(2)")
+conn.execute("create table if not exists tb3 using stb1 tags(3)")
+
+print("create topic")
+conn.execute("drop topic if exists topic_ctb_column")
+conn.execute(
+    "create topic if not exists topic_ctb_column as select ts, c1, c2, c3 from stb1"
+)
+
+print("build consumer")
+conf = TaosTmqConf()
+conf.set("group.id", "tg2")
+conf.set("td.connect.user", "root")
+conf.set("td.connect.pass", "taosdata")
+conf.set("enable.auto.commit", "true")
+
+
+def tmq_commit_cb_print(tmq, resp, offset, param=None):
+    print(f"commit: {resp}, tmq: {tmq}, offset: {offset}, param: {param}")
+
+
+conf.set_auto_commit_cb(tmq_commit_cb_print, None)
+tmq = conf.new_consumer()
+
+print("build topic list")
+
+topic_list = TaosTmqList()
+topic_list.append("topic_ctb_column")
+
+print("basic consume loop")
+tmq.subscribe(topic_list)
+
+sub_list = tmq.subscription()
+
+print("subscribed topics: ", sub_list)
+
+while 1:
+    res = tmq.poll(1000)
+    if res:
+        topic = res.get_topic_name()
+        vg = res.get_vgroup_id()
+        db = res.get_db_name()
+        print(f"topic: {topic}\nvgroup id: {vg}\ndb: {db}")
+        for row in res:
+            print(row)

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

@@ -188,13 +188,13 @@ Active: inactive (dead)
 
 <TabItem label="macOS 系统" value="macos">
 
-安装后，在应用程序目录下，双击 TDengine 图标来启动程序，也可以运行 `launchctl start taosd` 来启动 TDengine 服务进程。
+安装后，在应用程序目录下，双击 TDengine 图标来启动程序，也可以运行 `launchctl start com.tdengine.taosd` 来启动 TDengine 服务进程。
 
 如下 `launchctl` 命令可以帮助你管理 TDengine 服务：
 
-- 启动服务进程：`launchctl start taosd`
+- 启动服务进程：`launchctl start com.tdengine.taosd`
 
-- 停止服务进程：`launchctl stop taosd`
+- 停止服务进程：`launchctl stop com.tdengine.taosd`
 
 - 查看服务状态：`launchctl list | grep taosd`
 

docs/zh/12-taos-sql/10-function.md

@@ -340,7 +340,7 @@ LTRIM(expr)
 #### RTRIM
 
 ```sql
-LTRIM(expr)
+RTRIM(expr)
 ```
 
 **功能说明**：返回清除右边空格后的字符串。

docs/zh/12-taos-sql/28-recovery.md

@@ -17,10 +17,10 @@ conn_id 可以通过 `SHOW CONNECTIONS` 获取。
 ## 终止查询
 
 ```sql
-SHOW QUERY query_id;
+KILL QUERY kill_id;
 ```
 
-query_id 可以通过 `SHOW QUERIES` 获取。
+kill_id 可以通过 `SHOW QUERIES` 获取。
 
 ## 终止事务
 

docs/zh/12-taos-sql/29-changes.md

@@ -94,3 +94,10 @@ description: "TDengine 3.0 版本的语法变更说明"
 | 9 | SAMPLE | 增强 | 可以直接用于超级表了。没有PARTITION BY时，超级表的数据会被合并成一条时间线。
 | 10 | STATECOUNT | 增强 | 可以直接用于超级表了。没有PARTITION BY时，超级表的数据会被合并成一条时间线。
 | 11 | STATEDURATION | 增强 | 可以直接用于超级表了。没有PARTITION BY时，超级表的数据会被合并成一条时间线。
+
+
+## SCHEMALESS 变更
+
+| # | **元素**  | **<div style={{width: 60}}>差异性</div>** | **说明** |
+| - | :------- | :-------- | :------- |
+| 1 | 主键ts 变更为 _ts | 变更 | schemaless自动建的列名用 _ 开头，不同于2.x。

docs/zh/14-reference/05-taosbenchmark.md

@@ -231,7 +231,7 @@ taosBenchmark -A INT,DOUBLE,NCHAR,BINARY\(16\)
 
 - **name** : 数据库名。
 
-- **drop** : 插入前是否删除数据库，默认为 true。
+- **drop** : 插入前是否删除数据库，可选项为 "yes" 或者 "no", 为 "no" 时不创建。默认删除。
 
 #### 流式计算相关配置参数
 

docs/zh/14-reference/13-schemaless/13-schemaless.md

@@ -115,7 +115,7 @@ st,t1=3,t2=4,t3=t3 c1=3i64,c3="passit",c2=false,c4=4f64 1626006833639000000
 
 ## 数据模式映射规则
 
-本节将说明行协议的数据如何映射成为具有模式的数据。每个行协议中数据 measurement 映射为
+本节将说明 InfluxDB 行协议(Line Protocol)的数据如何映射成为具有模式的数据。每个行协议中数据 measurement 映射为
 超级表名称。tag_set 中的 标签名称为 数据模式中的标签名，field_set 中的名称为列名称。以如下数据为例，说明映射规则：
 
 ```json

examples/c/demoapi.c

-// C api call sequence demo
-// to compile: gcc -o apidemo apidemo.c -ltaos
+/*
+ * 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 <stdio.h>
 #include <stdlib.h>
@@ -13,23 +25,23 @@
 #include "taos.h"
 
 #define debugPrint(fmt, ...) \
-    do { if (g_args.debug_print || g_args.verbose_print) \
-      fprintf(stdout, "DEBG: "fmt, __VA_ARGS__); } while(0)
+    do { if (g_args.debug_print || g_args.verbose_print) {\
+      fprintf(stdout, "DEBG: "fmt, __VA_ARGS__); }} while (0)
 
 #define warnPrint(fmt, ...) \
     do { fprintf(stderr, "\033[33m"); \
         fprintf(stderr, "WARN: "fmt, __VA_ARGS__); \
-        fprintf(stderr, "\033[0m"); } while(0)
+        fprintf(stderr, "\033[0m"); } while (0)
 
 #define errorPrint(fmt, ...) \
     do { fprintf(stderr, "\033[31m"); \
         fprintf(stderr, "ERROR: "fmt, __VA_ARGS__); \
-        fprintf(stderr, "\033[0m"); } while(0)
+        fprintf(stderr, "\033[0m"); } while (0)
 
 #define okPrint(fmt, ...) \
     do { fprintf(stderr, "\033[32m"); \
         fprintf(stderr, "OK: "fmt, __VA_ARGS__); \
-        fprintf(stderr, "\033[0m"); } while(0)
+        fprintf(stderr, "\033[0m"); } while (0)
 
 int64_t g_num_of_tb = 2;
 int64_t g_num_of_rec = 3;
@@ -74,10 +86,16 @@ static void prepare_data(TAOS* taos) {
     res = taos_query(taos, "create database test;");
     taos_free_result(res);
     taosMsleep(100);
-    taos_select_db(taos, "test");
+    if (taos_select_db(taos, "test")) {
+        errorPrint("%s() LN%d: taos_select_db() failed\n",
+                __func__, __LINE__);
+        return;
+    }
 
     char command[1024] = {0};
-    sprintf(command, "%s", "create table meters(ts timestamp, f float, n int, bin1 binary(20), c nchar(20), bin2 binary(20)) tags(area int, city binary(20), dist nchar(20), street binary(20));");
+    sprintf(command, "%s", "create table meters(ts timestamp, f float, n int, "
+            "bin1 binary(20), c nchar(20), bin2 binary(20)) tags(area int, "
+            "city binary(20), dist nchar(20), street binary(20));");
     res = taos_query(taos, command);
     if ((res) && (0 == taos_errno(res))) {
         okPrint("%s created\n", "meters");
@@ -117,12 +135,11 @@ static void prepare_data(TAOS* taos) {
             sprintf(command, "insert into t%"PRId64" "
                     "values(%" PRId64 ", %f, %"PRId64", "
                     "'%c%d', '%s%c%d', '%c%d')",
-                    i, 1650000000000+j, (float)j, j,
+                    i, 1650000000000+j, j * 1.0, j,
                     'a'+(int)j%25, rand(),
                     // "涛思", 'z' - (int)j%25, rand(),
                     "TAOS", 'z' - (int)j%25, rand(),
-                    'b' - (int)j%25, rand()
-                    );
+                    'b' - (int)j%25, rand());
             res = taos_query(taos,  command);
             if ((res) && (0 == taos_errno(res))) {
                 affected = taos_affected_rows(res);
@@ -150,8 +167,7 @@ static void prepare_data(TAOS* taos) {
                 i, 1650000000000+j+1, (float)j, j,
                 'a'+(int)j%25, rand(),
                 "数据", 'z' - (int)j%25, rand(),
-                'b' - (int)j%25, rand()
-               );
+                'b' - (int)j%25, rand());
         res = taos_query(taos,  command);
         if ((res) && (0 == taos_errno(res))) {
             affected = taos_affected_rows(res);
@@ -179,7 +195,8 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
     }
 
     if (block) {
-        warnPrint("%s", "call taos_fetch_block(), don't call taos_fetch_lengths()\n");
+        warnPrint("%s", "call taos_fetch_block(), "
+                "don't call taos_fetch_lengths()\n");
         int rows = 0;
         while ((rows = taos_fetch_block(res, &row))) {
             int *lengths = taos_fetch_lengths(res);
@@ -190,7 +207,7 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
                     printf("col%d type is %d, no need get offset\n",
                             f, fields[f].type);
                     for (int64_t c = 0; c < rows; c++) {
-                        switch(fields[f].type) {
+                        switch (fields[f].type) {
                             case TSDB_DATA_TYPE_TIMESTAMP:
                                 if (taos_is_null(res, c, f)) {
                                     printf("col%d, row: %"PRId64" "
@@ -199,7 +216,8 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
                                     printf("col%d, row: %"PRId64", "
                                             "value: %"PRId64"\n",
                                             f, c,
-                                            *(int64_t*)((char*)(row[f])+c*sizeof(int64_t)));
+                                            *(int64_t*)((char*)(row[f])
+                                                +c*sizeof(int64_t)));
                                 }
                                 break;
 
@@ -211,7 +229,8 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
                                     printf("col%d, row: %"PRId64", "
                                             "value: %d\n",
                                             f, c,
-                                            *(int32_t*)((char*)(row[f])+c*sizeof(int32_t)));
+                                            *(int32_t*)((char*)(row[f])
+                                                +c*sizeof(int32_t)));
                                 }
                                 break;
 
@@ -223,7 +242,8 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
                                     printf("col%d, row: %"PRId64", "
                                             "value: %f\n",
                                             f, c,
-                                            *(float*)((char*)(row[f])+c*sizeof(float)));
+                                            *(float*)((char*)(row[f])
+                                                +c*sizeof(float)));
                                 }
                                 break;
 
@@ -238,14 +258,18 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
                     if (offsets) {
                         for (int c = 0; c < rows; c++) {
                             if (offsets[c] != -1) {
-                                int length = *(int16_t*)((char*)(row[f]) + offsets[c]);
+                                int length = *(int16_t*)((char*)(row[f])
+                                        + offsets[c]);
                                 char *buf = calloc(1, length + 1);
-                                strncpy(buf, (char *)((char*)(row[f]) + offsets[c] + 2), length);
-                                printf("row: %d, col: %d, offset: %d, length: %d, content: %s\n",
+                                strncpy(buf, (char *)((char*)(row[f])
+                                            + offsets[c] + 2), length);
+                                printf("row: %d, col: %d, offset: %d, "
+                                        "length: %d, content: %s\n",
                                         c, f, offsets[c], length, buf);
                                 free(buf);
                             } else {
-                                printf("row: %d, col: %d, offset: -1, means content is NULL\n",
+                                printf("row: %d, col: %d, offset: -1, "
+                                        "means content is NULL\n",
                                         c, f);
                             }
                         }
@@ -267,7 +291,8 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
             int* lengths = taos_fetch_lengths(res);
             if (lengths) {
                 for (int c = 0; c < num_fields; c++) {
-                    printf("row: %"PRId64", col: %d, is_null: %s, length of column %d is %d\n",
+                    printf("row: %"PRId64", col: %d, is_null: %s, "
+                            "length of column %d is %d\n",
                             num_rows, c,
                             taos_is_null(res, num_rows, c)?"True":"False",
                             c, lengths[c]);
@@ -277,7 +302,7 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
                         __func__, __LINE__, tbname);
             }
 
-            num_rows ++;
+            num_rows++;
         }
     }
 
@@ -322,7 +347,8 @@ int main(int argc, char *argv[]) {
 #endif
     TAOS* taos = taos_connect(host, user, passwd, "", 0);
     if (taos == NULL) {
-        printf("\033[31mfailed to connect to db, reason:%s\033[0m\n", taos_errstr(taos));
+        printf("\033[31mfailed to connect to db, reason:%s\033[0m\n",
+                taos_errstr(taos));
         exit(1);
     }
 

include/common/tglobal.h

@@ -93,6 +93,7 @@ extern int64_t tsQueryBufferSizeBytes;  // maximum allowed usage buffer size in
 
 // query client
 extern int32_t tsQueryPolicy;
+extern int32_t tsQueryRspPolicy;
 extern int32_t tsQuerySmaOptimize;
 extern int32_t tsQueryRsmaTolerance;
 extern bool    tsQueryPlannerTrace;

include/dnode/mnode/mnode.h

@@ -30,9 +30,10 @@ typedef struct SMnode SMnode;
 
 typedef struct {
   int32_t  dnodeId;
-  bool     standby;
   bool     deploy;
-  SReplica replica;
+  int8_t   selfIndex;
+  int8_t   numOfReplicas;
+  SReplica replicas[TSDB_MAX_REPLICA];
   SMsgCb   msgCb;
 } SMnodeOpt;
 

include/libs/qcom/query.h

@@ -54,6 +54,9 @@ typedef enum {
 #define QUERY_POLICY_QNODE  3
 #define QUERY_POLICY_CLIENT 4
 
+#define QUERY_RSP_POLICY_DELAY 0
+#define QUERY_RSP_POLICY_QUICK 1
+
 typedef struct STableComInfo {
   uint8_t  numOfTags;     // the number of tags in schema
   uint8_t  precision;     // the number of precision
@@ -230,7 +233,7 @@ int32_t cloneTableMeta(STableMeta* pSrc, STableMeta** pDst);
 int32_t cloneDbVgInfo(SDBVgInfo* pSrc, SDBVgInfo** pDst);
 
 extern int32_t (*queryBuildMsg[TDMT_MAX])(void* input, char** msg, int32_t msgSize, int32_t* msgLen,
-                                          void* (*mallocFp)(int32_t));
+                                          void* (*mallocFp)(int64_t));
 extern int32_t (*queryProcessMsgRsp[TDMT_MAX])(void* output, char* msg, int32_t msgSize);
 
 #define SET_META_TYPE_NULL(t)       (t) = META_TYPE_NULL_TABLE

include/libs/scalar/scalar.h

@@ -43,7 +43,7 @@ int32_t scalarGetOperatorParamNum(EOperatorType type);
 int32_t scalarGenerateSetFromList(void **data, void *pNode, uint32_t type);
 
 int32_t vectorGetConvertType(int32_t type1, int32_t type2);
-int32_t vectorConvertImpl(const SScalarParam *pIn, SScalarParam *pOut, int32_t *overflow);
+int32_t vectorConvertSingleColImpl(const SScalarParam *pIn, SScalarParam *pOut, int32_t *overflow, int32_t startIndex, int32_t numOfRows);
 
 /* Math functions */
 int32_t absFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput);

include/libs/stream/streamState.h

@@ -36,7 +36,7 @@ typedef struct {
   int32_t      number;
 } SStreamState;
 
-SStreamState* streamStateOpen(char* path, SStreamTask* pTask, bool specPath);
+SStreamState* streamStateOpen(char* path, SStreamTask* pTask, bool specPath, int32_t szPage, int32_t pages);
 void          streamStateClose(SStreamState* pState);
 int32_t       streamStateBegin(SStreamState* pState);
 int32_t       streamStateCommit(SStreamState* pState);

include/libs/sync/sync.h

@@ -211,7 +211,7 @@ typedef struct SSyncInfo {
 
 int32_t     syncInit();
 void        syncCleanUp();
-int64_t     syncOpen(const SSyncInfo* pSyncInfo);
+int64_t     syncOpen(SSyncInfo* pSyncInfo);
 void        syncStart(int64_t rid);
 void        syncStop(int64_t rid);
 int32_t     syncSetStandby(int64_t rid);
@@ -233,7 +233,7 @@ const char* syncStr(ESyncState state);
 bool        syncIsRestoreFinish(int64_t rid);
 int32_t     syncGetSnapshotByIndex(int64_t rid, SyncIndex index, SSnapshot* pSnapshot);
 
-int32_t syncReconfig(int64_t rid, const SSyncCfg* pNewCfg);
+int32_t syncReconfig(int64_t rid, SSyncCfg* pNewCfg);
 
 // build SRpcMsg, need to call syncPropose with SRpcMsg
 int32_t syncReconfigBuild(int64_t rid, const SSyncCfg* pNewCfg, SRpcMsg* pRpcMsg);

include/libs/transport/trpc.h

@@ -123,9 +123,9 @@ void    rpcCleanup();
 void *rpcOpen(const SRpcInit *pRpc);
 void  rpcClose(void *);
 void  rpcCloseImpl(void *);
-void *rpcMallocCont(int32_t contLen);
+void *rpcMallocCont(int64_t contLen);
 void  rpcFreeCont(void *pCont);
-void *rpcReallocCont(void *ptr, int32_t contLen);
+void *rpcReallocCont(void *ptr, int64_t contLen);
 
 // Because taosd supports multi-process mode
 // These functions should not be used on the server side

include/libs/wal/wal.h

@@ -43,6 +43,7 @@ extern "C" {
 #define WAL_FILE_LEN      (WAL_PATH_LEN + 32)
 #define WAL_MAGIC         0xFAFBFCFDF4F3F2F1ULL
 #define WAL_SCAN_BUF_SIZE (1024 * 1024 * 3)
+#define WAL_RECOV_SIZE_LIMIT (100 * WAL_SCAN_BUF_SIZE)
 
 typedef enum {
   TAOS_WAL_WRITE = 1,

include/os/osEnv.h

@@ -49,9 +49,15 @@ extern SDiskSpace tsTempSpace;
 void osDefaultInit();
 void osUpdate();
 void osCleanup();
+
 bool osLogSpaceAvailable();
 bool osDataSpaceAvailable();
 bool osTempSpaceAvailable();
+
+bool osLogSpaceSufficient();
+bool osDataSpaceSufficient();
+bool osTempSpaceSufficient();
+
 void osSetTimezone(const char *timezone);
 void osSetSystemLocale(const char *inLocale, const char *inCharSet);
 
@@ -59,4 +65,4 @@ void osSetSystemLocale(const char *inLocale, const char *inCharSet);
 }
 #endif
 
-#endif /*_TD_OS_ENV_H_*/
+#endif /*_TD_OS_ENV_H_*/
\ No newline at end of file

include/os/osMemory.h

@@ -29,12 +29,12 @@ extern "C" {
 #define free    FREE_FUNC_TAOS_FORBID
 #endif
 
-void   *taosMemoryMalloc(int32_t size);
-void   *taosMemoryCalloc(int32_t num, int32_t size);
-void   *taosMemoryRealloc(void *ptr, int32_t size);
+void   *taosMemoryMalloc(int64_t size);
+void   *taosMemoryCalloc(int64_t num, int64_t size);
+void   *taosMemoryRealloc(void *ptr, int64_t size);
 void   *taosMemoryStrDup(const char *ptr);
 void    taosMemoryFree(void *ptr);
-int32_t taosMemorySize(void *ptr);
+int64_t taosMemorySize(void *ptr);
 void    taosPrintBackTrace();
 
 #define taosMemoryFreeClear(ptr)   \

include/util/taoserror.h

@@ -450,6 +450,7 @@ int32_t* taosGetErrno();
 #define TSDB_CODE_WAL_INVALID_VER               TAOS_DEF_ERROR_CODE(0, 0x1003)
 #define TSDB_CODE_WAL_OUT_OF_MEMORY             TAOS_DEF_ERROR_CODE(0, 0x1004)
 #define TSDB_CODE_WAL_LOG_NOT_EXIST             TAOS_DEF_ERROR_CODE(0, 0x1005)
+#define TSDB_CODE_WAL_CHKSUM_MISMATCH           TAOS_DEF_ERROR_CODE(0, 0x1006)
 
 // tfs
 #define TSDB_CODE_FS_INVLD_CFG                  TAOS_DEF_ERROR_CODE(0, 0x2201)

include/util/trbtree.h

@@ -26,7 +26,7 @@ typedef struct SRBTree     SRBTree;
 typedef struct SRBTreeNode SRBTreeNode;
 typedef struct SRBTreeIter SRBTreeIter;
 
-typedef int32_t (*tRBTreeCmprFn)(const void *, const void *);
+typedef int32_t (*tRBTreeCmprFn)(const SRBTreeNode *, const SRBTreeNode *);
 
 // SRBTree =============================================
 #define tRBTreeMin(T) ((T)->min == ((T)->NIL) ? NULL : (T)->min)
@@ -36,7 +36,7 @@ void         tRBTreeCreate(SRBTree *pTree, tRBTreeCmprFn cmprFn);
 SRBTreeNode *tRBTreePut(SRBTree *pTree, SRBTreeNode *z);
 void         tRBTreeDrop(SRBTree *pTree, SRBTreeNode *z);
 SRBTreeNode *tRBTreeDropByKey(SRBTree *pTree, void *pKey);
-SRBTreeNode *tRBTreeGet(SRBTree *pTree, void *pKey);
+SRBTreeNode *tRBTreeGet(SRBTree *pTree, const SRBTreeNode *pKeyNode);
 
 // SRBTreeIter =============================================
 #define tRBTreeIterCreate(tree, ascend) \
@@ -53,8 +53,6 @@ struct SRBTreeNode {
   SRBTreeNode *right;
 };
 
-#define RBTREE_NODE_PAYLOAD(N) ((const void *)&(N)[1])
-
 struct SRBTree {
   tRBTreeCmprFn cmprFn;
   int64_t       n;

packaging/release.sh

@@ -260,7 +260,7 @@ if [ "$osType" != "Darwin" ]; then
       if [[ "$pagMode" == "full" ]]; then
         if [ -d ${top_dir}/tools/taos-tools/packaging/deb ]; then
           cd ${top_dir}/tools/taos-tools/packaging/deb
-          taos_tools_ver=$(git describe --tags | sed -e 's/ver-//g' | awk -F '-' '{print $1}')
+          taos_tools_ver=$(git tag |grep -v taos | sort | tail -1)
           [ -z "$taos_tools_ver" ] && taos_tools_ver="0.1.0"
 
           ${csudo}./make-taos-tools-deb.sh ${top_dir} \
@@ -285,7 +285,7 @@ if [ "$osType" != "Darwin" ]; then
       if [[ "$pagMode" == "full" ]]; then
         if [ -d ${top_dir}/tools/taos-tools/packaging/rpm ]; then
           cd ${top_dir}/tools/taos-tools/packaging/rpm
-          taos_tools_ver=$(git describe --tags | sed -e 's/ver-//g' | awk -F '-' '{print $1}' | sed -e 's/-/_/g')
+          taos_tools_ver=$(git tag |grep -v taos | sort | tail -1)
           [ -z "$taos_tools_ver" ] && taos_tools_ver="0.1.0"
 
           ${csudo}./make-taos-tools-rpm.sh ${top_dir} \

packaging/tools/TDengine

@@ -17,12 +17,12 @@ EOF
 
 taosd_status=`Launchctl list | grep taosd | head -n 1 | awk '{print $1}'`
 if [ "$taosd_status"x = "-"x ]; then
-    launchctl start taosd
+    launchctl start com.tdengine.taosd
     showAlertMessage "Taosd is running!" "TDengine" "ok" "note"
 else
     choose_result=`showAlertMessage "Taosd is running!\nDo you want to close it?" "TDengine" "yes,cancel" "stop"`
     if [ "$choose_result"x = "button returned:yes"x ]; then
-        launchctl stop taosd
+        launchctl stop com.tdengine.taosd
     fi
 fi
 

packaging/tools/com.taosdata.taosd.plist

@@ -3,7 +3,7 @@
 <plist version="1.0">
 <dict>
     <key>Label</key>
-    <string>taosd</string>
+    <string>com.tdengine.taosd</string>
     <key>ProgramArguments</key>
     <array>
         <string>/usr/local/bin/taosd</string>

packaging/tools/mac_before_install.txt

 TDengine is a high-efficient, scalable, high-available distributed time-series database, which makes a lot of optimizations on inserting and querying data, which is far more efficient than normal regular databases. So TDengine can meet the high requirements of IOT and other areas on storing and querying a large amount of data. 
 
 To configure TDengine : edit /etc/taos/taos.cfg
-To start service : launchctl start taosd
+To start service : launchctl start com.tdengine.taosd
 To access TDengine : use taos in shell
\ No newline at end of file

packaging/tools/make_install.sh

@@ -615,7 +615,7 @@ function update_TDengine() {
     if [ "$osType" != "Darwin" ]; then
       echo -e "${GREEN_DARK}To start ${productName}     ${NC}: ${serverName}${NC}"
     else
-      echo -e "${GREEN_DARK}To start service      ${NC}: launchctl start ${serverName}${NC}"
+      echo -e "${GREEN_DARK}To start service      ${NC}: launchctl start com.tdengine.taosd${NC}"
     fi
     [ -f ${installDir}/bin/taosadapter ] && \
       echo -e "${GREEN_DARK}To start Taos Adapter ${NC}: taosadapter &${NC}"
@@ -666,7 +666,7 @@ function install_TDengine() {
     if [ "$osType" != "Darwin" ]; then
       echo -e "${GREEN_DARK}To start ${productName}     ${NC}: ${serverName}${NC}"
     else
-      echo -e "${GREEN_DARK}To start service      ${NC}: launchctl start ${serverName}${NC}"
+      echo -e "${GREEN_DARK}To start service      ${NC}: launchctl start com.tdengine.taosd${NC}"
     fi
     [ -f ${installDir}/bin/taosadapter ] && \
       echo -e "${GREEN_DARK}To start Taos Adapter ${NC}: taosadapter &${NC}"

packaging/tools/post.sh

@@ -18,6 +18,7 @@ if [ "$osType" != "Darwin" ]; then
   script_dir=$(dirname $(readlink -f "$0"))
   verNumber=""
   lib_file_ext="so"
+  lib_file_ext_1="so.1"
 
   bin_link_dir="/usr/bin"
   lib_link_dir="/usr/lib"
@@ -29,6 +30,7 @@ else
   script_dir=${source_dir}/packaging/tools
   verNumber=`ls tdengine/driver | grep -E "libtaos\.[0-9]\.[0-9]" | sed "s/libtaos.//g" |  sed "s/.dylib//g" | head -n 1`
   lib_file_ext="dylib"
+  lib_file_ext_1="1.dylib"
 
   bin_link_dir="/usr/local/bin"
   lib_link_dir="/usr/local/lib"
@@ -134,14 +136,14 @@ function install_lib() {
     [ -f ${lib_link_dir}/libtaosws.${lib_file_ext} ] && ${csudo}rm -f ${lib_link_dir}/libtaosws.${lib_file_ext} || :
     [ -f ${lib64_link_dir}/libtaosws.${lib_file_ext} ] && ${csudo}rm -f ${lib64_link_dir}/libtaosws.${lib_file_ext} || :
 
-    ${csudo}ln -s ${lib_dir}/libtaos.* ${lib_link_dir}/libtaos.so.1
-    ${csudo}ln -s ${lib_link_dir}/libtaos.so.1 ${lib_link_dir}/libtaos.so
+    ${csudo}ln -s ${lib_dir}/libtaos.* ${lib_link_dir}/libtaos.${lib_file_ext_1}
+    ${csudo}ln -s ${lib_link_dir}/libtaos.${lib_file_ext_1} ${lib_link_dir}/libtaos.${lib_file_ext}
 
     [ -f ${lib_dir}/libtaosws.${lib_file_ext} ] && ${csudo}ln -sf ${lib_dir}/libtaosws.${lib_file_ext} ${lib_link_dir}/libtaosws.${lib_file_ext} ||:
 
-    if [[ -d ${lib64_link_dir} && ! -e ${lib64_link_dir}/libtaos.so ]]; then
-      ${csudo}ln -s ${lib_dir}/libtaos.* ${lib64_link_dir}/libtaos.so.1           || :
-      ${csudo}ln -s ${lib64_link_dir}/libtaos.so.1 ${lib64_link_dir}/libtaos.so   || :
+    if [[ -d ${lib64_link_dir} && ! -e ${lib64_link_dir}/libtaos.${lib_file_ext} ]]; then
+      ${csudo}ln -s ${lib_dir}/libtaos.* ${lib64_link_dir}/libtaos.${lib_file_ext_1}          || :
+      ${csudo}ln -s ${lib64_link_dir}/libtaos.${lib_file_ext_1} ${lib64_link_dir}/libtaos.${lib_file_ext}   || :
 
       [ -f ${lib_dir}/libtaosws.${lib_file_ext} ] && ${csudo}ln -sf ${lib_dir}/libtaosws.${lib_file_ext} ${lib64_link_dir}/libtaosws.${lib_file_ext} || :
     fi

packaging/tools/remove.sh

@@ -117,7 +117,7 @@ function clean_local_bin() {
 function clean_lib() {
   # Remove link
   ${csudo}rm -f ${lib_link_dir}/libtaos.* || :
-  [ -f ${lib_link_dir}/libtaosws.so ] && ${csudo}rm -f ${lib_link_dir}/libtaosws.so || :
+  [ -f ${lib_link_dir}/libtaosws.* ] && ${csudo}rm -f ${lib_link_dir}/libtaosws.* || :
 
   ${csudo}rm -f ${lib64_link_dir}/libtaos.* || :
   [ -f ${lib64_link_dir}/libtaosws.* ] && ${csudo}rm -f ${lib64_link_dir}/libtaosws.* || :

source/client/CMakeLists.txt

@@ -19,6 +19,13 @@ target_link_libraries(
     PRIVATE os util common transport nodes parser command planner catalog scheduler function qcom
 )
 
+if(TD_DARWIN_ARM64)
+    target_link_libraries(
+        taos
+        PRIVATE  "-arch x86_64"
+    )
+endif()
+
 if(TD_WINDOWS)
     INCLUDE_DIRECTORIES(jni/windows)
     INCLUDE_DIRECTORIES(jni/windows/win32)

source/client/inc/clientLog.h

@@ -22,6 +22,7 @@ extern "C" {
 
 #include "tlog.h"
 
+// clang-format off
 #define tscFatal(...)  do { if (cDebugFlag & DEBUG_FATAL) { taosPrintLog("TSC FATAL ", DEBUG_FATAL, cDebugFlag, __VA_ARGS__); }} while(0)
 #define tscError(...)  do { if (cDebugFlag & DEBUG_ERROR) { taosPrintLog("TSC ERROR ", DEBUG_ERROR, cDebugFlag, __VA_ARGS__); }} while(0)
 #define tscWarn(...)   do { if (cDebugFlag & DEBUG_WARN)  { taosPrintLog("TSC WARN ", DEBUG_WARN, cDebugFlag, __VA_ARGS__); }}  while(0)
@@ -30,6 +31,7 @@ extern "C" {
 #define tscTrace(...)  do { if (cDebugFlag & DEBUG_TRACE) { taosPrintLog("TSC ", DEBUG_TRACE, cDebugFlag, __VA_ARGS__); }} while(0)
 #define tscDebugL(...) do { if (cDebugFlag & DEBUG_DEBUG) { taosPrintLongString("TSC ", DEBUG_DEBUG, cDebugFlag, __VA_ARGS__); }} while(0)
 #define tscPerf(...)   do { taosPrintLog("TSC ", 0, cDebugFlag, __VA_ARGS__); } while(0)
+// clang-format on
 
 #ifdef __cplusplus
 }

source/client/inc/clientStmt.h

@@ -42,37 +42,37 @@ typedef enum {
 } STMT_STATUS;
 
 typedef struct SStmtTableCache {
-  STableDataBlocks* pDataBlock;
-  void*             boundTags;
+  STableDataBlocks *pDataBlock;
+  void             *boundTags;
 } SStmtTableCache;
 
 typedef struct SStmtQueryResInfo {
-  TAOS_FIELD*    fields;
-  TAOS_FIELD*    userFields;
-  uint32_t       numOfCols;
-  int32_t        precision;
+  TAOS_FIELD *fields;
+  TAOS_FIELD *userFields;
+  uint32_t    numOfCols;
+  int32_t     precision;
 } SStmtQueryResInfo;
 
 typedef struct SStmtBindInfo {
-  bool         needParse;
-  bool         inExecCache;
-  uint64_t     tbUid;
-  uint64_t     tbSuid;
-  int32_t      sBindRowNum;
-  int32_t      sBindLastIdx;
-  int8_t       tbType;
-  bool         tagsCached;
-  void*        boundTags;  
-  char         tbName[TSDB_TABLE_FNAME_LEN];
-  char         tbFName[TSDB_TABLE_FNAME_LEN];
-  char         stbFName[TSDB_TABLE_FNAME_LEN];
-  SName        sname;
+  bool     needParse;
+  bool     inExecCache;
+  uint64_t tbUid;
+  uint64_t tbSuid;
+  int32_t  sBindRowNum;
+  int32_t  sBindLastIdx;
+  int8_t   tbType;
+  bool     tagsCached;
+  void    *boundTags;
+  char     tbName[TSDB_TABLE_FNAME_LEN];
+  char     tbFName[TSDB_TABLE_FNAME_LEN];
+  char     stbFName[TSDB_TABLE_FNAME_LEN];
+  SName    sname;
 } SStmtBindInfo;
 
 typedef struct SStmtExecInfo {
   int32_t      affectedRows;
-  SRequestObj* pRequest;
-  SHashObj*    pBlockHash;
+  SRequestObj *pRequest;
+  SHashObj    *pBlockHash;
   bool         autoCreateTbl;
 } SStmtExecInfo;
 
@@ -80,20 +80,20 @@ typedef struct SStmtSQLInfo {
   STMT_TYPE         type;
   STMT_STATUS       status;
   uint64_t          runTimes;
-  SHashObj*         pTableCache;   //SHash<SStmtTableCache>
-  SQuery*           pQuery;
-  char*             sqlStr;
+  SHashObj         *pTableCache;  // SHash<SStmtTableCache>
+  SQuery           *pQuery;
+  char             *sqlStr;
   int32_t           sqlLen;
-  SArray*           nodeList;
+  SArray           *nodeList;
   SStmtQueryResInfo queryRes;
   bool              autoCreateTbl;
-  SHashObj*         pVgHash;
+  SHashObj         *pVgHash;
 } SStmtSQLInfo;
 
 typedef struct STscStmt {
-  STscObj*      taos;
-  SCatalog*     pCatalog;
-  int32_t       affectedRows;
+  STscObj  *taos;
+  SCatalog *pCatalog;
+  int32_t   affectedRows;
 
   SStmtSQLInfo  sql;
   SStmtExecInfo exec;
@@ -103,28 +103,48 @@ typedef struct STscStmt {
 #define STMT_STATUS_NE(S) (pStmt->sql.status != STMT_##S)
 #define STMT_STATUS_EQ(S) (pStmt->sql.status == STMT_##S)
 
-#define STMT_ERR_RET(c) do { int32_t _code = c; if (_code != TSDB_CODE_SUCCESS) { terrno = _code; return _code; } } while (0)
-#define STMT_RET(c) do { int32_t _code = c; if (_code != TSDB_CODE_SUCCESS) { terrno = _code; } return _code; } while (0)
-#define STMT_ERR_JRET(c) do { code = c; if (code != TSDB_CODE_SUCCESS) { terrno = code; goto _return; } } while (0)
-
-TAOS_STMT *stmtInit(STscObj* taos);
-int stmtClose(TAOS_STMT *stmt);
-int stmtExec(TAOS_STMT *stmt);
+#define STMT_ERR_RET(c)               \
+  do {                                \
+    int32_t _code = c;                \
+    if (_code != TSDB_CODE_SUCCESS) { \
+      terrno = _code;                 \
+      return _code;                   \
+    }                                 \
+  } while (0)
+#define STMT_RET(c)                   \
+  do {                                \
+    int32_t _code = c;                \
+    if (_code != TSDB_CODE_SUCCESS) { \
+      terrno = _code;                 \
+    }                                 \
+    return _code;                     \
+  } while (0)
+#define STMT_ERR_JRET(c)             \
+  do {                               \
+    code = c;                        \
+    if (code != TSDB_CODE_SUCCESS) { \
+      terrno = code;                 \
+      goto _return;                  \
+    }                                \
+  } while (0)
+
+TAOS_STMT  *stmtInit(STscObj *taos);
+int         stmtClose(TAOS_STMT *stmt);
+int         stmtExec(TAOS_STMT *stmt);
 const char *stmtErrstr(TAOS_STMT *stmt);
-int stmtAffectedRows(TAOS_STMT *stmt);
-int stmtAffectedRowsOnce(TAOS_STMT *stmt);
-int stmtPrepare(TAOS_STMT *stmt, const char *sql, unsigned long length);
-int stmtSetTbName(TAOS_STMT *stmt, const char *tbName);
-int stmtSetTbTags(TAOS_STMT *stmt, TAOS_MULTI_BIND *tags);
-int stmtGetTagFields(TAOS_STMT* stmt, int* nums, TAOS_FIELD_E** fields);
-int stmtGetColFields(TAOS_STMT* stmt, int* nums, TAOS_FIELD_E** fields);
-int stmtIsInsert(TAOS_STMT *stmt, int *insert);
-int stmtGetParamNum(TAOS_STMT *stmt, int *nums);
-int stmtGetParam(TAOS_STMT *stmt, int idx, int *type, int *bytes);
-int stmtAddBatch(TAOS_STMT *stmt);
-TAOS_RES *stmtUseResult(TAOS_STMT *stmt);
-int stmtBindBatch(TAOS_STMT *stmt, TAOS_MULTI_BIND *bind, int32_t colIdx);
-
+int         stmtAffectedRows(TAOS_STMT *stmt);
+int         stmtAffectedRowsOnce(TAOS_STMT *stmt);
+int         stmtPrepare(TAOS_STMT *stmt, const char *sql, unsigned long length);
+int         stmtSetTbName(TAOS_STMT *stmt, const char *tbName);
+int         stmtSetTbTags(TAOS_STMT *stmt, TAOS_MULTI_BIND *tags);
+int         stmtGetTagFields(TAOS_STMT *stmt, int *nums, TAOS_FIELD_E **fields);
+int         stmtGetColFields(TAOS_STMT *stmt, int *nums, TAOS_FIELD_E **fields);
+int         stmtIsInsert(TAOS_STMT *stmt, int *insert);
+int         stmtGetParamNum(TAOS_STMT *stmt, int *nums);
+int         stmtGetParam(TAOS_STMT *stmt, int idx, int *type, int *bytes);
+int         stmtAddBatch(TAOS_STMT *stmt);
+TAOS_RES   *stmtUseResult(TAOS_STMT *stmt);
+int         stmtBindBatch(TAOS_STMT *stmt, TAOS_MULTI_BIND *bind, int32_t colIdx);
 
 #ifdef __cplusplus
 }

source/client/src/clientImpl.c

@@ -1036,30 +1036,38 @@ static int32_t asyncExecSchQuery(SRequestObj* pRequest, SQuery* pQuery, SMetaDat
   pRequest->type = pQuery->msgType;
 
   SArray*      pMnodeList = taosArrayInit(4, sizeof(SQueryNodeLoad));
-  SPlanContext cxt = {.queryId = pRequest->requestId,
-                      .acctId = pRequest->pTscObj->acctId,
-                      .mgmtEpSet = getEpSet_s(&pRequest->pTscObj->pAppInfo->mgmtEp),
-                      .pAstRoot = pQuery->pRoot,
-                      .showRewrite = pQuery->showRewrite,
-                      .pMsg = pRequest->msgBuf,
-                      .msgLen = ERROR_MSG_BUF_DEFAULT_SIZE,
-                      .pUser = pRequest->pTscObj->user,
-                      .sysInfo = pRequest->pTscObj->sysInfo,
-                      .allocatorId = pRequest->allocatorRefId};
-  SQueryPlan*  pDag = NULL;
-  int32_t      code = qCreateQueryPlan(&cxt, &pDag, pMnodeList);
-  if (code) {
-    tscError("0x%" PRIx64 " failed to create query plan, code:%s 0x%" PRIx64, pRequest->self, tstrerror(code),
-             pRequest->requestId);
-  } else {
-    pRequest->body.subplanNum = pDag->numOfSubplans;
-  }
 
-  pRequest->metric.planEnd = taosGetTimestampUs();
+  SPlanContext cxt = {.queryId = pRequest->requestId,
+                          .acctId = pRequest->pTscObj->acctId,
+                          .mgmtEpSet = getEpSet_s(&pRequest->pTscObj->pAppInfo->mgmtEp),
+                          .pAstRoot = pQuery->pRoot,
+                          .showRewrite = pQuery->showRewrite,
+                          .pMsg = pRequest->msgBuf,
+                          .msgLen = ERROR_MSG_BUF_DEFAULT_SIZE,
+                          .pUser = pRequest->pTscObj->user,
+                          .sysInfo = pRequest->pTscObj->sysInfo,
+                          .allocatorId = pRequest->allocatorRefId};
+
+      SAppInstInfo* pAppInfo = getAppInfo(pRequest);
+      SQueryPlan*   pDag = NULL;
+
+  int64_t st = taosGetTimestampUs();
+  int32_t code = qCreateQueryPlan(&cxt, &pDag, pMnodeList);
+      if (code) {
+        tscError("0x%" PRIx64 " failed to create query plan, code:%s 0x%" PRIx64, pRequest->self, tstrerror(code),
+                 pRequest->requestId);
+      } else {
+        pRequest->body.subplanNum = pDag->numOfSubplans;
+      }
 
-  if (TSDB_CODE_SUCCESS == code && !pRequest->validateOnly) {
-    SArray* pNodeList = NULL;
-    buildAsyncExecNodeList(pRequest, &pNodeList, pMnodeList, pResultMeta);
+      pRequest->metric.planEnd = taosGetTimestampUs();
+  if (code == TSDB_CODE_SUCCESS) {
+    tscDebug("0x%" PRIx64 " create query plan success, elapsed time:%.2f ms, 0x%" PRIx64, pRequest->self,
+             (pRequest->metric.planEnd - st)/1000.0, pRequest->requestId);
+  }
+      if (TSDB_CODE_SUCCESS == code && !pRequest->validateOnly) {
+        SArray* pNodeList = NULL;
+        buildAsyncExecNodeList(pRequest, &pNodeList, pMnodeList, pResultMeta);
 
     SRequestConnInfo conn = {.pTrans = getAppInfo(pRequest)->pTransporter,
                              .requestId = pRequest->requestId,

source/client/src/clientMain.c

@@ -146,7 +146,7 @@ void taos_close(TAOS *taos) {
 
 int taos_errno(TAOS_RES *res) {
   if (res == NULL || TD_RES_TMQ_META(res)) {
-    if (terrno == TSDB_CODE_RPC_REDIRECT) terrno = TSDB_CODE_RPC_NETWORK_UNAVAIL;
+    if (terrno == TSDB_CODE_RPC_REDIRECT) terrno = TSDB_CODE_QRY_NOT_READY;
     return terrno;
   }
 
@@ -154,13 +154,12 @@ int taos_errno(TAOS_RES *res) {
     return 0;
   }
 
-  return ((SRequestObj *)res)->code == TSDB_CODE_RPC_REDIRECT ? TSDB_CODE_RPC_NETWORK_UNAVAIL
-                                                              : ((SRequestObj *)res)->code;
+  return ((SRequestObj *)res)->code == TSDB_CODE_RPC_REDIRECT ? TSDB_CODE_QRY_NOT_READY : ((SRequestObj *)res)->code;
 }
 
 const char *taos_errstr(TAOS_RES *res) {
   if (res == NULL || TD_RES_TMQ_META(res)) {
-    if (terrno == TSDB_CODE_RPC_REDIRECT) terrno = TSDB_CODE_RPC_NETWORK_UNAVAIL;
+    if (terrno == TSDB_CODE_RPC_REDIRECT) terrno = TSDB_CODE_QRY_NOT_READY;
     return (const char *)tstrerror(terrno);
   }
 
@@ -172,7 +171,7 @@ const char *taos_errstr(TAOS_RES *res) {
   if (NULL != pRequest->msgBuf && (strlen(pRequest->msgBuf) > 0 || pRequest->code == TSDB_CODE_RPC_FQDN_ERROR)) {
     return pRequest->msgBuf;
   } else {
-    return pRequest->code == TSDB_CODE_RPC_REDIRECT ? (const char *)tstrerror(TSDB_CODE_RPC_NETWORK_UNAVAIL)
+    return pRequest->code == TSDB_CODE_RPC_REDIRECT ? (const char *)tstrerror(TSDB_CODE_QRY_NOT_READY)
                                                     : (const char *)tstrerror(pRequest->code);
   }
 }
@@ -222,7 +221,9 @@ void taos_kill_query(TAOS *taos) {
 
   int64_t  rid = *(int64_t *)taos;
   STscObj *pTscObj = acquireTscObj(rid);
-  stopAllRequests(pTscObj->pRequests);
+  if (pTscObj) {
+    stopAllRequests(pTscObj->pRequests);
+  }
   releaseTscObj(rid);
 }
 

source/client/src/clientSml.c

@@ -2477,11 +2477,12 @@ static void smlInsertCallback(void *param, void *res, int32_t code) {
   } else {
     pParam->request->body.resInfo.numOfRows += info->affectedRows;
   }
+  // unlock
+  taosThreadSpinUnlock(&pParam->lock);
+
   if (pParam->cnt == pParam->total) {
     tsem_post(&pParam->sem);
   }
-  taosThreadSpinUnlock(&pParam->lock);
-  // unlock
   uDebug("SML:0x%" PRIx64 " insert finished, code: %d, rows: %d, total: %d", info->id, code, rows, info->affectedRows);
   info->cost.endTime = taosGetTimestampUs();
   info->cost.code = code;

source/client/src/clientStmt.c

@@ -7,8 +7,8 @@
 
 static int32_t stmtCreateRequest(STscStmt* pStmt) {
   int32_t code = 0;
-  
-  if (pStmt->exec.pRequest == NULL)  {
+
+  if (pStmt->exec.pRequest == NULL) {
     code = buildRequest(pStmt->taos->id, pStmt->sql.sqlStr, pStmt->sql.sqlLen, NULL, false, &pStmt->exec.pRequest);
     if (TSDB_CODE_SUCCESS == code) {
       pStmt->exec.pRequest->syncQuery = true;
@@ -232,7 +232,7 @@ int32_t stmtParseSql(STscStmt* pStmt) {
   };
 
   STMT_ERR_RET(stmtCreateRequest(pStmt));
-  
+
   STMT_ERR_RET(parseSql(pStmt->exec.pRequest, false, &pStmt->sql.pQuery, &stmtCb));
 
   pStmt->bInfo.needParse = false;
@@ -402,7 +402,7 @@ int32_t stmtGetFromCache(STscStmt* pStmt) {
         STMT_ERR_RET(TSDB_CODE_OUT_OF_MEMORY);
       }
 
-      tscDebug("reuse stmt block for tb %s in sqlBlock, suid:0x%" PRIx64 , pStmt->bInfo.tbFName, pStmt->bInfo.tbSuid);
+      tscDebug("reuse stmt block for tb %s in sqlBlock, suid:0x%" PRIx64, pStmt->bInfo.tbFName, pStmt->bInfo.tbSuid);
 
       return TSDB_CODE_SUCCESS;
     }
@@ -600,8 +600,9 @@ int stmtSetTbTags(TAOS_STMT* stmt, TAOS_MULTI_BIND* tags) {
   }
 
   tscDebug("start to bind stmt tag values");
-  STMT_ERR_RET(qBindStmtTagsValue(*pDataBlock, pStmt->bInfo.boundTags, pStmt->bInfo.tbSuid, pStmt->bInfo.stbFName, pStmt->bInfo.sname.tname,
-                                  tags, pStmt->exec.pRequest->msgBuf, pStmt->exec.pRequest->msgBufLen));
+  STMT_ERR_RET(qBindStmtTagsValue(*pDataBlock, pStmt->bInfo.boundTags, pStmt->bInfo.tbSuid, pStmt->bInfo.stbFName,
+                                  pStmt->bInfo.sname.tname, tags, pStmt->exec.pRequest->msgBuf,
+                                  pStmt->exec.pRequest->msgBufLen));
 
   return TSDB_CODE_SUCCESS;
 }

source/client/src/clientTmq.c

@@ -972,6 +972,8 @@ int32_t tmq_subscribe(tmq_t* tmq, const tmq_list_t* topic_list) {
   req.topicNames = taosArrayInit(sz, sizeof(void*));
   if (req.topicNames == NULL) goto FAIL;
 
+  tscDebug("call tmq subscribe, consumer: %ld, topic num %d", tmq->consumerId, sz);
+
   for (int32_t i = 0; i < sz; i++) {
     char* topic = taosArrayGetP(container, i);
 
@@ -1297,7 +1299,8 @@ int32_t tmqAskEpCb(void* param, SDataBuf* pMsg, int32_t code) {
 
   pParam->code = code;
   if (code != 0) {
-    tscError("consumer:%" PRId64 ", get topic endpoint error, not ready, wait:%d", tmq->consumerId, pParam->async);
+    tscError("consumer:%" PRId64 ", get topic endpoint error, not ready, wait:%d, code %x", tmq->consumerId,
+             pParam->async, code);
     goto END;
   }
 

source/client/src/clientTmqConnector.c

@@ -212,7 +212,8 @@ JNIEXPORT void JNICALL Java_com_taosdata_jdbc_tmq_TMQConnector_tmqCommitAsync(JN
   tmq_commit_async(tmq, res, commit_cb, consumer);
 }
 
-JNIEXPORT jint JNICALL Java_com_taosdata_jdbc_tmq_TMQConnector_tmqUnsubscribeImp(JNIEnv *env, jobject jobj, jlong jtmq) {
+JNIEXPORT jint JNICALL Java_com_taosdata_jdbc_tmq_TMQConnector_tmqUnsubscribeImp(JNIEnv *env, jobject jobj,
+                                                                                 jlong jtmq) {
   tmq_t *tmq = (tmq_t *)jtmq;
   if (tmq == NULL) {
     jniError("jobj:%p, tmq is closed", jobj);
@@ -223,7 +224,7 @@ JNIEXPORT jint JNICALL Java_com_taosdata_jdbc_tmq_TMQConnector_tmqUnsubscribeImp
 }
 
 JNIEXPORT jint JNICALL Java_com_taosdata_jdbc_tmq_TMQConnector_tmqConsumerCloseImp(JNIEnv *env, jobject jobj,
-                                                                                  jlong jtmq) {
+                                                                                   jlong jtmq) {
   tmq_t *tmq = (tmq_t *)jtmq;
   if (tmq == NULL) {
     jniDebug("jobj:%p, tmq is closed", jobj);

source/client/test/clientTests.cpp

@@ -13,12 +13,12 @@
  * along with this program. If not, see <http://www.gnu.org/licenses/>.
  */
 
-#include <iostream>
 #include <gtest/gtest.h>
+#include <iostream>
+#include "clientInt.h"
 #include "taoserror.h"
 #include "tglobal.h"
 #include "thash.h"
-#include "clientInt.h"
 
 #pragma GCC diagnostic push
 #pragma GCC diagnostic ignored "-Wwrite-strings"
@@ -26,8 +26,8 @@
 #pragma GCC diagnostic ignored "-Wunused-variable"
 #pragma GCC diagnostic ignored "-Wsign-compare"
 
-#include "taos.h"
 #include "executor.h"
+#include "taos.h"
 
 namespace {
 void showDB(TAOS* pConn) {
@@ -50,17 +50,17 @@ void printResult(TAOS_RES* pRes) {
   int32_t     numOfFields = taos_num_fields(pRes);
 
   int32_t n = 0;
-  char str[512] = {0};
+  char    str[512] = {0};
   while ((pRow = taos_fetch_row(pRes)) != NULL) {
-    int32_t* length = taos_fetch_lengths(pRes);
-    for(int32_t i = 0; i < numOfFields; ++i) {
-      printf("(%d):%d " , i, length[i]);
-    }
-    printf("\n");
-
-    int32_t code = taos_print_row(str, pRow, pFields, numOfFields);
-    printf("%s\n", str);
-    memset(str, 0, sizeof(str));
+//    int32_t* length = taos_fetch_lengths(pRes);
+//    for(int32_t i = 0; i < numOfFields; ++i) {
+//      printf("(%d):%d " , i, length[i]);
+//    }
+//    printf("\n");
+//
+//    int32_t code = taos_print_row(str, pRow, pFields, numOfFields);
+//    printf("%s\n", str);
+//    memset(str, 0, sizeof(str));
   }
 }
 
@@ -102,17 +102,6 @@ void queryCallback(void* param, void* res, int32_t code) {
   taos_fetch_raw_block_a(res, fetchCallback, param);
 }
 
-void queryCallback1(void* param, void* res, int32_t code) {
-  if (code != TSDB_CODE_SUCCESS) {
-    printf("failed to execute, reason:%s\n", taos_errstr(res));
-  }
-
-  taos_free_result(res);
-
-  printf("exec query:\n");
-  taos_query_a(param, "select * from tm1", queryCallback, param);
-}
-
 void createNewTable(TAOS* pConn, int32_t index) {
   char str[1024] = {0};
   sprintf(str, "create table tu%d using st2 tags(%d)", index, index);
@@ -123,16 +112,16 @@ void createNewTable(TAOS* pConn, int32_t index) {
   }
   taos_free_result(pRes);
 
-  for(int32_t i = 0; i < 3280; i += 20) {
+  for(int32_t i = 0; i < 10000; 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)"
             "(now+%da, %d)(now+%da, %d)(now+%da, %d)(now+%da, %d)(now+%da, %d)(now+%da, %d)"
             "(now+%da, %d)(now+%da, %d)(now+%da, %d)(now+%da, %d)"
-            "(now+%da, %d)(now+%da, %d)(now+%da, %d)(now+%da, %d)(now+%da, %d)(now+%da, %d)", index,
-            i, i, i + 1, i + 1, i + 2, i + 2, i + 3, i + 3, i + 4, i + 4, i + 5, i + 5, i + 6, i + 6, i + 7, i + 7,
-            i + 8, i + 8, i + 9, i + 9, i + 10, i + 10, i + 11, i + 11, i + 12, i + 12, i + 13, i + 13, i + 14, i + 14,
-            i + 15, i + 15, i + 16, i + 16, i + 17, i + 17, i + 18, i + 18, i + 19, i + 19);
+            "(now+%da, %d)(now+%da, %d)(now+%da, %d)(now+%da, %d)(now+%da, %d)(now+%da, %d)",
+            index, i, i, i + 1, i + 1, i + 2, i + 2, i + 3, i + 3, i + 4, i + 4, i + 5, i + 5, i + 6, i + 6, i + 7,
+            i + 7, i + 8, i + 8, i + 9, i + 9, i + 10, i + 10, i + 11, i + 11, i + 12, i + 12, i + 13, i + 13, i + 14,
+            i + 14, i + 15, i + 15, i + 16, i + 16, i + 17, i + 17, i + 18, i + 18, i + 19, i + 19);
     TAOS_RES* p = taos_query(pConn, sql);
     if (taos_errno(p) != 0) {
       printf("failed to insert data, reason:%s\n", taos_errstr(p));
@@ -141,16 +130,55 @@ void createNewTable(TAOS* pConn, int32_t index) {
     taos_free_result(p);
   }
 }
+
+void *queryThread(void *arg) {
+  TAOS* pConn = taos_connect("192.168.0.209", "root", "taosdata", NULL, 0);
+  if (pConn == NULL) {
+    printf("failed to connect to db, reason:%s", taos_errstr(pConn));
+    return NULL;
+  }
+
+  int64_t el = 0;
+
+  for (int32_t i = 0; i < 5000000; ++i) {
+    int64_t st = taosGetTimestampUs();
+    TAOS_RES* pRes = taos_query(pConn,
+                      "SELECT _wstart as ts,max(usage_user) FROM benchmarkcpu.host_49 WHERE  ts >= 1451618560000 AND ts < 1451622160000 INTERVAL(1m) ;");
+    if (taos_errno(pRes) != 0) {
+      printf("failed, reason:%s\n", taos_errstr(pRes));
+    } else {
+      printResult(pRes);
+    }
+
+    taos_free_result(pRes);
+    el += (taosGetTimestampUs() - st);
+    if (i % 1000 == 0 && i != 0) {
+      printf("total:%d, avg time:%.2fms\n", i, el/(double)(i*1000));
+    }
+  }
+
+  taos_close(pConn);
+  return NULL;
+}
+
+static int32_t numOfThreads = 1;
 }  // namespace
 
 int main(int argc, char** argv) {
   testing::InitGoogleTest(&argc, argv);
+  if (argc > 1) {
+    numOfThreads = atoi(argv[1]);
+  }
+
+  numOfThreads = TMAX(numOfThreads, 1);
+  printf("the runing threads is:%d", numOfThreads);
+
   return RUN_ALL_TESTS();
 }
 
 TEST(testCase, driverInit_Test) {
   // taosInitGlobalCfg();
-//  taos_init();
+  //  taos_init();
 }
 
 TEST(testCase, connect_Test) {
@@ -664,25 +692,25 @@ TEST(testCase, insert_test) {
   taos_free_result(pRes);
   taos_close(pConn);
 }
-#endif
 
 TEST(testCase, projection_query_tables) {
   TAOS* pConn = taos_connect("localhost", "root", "taosdata", NULL, 0);
   ASSERT_NE(pConn, nullptr);
 
-//  TAOS_RES* pRes = taos_query(pConn, "create database if not exists abc1 vgroups 1");
-//  if (taos_errno(pRes) != 0) {
-//    printf("error in create db, reason:%s\n", taos_errstr(pRes));
-//  }
-//  taos_free_result(pRes);
+  //  TAOS_RES* pRes = taos_query(pConn, "create database if not exists abc1 vgroups 1");
+  //  if (taos_errno(pRes) != 0) {
+  //    printf("error in create db, reason:%s\n", taos_errstr(pRes));
+  //  }
+  //  taos_free_result(pRes);
 
-  TAOS_RES* pRes = taos_query(pConn, "use abc1");
+  TAOS_RES* pRes = taos_query(pConn, "use benchmarkcpu");
   taos_free_result(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)");
@@ -697,31 +725,41 @@ TEST(testCase, projection_query_tables) {
   }
   taos_free_result(pRes);
 
-  for(int32_t i = 0; i < 2; ++i) {
+  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) {
-//    printf("failed to select from table, reason:%s\n", taos_errstr(pRes));
-//    taos_free_result(pRes);
-//    ASSERT_TRUE(false);
-//  }
-//
-//  TAOS_ROW    pRow = NULL;
-//  TAOS_FIELD* pFields = taos_fetch_fields(pRes);
-//  int32_t     numOfFields = taos_num_fields(pRes);
-//
-//  char str[512] = {0};
-//  while ((pRow = taos_fetch_row(pRes)) != NULL) {
-//    int32_t code = taos_print_row(str, pRow, pFields, numOfFields);
-//    printf("%s\n", str);
-//  }
+  //
+  //  pRes = taos_query(pConn, "select * from tu");
+  //  if (taos_errno(pRes) != 0) {
+  //    printf("failed to select from table, reason:%s\n", taos_errstr(pRes));
+  //    taos_free_result(pRes);
+  //    ASSERT_TRUE(false);
+  //  }
+  //
+  //  TAOS_ROW    pRow = NULL;
+  //  TAOS_FIELD* pFields = taos_fetch_fields(pRes);
+  //  int32_t     numOfFields = taos_num_fields(pRes);
+  //
+  //  char str[512] = {0};
+  //  while ((pRow = taos_fetch_row(pRes)) != NULL) {
+  //    int32_t code = taos_print_row(str, pRow, pFields, numOfFields);
+  //    printf("%s\n", str);
+  //  }
 
   taos_free_result(pRes);
   taos_close(pConn);
 }
+#endif
+
+TEST(testCase, tsbs_perf_test) {
+  TdThread qid[20] = {0};
+
+  for(int32_t i = 0; i < numOfThreads; ++i) {
+    taosThreadCreate(&qid[i], NULL, queryThread, NULL);
+  }
+  getchar();
+}
 
 #if 0
 TEST(testCase, projection_query_stables) {
@@ -860,5 +898,4 @@ TEST(testCase, update_test) {
 
 #endif
 
-
 #pragma GCC diagnostic pop

source/common/src/tglobal.c

@@ -90,6 +90,7 @@ bool tsSmlDataFormat = false;
 
 // query
 int32_t tsQueryPolicy = 1;
+int32_t tsQueryRspPolicy = 0;
 int32_t tsQuerySmaOptimize = 0;
 int32_t tsQueryRsmaTolerance = 1000;  // the tolerance time (ms) to judge from which level to query rsma data.
 bool    tsQueryPlannerTrace = false;
@@ -350,6 +351,7 @@ static int32_t taosAddServerCfg(SConfig *pCfg) {
   if (cfgAddInt32(pCfg, "countAlwaysReturnValue", tsCountAlwaysReturnValue, 0, 1, 0) != 0) return -1;
   if (cfgAddInt32(pCfg, "queryBufferSize", tsQueryBufferSize, -1, 500000000000, 0) != 0) return -1;
   if (cfgAddBool(pCfg, "printAuth", tsPrintAuth, 0) != 0) return -1;
+  if (cfgAddInt32(pCfg, "queryRspPolicy", tsQueryRspPolicy, 0, 1, 0) != 0) return -1;
 
   if (cfgAddInt32(pCfg, "multiProcess", tsMultiProcess, 0, 2, 0) != 0) return -1;
   if (cfgAddInt32(pCfg, "mnodeShmSize", tsMnodeShmSize, TSDB_MAX_MSG_SIZE * 2 + 1024, INT32_MAX, 0) != 0) return -1;
@@ -728,6 +730,7 @@ static int32_t taosSetServerCfg(SConfig *pCfg) {
   tsMonitorPort = (uint16_t)cfgGetItem(pCfg, "monitorPort")->i32;
   tsMonitorMaxLogs = cfgGetItem(pCfg, "monitorMaxLogs")->i32;
   tsMonitorComp = cfgGetItem(pCfg, "monitorComp")->bval;
+  tsQueryRspPolicy = cfgGetItem(pCfg, "queryRspPolicy")->i32;
 
   tsEnableTelem = cfgGetItem(pCfg, "telemetryReporting")->bval;
   tsTelemInterval = cfgGetItem(pCfg, "telemetryInterval")->i32;

source/dnode/mgmt/mgmt_mnode/inc/mmInt.h

@@ -41,8 +41,8 @@ typedef struct SMnodeMgmt {
 } SMnodeMgmt;
 
 // mmFile.c
-int32_t mmReadFile(SMnodeMgmt *pMgmt, SReplica *pReplica, bool *pDeployed);
-int32_t mmWriteFile(SMnodeMgmt *pMgmt, const SReplica *pReplica, bool deployed);
+int32_t mmReadFile(const char *path, SMnodeOpt *pOption);
+int32_t mmWriteFile(const char *path, const SMnodeOpt *pOption);
 
 // mmHandle.c
 SArray *mmGetMsgHandles();

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

@@ -16,7 +16,7 @@
 #define _DEFAULT_SOURCE
 #include "mmInt.h"
 
-int32_t mmReadFile(SMnodeMgmt *pMgmt, SReplica *pReplica, bool *pDeployed) {
+int32_t mmReadFile(const char *path, SMnodeOpt *pOption) {
   int32_t   code = TSDB_CODE_INVALID_JSON_FORMAT;
   int32_t   len = 0;
   int32_t   maxLen = 4096;
@@ -25,7 +25,7 @@ int32_t mmReadFile(SMnodeMgmt *pMgmt, SReplica *pReplica, bool *pDeployed) {
   char      file[PATH_MAX] = {0};
   TdFilePtr pFile = NULL;
 
-  snprintf(file, sizeof(file), "%s%smnode.json", pMgmt->path, TD_DIRSEP);
+  snprintf(file, sizeof(file), "%s%smnode.json", path, TD_DIRSEP);
   pFile = taosOpenFile(file, TD_FILE_READ);
   if (pFile == NULL) {
     code = 0;
@@ -50,38 +50,69 @@ int32_t mmReadFile(SMnodeMgmt *pMgmt, SReplica *pReplica, bool *pDeployed) {
     dError("failed to read %s since deployed not found", file);
     goto _OVER;
   }
-  *pDeployed = deployed->valueint;
+  pOption->deploy = deployed->valueint;
 
-  cJSON *id = cJSON_GetObjectItem(root, "id");
-  if (id) {
-    if (id->type != cJSON_Number) {
-      dError("failed to read %s since id not found", file);
+  cJSON *selfIndex = cJSON_GetObjectItem(root, "selfIndex");
+  if (selfIndex) {
+    if (selfIndex->type != cJSON_Number) {
+      dError("failed to read %s since selfIndex not found", file);
       goto _OVER;
     }
-    if (pReplica) {
-      pReplica->id = id->valueint;
-    }
+    pOption->selfIndex = selfIndex->valueint;
   }
 
-  cJSON *fqdn = cJSON_GetObjectItem(root, "fqdn");
-  if (fqdn) {
-    if (fqdn->type != cJSON_String || fqdn->valuestring == NULL) {
-      dError("failed to read %s since fqdn not found", file);
+  cJSON *replicas = cJSON_GetObjectItem(root, "replicas");
+  if (replicas) {
+    if (replicas->type != cJSON_Array) {
+      dError("failed to read %s since replicas not found", file);
       goto _OVER;
     }
-    if (pReplica) {
-      tstrncpy(pReplica->fqdn, fqdn->valuestring, TSDB_FQDN_LEN);
-    }
-  }
 
-  cJSON *port = cJSON_GetObjectItem(root, "port");
-  if (port) {
-    if (port->type != cJSON_Number) {
-      dError("failed to read %s since port not found", file);
+    int32_t numOfReplicas = cJSON_GetArraySize(replicas);
+    if (numOfReplicas <= 0) {
+      dError("failed to read %s since numOfReplicas:%d invalid", file, numOfReplicas);
       goto _OVER;
     }
-    if (pReplica) {
-      pReplica->port = (uint16_t)port->valueint;
+    pOption->numOfReplicas = numOfReplicas;
+
+    for (int32_t i = 0; i < numOfReplicas; ++i) {
+      SReplica *pReplica = pOption->replicas + i;
+
+      cJSON *replica = cJSON_GetArrayItem(replicas, i);
+      if (replica == NULL) break;
+
+      cJSON *id = cJSON_GetObjectItem(replica, "id");
+      if (id) {
+        if (id->type != cJSON_Number) {
+          dError("failed to read %s since id not found", file);
+          goto _OVER;
+        }
+        if (pReplica) {
+          pReplica->id = id->valueint;
+        }
+      }
+
+      cJSON *fqdn = cJSON_GetObjectItem(replica, "fqdn");
+      if (fqdn) {
+        if (fqdn->type != cJSON_String || fqdn->valuestring == NULL) {
+          dError("failed to read %s since fqdn not found", file);
+          goto _OVER;
+        }
+        if (pReplica) {
+          tstrncpy(pReplica->fqdn, fqdn->valuestring, TSDB_FQDN_LEN);
+        }
+      }
+
+      cJSON *port = cJSON_GetObjectItem(replica, "port");
+      if (port) {
+        if (port->type != cJSON_Number) {
+          dError("failed to read %s since port not found", file);
+          goto _OVER;
+        }
+        if (pReplica) {
+          pReplica->port = (uint16_t)port->valueint;
+        }
+      }
     }
   }
 
@@ -92,18 +123,18 @@ _OVER:
   if (root != NULL) cJSON_Delete(root);
   if (pFile != NULL) taosCloseFile(&pFile);
   if (code == 0) {
-    dDebug("succcessed to read file %s, deployed:%d", file, *pDeployed);
+    dDebug("succcessed to read file %s, deployed:%d", file, pOption->deploy);
   }
 
   terrno = code;
   return code;
 }
 
-int32_t mmWriteFile(SMnodeMgmt *pMgmt, const SReplica *pReplica, bool deployed) {
+int32_t mmWriteFile(const char *path, const SMnodeOpt *pOption) {
   char file[PATH_MAX] = {0};
   char realfile[PATH_MAX] = {0};
-  snprintf(file, sizeof(file), "%s%smnode.json.bak", pMgmt->path, TD_DIRSEP);
-  snprintf(realfile, sizeof(realfile), "%s%smnode.json", pMgmt->path, TD_DIRSEP);
+  snprintf(file, sizeof(file), "%s%smnode.json.bak", path, TD_DIRSEP);
+  snprintf(realfile, sizeof(realfile), "%s%smnode.json", path, TD_DIRSEP);
 
   TdFilePtr pFile = taosOpenFile(file, TD_FILE_CREATE | TD_FILE_WRITE | TD_FILE_TRUNC);
   if (pFile == NULL) {
@@ -117,12 +148,25 @@ int32_t mmWriteFile(SMnodeMgmt *pMgmt, const SReplica *pReplica, bool deployed)
   char   *content = taosMemoryCalloc(1, maxLen + 1);
 
   len += snprintf(content + len, maxLen - len, "{\n");
-  if (pReplica != NULL && pReplica->id > 0) {
-    len += snprintf(content + len, maxLen - len, "  \"id\": %d,\n", pReplica->id);
-    len += snprintf(content + len, maxLen - len, "  \"fqdn\": \"%s\",\n", pReplica->fqdn);
-    len += snprintf(content + len, maxLen - len, "  \"port\": %u\n,", pReplica->port);
+  if (pOption->deploy && pOption->numOfReplicas > 0) {
+    len += snprintf(content + len, maxLen - len, "  \"selfIndex\": %d,\n", pOption->selfIndex);
+    len += snprintf(content + len, maxLen - len, "  \"replicas\": [{\n");
+
+    for (int32_t i = 0; i < pOption->numOfReplicas; ++i) {
+      const SReplica *pReplica = pOption->replicas + i;
+      if (pReplica != NULL && pReplica->id > 0) {
+        len += snprintf(content + len, maxLen - len, "    \"id\": %d,\n", pReplica->id);
+        len += snprintf(content + len, maxLen - len, "    \"fqdn\": \"%s\",\n", pReplica->fqdn);
+        len += snprintf(content + len, maxLen - len, "    \"port\": %u\n", pReplica->port);
+      }
+      if (i < pOption->numOfReplicas - 1) {
+        len += snprintf(content + len, maxLen - len, "  },{\n");
+      } else {
+        len += snprintf(content + len, maxLen - len, "  }],\n");
+      }
+    }
   }
-  len += snprintf(content + len, maxLen - len, "  \"deployed\": %d\n", deployed);
+  len += snprintf(content + len, maxLen - len, "  \"deployed\": %d\n", pOption->deploy);
   len += snprintf(content + len, maxLen - len, "}\n");
 
   taosWriteFile(pFile, content, len);
@@ -136,6 +180,6 @@ int32_t mmWriteFile(SMnodeMgmt *pMgmt, const SReplica *pReplica, bool deployed)
     return -1;
   }
 
-  dDebug("successed to write %s, deployed:%d", realfile, deployed);
+  dDebug("successed to write %s, deployed:%d", realfile, pOption->deploy);
   return 0;
 }

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

@@ -80,18 +80,21 @@ int32_t mmProcessCreateReq(const SMgmtInputOpt *pInput, SRpcMsg *pMsg) {
     return -1;
   }
 
-  if (createReq.replica != 1) {
+  SMnodeOpt option = {.deploy = true, .numOfReplicas = createReq.replica, .selfIndex = -1};
+  memcpy(option.replicas, createReq.replicas, sizeof(createReq.replicas));
+  for (int32_t i = 0; i < option.numOfReplicas; ++i) {
+    if (createReq.replicas[i].id == pInput->pData->dnodeId) {
+      option.selfIndex = i;
+    }
+  }
+
+  if (option.selfIndex == -1) {
     terrno = TSDB_CODE_INVALID_OPTION;
-    dGError("failed to create mnode since %s", terrstr());
+    dGError("failed to create mnode since %s, selfIndex is -1", terrstr());
     return -1;
   }
 
-  bool deployed = true;
-
-  SMnodeMgmt mgmt = {0};
-  mgmt.path = pInput->path;
-  mgmt.name = pInput->name;
-  if (mmWriteFile(&mgmt, &createReq.replicas[0], deployed) != 0) {
+  if (mmWriteFile(pInput->path, &option) != 0) {
     dGError("failed to write mnode file since %s", terrstr());
     return -1;
   }
@@ -113,12 +116,8 @@ int32_t mmProcessDropReq(const SMgmtInputOpt *pInput, SRpcMsg *pMsg) {
     return -1;
   }
 
-  bool deployed = false;
-
-  SMnodeMgmt mgmt = {0};
-  mgmt.path = pInput->path;
-  mgmt.name = pInput->name;
-  if (mmWriteFile(&mgmt, NULL, deployed) != 0) {
+  SMnodeOpt option = {.deploy = false};
+  if (mmWriteFile(pInput->path, &option) != 0) {
     dGError("failed to write mnode file since %s", terrstr());
     return -1;
   }
@@ -207,7 +206,6 @@ SArray *mmGetMsgHandles() {
   if (dmSetMgmtHandle(pArray, TDMT_MND_HEARTBEAT, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_STATUS, mmPutMsgToReadQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_SYSTABLE_RETRIEVE, mmPutMsgToReadQueue, 0) == NULL) goto _OVER;
-  // if (dmSetMgmtHandle(pArray, TDMT_MND_GRANT, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_AUTH, mmPutMsgToReadQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_SHOW_VARIABLES, mmPutMsgToReadQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_SERVER_VERSION, mmPutMsgToReadQueue, 0) == NULL) goto _OVER;

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

@@ -25,38 +25,35 @@ static bool mmDeployRequired(const SMgmtInputOpt *pInput) {
 }
 
 static int32_t mmRequire(const SMgmtInputOpt *pInput, bool *required) {
-  SMnodeMgmt mgmt = {0};
-  mgmt.path = pInput->path;
-  if (mmReadFile(&mgmt, NULL, required) != 0) {
+  SMnodeOpt option = {0};
+  if (mmReadFile(pInput->path, &option) != 0) {
     return -1;
   }
 
-  if (!(*required)) {
+  if (!option.deploy) {
     *required = mmDeployRequired(pInput);
+  } else {
+    *required = true;
   }
 
   return 0;
 }
 
 static void mmBuildOptionForDeploy(SMnodeMgmt *pMgmt, const SMgmtInputOpt *pInput, SMnodeOpt *pOption) {
-  pOption->standby = false;
   pOption->deploy = true;
   pOption->msgCb = pMgmt->msgCb;
   pOption->dnodeId = pMgmt->pData->dnodeId;
-  pOption->replica.id = 1;
-  pOption->replica.port = tsServerPort;
-  tstrncpy(pOption->replica.fqdn, tsLocalFqdn, TSDB_FQDN_LEN);
+  pOption->selfIndex = 0;
+  pOption->numOfReplicas = 1;
+  pOption->replicas[0].id = 1;
+  pOption->replicas[0].port = tsServerPort;
+  tstrncpy(pOption->replicas[0].fqdn, tsLocalFqdn, TSDB_FQDN_LEN);
 }
 
-static void mmBuildOptionForOpen(SMnodeMgmt *pMgmt, const SReplica *pReplica, SMnodeOpt *pOption) {
-  pOption->standby = false;
+static void mmBuildOptionForOpen(SMnodeMgmt *pMgmt, SMnodeOpt *pOption) {
   pOption->deploy = false;
   pOption->msgCb = pMgmt->msgCb;
   pOption->dnodeId = pMgmt->pData->dnodeId;
-  if (pReplica->id > 0) {
-    pOption->standby = true;
-    pOption->replica = *pReplica;
-  }
 }
 
 static void mmClose(SMnodeMgmt *pMgmt) {
@@ -95,22 +92,20 @@ static int32_t mmOpen(SMgmtInputOpt *pInput, SMgmtOutputOpt *pOutput) {
   pMgmt->msgCb.mgmt = pMgmt;
   taosThreadRwlockInit(&pMgmt->lock, NULL);
 
-  bool     deployed = false;
-  SReplica replica = {0};
-  if (mmReadFile(pMgmt, &replica, &deployed) != 0) {
+  SMnodeOpt option = {0};
+  if (mmReadFile(pMgmt->path, &option) != 0) {
     dError("failed to read file since %s", terrstr());
     mmClose(pMgmt);
     return -1;
   }
 
-  SMnodeOpt option = {0};
-  if (!deployed) {
+  if (!option.deploy) {
     dInfo("mnode start to deploy");
     pMgmt->pData->dnodeId = 1;
     mmBuildOptionForDeploy(pMgmt, pInput, &option);
   } else {
     dInfo("mnode start to open");
-    mmBuildOptionForOpen(pMgmt, &replica, &option);
+    mmBuildOptionForOpen(pMgmt, &option);
   }
 
   pMgmt->pMnode = mndOpen(pMgmt->path, &option);
@@ -128,9 +123,10 @@ static int32_t mmOpen(SMgmtInputOpt *pInput, SMgmtOutputOpt *pOutput) {
   }
   tmsgReportStartup("mnode-worker", "initialized");
 
-  if (!deployed || replica.id > 0) {
-    deployed = true;
-    if (mmWriteFile(pMgmt, NULL, deployed) != 0) {
+  if (option.numOfReplicas > 0) {
+    option.deploy = true;
+    option.numOfReplicas = 0;
+    if (mmWriteFile(pMgmt->path, &option) != 0) {
       dError("failed to write mnode file since %s", terrstr());
       return -1;
     }

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

@@ -72,7 +72,7 @@ static void mmProcessRpcMsg(SQueueInfo *pInfo, SRpcMsg *pMsg) {
     mndPostProcessQueryMsg(pMsg);
   }
 
-  dGTrace("msg:%p, is freed, code:0x%x", pMsg, code);
+  dGTrace("msg:%p is freed, code:%s", pMsg, tstrerror(code));
   rpcFreeCont(pMsg->pCont);
   taosFreeQitem(pMsg);
 }

source/dnode/mgmt/node_mgmt/src/dmEnv.c

@@ -51,26 +51,14 @@ static int32_t dmInitMonitor() {
 
 static bool dmCheckDiskSpace() {
   osUpdate();
-  if (!osDataSpaceAvailable()) {
-    dError("free disk size: %f GB, too little, require %f GB at least at least , quit",
-           (double)tsDataSpace.size.avail / 1024.0 / 1024.0 / 1024.0,
-           (double)tsDataSpace.reserved / 1024.0 / 1024.0 / 1024.0);
-    terrno = TSDB_CODE_NO_AVAIL_DISK;
-    return false;
+  if (!osDataSpaceSufficient()) {
+    dWarn("free data disk size: %f GB, not sufficient, expected %f GB at least", (double)tsDataSpace.size.avail / 1024.0 / 1024.0 / 1024.0, (double)tsDataSpace.reserved / 1024.0 / 1024.0 / 1024.0);
   }
-  if (!osLogSpaceAvailable()) {
-    dError("free disk size: %f GB, too little, require %f GB at least at least, quit",
-           (double)tsLogSpace.size.avail / 1024.0 / 1024.0 / 1024.0,
-           (double)tsLogSpace.reserved / 1024.0 / 1024.0 / 1024.0);
-    terrno = TSDB_CODE_NO_AVAIL_DISK;
-    return false;
+  if (!osLogSpaceSufficient()) {
+    dWarn("free log disk size: %f GB, not sufficient, expected %f GB at least", (double)tsLogSpace.size.avail / 1024.0 / 1024.0 / 1024.0, (double)tsLogSpace.reserved / 1024.0 / 1024.0 / 1024.0);
   }
-  if (!osTempSpaceAvailable()) {
-    dError("free disk size: %f GB, too little, require %f GB at least at least, quit",
-           (double)tsTempSpace.size.avail / 1024.0 / 1024.0 / 1024.0,
-           (double)tsTempSpace.reserved / 1024.0 / 1024.0 / 1024.0);
-    terrno = TSDB_CODE_NO_AVAIL_DISK;
-    return false;
+  if (!osTempSpaceSufficient()) {
+    dWarn("free temp disk size: %f GB, not sufficient, expected %f GB at least", (double)tsTempSpace.size.avail / 1024.0 / 1024.0 / 1024.0, (double)tsTempSpace.reserved / 1024.0 / 1024.0 / 1024.0);
   }
   return true;
 }

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

@@ -87,12 +87,13 @@ typedef struct {
 typedef struct {
   tsem_t   syncSem;
   int64_t  sync;
-  SReplica replica;
   int32_t  errCode;
   int32_t  transId;
   SRWLatch lock;
-  int8_t   standby;
   int8_t   leaderTransferFinish;
+  int8_t   selfIndex;
+  int8_t   numOfReplicas;
+  SReplica replicas[TSDB_MAX_REPLICA];
 } SSyncMgmt;
 
 typedef struct {
@@ -130,11 +131,10 @@ typedef struct SMnode {
 void    mndSetMsgHandle(SMnode *pMnode, tmsg_t msgType, MndMsgFp fp);
 int64_t mndGenerateUid(const char *name, int32_t len);
 
-int32_t mndAcquireRpcRef(SMnode *pMnode);
-void    mndReleaseRpcRef(SMnode *pMnode);
-void    mndSetRestore(SMnode *pMnode, bool restored);
-void    mndSetStop(SMnode *pMnode);
-bool    mndGetStop(SMnode *pMnode);
+void mndSetRestored(SMnode *pMnode, bool restored);
+bool mndGetRestored(SMnode *pMnode);
+void mndSetStop(SMnode *pMnode);
+bool mndGetStop(SMnode *pMnode);
 
 #ifdef __cplusplus
 }

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

@@ -24,7 +24,7 @@ extern "C" {
 
 int32_t mndInitSync(SMnode *pMnode);
 void    mndCleanupSync(SMnode *pMnode);
-bool    mndIsMaster(SMnode *pMnode);
+bool    mndIsLeader(SMnode *pMnode);
 int32_t mndSyncPropose(SMnode *pMnode, SSdbRaw *pRaw, int32_t transId);
 void    mndSyncStart(SMnode *pMnode);
 void    mndSyncStop(SMnode *pMnode);

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

@@ -116,11 +116,13 @@ static int32_t mndProcessConsumerLostMsg(SRpcMsg *pMsg) {
   if (mndSetConsumerCommitLogs(pMnode, pTrans, pConsumerNew) != 0) goto FAIL;
   if (mndTransPrepare(pMnode, pTrans) != 0) goto FAIL;
 
-  mndTransDrop(pTrans);
   tDeleteSMqConsumerObj(pConsumerNew);
+  taosMemoryFree(pConsumerNew);
+  mndTransDrop(pTrans);
   return 0;
 FAIL:
   tDeleteSMqConsumerObj(pConsumerNew);
+  taosMemoryFree(pConsumerNew);
   mndTransDrop(pTrans);
   return -1;
 }

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

@@ -1809,7 +1809,7 @@ static int32_t mndRetrieveDbs(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pBloc
   }
 
   while (numOfRows < rowsCapacity) {
-    pShow->pIter = sdbFetchAll(pSdb, SDB_DB, pShow->pIter, (void **)&pDb, &objStatus);
+    pShow->pIter = sdbFetchAll(pSdb, SDB_DB, pShow->pIter, (void **)&pDb, &objStatus, true);
     if (pShow->pIter == NULL) break;
 
     if (mndCheckDbPrivilege(pMnode, pReq->info.conn.user, MND_OPER_READ_OR_WRITE_DB, pDb) == 0) {

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

@@ -423,7 +423,7 @@ static int32_t mndProcessStatusReq(SRpcMsg *pReq) {
         goto _OVER;
       } else {
         pDnode->accessTimes++;
-        mTrace("dnode:%d, status received, access times %d", pDnode->id, pDnode->accessTimes);
+        mDebug("dnode:%d, status received, access times %d", pDnode->id, pDnode->accessTimes);
       }
     }
 
@@ -471,6 +471,7 @@ static int32_t mndProcessStatusReq(SRpcMsg *pReq) {
   }
 
   pDnode->lastAccessTime = curMs;
+  pDnode->accessTimes++;
   code = 0;
 
 _OVER:

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

@@ -43,6 +43,37 @@
 #include "mndUser.h"
 #include "mndVgroup.h"
 
+static inline int32_t mndAcquireRpc(SMnode *pMnode) {
+  int32_t code = 0;
+  taosThreadRwlockRdlock(&pMnode->lock);
+  if (pMnode->stopped) {
+    terrno = TSDB_CODE_APP_NOT_READY;
+    code = -1;
+  } else if (!mndIsLeader(pMnode)) {
+    code = -1;
+  } else {
+#if 1
+    atomic_add_fetch_32(&pMnode->rpcRef, 1);
+#else
+    int32_t ref = atomic_add_fetch_32(&pMnode->rpcRef, 1);
+    mTrace("mnode rpc is acquired, ref:%d", ref);
+#endif
+  }
+  taosThreadRwlockUnlock(&pMnode->lock);
+  return code;
+}
+
+static inline void mndReleaseRpc(SMnode *pMnode) {
+  taosThreadRwlockRdlock(&pMnode->lock);
+#if 1
+  atomic_sub_fetch_32(&pMnode->rpcRef, 1);
+#else
+  int32_t ref = atomic_sub_fetch_32(&pMnode->rpcRef, 1);
+  mTrace("mnode rpc is released, ref:%d", ref);
+#endif
+  taosThreadRwlockUnlock(&pMnode->lock);
+}
+
 static void *mndBuildTimerMsg(int32_t *pContLen) {
   SMTimerReq timerReq = {0};
 
@@ -201,7 +232,7 @@ static int32_t mndInitWal(SMnode *pMnode) {
 
   pMnode->pWal = walOpen(path, &cfg);
   if (pMnode->pWal == NULL) {
-    mError("failed to open wal since %s", terrstr());
+    mError("failed to open wal since %s. wal:%s", terrstr(), path);
     return -1;
   }
 
@@ -338,8 +369,9 @@ static int32_t mndExecSteps(SMnode *pMnode) {
 static void mndSetOptions(SMnode *pMnode, const SMnodeOpt *pOption) {
   pMnode->msgCb = pOption->msgCb;
   pMnode->selfDnodeId = pOption->dnodeId;
-  pMnode->syncMgmt.replica = pOption->replica;
-  pMnode->syncMgmt.standby = pOption->standby;
+  pMnode->syncMgmt.selfIndex = pOption->selfIndex;
+  pMnode->syncMgmt.numOfReplicas = pOption->numOfReplicas;
+  memcpy(pMnode->syncMgmt.replicas, pOption->replicas, sizeof(pOption->replicas));
 }
 
 SMnode *mndOpen(const char *path, const SMnodeOpt *pOption) {
@@ -430,7 +462,7 @@ int32_t mndStart(SMnode *pMnode) {
       mError("failed to deploy sdb while start mnode");
       return -1;
     }
-    mndSetRestore(pMnode, true);
+    mndSetRestored(pMnode, true);
   }
 
   grantReset(pMnode, TSDB_GRANT_ALL, 0);
@@ -570,23 +602,27 @@ static int32_t mndCheckMnodeState(SRpcMsg *pMsg) {
       pMsg->msgType == TDMT_SCH_FETCH || pMsg->msgType == TDMT_SCH_MERGE_FETCH || pMsg->msgType == TDMT_SCH_DROP_TASK) {
     return 0;
   }
-  if (mndAcquireRpcRef(pMsg->info.node) == 0) return 0;
+  if (mndAcquireRpc(pMsg->info.node) == 0) return 0;
   if (pMsg->msgType == TDMT_MND_MQ_TIMER || pMsg->msgType == TDMT_MND_TELEM_TIMER ||
       pMsg->msgType == TDMT_MND_TRANS_TIMER || pMsg->msgType == TDMT_MND_TTL_TIMER ||
       pMsg->msgType == TDMT_MND_UPTIME_TIMER) {
     return -1;
   }
 
-  SEpSet epSet = {0};
-  mndGetMnodeEpSet(pMsg->info.node, &epSet);
+  SEpSet  epSet = {0};
+  SMnode *pMnode = pMsg->info.node;
+  mndGetMnodeEpSet(pMnode, &epSet);
 
   const STraceId *trace = &pMsg->info.traceId;
-  mError("msg:%p, failed to check mnode state since %s, type:%s, numOfMnodes:%d inUse:%d", pMsg, terrstr(),
-         TMSG_INFO(pMsg->msgType), epSet.numOfEps, epSet.inUse);
+  mDebug(
+      "msg:%p, failed to check mnode state since %s, mnode restored:%d stopped:%d, sync restored:%d role:%s type:%s "
+      "numOfEps:%d inUse:%d",
+      pMsg, terrstr(), pMnode->restored, pMnode->stopped, syncIsRestoreFinish(pMnode->syncMgmt.sync),
+      syncGetMyRoleStr(pMnode->syncMgmt.sync), TMSG_INFO(pMsg->msgType), epSet.numOfEps, epSet.inUse);
 
   if (epSet.numOfEps > 0) {
     for (int32_t i = 0; i < epSet.numOfEps; ++i) {
-      mInfo("mnode index:%d, ep:%s:%u", i, epSet.eps[i].fqdn, epSet.eps[i].port);
+      mDebug("mnode index:%d, ep:%s:%u", i, epSet.eps[i].fqdn, epSet.eps[i].port);
     }
 
     int32_t contLen = tSerializeSEpSet(NULL, 0, &epSet);
@@ -633,7 +669,7 @@ int32_t mndProcessRpcMsg(SRpcMsg *pMsg) {
 
   mGTrace("msg:%p, start to process in mnode, app:%p type:%s", pMsg, pMsg->info.ahandle, TMSG_INFO(pMsg->msgType));
   int32_t code = (*fp)(pMsg);
-  mndReleaseRpcRef(pMnode);
+  mndReleaseRpc(pMnode);
 
   if (code == TSDB_CODE_ACTION_IN_PROGRESS) {
     mGTrace("msg:%p, won't response immediately since in progress", pMsg);
@@ -669,7 +705,7 @@ int64_t mndGenerateUid(const char *name, int32_t len) {
 
 int32_t mndGetMonitorInfo(SMnode *pMnode, SMonClusterInfo *pClusterInfo, SMonVgroupInfo *pVgroupInfo,
                           SMonStbInfo *pStbInfo, SMonGrantInfo *pGrantInfo) {
-  if (mndAcquireRpcRef(pMnode) != 0) return -1;
+  if (mndAcquireRpc(pMnode) != 0) return -1;
 
   SSdb   *pSdb = pMnode->pSdb;
   int64_t ms = taosGetTimestampMs();
@@ -680,7 +716,7 @@ int32_t mndGetMonitorInfo(SMnode *pMnode, SMonClusterInfo *pClusterInfo, SMonVgr
   pStbInfo->stbs = taosArrayInit(sdbGetSize(pSdb, SDB_STB), sizeof(SMonStbDesc));
   if (pClusterInfo->dnodes == NULL || pClusterInfo->mnodes == NULL || pVgroupInfo->vgroups == NULL ||
       pStbInfo->stbs == NULL) {
-    mndReleaseRpcRef(pMnode);
+    mndReleaseRpc(pMnode);
     return -1;
   }
 
@@ -800,7 +836,7 @@ int32_t mndGetMonitorInfo(SMnode *pMnode, SMonClusterInfo *pClusterInfo, SMonVgr
     pGrantInfo->timeseries_total = INT32_MAX;
   }
 
-  mndReleaseRpcRef(pMnode);
+  mndReleaseRpc(pMnode);
   return 0;
 }
 
@@ -810,32 +846,7 @@ int32_t mndGetLoad(SMnode *pMnode, SMnodeLoad *pLoad) {
   return 0;
 }
 
-int32_t mndAcquireRpcRef(SMnode *pMnode) {
-  int32_t code = 0;
-  taosThreadRwlockRdlock(&pMnode->lock);
-  if (pMnode->stopped) {
-    mTrace("mnode not running");
-    terrno = TSDB_CODE_APP_NOT_READY;
-    code = -1;
-  } else if (!mndIsMaster(pMnode)) {
-    mTrace("mnode not ready, role:%s restored:%d", syncGetMyRoleStr(pMnode->syncMgmt.sync), pMnode->restored);
-    code = -1;
-  } else {
-    int32_t ref = atomic_add_fetch_32(&pMnode->rpcRef, 1);
-    // mTrace("mnode rpc is acquired, ref:%d", ref);
-  }
-  taosThreadRwlockUnlock(&pMnode->lock);
-  return code;
-}
-
-void mndReleaseRpcRef(SMnode *pMnode) {
-  taosThreadRwlockRdlock(&pMnode->lock);
-  int32_t ref = atomic_sub_fetch_32(&pMnode->rpcRef, 1);
-  // mTrace("mnode rpc is released, ref:%d", ref);
-  taosThreadRwlockUnlock(&pMnode->lock);
-}
-
-void mndSetRestore(SMnode *pMnode, bool restored) {
+void mndSetRestored(SMnode *pMnode, bool restored) {
   if (restored) {
     taosThreadRwlockWrlock(&pMnode->lock);
     pMnode->restored = true;

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

@@ -36,6 +36,7 @@ static int32_t  mndProcessAlterMnodeReq(SRpcMsg *pReq);
 static int32_t  mndProcessDropMnodeReq(SRpcMsg *pReq);
 static int32_t  mndRetrieveMnodes(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pBlock, int32_t rows);
 static void     mndCancelGetNextMnode(SMnode *pMnode, void *pIter);
+static void     mndReloadSyncConfig(SMnode *pMnode);
 
 int32_t mndInitMnode(SMnode *pMnode) {
   SSdbTable table = {
@@ -187,6 +188,7 @@ static int32_t mndMnodeActionInsert(SSdb *pSdb, SMnodeObj *pObj) {
   }
 
   pObj->state = TAOS_SYNC_STATE_ERROR;
+  mndReloadSyncConfig(pSdb->pMnode);
   return 0;
 }
 
@@ -203,6 +205,8 @@ static int32_t mndMnodeActionDelete(SSdb *pSdb, SMnodeObj *pObj) {
 static int32_t mndMnodeActionUpdate(SSdb *pSdb, SMnodeObj *pOld, SMnodeObj *pNew) {
   mTrace("mnode:%d, perform update action, old row:%p new row:%p", pOld->id, pOld, pNew);
   pOld->updateTime = pNew->updateTime;
+  mndReloadSyncConfig(pSdb->pMnode);
+
   return 0;
 }
 
@@ -233,7 +237,7 @@ void mndGetMnodeEpSet(SMnode *pMnode, SEpSet *pEpSet) {
     if (pIter == NULL) break;
 
     if (pObj->id == pMnode->selfDnodeId) {
-      if (mndIsMaster(pMnode)) {
+      if (mndIsLeader(pMnode)) {
         pEpSet->inUse = pEpSet->numOfEps;
       } else {
         pEpSet->inUse = (pEpSet->numOfEps + 1) % totalMnodes;
@@ -248,6 +252,10 @@ void mndGetMnodeEpSet(SMnode *pMnode, SEpSet *pEpSet) {
   if (pEpSet->numOfEps == 0) {
     syncGetRetryEpSet(pMnode->syncMgmt.sync, pEpSet);
   }
+
+  if (pEpSet->inUse >= pEpSet->numOfEps) {
+    pEpSet->inUse = 0;
+  }
 }
 
 static int32_t mndSetCreateMnodeRedoLogs(SMnode *pMnode, STrans *pTrans, SMnodeObj *pObj) {
@@ -274,13 +282,72 @@ static int32_t mndSetCreateMnodeCommitLogs(SMnode *pMnode, STrans *pTrans, SMnod
   return 0;
 }
 
+static int32_t mndBuildCreateMnodeRedoAction(STrans *pTrans, SDCreateMnodeReq *pCreateReq, SEpSet *pCreateEpSet) {
+  int32_t contLen = tSerializeSDCreateMnodeReq(NULL, 0, pCreateReq);
+  void   *pReq = taosMemoryMalloc(contLen);
+  tSerializeSDCreateMnodeReq(pReq, contLen, pCreateReq);
+
+  STransAction action = {
+      .epSet = *pCreateEpSet,
+      .pCont = pReq,
+      .contLen = contLen,
+      .msgType = TDMT_DND_CREATE_MNODE,
+      .acceptableCode = TSDB_CODE_NODE_ALREADY_DEPLOYED,
+  };
+
+  if (mndTransAppendRedoAction(pTrans, &action) != 0) {
+    taosMemoryFree(pReq);
+    return -1;
+  }
+  return 0;
+}
+
+static int32_t mndBuildAlterMnodeRedoAction(STrans *pTrans, SDCreateMnodeReq *pAlterReq, SEpSet *pAlterEpSet) {
+  int32_t contLen = tSerializeSDCreateMnodeReq(NULL, 0, pAlterReq);
+  void   *pReq = taosMemoryMalloc(contLen);
+  tSerializeSDCreateMnodeReq(pReq, contLen, pAlterReq);
+
+  STransAction action = {
+      .epSet = *pAlterEpSet,
+      .pCont = pReq,
+      .contLen = contLen,
+      .msgType = TDMT_MND_ALTER_MNODE,
+      .acceptableCode = 0,
+  };
+
+  if (mndTransAppendRedoAction(pTrans, &action) != 0) {
+    taosMemoryFree(pReq);
+    return -1;
+  }
+
+  return 0;
+}
+
+static int32_t mndBuildDropMnodeRedoAction(STrans *pTrans, SDDropMnodeReq *pDropReq, SEpSet *pDroprEpSet) {
+  int32_t contLen = tSerializeSCreateDropMQSBNodeReq(NULL, 0, pDropReq);
+  void   *pReq = taosMemoryMalloc(contLen);
+  tSerializeSCreateDropMQSBNodeReq(pReq, contLen, pDropReq);
+
+  STransAction action = {
+      .epSet = *pDroprEpSet,
+      .pCont = pReq,
+      .contLen = contLen,
+      .msgType = TDMT_DND_DROP_MNODE,
+      .acceptableCode = TSDB_CODE_NODE_NOT_DEPLOYED,
+  };
+
+  if (mndTransAppendRedoAction(pTrans, &action) != 0) {
+    taosMemoryFree(pReq);
+    return -1;
+  }
+  return 0;
+}
+
 static int32_t mndSetCreateMnodeRedoActions(SMnode *pMnode, STrans *pTrans, SDnodeObj *pDnode, SMnodeObj *pObj) {
   SSdb            *pSdb = pMnode->pSdb;
   void            *pIter = NULL;
   int32_t          numOfReplicas = 0;
-  SDAlterMnodeReq  alterReq = {0};
   SDCreateMnodeReq createReq = {0};
-  SEpSet           alterEpset = {0};
   SEpSet           createEpset = {0};
 
   while (1) {
@@ -288,75 +355,25 @@ static int32_t mndSetCreateMnodeRedoActions(SMnode *pMnode, STrans *pTrans, SDno
     pIter = sdbFetch(pSdb, SDB_MNODE, pIter, (void **)&pMObj);
     if (pIter == NULL) break;
 
-    alterReq.replicas[numOfReplicas].id = pMObj->id;
-    alterReq.replicas[numOfReplicas].port = pMObj->pDnode->port;
-    memcpy(alterReq.replicas[numOfReplicas].fqdn, pMObj->pDnode->fqdn, TSDB_FQDN_LEN);
-
-    alterEpset.eps[numOfReplicas].port = pMObj->pDnode->port;
-    memcpy(alterEpset.eps[numOfReplicas].fqdn, pMObj->pDnode->fqdn, TSDB_FQDN_LEN);
-    if (pMObj->state == TAOS_SYNC_STATE_LEADER) {
-      alterEpset.inUse = numOfReplicas;
-    }
+    createReq.replicas[numOfReplicas].id = pMObj->id;
+    createReq.replicas[numOfReplicas].port = pMObj->pDnode->port;
+    memcpy(createReq.replicas[numOfReplicas].fqdn, pMObj->pDnode->fqdn, TSDB_FQDN_LEN);
 
     numOfReplicas++;
     sdbRelease(pSdb, pMObj);
   }
 
-  alterReq.replica = numOfReplicas + 1;
-  alterReq.replicas[numOfReplicas].id = pDnode->id;
-  alterReq.replicas[numOfReplicas].port = pDnode->port;
-  memcpy(alterReq.replicas[numOfReplicas].fqdn, pDnode->fqdn, TSDB_FQDN_LEN);
-
-  alterEpset.numOfEps = numOfReplicas + 1;
-  alterEpset.eps[numOfReplicas].port = pDnode->port;
-  memcpy(alterEpset.eps[numOfReplicas].fqdn, pDnode->fqdn, TSDB_FQDN_LEN);
-
-  createReq.replica = 1;
-  createReq.replicas[0].id = pDnode->id;
-  createReq.replicas[0].port = pDnode->port;
-  memcpy(createReq.replicas[0].fqdn, pDnode->fqdn, TSDB_FQDN_LEN);
+  createReq.replica = numOfReplicas + 1;
+  createReq.replicas[numOfReplicas].id = pDnode->id;
+  createReq.replicas[numOfReplicas].port = pDnode->port;
+  memcpy(createReq.replicas[numOfReplicas].fqdn, pDnode->fqdn, TSDB_FQDN_LEN);
 
+  createEpset.inUse = 0;
   createEpset.numOfEps = 1;
   createEpset.eps[0].port = pDnode->port;
   memcpy(createEpset.eps[0].fqdn, pDnode->fqdn, TSDB_FQDN_LEN);
 
-  {
-    int32_t contLen = tSerializeSDCreateMnodeReq(NULL, 0, &createReq);
-    void   *pReq = taosMemoryMalloc(contLen);
-    tSerializeSDCreateMnodeReq(pReq, contLen, &createReq);
-
-    STransAction action = {
-        .epSet = createEpset,
-        .pCont = pReq,
-        .contLen = contLen,
-        .msgType = TDMT_DND_CREATE_MNODE,
-        .acceptableCode = TSDB_CODE_NODE_ALREADY_DEPLOYED,
-    };
-
-    if (mndTransAppendRedoAction(pTrans, &action) != 0) {
-      taosMemoryFree(pReq);
-      return -1;
-    }
-  }
-
-  {
-    int32_t contLen = tSerializeSDCreateMnodeReq(NULL, 0, &alterReq);
-    void   *pReq = taosMemoryMalloc(contLen);
-    tSerializeSDCreateMnodeReq(pReq, contLen, &alterReq);
-
-    STransAction action = {
-        .epSet = alterEpset,
-        .pCont = pReq,
-        .contLen = contLen,
-        .msgType = TDMT_MND_ALTER_MNODE,
-        .acceptableCode = 0,
-    };
-
-    if (mndTransAppendRedoAction(pTrans, &action) != 0) {
-      taosMemoryFree(pReq);
-      return -1;
-    }
-  }
+  if (mndBuildCreateMnodeRedoAction(pTrans, &createReq, &createEpset) != 0) return -1;
 
   return 0;
 }
@@ -374,9 +391,9 @@ static int32_t mndCreateMnode(SMnode *pMnode, SRpcMsg *pReq, SDnodeObj *pDnode,
   mndTransSetSerial(pTrans);
   mInfo("trans:%d, used to create mnode:%d", pTrans->id, pCreate->dnodeId);
 
+  if (mndSetCreateMnodeRedoActions(pMnode, pTrans, pDnode, &mnodeObj) != 0) goto _OVER;
   if (mndSetCreateMnodeRedoLogs(pMnode, pTrans, &mnodeObj) != 0) goto _OVER;
   if (mndSetCreateMnodeCommitLogs(pMnode, pTrans, &mnodeObj) != 0) goto _OVER;
-  if (mndSetCreateMnodeRedoActions(pMnode, pTrans, pDnode, &mnodeObj) != 0) goto _OVER;
   if (mndTransAppendNullLog(pTrans) != 0) goto _OVER;
   if (mndTransPrepare(pMnode, pTrans) != 0) goto _OVER;
 
@@ -459,107 +476,28 @@ static int32_t mndSetDropMnodeCommitLogs(SMnode *pMnode, STrans *pTrans, SMnodeO
 }
 
 static int32_t mndSetDropMnodeRedoActions(SMnode *pMnode, STrans *pTrans, SDnodeObj *pDnode, SMnodeObj *pObj) {
-  SSdb           *pSdb = pMnode->pSdb;
-  void           *pIter = NULL;
-  int32_t         numOfReplicas = 0;
-  SDAlterMnodeReq alterReq = {0};
-  SDDropMnodeReq  dropReq = {0};
-  SSetStandbyReq  standbyReq = {0};
-  SEpSet          alterEpset = {0};
-  SEpSet          dropEpSet = {0};
-
-  while (1) {
-    SMnodeObj *pMObj = NULL;
-    pIter = sdbFetch(pSdb, SDB_MNODE, pIter, (void **)&pMObj);
-    if (pIter == NULL) break;
-    if (pMObj->id == pObj->id) {
-      sdbRelease(pSdb, pMObj);
-      continue;
-    }
-
-    alterReq.replicas[numOfReplicas].id = pMObj->id;
-    alterReq.replicas[numOfReplicas].port = pMObj->pDnode->port;
-    memcpy(alterReq.replicas[numOfReplicas].fqdn, pMObj->pDnode->fqdn, TSDB_FQDN_LEN);
-
-    alterEpset.eps[numOfReplicas].port = pMObj->pDnode->port;
-    memcpy(alterEpset.eps[numOfReplicas].fqdn, pMObj->pDnode->fqdn, TSDB_FQDN_LEN);
-    if (pMObj->state == TAOS_SYNC_STATE_LEADER) {
-      alterEpset.inUse = numOfReplicas;
-    }
-
-    numOfReplicas++;
-    sdbRelease(pSdb, pMObj);
-  }
-
-  alterReq.replica = numOfReplicas;
-  alterEpset.numOfEps = numOfReplicas;
+  SSdb          *pSdb = pMnode->pSdb;
+  void          *pIter = NULL;
+  int32_t        numOfReplicas = 0;
+  SDDropMnodeReq dropReq = {0};
+  SEpSet         dropEpSet = {0};
 
   dropReq.dnodeId = pDnode->id;
   dropEpSet.numOfEps = 1;
   dropEpSet.eps[0].port = pDnode->port;
   memcpy(dropEpSet.eps[0].fqdn, pDnode->fqdn, TSDB_FQDN_LEN);
 
-  standbyReq.dnodeId = pDnode->id;
-  standbyReq.standby = 1;
-
-  {
-    int32_t contLen = tSerializeSSetStandbyReq(NULL, 0, &standbyReq) + sizeof(SMsgHead);
-    void   *pReq = taosMemoryMalloc(contLen);
-    tSerializeSSetStandbyReq((char *)pReq + sizeof(SMsgHead), contLen, &standbyReq);
-    SMsgHead *pHead = pReq;
-    pHead->contLen = htonl(contLen);
-    pHead->vgId = htonl(MNODE_HANDLE);
-
-    STransAction action = {
-        .epSet = dropEpSet,
-        .pCont = pReq,
-        .contLen = contLen,
-        .msgType = TDMT_SYNC_SET_MNODE_STANDBY,
-        .acceptableCode = TSDB_CODE_NODE_NOT_DEPLOYED,
-    };
-
-    if (mndTransAppendRedoAction(pTrans, &action) != 0) {
-      taosMemoryFree(pReq);
-      return -1;
-    }
-  }
-
-  {
-    int32_t contLen = tSerializeSDCreateMnodeReq(NULL, 0, &alterReq);
-    void   *pReq = taosMemoryMalloc(contLen);
-    tSerializeSDCreateMnodeReq(pReq, contLen, &alterReq);
-
-    STransAction action = {
-        .epSet = alterEpset,
-        .pCont = pReq,
-        .contLen = contLen,
-        .msgType = TDMT_MND_ALTER_MNODE,
-        .acceptableCode = 0,
-    };
-
-    if (mndTransAppendRedoAction(pTrans, &action) != 0) {
-      taosMemoryFree(pReq);
-      return -1;
-    }
-  }
-
-  {
-    int32_t contLen = tSerializeSCreateDropMQSBNodeReq(NULL, 0, &dropReq);
-    void   *pReq = taosMemoryMalloc(contLen);
-    tSerializeSCreateDropMQSBNodeReq(pReq, contLen, &dropReq);
-
-    STransAction action = {
-        .epSet = dropEpSet,
-        .pCont = pReq,
-        .contLen = contLen,
-        .msgType = TDMT_DND_DROP_MNODE,
-        .acceptableCode = TSDB_CODE_NODE_NOT_DEPLOYED,
-    };
-
-    if (mndTransAppendRedoAction(pTrans, &action) != 0) {
-      taosMemoryFree(pReq);
-      return -1;
-    }
+  int32_t totalMnodes = sdbGetSize(pSdb, SDB_MNODE);
+  if (totalMnodes == 2) {
+    mInfo("vgId:1, has %d mnodes, exec redo log first", totalMnodes);
+    if (mndSetDropMnodeRedoLogs(pMnode, pTrans, pObj) != 0) return -1;
+    if (mndBuildDropMnodeRedoAction(pTrans, &dropReq, &dropEpSet) != 0) return -1;
+  } else if (totalMnodes == 3) {
+    mInfo("vgId:1, has %d mnodes, exec redo action first", totalMnodes);
+    if (mndBuildDropMnodeRedoAction(pTrans, &dropReq, &dropEpSet) != 0) return -1;
+    if (mndSetDropMnodeRedoLogs(pMnode, pTrans, pObj) != 0) return -1;
+  } else {
+    return -1;
   }
 
   return 0;
@@ -567,7 +505,6 @@ static int32_t mndSetDropMnodeRedoActions(SMnode *pMnode, STrans *pTrans, SDnode
 
 int32_t mndSetDropMnodeInfoToTrans(SMnode *pMnode, STrans *pTrans, SMnodeObj *pObj) {
   if (pObj == NULL) return 0;
-  if (mndSetDropMnodeRedoLogs(pMnode, pTrans, pObj) != 0) return -1;
   if (mndSetDropMnodeCommitLogs(pMnode, pTrans, pObj) != 0) return -1;
   if (mndSetDropMnodeRedoActions(pMnode, pTrans, pObj->pDnode, pObj) != 0) return -1;
   if (mndTransAppendNullLog(pTrans) != 0) return -1;
@@ -657,7 +594,7 @@ static int32_t mndRetrieveMnodes(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pB
   int64_t    curMs = taosGetTimestampMs();
 
   while (numOfRows < rows) {
-    pShow->pIter = sdbFetchAll(pSdb, SDB_MNODE, pShow->pIter, (void **)&pObj, &objStatus);
+    pShow->pIter = sdbFetchAll(pSdb, SDB_MNODE, pShow->pIter, (void **)&pObj, &objStatus, true);
     if (pShow->pIter == NULL) break;
 
     cols = 0;
@@ -712,6 +649,9 @@ static void mndCancelGetNextMnode(SMnode *pMnode, void *pIter) {
 }
 
 static int32_t mndProcessAlterMnodeReq(SRpcMsg *pReq) {
+#if 1
+  return 0;
+#else
   SMnode         *pMnode = pReq->info.node;
   SDAlterMnodeReq alterReq = {0};
 
@@ -720,41 +660,107 @@ static int32_t mndProcessAlterMnodeReq(SRpcMsg *pReq) {
     return -1;
   }
 
+  SMnodeOpt option = {.deploy = true, .numOfReplicas = alterReq.replica, .selfIndex = -1};
+  memcpy(option.replicas, alterReq.replicas, sizeof(alterReq.replicas));
+  for (int32_t i = 0; i < option.numOfReplicas; ++i) {
+    if (alterReq.replicas[i].id == pMnode->selfDnodeId) {
+      option.selfIndex = i;
+    }
+  }
+
+  if (option.selfIndex == -1) {
+    mInfo("alter mnode not processed since selfIndex is -1", terrstr());
+    return 0;
+  }
+
+  if (mndWriteFile(pMnode->path, &option) != 0) {
+    mError("failed to write mnode file since %s", terrstr());
+    return -1;
+  }
+
   SSyncCfg cfg = {.replicaNum = alterReq.replica, .myIndex = -1};
   for (int32_t i = 0; i < alterReq.replica; ++i) {
     SNodeInfo *pNode = &cfg.nodeInfo[i];
     tstrncpy(pNode->nodeFqdn, alterReq.replicas[i].fqdn, sizeof(pNode->nodeFqdn));
     pNode->nodePort = alterReq.replicas[i].port;
-    if (alterReq.replicas[i].id == pMnode->selfDnodeId) cfg.myIndex = i;
+    if (alterReq.replicas[i].id == pMnode->selfDnodeId) {
+      cfg.myIndex = i;
+    }
   }
 
   if (cfg.myIndex == -1) {
     mError("failed to alter mnode since myindex is -1");
     return -1;
   } else {
-    mInfo("start to alter mnode sync, replica:%d myindex:%d", cfg.replicaNum, cfg.myIndex);
+    mInfo("start to alter mnode sync, replica:%d myIndex:%d", cfg.replicaNum, cfg.myIndex);
     for (int32_t i = 0; i < alterReq.replica; ++i) {
       SNodeInfo *pNode = &cfg.nodeInfo[i];
       mInfo("index:%d, fqdn:%s port:%d", i, pNode->nodeFqdn, pNode->nodePort);
     }
   }
 
-  mInfo("trans:-1, sync reconfig will be proposed");
-
-  SSyncMgmt *pMgmt = &pMnode->syncMgmt;
-  pMgmt->standby = 0;
-  int32_t code = syncReconfig(pMgmt->sync, &cfg);
+  int32_t code = syncReconfig(pMnode->syncMgmt.sync, &cfg);
   if (code != 0) {
-    mError("trans:-1, failed to propose sync reconfig since %s", terrstr());
-    return code;
+    mError("failed to sync reconfig since %s", terrstr());
   } else {
-    pMgmt->errCode = 0;
-    taosWLockLatch(&pMgmt->lock);
-    pMgmt->transId = -1;
-    taosWUnLockLatch(&pMgmt->lock);
-    tsem_wait(&pMgmt->syncSem);
-    mInfo("alter mnode sync result:0x%x %s", pMgmt->errCode, tstrerror(pMgmt->errCode));
-    terrno = pMgmt->errCode;
-    return pMgmt->errCode;
+    mInfo("alter mnode sync success");
+  }
+
+  return code;
+#endif
+}
+
+static void mndReloadSyncConfig(SMnode *pMnode) {
+  SSdb      *pSdb = pMnode->pSdb;
+  SMnodeObj *pObj = NULL;
+  ESdbStatus objStatus = 0;
+  void      *pIter = NULL;
+  bool       hasUpdatingMnode = false;
+  SSyncCfg   cfg = {.myIndex = -1};
+
+  while (1) {
+    pIter = sdbFetchAll(pSdb, SDB_MNODE, pIter, (void **)&pObj, &objStatus, false);
+    if (pIter == NULL) break;
+    if (objStatus == SDB_STATUS_CREATING || objStatus == SDB_STATUS_DROPPING) {
+      mInfo("vgId:1, has updating mnode:%d, status:%s", pObj->id, sdbStatusName(objStatus));
+      hasUpdatingMnode = true;
+    }
+
+    if (objStatus == SDB_STATUS_READY || objStatus == SDB_STATUS_CREATING) {
+      SNodeInfo *pNode = &cfg.nodeInfo[cfg.replicaNum];
+      tstrncpy(pNode->nodeFqdn, pObj->pDnode->fqdn, sizeof(pNode->nodeFqdn));
+      pNode->nodePort = pObj->pDnode->port;
+      if (pObj->pDnode->id == pMnode->selfDnodeId) {
+        cfg.myIndex = cfg.replicaNum;
+      }
+      cfg.replicaNum++;
+    }
+
+    sdbReleaseLock(pSdb, pObj, false);
+  }
+
+  if (cfg.myIndex == -1) {
+    mInfo("vgId:1, mnode not reload since selfIndex is -1");
+    return;
+  }
+
+  if (!mndGetRestored(pMnode)) {
+    mInfo("vgId:1, mnode not reload since restore not finished");
+    return;
+  }
+
+  if (hasUpdatingMnode) {
+    mInfo("vgId:1, start to reload mnode sync, replica:%d myIndex:%d", cfg.replicaNum, cfg.myIndex);
+    for (int32_t i = 0; i < cfg.replicaNum; ++i) {
+      SNodeInfo *pNode = &cfg.nodeInfo[i];
+      mInfo("vgId:1, index:%d, fqdn:%s port:%d", i, pNode->nodeFqdn, pNode->nodePort);
+    }
+
+    int32_t code = syncReconfig(pMnode->syncMgmt.sync, &cfg);
+    if (code != 0) {
+      mError("vgId:1, failed to reconfig mnode sync since %s", terrstr());
+    } else {
+      mInfo("vgId:1, reconfig mnode sync success");
+    }
   }
 }

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

@@ -107,7 +107,7 @@ void mndRestoreFinish(struct SSyncFSM *pFsm) {
   if (!pMnode->deploy) {
     mInfo("vgId:1, sync restore finished, and will handle outstanding transactions");
     mndTransPullup(pMnode);
-    mndSetRestore(pMnode, true);
+    mndSetRestored(pMnode, true);
   } else {
     mInfo("vgId:1, sync restore finished");
   }
@@ -225,18 +225,17 @@ int32_t mndInitSync(SMnode *pMnode) {
   snprintf(syncInfo.path, sizeof(syncInfo.path), "%s%ssync", pMnode->path, TD_DIRSEP);
   syncInfo.pWal = pMnode->pWal;
   syncInfo.pFsm = mndSyncMakeFsm(pMnode);
-  syncInfo.isStandBy = pMgmt->standby;
   syncInfo.snapshotStrategy = SYNC_STRATEGY_STANDARD_SNAPSHOT;
 
-  mInfo("vgId:1, start to open sync, standby:%d", pMgmt->standby);
-  if (pMgmt->standby || pMgmt->replica.id > 0) {
-    SSyncCfg *pCfg = &syncInfo.syncCfg;
-    pCfg->replicaNum = 1;
-    pCfg->myIndex = 0;
-    SNodeInfo *pNode = &pCfg->nodeInfo[0];
-    tstrncpy(pNode->nodeFqdn, pMgmt->replica.fqdn, sizeof(pNode->nodeFqdn));
-    pNode->nodePort = pMgmt->replica.port;
-    mInfo("vgId:1, ep:%s:%u", pNode->nodeFqdn, pNode->nodePort);
+  mInfo("vgId:1, start to open sync, selfIndex:%d replica:%d", pMgmt->selfIndex, pMgmt->numOfReplicas);
+  SSyncCfg *pCfg = &syncInfo.syncCfg;
+  pCfg->replicaNum = pMgmt->numOfReplicas;
+  pCfg->myIndex = pMgmt->selfIndex;
+  for (int32_t i = 0; i < pMgmt->numOfReplicas; ++i) {
+    SNodeInfo *pNode = &pCfg->nodeInfo[i];
+    tstrncpy(pNode->nodeFqdn, pMgmt->replicas[i].fqdn, sizeof(pNode->nodeFqdn));
+    pNode->nodePort = pMgmt->replicas[i].port;
+    mInfo("vgId:1, index:%d ep:%s:%u", i, pNode->nodeFqdn, pNode->nodePort);
   }
 
   tsem_init(&pMgmt->syncSem, 0, 0);
@@ -250,10 +249,6 @@ int32_t mndInitSync(SMnode *pMnode) {
   setPingTimerMS(pMgmt->sync, 5000);
   setElectTimerMS(pMgmt->sync, 3000);
   setHeartbeatTimerMS(pMgmt->sync, 500);
-  /*
-    setElectTimerMS(pMgmt->sync, 600);
-    setHeartbeatTimerMS(pMgmt->sync, 300);
-  */
 
   mInfo("mnode-sync is opened, id:%" PRId64, pMgmt->sync);
   return 0;
@@ -319,7 +314,7 @@ void mndSyncStart(SMnode *pMnode) {
   SSyncMgmt *pMgmt = &pMnode->syncMgmt;
   syncSetMsgCb(pMgmt->sync, &pMnode->msgCb);
   syncStart(pMgmt->sync);
-  mInfo("vgId:1, sync started, id:%" PRId64 " standby:%d", pMgmt->sync, pMgmt->standby);
+  mInfo("vgId:1, sync started, id:%" PRId64, pMgmt->sync);
 }
 
 void mndSyncStop(SMnode *pMnode) {
@@ -331,7 +326,7 @@ void mndSyncStop(SMnode *pMnode) {
   taosWUnLockLatch(&pMnode->syncMgmt.lock);
 }
 
-bool mndIsMaster(SMnode *pMnode) {
+bool mndIsLeader(SMnode *pMnode) {
   SSyncMgmt *pMgmt = &pMnode->syncMgmt;
 
   if (!syncIsReady(pMgmt->sync)) {
@@ -340,7 +335,7 @@ bool mndIsMaster(SMnode *pMnode) {
     return false;
   }
 
-  if (!pMnode->restored) {
+  if (!mndGetRestored(pMnode)) {
     terrno = TSDB_CODE_APP_NOT_READY;
     return false;
   }

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

@@ -154,6 +154,8 @@ TOPIC_ENCODE_OVER:
 
 SSdbRow *mndTopicActionDecode(SSdbRaw *pRaw) {
   terrno = TSDB_CODE_OUT_OF_MEMORY;
+
+  void  *buf = NULL;
   int8_t sver = 0;
   if (sdbGetRawSoftVer(pRaw, &sver) != 0) goto TOPIC_DECODE_OVER;
 
@@ -215,17 +217,15 @@ SSdbRow *mndTopicActionDecode(SSdbRaw *pRaw) {
 
   SDB_GET_INT32(pRaw, dataPos, &len, TOPIC_DECODE_OVER);
   if (len) {
-    void *buf = taosMemoryMalloc(len);
+    buf = taosMemoryMalloc(len);
     if (buf == NULL) {
       terrno = TSDB_CODE_OUT_OF_MEMORY;
       goto TOPIC_DECODE_OVER;
     }
     SDB_GET_BINARY(pRaw, dataPos, buf, len, TOPIC_DECODE_OVER);
     if (taosDecodeSSchemaWrapper(buf, &pTopic->schema) == NULL) {
-      taosMemoryFree(buf);
       goto TOPIC_DECODE_OVER;
     }
-    taosMemoryFree(buf);
   } else {
     pTopic->schema.nCols = 0;
     pTopic->schema.version = 0;
@@ -251,6 +251,7 @@ SSdbRow *mndTopicActionDecode(SSdbRaw *pRaw) {
   terrno = TSDB_CODE_SUCCESS;
 
 TOPIC_DECODE_OVER:
+  taosMemoryFreeClear(buf);
   if (terrno != TSDB_CODE_SUCCESS) {
     mError("topic:%s, failed to decode from raw:%p since %s", pTopic->name, pRaw, terrstr());
     taosMemoryFreeClear(pRow);

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

@@ -53,7 +53,7 @@ static bool    mndTransPerformCommitActionStage(SMnode *pMnode, STrans *pTrans);
 static bool    mndTransPerformCommitStage(SMnode *pMnode, STrans *pTrans);
 static bool    mndTransPerformRollbackStage(SMnode *pMnode, STrans *pTrans);
 static bool    mndTransPerfromFinishedStage(SMnode *pMnode, STrans *pTrans);
-static bool    mndCannotExecuteTransAction(SMnode *pMnode) { return !pMnode->deploy && !mndIsMaster(pMnode); }
+static bool    mndCannotExecuteTransAction(SMnode *pMnode) { return !pMnode->deploy && !mndIsLeader(pMnode); }
 
 static void    mndTransSendRpcRsp(SMnode *pMnode, STrans *pTrans);
 static int32_t mndProcessTransTimer(SRpcMsg *pReq);

source/dnode/mnode/sdb/inc/sdb.h

@@ -298,6 +298,7 @@ void *sdbAcquire(SSdb *pSdb, ESdbType type, const void *pKey);
  * @param pObj The object of the row.
  */
 void sdbRelease(SSdb *pSdb, void *pObj);
+void sdbReleaseLock(SSdb *pSdb, void *pObj, bool lock);
 
 /**
  * @brief Traverse a sdb table
@@ -309,7 +310,7 @@ void sdbRelease(SSdb *pSdb, void *pObj);
  * @return void* The next iterator of the table.
  */
 void *sdbFetch(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj);
-void *sdbFetchAll(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj, ESdbStatus *status);
+void *sdbFetchAll(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj, ESdbStatus *status, bool lock);
 
 /**
  * @brief Cancel a traversal

source/dnode/mnode/sdb/src/sdbHash.c

@@ -327,14 +327,16 @@ static void sdbCheckRow(SSdb *pSdb, SSdbRow *pRow) {
   taosThreadRwlockUnlock(pLock);
 }
 
-void sdbRelease(SSdb *pSdb, void *pObj) {
+void sdbReleaseLock(SSdb *pSdb, void *pObj, bool lock) {
   if (pObj == NULL) return;
 
   SSdbRow *pRow = (SSdbRow *)((char *)pObj - sizeof(SSdbRow));
   if (pRow->type >= SDB_MAX) return;
 
   TdThreadRwlock *pLock = &pSdb->locks[pRow->type];
-  taosThreadRwlockWrlock(pLock);
+  if (lock) {
+    taosThreadRwlockWrlock(pLock);
+  }
 
   int32_t ref = atomic_sub_fetch_32(&pRow->refCount, 1);
   sdbPrintOper(pSdb, pRow, "release");
@@ -342,9 +344,13 @@ void sdbRelease(SSdb *pSdb, void *pObj) {
     sdbFreeRow(pSdb, pRow, true);
   }
 
-  taosThreadRwlockUnlock(pLock);
+  if (lock) {
+    taosThreadRwlockUnlock(pLock);
+  }
 }
 
+void sdbRelease(SSdb *pSdb, void *pObj) { sdbReleaseLock(pSdb, pObj, true); }
+
 void *sdbFetch(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj) {
   *ppObj = NULL;
 
@@ -372,14 +378,16 @@ void *sdbFetch(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj) {
   return ppRow;
 }
 
-void *sdbFetchAll(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj, ESdbStatus *status) {
+void *sdbFetchAll(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj, ESdbStatus *status, bool lock) {
   *ppObj = NULL;
 
   SHashObj *hash = sdbGetHash(pSdb, type);
   if (hash == NULL) return NULL;
 
   TdThreadRwlock *pLock = &pSdb->locks[type];
-  taosThreadRwlockRdlock(pLock);
+  if (lock) {
+    taosThreadRwlockRdlock(pLock);
+  }
 
   SSdbRow **ppRow = taosHashIterate(hash, pIter);
   while (ppRow != NULL) {
@@ -395,7 +403,9 @@ void *sdbFetchAll(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj, ESdbStat
     *status = pRow->status;
     break;
   }
-  taosThreadRwlockUnlock(pLock);
+  if (lock) {
+    taosThreadRwlockUnlock(pLock);
+  }
 
   return ppRow;
 }

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

@@ -51,7 +51,7 @@ static int metaUpdateMetaRsp(tb_uid_t uid, char *tbName, SSchemaWrapper *pSchema
     return -1;
   }
 
-  strcpy(pMetaRsp->tbName, tbName);
+  strncpy(pMetaRsp->tbName, tbName, TSDB_TABLE_NAME_LEN);
   pMetaRsp->numOfColumns = pSchema->nCols;
   pMetaRsp->tableType = TSDB_NORMAL_TABLE;
   pMetaRsp->sversion = pSchema->version;
@@ -693,6 +693,7 @@ static int metaAlterTableColumn(SMeta *pMeta, int64_t version, SVAlterTbReq *pAl
     if (iCol >= pSchema->nCols) break;
     pColumn = &pSchema->pSchema[iCol];
 
+    ASSERT(pAlterTbReq->colName);
     if (strcmp(pColumn->name, pAlterTbReq->colName) == 0) break;
     iCol++;
   }
@@ -816,6 +817,11 @@ static int metaUpdateTableTagVal(SMeta *pMeta, int64_t version, SVAlterTbReq *pA
   const void *pData = NULL;
   int         nData = 0;
 
+  if (pAlterTbReq->tagName == NULL) {
+    terrno = TSDB_CODE_INVALID_MSG;
+    return -1;
+  }
+
   // search name index
   ret = tdbTbGet(pMeta->pNameIdx, pAlterTbReq->tbName, strlen(pAlterTbReq->tbName) + 1, &pVal, &nVal);
   if (ret < 0) {

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

@@ -321,7 +321,7 @@ static int32_t tdSetRSmaInfoItemParams(SSma *pSma, SRSmaParam *param, SRSmaStat
       }
       taosMemoryFree(s);
     }
-    pStreamState = streamStateOpen(taskInfDir, NULL, true);
+    pStreamState = streamStateOpen(taskInfDir, NULL, true, -1, -1);
     if (!pStreamState) {
       terrno = TSDB_CODE_RSMA_STREAM_STATE_OPEN;
       return TSDB_CODE_FAILED;

source/dnode/vnode/src/tq/tq.c

@@ -171,11 +171,11 @@ int32_t tqPushDataRsp(STQ* pTq, STqPushEntry* pPushEntry) {
   ASSERT(taosArrayGetSize(pRsp->blockSchema) == 0);
 
   if (pRsp->reqOffset.type == TMQ_OFFSET__LOG) {
-    if (pRsp->blockNum > 0) {
-      ASSERT(pRsp->rspOffset.version > pRsp->reqOffset.version);
-    } else {
-      ASSERT(pRsp->rspOffset.version >= pRsp->reqOffset.version);
-    }
+    /*if (pRsp->blockNum > 0) {*/
+    /*ASSERT(pRsp->rspOffset.version > pRsp->reqOffset.version);*/
+    /*} else {*/
+    ASSERT(pRsp->rspOffset.version > pRsp->reqOffset.version);
+    /*}*/
   }
 
   int32_t len = 0;
@@ -890,7 +890,7 @@ int32_t tqExpandTask(STQ* pTq, SStreamTask* pTask) {
 
   // expand executor
   if (pTask->taskLevel == TASK_LEVEL__SOURCE) {
-    pTask->pState = streamStateOpen(pTq->pStreamMeta->path, pTask, false);
+    pTask->pState = streamStateOpen(pTq->pStreamMeta->path, pTask, false, -1, -1);
     if (pTask->pState == NULL) {
       return -1;
     }
@@ -904,7 +904,7 @@ int32_t tqExpandTask(STQ* pTq, SStreamTask* pTask) {
     pTask->exec.executor = qCreateStreamExecTaskInfo(pTask->exec.qmsg, &handle);
     ASSERT(pTask->exec.executor);
   } else if (pTask->taskLevel == TASK_LEVEL__AGG) {
-    pTask->pState = streamStateOpen(pTq->pStreamMeta->path, pTask, false);
+    pTask->pState = streamStateOpen(pTq->pStreamMeta->path, pTask, false, -1, -1);
     if (pTask->pState == NULL) {
       return -1;
     }

source/dnode/vnode/src/tq/tqMeta.c

@@ -177,6 +177,7 @@ int32_t tqMetaRestoreCheckInfo(STQ* pTq) {
     tDecoderClear(&decoder);
     if (taosHashPut(pTq->pCheckInfo, info.topic, strlen(info.topic), &info, sizeof(STqCheckInfo)) < 0) {
       terrno = TSDB_CODE_OUT_OF_MEMORY;
+      tdbFree(pKey);
       tdbTbcClose(pCur);
       return -1;
     }

source/dnode/vnode/src/tq/tqPush.c

@@ -245,7 +245,7 @@ int tqPushMsg(STQ* pTq, void* msg, int32_t msgLen, tmsg_t msgType, int64_t ver)
           tqDebug("vgId:%d cannot find handle %s", pTq->pVnode->config.vgId, pPushEntry->subKey);
           continue;
         }
-        if (pPushEntry->dataRsp.reqOffset.version > ver) {
+        if (pPushEntry->dataRsp.reqOffset.version >= ver) {
           tqDebug("vgId:%d push entry req version %ld, while push version %ld, skip", pTq->pVnode->config.vgId,
                   pPushEntry->dataRsp.reqOffset.version, ver);
           continue;

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

@@ -378,10 +378,10 @@ static int32_t getTableDelSkyline(STbData *pMem, STbData *pIMem, SDelFReader *pD
     if (code) goto _err;
   }
 
+_err:
   if (aDelData) {
     taosArrayDestroy(aDelData);
   }
-_err:
   return code;
 }
 
@@ -399,14 +399,13 @@ static int32_t getTableDelIdx(SDelFReader *pDelFReader, tb_uid_t suid, tb_uid_t
 
   // code = tMapDataSearch(&delIdxMap, &idx, tGetDelIdx, tCmprDelIdx, pDelIdx);
   SDelIdx *pIdx = taosArraySearch(pDelIdxArray, &idx, tCmprDelIdx, TD_EQ);
-  if (code) goto _err;
 
   *pDelIdx = *pIdx;
 
+_err:
   if (pDelIdxArray) {
     taosArrayDestroy(pDelIdxArray);
   }
-_err:
   return code;
 }
 
@@ -429,7 +428,8 @@ typedef struct {
   SDataFReader        *pDataFReader;
   TSDBROW              row;
 
-  SMergeTree mergeTree;
+  SMergeTree  mergeTree;
+  SMergeTree *pMergeTree;
 } SFSLastNextRowIter;
 
 static int32_t getNextRowFromFSLast(void *iter, TSDBROW **ppRow) {
@@ -444,11 +444,14 @@ static int32_t getNextRowFromFSLast(void *iter, TSDBROW **ppRow) {
     case SFSLASTNEXTROW_FILESET: {
       SDFileSet *pFileSet = NULL;
     _next_fileset:
+      if (state->pMergeTree != NULL) {
+        tMergeTreeClose(state->pMergeTree);
+        state->pMergeTree = NULL;
+      }
+
       if (--state->iFileSet >= 0) {
         pFileSet = (SDFileSet *)taosArrayGet(state->aDFileSet, state->iFileSet);
       } else {
-        tMergeTreeClose(&state->mergeTree);
-
         *ppRow = NULL;
         return code;
       }
@@ -460,10 +463,10 @@ static int32_t getNextRowFromFSLast(void *iter, TSDBROW **ppRow) {
       tMergeTreeOpen(&state->mergeTree, 1, state->pDataFReader, state->suid, state->uid,
                      &(STimeWindow){.skey = TSKEY_MIN, .ekey = TSKEY_MAX},
                      &(SVersionRange){.minVer = 0, .maxVer = UINT64_MAX}, pLoadInfo, true, NULL);
+      state->pMergeTree = &state->mergeTree;
       bool hasVal = tMergeTreeNext(&state->mergeTree);
       if (!hasVal) {
         state->state = SFSLASTNEXTROW_FILESET;
-        tMergeTreeClose(&state->mergeTree);
         goto _next_fileset;
       }
       state->state = SFSLASTNEXTROW_BLOCKROW;
@@ -475,6 +478,7 @@ static int32_t getNextRowFromFSLast(void *iter, TSDBROW **ppRow) {
       if (!hasVal) {
         state->state = SFSLASTNEXTROW_FILESET;
       }
+
       return code;
     default:
       ASSERT(0);
@@ -486,6 +490,11 @@ _err:
     tsdbDataFReaderClose(&state->pDataFReader);
     state->pDataFReader = NULL;
   }
+  if (state->pMergeTree != NULL) {
+    tMergeTreeClose(state->pMergeTree);
+    state->pMergeTree = NULL;
+  }
+
   *ppRow = NULL;
 
   return code;
@@ -504,6 +513,11 @@ int32_t clearNextRowFromFSLast(void *iter) {
     state->pDataFReader = NULL;
   }
 
+  if (state->pMergeTree != NULL) {
+    tMergeTreeClose(state->pMergeTree);
+    state->pMergeTree = NULL;
+  }
+
   return code;
 }
 
@@ -584,8 +598,6 @@ static int32_t getNextRowFromFS(void *iter, TSDBROW **ppRow) {
        * &state->blockIdx);
        */
       state->pBlockIdx = taosArraySearch(state->aBlockIdx, state->pBlockIdxExp, tCmprBlockIdx, TD_EQ);
-      if (code) goto _err;
-
       if (!state->pBlockIdx) {
         goto _next_fileset;
       }
@@ -883,10 +895,16 @@ static int32_t nextRowIterOpen(CacheNextRowIter *pIter, tb_uid_t uid, STsdb *pTs
     if (code) goto _err;
 
     code = getTableDelIdx(pDelFReader, suid, uid, &delIdx);
-    if (code) goto _err;
+    if (code) {
+      tsdbDelFReaderClose(&pDelFReader);
+      goto _err;
+    }
 
     code = getTableDelSkyline(pMem, pIMem, pDelFReader, &delIdx, pIter->pSkyline);
-    if (code) goto _err;
+    if (code) {
+      tsdbDelFReaderClose(&pDelFReader);
+      goto _err;
+    }
 
     tsdbDelFReaderClose(&pDelFReader);
   } else {
@@ -1216,6 +1234,8 @@ static int32_t mergeLast(tb_uid_t uid, STsdb *pTsdb, SArray **ppLastArray) {
 _err:
   nextRowIterClose(&iter);
   taosMemoryFreeClear(pTSchema);
+  *ppLastArray = NULL;
+  taosArrayDestroy(pColArray);
   return code;
 }
 

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

@@ -396,12 +396,19 @@ _exit:
   return code;
 }
 
+static int32_t tDataIterCmprFn(const SRBTreeNode *n1, const SRBTreeNode *n2) {
+  SDataIter *pIter1 = (SDataIter *)((uint8_t *)n1 - offsetof(SDataIter, n));
+  SDataIter *pIter2 = (SDataIter *)((uint8_t *)n2 - offsetof(SDataIter, n));
+
+  return tRowInfoCmprFn(&pIter1->r, &pIter2->r);
+}
+
 static int32_t tsdbOpenCommitIter(SCommitter *pCommitter) {
   int32_t code = 0;
   int32_t lino = 0;
 
   pCommitter->pIter = NULL;
-  tRBTreeCreate(&pCommitter->rbt, tRowInfoCmprFn);
+  tRBTreeCreate(&pCommitter->rbt, tDataIterCmprFn);
 
   // memory
   TSDBKEY    tKey = {.ts = pCommitter->minKey, .version = VERSION_MIN};
@@ -1038,7 +1045,9 @@ static int32_t tsdbEndCommit(SCommitter *pCommitter, int32_t eno) {
   STsdb     *pTsdb = pCommitter->pTsdb;
   SMemTable *pMemTable = pTsdb->imem;
 
-  ASSERT(eno == 0);
+  ASSERT(eno == 0 &&
+    "tsdbCommit failure"
+    "Restart taosd");
 
   code = tsdbFSCommit1(pTsdb, &pCommitter->fs);
   TSDB_CHECK_CODE(code, lino, _exit);

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

@@ -610,9 +610,6 @@ int32_t tsdbFSRollback(STsdbFS *pFS) {
   ASSERT(0);
 
   return code;
-
-_err:
-  return code;
 }
 
 int32_t tsdbFSUpsertDelFile(STsdbFS *pFS, SDelFile *pDelFile) {
@@ -822,7 +819,7 @@ int32_t tsdbFSCommit2(STsdb *pTsdb, STsdbFS *pFSNew) {
       nRef = atomic_sub_fetch_32(&fSet.pHeadF->nRef, 1);
       if (nRef == 0) {
         tsdbHeadFileName(pTsdb, pSetOld->diskId, pSetOld->fid, fSet.pHeadF, fname);
-        taosRemoveFile(fname);
+        (void)taosRemoveFile(fname);
         taosMemoryFree(fSet.pHeadF);
       }
     } else {
@@ -866,7 +863,7 @@ int32_t tsdbFSCommit2(STsdb *pTsdb, STsdbFS *pFSNew) {
       nRef = atomic_sub_fetch_32(&fSet.pSmaF->nRef, 1);
       if (nRef == 0) {
         tsdbSmaFileName(pTsdb, pSetOld->diskId, pSetOld->fid, fSet.pSmaF, fname);
-        taosRemoveFile(fname);
+        (void)taosRemoveFile(fname);
         taosMemoryFree(fSet.pSmaF);
       }
     } else {
@@ -877,7 +874,7 @@ int32_t tsdbFSCommit2(STsdb *pTsdb, STsdbFS *pFSNew) {
     // stt
     if (sameDisk) {
       if (pSetNew->nSttF > pSetOld->nSttF) {
-        ASSERT(pSetNew->nSttF = pSetOld->nSttF + 1);
+        ASSERT(pSetNew->nSttF == pSetOld->nSttF + 1);
         pSetOld->aSttF[pSetOld->nSttF] = (SSttFile *)taosMemoryMalloc(sizeof(SSttFile));
         if (pSetOld->aSttF[pSetOld->nSttF] == NULL) {
           code = TSDB_CODE_OUT_OF_MEMORY;
@@ -964,7 +961,7 @@ int32_t tsdbFSCommit2(STsdb *pTsdb, STsdbFS *pFSNew) {
     nRef = atomic_sub_fetch_32(&pSetOld->pHeadF->nRef, 1);
     if (nRef == 0) {
       tsdbHeadFileName(pTsdb, pSetOld->diskId, pSetOld->fid, pSetOld->pHeadF, fname);
-      taosRemoveFile(fname);
+      (void)taosRemoveFile(fname);
       taosMemoryFree(pSetOld->pHeadF);
     }
 
@@ -1104,7 +1101,7 @@ void tsdbFSUnref(STsdb *pTsdb, STsdbFS *pFS) {
     ASSERT(nRef >= 0);
     if (nRef == 0) {
       tsdbDelFileName(pTsdb, pFS->pDelFile, fname);
-      taosRemoveFile(fname);
+      (void)taosRemoveFile(fname);
       taosMemoryFree(pFS->pDelFile);
     }
   }
@@ -1117,7 +1114,7 @@ void tsdbFSUnref(STsdb *pTsdb, STsdbFS *pFS) {
     ASSERT(nRef >= 0);
     if (nRef == 0) {
       tsdbHeadFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pHeadF, fname);
-      taosRemoveFile(fname);
+      (void)taosRemoveFile(fname);
       taosMemoryFree(pSet->pHeadF);
     }
 

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

@@ -160,6 +160,7 @@ int32_t tsdbDeleteTableData(STsdb *pTsdb, int64_t version, tb_uid_t suid, tb_uid
     goto _err;
   }
 
+  ASSERT(pPool != NULL);
   // do delete
   SDelData *pDelData = (SDelData *)vnodeBufPoolMalloc(pPool, sizeof(*pDelData));
   if (pDelData == NULL) {
@@ -353,6 +354,7 @@ static int32_t tsdbGetOrCreateTbData(SMemTable *pMemTable, tb_uid_t suid, tb_uid
   SVBufPool *pPool = pMemTable->pTsdb->pVnode->inUse;
   int8_t     maxLevel = pMemTable->pTsdb->pVnode->config.tsdbCfg.slLevel;
 
+  ASSERT(pPool != NULL);
   pTbData = vnodeBufPoolMalloc(pPool, sizeof(*pTbData) + SL_NODE_SIZE(maxLevel) * 2);
   if (pTbData == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;
@@ -492,6 +494,7 @@ static int32_t tbDataDoPut(SMemTable *pMemTable, STbData *pTbData, SMemSkipListN
 
   // node
   level = tsdbMemSkipListRandLevel(&pTbData->sl);
+  ASSERT(pPool != NULL);
   pNode = (SMemSkipListNode *)vnodeBufPoolMalloc(pPool, SL_NODE_SIZE(level) + tPutTSDBRow(NULL, pRow));
   if (pNode == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;
@@ -559,6 +562,8 @@ static int32_t tsdbInsertTableDataImpl(SMemTable *pMemTable, STbData *pTbData, i
 
   // backward put first data
   row.pTSRow = tGetSubmitBlkNext(&blkIter);
+  if (row.pTSRow == NULL) return code;
+
   key.ts = row.pTSRow->ts;
   nRow++;
   tbDataMovePosTo(pTbData, pos, &key, SL_MOVE_BACKWARD);

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

@@ -103,10 +103,20 @@ static SBlockData *loadLastBlock(SLDataIter *pIter, const char *idStr) {
 
   SSttBlockLoadInfo *pInfo = pIter->pBlockLoadInfo;
   if (pInfo->blockIndex[0] == pIter->iSttBlk) {
+    if (pInfo->currentLoadBlockIndex != 0) {
+      tsdbDebug("current load index is set to 0, block index:%d, file index:%d, due to uid:%" PRIu64 ", load data, %s",
+                pIter->iSttBlk, pIter->iStt, pIter->uid, idStr);
+      pInfo->currentLoadBlockIndex = 0;
+    }
     return &pInfo->blockData[0];
   }
 
   if (pInfo->blockIndex[1] == pIter->iSttBlk) {
+    if (pInfo->currentLoadBlockIndex != 1) {
+      tsdbDebug("current load index is set to 1, block index:%d, file index:%d, due to uid:%" PRIu64 ", load data, %s",
+                pIter->iSttBlk, pIter->iStt, pIter->uid, idStr);
+      pInfo->currentLoadBlockIndex = 1;
+    }
     return &pInfo->blockData[1];
   }
 
@@ -123,20 +133,31 @@ static SBlockData *loadLastBlock(SLDataIter *pIter, const char *idStr) {
       id.uid = pIter->uid;
     }
 
-    tBlockDataInit(pBlock, &id, pInfo->pSchema, pInfo->colIds, pInfo->numOfCols);
+    code = tBlockDataInit(pBlock, &id, pInfo->pSchema, pInfo->colIds, pInfo->numOfCols);
+    if (code != TSDB_CODE_SUCCESS) {
+      goto _exit;
+    }
+
     code = tsdbReadSttBlock(pIter->pReader, pIter->iStt, pIter->pSttBlk, pBlock);
+    if (code != TSDB_CODE_SUCCESS) {
+      goto _exit;
+    }
 
     double el = (taosGetTimestampUs() - st) / 1000.0;
     pInfo->elapsedTime += el;
     pInfo->loadBlocks += 1;
 
-    tsdbDebug("read last block, index:%d, last file index:%d, elapsed time:%.2f ms, %s", pIter->iSttBlk, pIter->iStt,
-              el, idStr);
+    tsdbDebug("read last block, total load:%d, trigger by uid:%" PRIu64
+              ", last file index:%d, last block index:%d, entry:%d, %p, elapsed time:%.2f ms, %s",
+              pInfo->loadBlocks, pIter->uid, pIter->iStt, pIter->iSttBlk, pInfo->currentLoadBlockIndex, pBlock, el,
+              idStr);
     if (code != TSDB_CODE_SUCCESS) {
       goto _exit;
     }
 
     pInfo->blockIndex[pInfo->currentLoadBlockIndex] = pIter->iSttBlk;
+    tsdbDebug("last block index list:%d, %d, %s", pInfo->blockIndex[0], pInfo->blockIndex[1], idStr);
+
     pIter->iRow = (pIter->backward) ? pInfo->blockData[pInfo->currentLoadBlockIndex].nRow : -1;
   }
 
@@ -324,8 +345,10 @@ _exit:
 
 void tLDataIterClose(SLDataIter *pIter) { taosMemoryFree(pIter); }
 
-void tLDataIterNextBlock(SLDataIter *pIter) {
+void tLDataIterNextBlock(SLDataIter *pIter, const char *idStr) {
   int32_t step = pIter->backward ? -1 : 1;
+  int32_t oldIndex = pIter->iSttBlk;
+
   pIter->iSttBlk += step;
 
   int32_t index = -1;
@@ -372,6 +395,10 @@ void tLDataIterNextBlock(SLDataIter *pIter) {
   if (index != -1) {
     pIter->iSttBlk = index;
     pIter->pSttBlk = (SSttBlk *)taosArrayGet(pIter->pBlockLoadInfo->aSttBlk, pIter->iSttBlk);
+    tsdbDebug("try next last file block:%d from %d, trigger by uid:%" PRIu64 ", file index:%d, %s", pIter->iSttBlk,
+              oldIndex, pIter->uid, pIter->iStt, idStr);
+  } else {
+    tsdbDebug("no more last block qualified, uid:%" PRIu64 ", file index::%d, %s", pIter->uid, oldIndex, idStr);
   }
 }
 
@@ -396,16 +423,10 @@ static void findNextValidRow(SLDataIter *pIter, const char *idStr) {
   for (; i < pBlockData->nRow && i >= 0; i += step) {
     if (pBlockData->aUid != NULL) {
       if (!pIter->backward) {
-        /*if (pBlockData->aUid[i] < pIter->uid) {
-          continue;
-        } else */
         if (pBlockData->aUid[i] > pIter->uid) {
           break;
         }
       } else {
-        /*if (pBlockData->aUid[i] > pIter->uid) {
-          continue;
-        } else */
         if (pBlockData->aUid[i] < pIter->uid) {
           break;
         }
@@ -461,7 +482,7 @@ bool tLDataIterNextRow(SLDataIter *pIter, const char *idStr) {
     findNextValidRow(pIter, idStr);
 
     if (pIter->iRow >= pBlockData->nRow || pIter->iRow < 0) {
-      tLDataIterNextBlock(pIter);
+      tLDataIterNextBlock(pIter, idStr);
       if (pIter->pSttBlk == NULL) {  // no more data
         goto _exit;
       }
@@ -490,9 +511,9 @@ _exit:
 SRowInfo *tLDataIterGet(SLDataIter *pIter) { return &pIter->rInfo; }
 
 // SMergeTree =================================================
-static FORCE_INLINE int32_t tLDataIterCmprFn(const void *p1, const void *p2) {
-  SLDataIter *pIter1 = (SLDataIter *)(((uint8_t *)p1) - sizeof(SRBTreeNode));
-  SLDataIter *pIter2 = (SLDataIter *)(((uint8_t *)p2) - sizeof(SRBTreeNode));
+static FORCE_INLINE int32_t tLDataIterCmprFn(const SRBTreeNode *p1, const SRBTreeNode *p2) {
+  SLDataIter *pIter1 = (SLDataIter *)(((uint8_t *)p1) - offsetof(SLDataIter, node));
+  SLDataIter *pIter2 = (SLDataIter *)(((uint8_t *)p2) - offsetof(SLDataIter, node));
 
   TSDBKEY key1 = TSDBROW_KEY(&pIter1->rInfo.row);
   TSDBKEY key2 = TSDBROW_KEY(&pIter2->rInfo.row);
@@ -529,6 +550,7 @@ int32_t tMergeTreeOpen(SMergeTree *pMTree, int8_t backward, SDataFReader *pFRead
 
   pMTree->pLoadInfo = pBlockLoadInfo;
   pMTree->destroyLoadInfo = destroyLoadInfo;
+  ASSERT(pMTree->pLoadInfo != NULL);
 
   for (int32_t i = 0; i < pFReader->pSet->nSttF; ++i) {  // open all last file
     struct SLDataIter *pIter = NULL;
@@ -569,7 +591,7 @@ bool tMergeTreeNext(SMergeTree *pMTree) {
     // compare with min in RB Tree
     pIter = (SLDataIter *)tRBTreeMin(&pMTree->rbt);
     if (pMTree->pIter && pIter) {
-      int32_t c = pMTree->rbt.cmprFn(RBTREE_NODE_PAYLOAD(&pMTree->pIter->node), RBTREE_NODE_PAYLOAD(&pIter->node));
+      int32_t c = pMTree->rbt.cmprFn(&pMTree->pIter->node, &pIter->node);
       if (c > 0) {
         tRBTreePut(&pMTree->rbt, (SRBTreeNode *)pMTree->pIter);
         pMTree->pIter = NULL;

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

@@ -87,9 +87,13 @@ _err:
 
 int tsdbClose(STsdb **pTsdb) {
   if (*pTsdb) {
-    taosThreadRwlockDestroy(&(*pTsdb)->rwLock);
+    taosThreadRwlockWrlock(&(*pTsdb)->rwLock);
     tsdbMemTableDestroy((*pTsdb)->mem);
     (*pTsdb)->mem = NULL;
+    taosThreadRwlockUnlock(&(*pTsdb)->rwLock);
+
+    taosThreadRwlockDestroy(&(*pTsdb)->rwLock);
+
     tsdbFSClose(*pTsdb);
     tsdbCloseCache(*pTsdb);
     taosMemoryFreeClear(*pTsdb);

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

@@ -1645,6 +1645,7 @@ static int32_t doMergeFileBlockAndLastBlock(SLastBlockReader* pLastBlockReader,
   STSRow*    pTSRow = NULL;
   SRowMerger merge = {0};
   TSDBROW    fRow = tMergeTreeGetRow(&pLastBlockReader->mergeTree);
+  tsdbTrace("fRow ptr:%p, %d, uid:%" PRIu64 ", %s", fRow.pBlockData, fRow.iRow, pLastBlockReader->uid, pReader->idStr);
 
   // only last block exists
   if ((!mergeBlockData) || (tsLastBlock != pBlockData->aTSKEY[pDumpInfo->rowIndex])) {
@@ -2675,9 +2676,9 @@ static STsdb* getTsdbByRetentions(SVnode* pVnode, TSKEY winSKey, SRetention* ret
     int8_t  level = 0;
     int8_t  precision = pVnode->config.tsdbCfg.precision;
     int64_t now = taosGetTimestamp(precision);
-    int64_t offset = tsQueryRsmaTolerance * ((precision == TSDB_TIME_PRECISION_MILLI)   ? 1
-                                             : (precision == TSDB_TIME_PRECISION_MICRO) ? 1000
-                                                                                        : 1000000);
+    int64_t offset = tsQueryRsmaTolerance * ((precision == TSDB_TIME_PRECISION_MILLI)   ? 1L
+                                             : (precision == TSDB_TIME_PRECISION_MICRO) ? 1000L
+                                                                                        : 1000000L);
 
     for (int8_t i = 0; i < TSDB_RETENTION_MAX; ++i) {
       SRetention* pRetention = retentions + level;
@@ -3371,14 +3372,13 @@ int32_t tsdbReaderOpen(SVnode* pVnode, SQueryTableDataCond* pCond, SArray* pTabl
   if (pCond->suid != 0) {
     pReader->pSchema = metaGetTbTSchema(pReader->pTsdb->pVnode->pMeta, pReader->suid, -1, 1);
     if (pReader->pSchema == NULL) {
-      tsdbError("failed to get table schema, suid:%" PRIu64 ", ver:%" PRId64 " , %s", pReader->suid, -1,
-                pReader->idStr);
+      tsdbError("failed to get table schema, suid:%" PRIu64 ", ver:-1, %s", pReader->suid, pReader->idStr);
     }
   } else if (taosArrayGetSize(pTableList) > 0) {
     STableKeyInfo* pKey = taosArrayGet(pTableList, 0);
     pReader->pSchema = metaGetTbTSchema(pReader->pTsdb->pVnode->pMeta, pKey->uid, -1, 1);
     if (pReader->pSchema == NULL) {
-      tsdbError("failed to get table schema, uid:%" PRIu64 ", ver:%" PRId64 " , %s", pKey->uid, -1, pReader->idStr);
+      tsdbError("failed to get table schema, uid:%" PRIu64 ", ver:-1, %s", pKey->uid, pReader->idStr);
     }
   }
 
@@ -3394,19 +3394,20 @@ int32_t tsdbReaderOpen(SVnode* pVnode, SQueryTableDataCond* pCond, SArray* pTabl
     goto _err;
   }
 
-  code = tsdbTakeReadSnap(pReader->pTsdb, &pReader->pReadSnap, pReader->idStr);
-  if (code != TSDB_CODE_SUCCESS) {
-    goto _err;
-  }
-
-  if (pReader->type == TIMEWINDOW_RANGE_CONTAINED) {
-    code = doOpenReaderImpl(pReader);
+  if (numOfTables > 0) {
+    code = tsdbTakeReadSnap(pReader->pTsdb, &pReader->pReadSnap, pReader->idStr);
     if (code != TSDB_CODE_SUCCESS) {
-      return code;
+      goto _err;
     }
-  } else {
-    STsdbReader* pPrevReader = pReader->innerReader[0];
-    STsdbReader* pNextReader = pReader->innerReader[1];
+
+    if (pReader->type == TIMEWINDOW_RANGE_CONTAINED) {
+      code = doOpenReaderImpl(pReader);
+      if (code != TSDB_CODE_SUCCESS) {
+        return code;
+      }
+    } else {
+      STsdbReader* pPrevReader = pReader->innerReader[0];
+      STsdbReader* pNextReader = pReader->innerReader[1];
 
     // we need only one row
     pPrevReader->capacity = 1;
@@ -3421,19 +3422,20 @@ int32_t tsdbReaderOpen(SVnode* pVnode, SQueryTableDataCond* pCond, SArray* pTabl
     pNextReader->pMemSchema = pReader->pMemSchema;
     pNextReader->pReadSnap = pReader->pReadSnap;
 
-    code = doOpenReaderImpl(pPrevReader);
-    if (code != TSDB_CODE_SUCCESS) {
-      return code;
-    }
+      code = doOpenReaderImpl(pPrevReader);
+      if (code != TSDB_CODE_SUCCESS) {
+        return code;
+      }
 
-    code = doOpenReaderImpl(pNextReader);
-    if (code != TSDB_CODE_SUCCESS) {
-      return code;
-    }
+      code = doOpenReaderImpl(pNextReader);
+      if (code != TSDB_CODE_SUCCESS) {
+        return code;
+      }
 
-    code = doOpenReaderImpl(pReader);
-    if (code != TSDB_CODE_SUCCESS) {
-      return code;
+      code = doOpenReaderImpl(pReader);
+      if (code != TSDB_CODE_SUCCESS) {
+        return code;
+      }
     }
   }
 
@@ -3512,6 +3514,10 @@ void tsdbReaderClose(STsdbReader* pReader) {
     taosMemoryFree(pLReader);
   }
 
+  if (pReader->innerReader[0] != 0) {
+    tsdbUntakeReadSnap(pReader->innerReader[0]->pTsdb, pReader->innerReader[0]->pReadSnap, pReader->idStr);
+  }
+
   tsdbDebug(
       "%p :io-cost summary: head-file:%" PRIu64 ", head-file time:%.2f ms, SMA:%" PRId64
       " SMA-time:%.2f ms, fileBlocks:%" PRId64

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

@@ -18,7 +18,7 @@
 // =============== PAGE-WISE FILE ===============
 static int32_t tsdbOpenFile(const char *path, int32_t szPage, int32_t flag, STsdbFD **ppFD) {
   int32_t  code = 0;
-  STsdbFD *pFD;
+  STsdbFD *pFD = NULL;
 
   *ppFD = NULL;
 
@@ -35,6 +35,7 @@ static int32_t tsdbOpenFile(const char *path, int32_t szPage, int32_t flag, STsd
   pFD->pFD = taosOpenFile(path, flag);
   if (pFD->pFD == NULL) {
     code = TAOS_SYSTEM_ERROR(errno);
+    taosMemoryFree(pFD);
     goto _exit;
   }
   pFD->szPage = szPage;
@@ -42,11 +43,15 @@ static int32_t tsdbOpenFile(const char *path, int32_t szPage, int32_t flag, STsd
   pFD->pBuf = taosMemoryCalloc(1, szPage);
   if (pFD->pBuf == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;
+    taosCloseFile(&pFD->pFD);
     taosMemoryFree(pFD);
     goto _exit;
   }
   if (taosStatFile(path, &pFD->szFile, NULL) < 0) {
     code = TAOS_SYSTEM_ERROR(errno);
+    taosMemoryFree(pFD->pBuf);
+    taosCloseFile(&pFD->pFD);
+    taosMemoryFree(pFD);
     goto _exit;
   }
   ASSERT(pFD->szFile % szPage == 0);
@@ -59,10 +64,12 @@ _exit:
 
 static void tsdbCloseFile(STsdbFD **ppFD) {
   STsdbFD *pFD = *ppFD;
-  taosMemoryFree(pFD->pBuf);
-  taosCloseFile(&pFD->pFD);
-  taosMemoryFree(pFD);
-  *ppFD = NULL;
+  if (pFD) {
+    taosMemoryFree(pFD->pBuf);
+    taosCloseFile(&pFD->pFD);
+    taosMemoryFree(pFD);
+    *ppFD = NULL;
+  }
 }
 
 static int32_t tsdbWriteFilePage(STsdbFD *pFD) {
@@ -117,7 +124,7 @@ static int32_t tsdbReadFilePage(STsdbFD *pFD, int64_t pgno) {
   }
 
   // check
-  if (!taosCheckChecksumWhole(pFD->pBuf, pFD->szPage)) {
+  if (pgno > 1 && !taosCheckChecksumWhole(pFD->pBuf, pFD->szPage)) {
     code = TSDB_CODE_FILE_CORRUPTED;
     goto _exit;
   }
@@ -443,7 +450,7 @@ int32_t tsdbWriteDataBlk(SDataFWriter *pWriter, SMapData *mDataBlk, SBlockIdx *p
   pBlockIdx->size = size;
   pHeadFile->size += size;
 
-  tsdbTrace("vgId:%d, write block, file ID:%d commit ID:%d suid:%" PRId64 " uid:%" PRId64 " offset:%" PRId64
+  tsdbTrace("vgId:%d, write block, file ID:%d commit ID:%" PRId64 " suid:%" PRId64 " uid:%" PRId64 " offset:%" PRId64
             " size:%" PRId64 " nItem:%d",
             TD_VID(pWriter->pTsdb->pVnode), pWriter->wSet.fid, pHeadFile->commitID, pBlockIdx->suid, pBlockIdx->uid,
             pBlockIdx->offset, pBlockIdx->size, mDataBlk->nItem);
@@ -457,7 +464,7 @@ _err:
 int32_t tsdbWriteSttBlk(SDataFWriter *pWriter, SArray *aSttBlk) {
   int32_t   code = 0;
   SSttFile *pSttFile = &pWriter->fStt[pWriter->wSet.nSttF - 1];
-  int64_t   size;
+  int64_t   size = 0;
   int64_t   n;
 
   // check
@@ -836,7 +843,8 @@ _err:
 // SDataFReader ====================================================
 int32_t tsdbDataFReaderOpen(SDataFReader **ppReader, STsdb *pTsdb, SDFileSet *pSet) {
   int32_t       code = 0;
-  SDataFReader *pReader;
+  int32_t       lino = 0;
+  SDataFReader *pReader = NULL;
   int32_t       szPage = pTsdb->pVnode->config.tsdbPageSize;
   char          fname[TSDB_FILENAME_LEN];
 
@@ -844,7 +852,7 @@ int32_t tsdbDataFReaderOpen(SDataFReader **ppReader, STsdb *pTsdb, SDFileSet *pS
   pReader = (SDataFReader *)taosMemoryCalloc(1, sizeof(*pReader));
   if (pReader == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;
-    goto _err;
+    TSDB_CHECK_CODE(code, lino, _exit);
   }
   pReader->pTsdb = pTsdb;
   pReader->pSet = pSet;
@@ -852,31 +860,40 @@ int32_t tsdbDataFReaderOpen(SDataFReader **ppReader, STsdb *pTsdb, SDFileSet *pS
   // head
   tsdbHeadFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pHeadF, fname);
   code = tsdbOpenFile(fname, szPage, TD_FILE_READ, &pReader->pHeadFD);
-  if (code) goto _err;
+  TSDB_CHECK_CODE(code, lino, _exit);
 
   // data
   tsdbDataFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pDataF, fname);
   code = tsdbOpenFile(fname, szPage, TD_FILE_READ, &pReader->pDataFD);
-  if (code) goto _err;
+  TSDB_CHECK_CODE(code, lino, _exit);
 
   // sma
   tsdbSmaFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pSmaF, fname);
   code = tsdbOpenFile(fname, szPage, TD_FILE_READ, &pReader->pSmaFD);
-  if (code) goto _err;
+  TSDB_CHECK_CODE(code, lino, _exit);
 
   // stt
   for (int32_t iStt = 0; iStt < pSet->nSttF; iStt++) {
     tsdbSttFileName(pTsdb, pSet->diskId, pSet->fid, pSet->aSttF[iStt], fname);
     code = tsdbOpenFile(fname, szPage, TD_FILE_READ, &pReader->aSttFD[iStt]);
-    if (code) goto _err;
+    TSDB_CHECK_CODE(code, lino, _exit);
   }
 
-  *ppReader = pReader;
-  return code;
-
-_err:
-  tsdbError("vgId:%d, tsdb data file reader open failed since %s", TD_VID(pTsdb->pVnode), tstrerror(code));
-  *ppReader = NULL;
+_exit:
+  if (code) {
+    *ppReader = NULL;
+    tsdbError("vgId:%d %s failed at line %d since %s", TD_VID(pTsdb->pVnode), __func__, lino, tstrerror(code));
+
+    if (pReader) {
+      for (int32_t iStt = 0; iStt < pSet->nSttF; iStt++) tsdbCloseFile(&pReader->aSttFD[iStt]);
+      tsdbCloseFile(&pReader->pSmaFD);
+      tsdbCloseFile(&pReader->pDataFD);
+      tsdbCloseFile(&pReader->pHeadFD);
+      taosMemoryFree(pReader);
+    }
+  } else {
+    *ppReader = pReader;
+  }
   return code;
 }
 
@@ -906,10 +923,6 @@ int32_t tsdbDataFReaderClose(SDataFReader **ppReader) {
   taosMemoryFree(*ppReader);
   *ppReader = NULL;
   return code;
-
-_err:
-  tsdbError("vgId:%d, data file reader close failed since %s", TD_VID((*ppReader)->pTsdb->pVnode), tstrerror(code));
-  return code;
 }
 
 int32_t tsdbReadBlockIdx(SDataFReader *pReader, SArray *aBlockIdx) {
@@ -1289,16 +1302,17 @@ _exit:
 // SDelFWriter ====================================================
 int32_t tsdbDelFWriterOpen(SDelFWriter **ppWriter, SDelFile *pFile, STsdb *pTsdb) {
   int32_t      code = 0;
+  int32_t      lino = 0;
   char         fname[TSDB_FILENAME_LEN];
   uint8_t      hdr[TSDB_FHDR_SIZE] = {0};
-  SDelFWriter *pDelFWriter;
+  SDelFWriter *pDelFWriter = NULL;
   int64_t      n;
 
   // alloc
   pDelFWriter = (SDelFWriter *)taosMemoryCalloc(1, sizeof(*pDelFWriter));
   if (pDelFWriter == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;
-    goto _err;
+    TSDB_CHECK_CODE(code, lino, _exit);
   }
   pDelFWriter->pTsdb = pTsdb;
   pDelFWriter->fDel = *pFile;
@@ -1306,21 +1320,28 @@ int32_t tsdbDelFWriterOpen(SDelFWriter **ppWriter, SDelFile *pFile, STsdb *pTsdb
   tsdbDelFileName(pTsdb, pFile, fname);
   code = tsdbOpenFile(fname, pTsdb->pVnode->config.tsdbPageSize, TD_FILE_READ | TD_FILE_WRITE | TD_FILE_CREATE,
                       &pDelFWriter->pWriteH);
-  if (code) goto _err;
+  TSDB_CHECK_CODE(code, lino, _exit);
 
   // update header
   code = tsdbWriteFile(pDelFWriter->pWriteH, 0, hdr, TSDB_FHDR_SIZE);
-  if (code) goto _err;
+  TSDB_CHECK_CODE(code, lino, _exit);
 
   pDelFWriter->fDel.size = TSDB_FHDR_SIZE;
   pDelFWriter->fDel.offset = 0;
 
   *ppWriter = pDelFWriter;
-  return code;
 
-_err:
-  tsdbError("vgId:%d, failed to open del file writer since %s", TD_VID(pTsdb->pVnode), tstrerror(code));
-  *ppWriter = NULL;
+_exit:
+  if (code) {
+    if (pDelFWriter) {
+      taosMemoryFree(pDelFWriter);
+      tsdbCloseFile(&pDelFWriter->pWriteH);
+    }
+    *ppWriter = NULL;
+    tsdbError("vgId:%d %s failed at line %d since %s", TD_VID(pTsdb->pVnode), __func__, lino, tstrerror(errno));
+  } else {
+    *ppWriter = pDelFWriter;
+  }
   return code;
 }
 
@@ -1456,15 +1477,15 @@ struct SDelFReader {
 
 int32_t tsdbDelFReaderOpen(SDelFReader **ppReader, SDelFile *pFile, STsdb *pTsdb) {
   int32_t      code = 0;
+  int32_t      lino = 0;
   char         fname[TSDB_FILENAME_LEN];
-  SDelFReader *pDelFReader;
-  int64_t      n;
+  SDelFReader *pDelFReader = NULL;
 
   // alloc
   pDelFReader = (SDelFReader *)taosMemoryCalloc(1, sizeof(*pDelFReader));
   if (pDelFReader == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;
-    goto _err;
+    goto _exit;
   }
 
   // open impl
@@ -1473,14 +1494,18 @@ int32_t tsdbDelFReaderOpen(SDelFReader **ppReader, SDelFile *pFile, STsdb *pTsdb
 
   tsdbDelFileName(pTsdb, pFile, fname);
   code = tsdbOpenFile(fname, pTsdb->pVnode->config.tsdbPageSize, TD_FILE_READ, &pDelFReader->pReadH);
-  if (code) goto _err;
-
-  *ppReader = pDelFReader;
-  return code;
+  if (code) {
+    taosMemoryFree(pDelFReader);
+    goto _exit;
+  }
 
-_err:
-  tsdbError("vgId:%d, del file reader open failed since %s", TD_VID(pTsdb->pVnode), tstrerror(code));
-  *ppReader = NULL;
+_exit:
+  if (code) {
+    *ppReader = NULL;
+    tsdbError("vgId:%d %s failed at %d since %s", TD_VID(pTsdb->pVnode), __func__, lino, tstrerror(code));
+  } else {
+    *ppReader = pDelFReader;
+  }
   return code;
 }
 
@@ -1573,4 +1598,4 @@ int32_t tsdbReadDelIdx(SDelFReader *pReader, SArray *aDelIdx) {
 _err:
   tsdbError("vgId:%d, read del idx failed since %s", TD_VID(pReader->pTsdb->pVnode), tstrerror(code));
   return code;
-}
+}
\ No newline at end of file

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

@@ -34,7 +34,9 @@ int tsdbInsertData(STsdb *pTsdb, int64_t version, SSubmitReq *pMsg, SSubmitRsp *
   }
 
   // loop to insert
-  tInitSubmitMsgIter(pMsg, &msgIter);
+  if (tInitSubmitMsgIter(pMsg, &msgIter) < 0) {
+    return -1;
+  }
   while (true) {
     SSubmitBlkRsp r = {0};
     tGetSubmitMsgNext(&msgIter, &pBlock);

source/dnode/vnode/src/vnd/vnodeBufPool.c

@@ -112,6 +112,8 @@ void vnodeBufPoolReset(SVBufPool *pPool) {
 void *vnodeBufPoolMalloc(SVBufPool *pPool, int size) {
   SVBufPoolNode *pNode;
   void          *p = NULL;
+  ASSERT(pPool != NULL);
+
   taosThreadSpinLock(&pPool->lock);
   if (pPool->node.size >= pPool->ptr - pPool->node.data + size) {
     // allocate from the anchor node
@@ -141,17 +143,17 @@ void *vnodeBufPoolMalloc(SVBufPool *pPool, int size) {
 }
 
 void vnodeBufPoolFree(SVBufPool *pPool, void *p) {
-  uint8_t       *ptr = (uint8_t *)p;
-  SVBufPoolNode *pNode;
+  // uint8_t       *ptr = (uint8_t *)p;
+  // SVBufPoolNode *pNode;
 
-  if (ptr < pPool->node.data || ptr >= pPool->node.data + pPool->node.size) {
-    pNode = &((SVBufPoolNode *)p)[-1];
-    *pNode->pnext = pNode->prev;
-    pNode->prev->pnext = pNode->pnext;
+  // if (ptr < pPool->node.data || ptr >= pPool->node.data + pPool->node.size) {
+  //   pNode = &((SVBufPoolNode *)p)[-1];
+  //   *pNode->pnext = pNode->prev;
+  //   pNode->prev->pnext = pNode->pnext;
 
-    pPool->size = pPool->size - sizeof(*pNode) - pNode->size;
-    taosMemoryFree(pNode);
-  }
+  //   pPool->size = pPool->size - sizeof(*pNode) - pNode->size;
+  //   taosMemoryFree(pNode);
+  // }
 }
 
 void vnodeBufPoolRef(SVBufPool *pPool) {

source/libs/catalog/inc/ctgRemote.h

@@ -20,8 +20,6 @@
 extern "C" {
 #endif
 
-
-
 #ifdef __cplusplus
 }
 #endif

source/libs/catalog/src/ctgAsync.c

@@ -636,7 +636,7 @@ int32_t ctgInitJob(SCatalog* pCtg, SRequestConnInfo* pConn, SCtgJob** job, const
 
   taosAcquireRef(gCtgMgmt.jobPool, pJob->refId);
 
-  double el = (taosGetTimestampUs() - st)/1000.0;
+  double el = (taosGetTimestampUs() - st) / 1000.0;
   qDebug("QID:0x%" PRIx64 ", jobId: 0x%" PRIx64 " initialized, task num %d, forceUpdate %d, elapsed time:%.2f ms",
          pJob->queryId, pJob->refId, taskNum, pReq->forceUpdate, el);
   return TSDB_CODE_SUCCESS;