other: merge 3.0

Jenkinsfile2

@@ -173,8 +173,8 @@ def pre_test_build_mac() {
     '''
     sh '''
         cd ${WK}/debug
-        cmake .. -DBUILD_TEST=true
+        cmake .. -DBUILD_TEST=true -DBUILD_HTTPS=false
-        make -j8
+        make -j10
     '''
     sh '''
         date
@@ -428,7 +428,7 @@ pipeline {
                                         date
                                         rm -rf ${WKC}/debug
                                         cd ${WKC}/tests/parallel_test
-                                        time ./container_build.sh -w ${WKDIR} -t 8 -e
+                                        time ./container_build.sh -w ${WKDIR} -t 10 -e
                                         rm -f /tmp/cases.task
                                         ./collect_cases.sh -e
                                     '''

cmake/cmake.define

@@ -100,6 +100,8 @@ IF (TD_WINDOWS)
     SET(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${COMMON_FLAGS}")
     SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${COMMON_FLAGS}")
 
+    SET(JEMALLOC_ENABLED OFF)
+
 ELSE ()
     IF (${TD_DARWIN})
         set(CMAKE_MACOSX_RPATH 0)

cmake/taosadapter_CMakeLists.txt.in

@@ -2,7 +2,7 @@
 # taosadapter
 ExternalProject_Add(taosadapter
         GIT_REPOSITORY https://github.com/taosdata/taosadapter.git
-        GIT_TAG cc43ef0
+        GIT_TAG a11131c
         SOURCE_DIR "${TD_SOURCE_DIR}/tools/taosadapter"
         BINARY_DIR ""
         #BUILD_IN_SOURCE TRUE

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 cc973e0
+        GIT_TAG 7321fbb
         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 --with-malloc-conf='background_thread:true,metadata_thp:auto'
+        CONFIGURE_COMMAND   ./autogen.sh COMMAND ./configure --prefix=${CMAKE_BINARY_DIR}/build/ --disable-initial-exec-tls
         BUILD_COMMAND       ${MAKE}
     )
     INCLUDE_DIRECTORIES(${CMAKE_BINARY_DIR}/build/include)

docs/en/12-taos-sql/21-node.md

@@ -40,25 +40,25 @@ ALTER ALL DNODES dnode_option
 
 dnode_option: {
     'resetLog'
-  | 'balance' value
+  | 'balance' 'value'
-  | 'monitor' value
+  | 'monitor' 'value'
-  | 'debugFlag' value
+  | 'debugFlag' 'value'
-  | 'monDebugFlag' value
+  | 'monDebugFlag' 'value'
-  | 'vDebugFlag' value
+  | 'vDebugFlag' 'value'
-  | 'mDebugFlag' value
+  | 'mDebugFlag' 'value'
-  | 'cDebugFlag' value
+  | 'cDebugFlag' 'value'
-  | 'httpDebugFlag' value
+  | 'httpDebugFlag' 'value'
-  | 'qDebugflag' value
+  | 'qDebugflag' 'value'
-  | 'sdbDebugFlag' value
+  | 'sdbDebugFlag' 'value'
-  | 'uDebugFlag' value
+  | 'uDebugFlag' 'value'
-  | 'tsdbDebugFlag' value
+  | 'tsdbDebugFlag' 'value'
-  | 'sDebugflag' value
+  | 'sDebugflag' 'value'
-  | 'rpcDebugFlag' value
+  | 'rpcDebugFlag' 'value'
-  | 'dDebugFlag' value
+  | 'dDebugFlag' 'value'
-  | 'mqttDebugFlag' value
+  | 'mqttDebugFlag' 'value'
-  | 'wDebugFlag' value
+  | 'wDebugFlag' 'value'
-  | 'tmrDebugFlag' value
+  | 'tmrDebugFlag' 'value'
-  | 'cqDebugFlag' value
+  | 'cqDebugFlag' 'value'
 }
 ```
 
@@ -127,11 +127,11 @@ ALTER LOCAL local_option
 
 local_option: {
     'resetLog'
-  | 'rpcDebugFlag' value
+  | 'rpcDebugFlag' 'value'
-  | 'tmrDebugFlag' value
+  | 'tmrDebugFlag' 'value'
-  | 'cDebugFlag' value
+  | 'cDebugFlag' 'value'
-  | 'uDebugFlag' value
+  | 'uDebugFlag' 'value'
-  | 'debugFlag' value
+  | 'debugFlag' 'value'
 }
 ```
 

docs/en/14-reference/02-rest-api/02-rest-api.mdx

@@ -67,7 +67,7 @@ The following return value results indicate that the verification passed.
 ## HTTP request URL format
 
 ```text
-http://<fqdn>:<port>/rest/sql/[db_name]
+http://<fqdn>:<port>/rest/sql/[db_name][?tz=timezone]
 ```
 
 Parameter Description:
@@ -75,6 +75,7 @@ Parameter Description:
 - fqnd: FQDN or IP address of any host in the cluster.
 - port: httpPort configuration item in the configuration file, default is 6041.
 - db_name: Optional parameter that specifies the default database name for the executed SQL command.
+- tz: Optional parameter that specifies the timezone of the returned time, following the IANA Time Zone rules, e.g. `America/New_York`.
 
 For example, `http://h1.taos.com:6041/rest/sql/test` is a URL to `h1.taos.com:6041` and sets the default database name to `test`.
 
@@ -97,13 +98,13 @@ The HTTP request's BODY is a complete SQL command, and the data table in the SQL
 Use `curl` to initiate an HTTP request with a custom authentication method, with the following syntax.
 
 ```bash
-curl -L -H "Authorization: Basic <TOKEN>" -d "<SQL>" <ip>:<PORT>/rest/sql/[db_name]
+curl -L -H "Authorization: Basic <TOKEN>" -d "<SQL>" <ip>:<PORT>/rest/sql/[db_name][?tz=timezone]
 ```
 
 or
 
 ```bash
-curl -L -u username:password -d "<SQL>" <ip>:<PORT>/rest/sql/[db_name]
+curl -L -u username:password -d "<SQL>" <ip>:<PORT>/rest/sql/[db_name][?tz=timezone]
 ```
 
 where `TOKEN` is the string after Base64 encoding of `{username}:{password}`, e.g. `root:taosdata` is encoded as `cm9vdDp0YW9zZGF0YQ==`..

docs/en/14-reference/03-connector/05-go.mdx

@@ -353,7 +353,7 @@ The `af` package encapsulates TDengine advanced functions such as connection man
 
   Closes the connection.
 
-#### Subscribe to
+#### Subscribe
 
 * `func NewConsumer(conf *Config) (*Consumer, error)`
 
@@ -387,7 +387,7 @@ Close consumer.
 
 * `func (conn *Connector) InfluxDBInsertLines(lines []string, precision string) error`
 
-  Write to influxDB line protocol.
+  Write to InfluxDB line protocol.
 
 * `func (conn *Connector) OpenTSDBInsertTelnetLines(lines []string) error`
 
@@ -413,7 +413,7 @@ Close consumer.
 
 * `func (stmt *InsertStmt) SetTableName(name string) error`
 
-  Bind the set table name parameter.
+  Bind the table name parameter.
 
 * `func (stmt *InsertStmt) SetSubTableName(name string) error`
 
@@ -463,6 +463,54 @@ Close consumer.
 
 For a complete example see [GitHub sample file](https://github.com/taosdata/driver-go/blob/3.0/examples/tmqoverws/main.go)
 
+### parameter binding via WebSocket
+
+* `func NewConnector(config *Config) (*Connector, error)`
+
+  Create a connection.
+
+* `func (c *Connector) Init() (*Stmt, error)`
+
+  Initialize the parameters.
+
+* `func (c *Connector) Close() error`
+
+  Close the connection.
+
+* `func (s *Stmt) Prepare(sql string) error`
+
+  Parameter binding preprocessing SQL statement.
+
+* `func (s *Stmt) SetTableName(name string) error`
+
+  Bind the table name parameter.
+
+* `func (s *Stmt) SetTags(tags *param.Param, bindType *param.ColumnType) error`
+
+  Set tags.
+
+* `func (s *Stmt) BindParam(params []*param.Param, bindType *param.ColumnType) error`
+
+  Parameter bind multiple rows of data.
+
+* `func (s *Stmt) AddBatch() error`
+
+  Add to a parameter-bound batch.
+
+* `func (s *Stmt) Exec() error`
+
+  Execute a parameter binding.
+
+* `func (s *Stmt) GetAffectedRows() int`
+
+  Gets the number of affected rows inserted by the parameter binding.
+
+* `func (s *Stmt) Close() error`
+
+  Closes the parameter binding.
+
+For a complete example see [GitHub sample file](https://github.com/taosdata/driver-go/blob/3.0/examples/stmtoverws/main.go)
+
 ## API Reference
 
 Full API see [driver-go documentation](https://pkg.go.dev/github.com/taosdata/driver-go/v3)

docs/en/27-train-faq/01-faq.md

@@ -106,11 +106,11 @@ ALTER LOCAL local_option
  
 local_option: {
     'resetLog'
-  | 'rpcDebugFlag' value
+  | 'rpcDebugFlag' 'value'
-  | 'tmrDebugFlag' value
+  | 'tmrDebugFlag' 'value'
-  | 'cDebugFlag' value
+  | 'cDebugFlag' 'value'
-  | 'uDebugFlag' value
+  | 'uDebugFlag' 'value'
-  | 'debugFlag' value
+  | 'debugFlag' 'value'
 }
 ```
 

docs/en/28-releases/index.md

@@ -0,0 +1,10 @@
+---
+title: Releases
+---
+
+```mdx-code-block
+import DocCardList from '@theme/DocCardList';
+import {useCurrentSidebarCategory} from '@docusaurus/theme-common';
+
+<DocCardList items={useCurrentSidebarCategory().items}/>
+```

docs/zh/08-connector/02-rest-api.mdx

@@ -69,7 +69,7 @@ curl -L -H "Authorization: Basic cm9vdDp0YW9zZGF0YQ==" \
 ## HTTP 请求格式
 
 ```text
-http://<fqdn>:<port>/rest/sql/[db_name]
+http://<fqdn>:<port>/rest/sql/[db_name][?tz=timezone]
 ```
 
 参数说明：
@@ -77,6 +77,7 @@ http://<fqdn>:<port>/rest/sql/[db_name]
 - fqdn: 集群中的任一台主机 FQDN 或 IP 地址。
 - port: 配置文件中 httpPort 配置项，缺省为 6041。
 - db_name: 可选参数，指定本次所执行的 SQL 语句的默认数据库库名。
+- tz: 可选参数，指定返回时间的时区，遵照 IANA Time Zone 规则，如 `America/New_York`。
 
 例如：`http://h1.taos.com:6041/rest/sql/test` 是指向地址为 `h1.taos.com:6041` 的 URL，并将默认使用的数据库库名设置为 `test`。
 
@@ -99,13 +100,13 @@ HTTP 请求的 BODY 里就是一个完整的 SQL 语句，SQL 语句中的数据
 使用 `curl` 通过自定义身份认证方式来发起一个 HTTP Request，语法如下：
 
 ```bash
-curl -L -H "Authorization: Basic <TOKEN>" -d "<SQL>" <ip>:<PORT>/rest/sql/[db_name]
+curl -L -H "Authorization: Basic <TOKEN>" -d "<SQL>" <ip>:<PORT>/rest/sql/[db_name][?tz=timezone]
 ```
 
 或者，
 
 ```bash
-curl -L -u username:password -d "<SQL>" <ip>:<PORT>/rest/sql/[db_name]
+curl -L -u username:password -d "<SQL>" <ip>:<PORT>/rest/sql/[db_name][?tz=timezone]
 ```
 
 其中，`TOKEN` 为 `{username}:{password}` 经过 Base64 编码之后的字符串，例如 `root:taosdata` 编码后为 `cm9vdDp0YW9zZGF0YQ==`。

docs/zh/08-connector/20-go.mdx

@@ -8,7 +8,7 @@ title: TDengine Go Connector
 import Tabs from '@theme/Tabs';
 import TabItem from '@theme/TabItem';
 
-import Preparition from "./_preparition.mdx"
+import Preparation from "./_preparation.mdx"
 import GoInsert from "../07-develop/03-insert-data/_go_sql.mdx"
 import GoInfluxLine from "../07-develop/03-insert-data/_go_line.mdx"
 import GoOpenTSDBTelnet from "../07-develop/03-insert-data/_go_opts_telnet.mdx"
@@ -389,7 +389,7 @@ func main() {
 
 * `func (conn *Connector) InfluxDBInsertLines(lines []string, precision string) error`
 
-  写入 influxDB 行协议。
+  写入 InfluxDB 行协议。
 
 * `func (conn *Connector) OpenTSDBInsertTelnetLines(lines []string) error`
 
@@ -465,6 +465,54 @@ func main() {
 
 完整订阅示例参见 [GitHub 示例文件](https://github.com/taosdata/driver-go/blob/3.0/examples/tmqoverws/main.go)
 
+### 通过 WebSocket 进行参数绑定
+
+* `func NewConnector(config *Config) (*Connector, error)`
+
+  创建连接。
+
+* `func (c *Connector) Init() (*Stmt, error)`
+
+  初始化参数。
+
+* `func (c *Connector) Close() error`
+
+  关闭连接。
+
+* `func (s *Stmt) Prepare(sql string) error`
+
+  参数绑定预处理 SQL 语句。
+
+* `func (s *Stmt) SetTableName(name string) error`
+
+  参数绑定设置表名。
+
+* `func (s *Stmt) SetTags(tags *param.Param, bindType *param.ColumnType) error`
+
+  参数绑定设置标签。
+
+* `func (s *Stmt) BindParam(params []*param.Param, bindType *param.ColumnType) error`
+
+  参数绑定多行数据。
+
+* `func (s *Stmt) AddBatch() error`
+
+  添加到参数绑定批处理。
+
+* `func (s *Stmt) Exec() error`
+
+  执行参数绑定。
+
+* `func (s *Stmt) GetAffectedRows() int`
+
+  获取参数绑定插入受影响行数。
+
+* `func (s *Stmt) Close() error`
+
+  结束参数绑定。
+
+完整参数绑定示例参见 [GitHub 示例文件](https://github.com/taosdata/driver-go/blob/3.0/examples/stmtoverws/main.go)
+
 ## API 参考
 
 全部 API 见 [driver-go 文档](https://pkg.go.dev/github.com/taosdata/driver-go/v3)

docs/zh/08-connector/26-rust.mdx

@@ -7,7 +7,7 @@ title: TDengine Rust Connector
 import Tabs from '@theme/Tabs';
 import TabItem from '@theme/TabItem';
 
-import Preparition from "./_preparition.mdx"
+import Preparation from "./_preparation.mdx"
 import RustInsert from "../07-develop/03-insert-data/_rust_sql.mdx"
 import RustBind from "../07-develop/03-insert-data/_rust_stmt.mdx"
 import RustQuery from "../07-develop/04-query-data/_rust.mdx"

docs/zh/08-connector/35-node.mdx

@@ -7,7 +7,7 @@ title: TDengine Node.js Connector
 import Tabs from "@theme/Tabs";
 import TabItem from "@theme/TabItem";
 
-import Preparition from "./_preparition.mdx";
+import Preparation from "./_preparation.mdx";
 import NodeInsert from "../07-develop/03-insert-data/_js_sql.mdx";
 import NodeInfluxLine from "../07-develop/03-insert-data/_js_line.mdx";
 import NodeOpenTSDBTelnet from "../07-develop/03-insert-data/_js_opts_telnet.mdx";

docs/zh/08-connector/40-csharp.mdx

@@ -7,7 +7,7 @@ title: C# Connector
 import Tabs from '@theme/Tabs';
 import TabItem from '@theme/TabItem';
 
-import Preparition from "./_preparition.mdx"
+import Preparation from "./_preparation.mdx"
 import CSInsert from "../07-develop/03-insert-data/_cs_sql.mdx"
 import CSInfluxLine from "../07-develop/03-insert-data/_cs_line.mdx"
 import CSOpenTSDBTelnet from "../07-develop/03-insert-data/_cs_opts_telnet.mdx"

docs/zh/12-taos-sql/21-node.md

@@ -41,25 +41,25 @@ ALTER ALL DNODES dnode_option
 
 dnode_option: {
     'resetLog'
-  | 'balance' value
+  | 'balance' 'value'
-  | 'monitor' value
+  | 'monitor' 'value'
-  | 'debugFlag' value
+  | 'debugFlag' 'value'
-  | 'monDebugFlag' value
+  | 'monDebugFlag' 'value'
-  | 'vDebugFlag' value
+  | 'vDebugFlag' 'value'
-  | 'mDebugFlag' value
+  | 'mDebugFlag' 'value'
-  | 'cDebugFlag' value
+  | 'cDebugFlag' 'value'
-  | 'httpDebugFlag' value
+  | 'httpDebugFlag' 'value'
-  | 'qDebugflag' value
+  | 'qDebugflag' 'value'
-  | 'sdbDebugFlag' value
+  | 'sdbDebugFlag' 'value'
-  | 'uDebugFlag' value
+  | 'uDebugFlag' 'value'
-  | 'tsdbDebugFlag' value
+  | 'tsdbDebugFlag' 'value'
-  | 'sDebugflag' value
+  | 'sDebugflag' 'value'
-  | 'rpcDebugFlag' value
+  | 'rpcDebugFlag' 'value'
-  | 'dDebugFlag' value
+  | 'dDebugFlag' 'value'
-  | 'mqttDebugFlag' value
+  | 'mqttDebugFlag' 'value'
-  | 'wDebugFlag' value
+  | 'wDebugFlag' 'value'
-  | 'tmrDebugFlag' value
+  | 'tmrDebugFlag' 'value'
-  | 'cqDebugFlag' value
+  | 'cqDebugFlag' 'value'
 }
 ```
 
@@ -128,11 +128,11 @@ ALTER LOCAL local_option
 
 local_option: {
     'resetLog'
-  | 'rpcDebugFlag' value
+  | 'rpcDebugFlag' 'value'
-  | 'tmrDebugFlag' value
+  | 'tmrDebugFlag' 'value'
-  | 'cDebugFlag' value
+  | 'cDebugFlag' 'value'
-  | 'uDebugFlag' value
+  | 'uDebugFlag' 'value'
-  | 'debugFlag' value
+  | 'debugFlag' 'value'
 }
 ```
 

docs/zh/12-taos-sql/24-show.md

@@ -201,7 +201,7 @@ SHOW TOPICS;
 SHOW TRANSACTIONS;
 ```
 
-显示当前系统中正在执行的事务的信息
+显示当前系统中正在执行的事务的信息(该事务仅针对除普通表以外的元数据级别)
 
 ## SHOW USERS
 

docs/zh/27-train-faq/01-faq.md

@@ -135,11 +135,11 @@ ALTER LOCAL local_option
  
 local_option: {
     'resetLog'
-  | 'rpcDebugFlag' value
+  | 'rpcDebugFlag' 'value'
-  | 'tmrDebugFlag' value
+  | 'tmrDebugFlag' 'value'
-  | 'cDebugFlag' value
+  | 'cDebugFlag' 'value'
-  | 'uDebugFlag' value
+  | 'uDebugFlag' 'value'
-  | 'debugFlag' value
+  | 'debugFlag' 'value'
 }
 ```
 

examples/c/tmq.c

@@ -47,7 +47,7 @@ static int32_t msg_process(TAOS_RES* msg) {
     int32_t     precision = taos_result_precision(msg);
     rows++;
     taos_print_row(buf, row, fields, numOfFields);
-    printf("row content: %s\n", buf);
+    printf("precision: %d, row content: %s\n", precision, buf);
   }
 
   return rows;
@@ -70,7 +70,7 @@ static int32_t init_env() {
   taos_free_result(pRes);
 
   // create database
-  pRes = taos_query(pConn, "create database tmqdb");
+  pRes = taos_query(pConn, "create database tmqdb precision 'ns'");
   if (taos_errno(pRes) != 0) {
     printf("error in create tmqdb, reason:%s\n", taos_errstr(pRes));
     return -1;

include/common/tcommon.h

@@ -161,6 +161,7 @@ typedef enum EStreamType {
   STREAM_RETRIEVE,
   STREAM_PULL_DATA,
   STREAM_PULL_OVER,
+  STREAM_FILL_OVER,
 } EStreamType;
 
 typedef struct {

include/common/tglobal.h

@@ -60,8 +60,8 @@ extern int32_t tsNumOfVnodeSyncThreads;
 extern int32_t tsNumOfVnodeRsmaThreads;
 extern int32_t tsNumOfQnodeQueryThreads;
 extern int32_t tsNumOfQnodeFetchThreads;
-extern int32_t tsNumOfSnodeSharedThreads;
+extern int32_t tsNumOfSnodeStreamThreads;
-extern int32_t tsNumOfSnodeUniqueThreads;
+extern int32_t tsNumOfSnodeWriteThreads;
 extern int64_t tsRpcQueueMemoryAllowed;
 
 // monitor
@@ -91,6 +91,7 @@ extern bool    tsQueryPlannerTrace;
 extern int32_t tsQueryNodeChunkSize;
 extern bool    tsQueryUseNodeAllocator;
 extern bool    tsKeepColumnName;
+extern bool    tsEnableQueryHb;
 
 // client
 extern int32_t tsMinSlidingTime;

include/common/tmsg.h

@@ -1726,6 +1726,9 @@ typedef struct {
 #define STREAM_TRIGGER_WINDOW_CLOSE   2
 #define STREAM_TRIGGER_MAX_DELAY      3
 #define STREAM_DEFAULT_IGNORE_EXPIRED 0
+#define STREAM_FILL_HISTORY_ON        1
+#define STREAM_FILL_HISTORY_OFF       0
+#define STREAM_DEFAULT_FILL_HISTORY   STREAM_FILL_HISTORY_OFF
 
 typedef struct {
   char    name[TSDB_STREAM_FNAME_LEN];

include/common/tmsgdef.h

@@ -198,6 +198,9 @@ enum {
   TD_DEF_MSG_TYPE(TDMT_VND_CONSUME, "vnode-consume", SMqPollReq, SMqDataBlkRsp)
   TD_DEF_MSG_TYPE(TDMT_VND_STREAM_TRIGGER, "vnode-stream-trigger", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_VND_STREAM_DISPATCH_WRITE, "vnode-stream-task-dispatch-write", NULL, NULL)
+  TD_DEF_MSG_TYPE(TDMT_VND_STREAM_RECOVER_STEP1, "vnode-stream-recover1", NULL, NULL)
+  TD_DEF_MSG_TYPE(TDMT_VND_STREAM_RECOVER_STEP2, "vnode-stream-recover2", NULL, NULL)
+  TD_DEF_MSG_TYPE(TDMT_VND_STREAM_RECOVER_FINISH, "vnode-stream-finish", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_VND_CREATE_SMA, "vnode-create-sma", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_VND_CANCEL_SMA, "vnode-cancel-sma", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_VND_DROP_SMA, "vnode-drop-sma", NULL, NULL)

include/common/trow.h

@@ -169,7 +169,7 @@ typedef struct {
 // #define TD_ROW_VER(r)      ((r)->ver)
 #define TD_ROW_KEY_ADDR(r) (r)
 
-// N.B. If without STSchema, getExtendedRowSize() is used to get the rowMaxBytes and
+// N.B. If without STSchema, insGetExtendedRowSize() is used to get the rowMaxBytes and
 // (int32_t)ceil((double)nCols/TD_VTYPE_PARTS) should be added if TD_SUPPORT_BITMAP defined.
 #define TD_ROW_MAX_BYTES_FROM_SCHEMA(s) ((s)->tlen + TD_ROW_HEAD_LEN)
 

include/common/ttokendef.h

@@ -164,9 +164,9 @@
 #define TK_SCORES               146
 #define TK_TOPICS               147
 #define TK_VARIABLES            148
-#define TK_BNODES               149
+#define TK_CLUSTER              149
-#define TK_SNODES               150
+#define TK_BNODES               150
-#define TK_CLUSTER              151
+#define TK_SNODES               151
 #define TK_TRANSACTIONS         152
 #define TK_DISTRIBUTED          153
 #define TK_CONSUMERS            154
@@ -203,135 +203,136 @@
 #define TK_WINDOW_CLOSE         185
 #define TK_IGNORE               186
 #define TK_EXPIRED              187
-#define TK_SUBTABLE             188
+#define TK_FILL_HISTORY         188
-#define TK_KILL                 189
+#define TK_SUBTABLE             189
-#define TK_CONNECTION           190
+#define TK_KILL                 190
-#define TK_TRANSACTION          191
+#define TK_CONNECTION           191
-#define TK_BALANCE              192
+#define TK_TRANSACTION          192
-#define TK_VGROUP               193
+#define TK_BALANCE              193
-#define TK_MERGE                194
+#define TK_VGROUP               194
-#define TK_REDISTRIBUTE         195
+#define TK_MERGE                195
-#define TK_SPLIT                196
+#define TK_REDISTRIBUTE         196
-#define TK_DELETE               197
+#define TK_SPLIT                197
-#define TK_INSERT               198
+#define TK_DELETE               198
-#define TK_NULL                 199
+#define TK_INSERT               199
-#define TK_NK_QUESTION          200
+#define TK_NULL                 200
-#define TK_NK_ARROW             201
+#define TK_NK_QUESTION          201
-#define TK_ROWTS                202
+#define TK_NK_ARROW             202
-#define TK_TBNAME               203
+#define TK_ROWTS                203
-#define TK_QSTART               204
+#define TK_TBNAME               204
-#define TK_QEND                 205
+#define TK_QSTART               205
-#define TK_QDURATION            206
+#define TK_QEND                 206
-#define TK_WSTART               207
+#define TK_QDURATION            207
-#define TK_WEND                 208
+#define TK_WSTART               208
-#define TK_WDURATION            209
+#define TK_WEND                 209
-#define TK_IROWTS               210
+#define TK_WDURATION            210
-#define TK_QTAGS                211
+#define TK_IROWTS               211
-#define TK_CAST                 212
+#define TK_QTAGS                212
-#define TK_NOW                  213
+#define TK_CAST                 213
-#define TK_TODAY                214
+#define TK_NOW                  214
-#define TK_TIMEZONE             215
+#define TK_TODAY                215
-#define TK_CLIENT_VERSION       216
+#define TK_TIMEZONE             216
-#define TK_SERVER_VERSION       217
+#define TK_CLIENT_VERSION       217
-#define TK_SERVER_STATUS        218
+#define TK_SERVER_VERSION       218
-#define TK_CURRENT_USER         219
+#define TK_SERVER_STATUS        219
-#define TK_COUNT                220
+#define TK_CURRENT_USER         220
-#define TK_LAST_ROW             221
+#define TK_COUNT                221
-#define TK_CASE                 222
+#define TK_LAST_ROW             222
-#define TK_END                  223
+#define TK_CASE                 223
-#define TK_WHEN                 224
+#define TK_END                  224
-#define TK_THEN                 225
+#define TK_WHEN                 225
-#define TK_ELSE                 226
+#define TK_THEN                 226
-#define TK_BETWEEN              227
+#define TK_ELSE                 227
-#define TK_IS                   228
+#define TK_BETWEEN              228
-#define TK_NK_LT                229
+#define TK_IS                   229
-#define TK_NK_GT                230
+#define TK_NK_LT                230
-#define TK_NK_LE                231
+#define TK_NK_GT                231
-#define TK_NK_GE                232
+#define TK_NK_LE                232
-#define TK_NK_NE                233
+#define TK_NK_GE                233
-#define TK_MATCH                234
+#define TK_NK_NE                234
-#define TK_NMATCH               235
+#define TK_MATCH                235
-#define TK_CONTAINS             236
+#define TK_NMATCH               236
-#define TK_IN                   237
+#define TK_CONTAINS             237
-#define TK_JOIN                 238
+#define TK_IN                   238
-#define TK_INNER                239
+#define TK_JOIN                 239
-#define TK_SELECT               240
+#define TK_INNER                240
-#define TK_DISTINCT             241
+#define TK_SELECT               241
-#define TK_WHERE                242
+#define TK_DISTINCT             242
-#define TK_PARTITION            243
+#define TK_WHERE                243
-#define TK_BY                   244
+#define TK_PARTITION            244
-#define TK_SESSION              245
+#define TK_BY                   245
-#define TK_STATE_WINDOW         246
+#define TK_SESSION              246
-#define TK_SLIDING              247
+#define TK_STATE_WINDOW         247
-#define TK_FILL                 248
+#define TK_SLIDING              248
-#define TK_VALUE                249
+#define TK_FILL                 249
-#define TK_NONE                 250
+#define TK_VALUE                250
-#define TK_PREV                 251
+#define TK_NONE                 251
-#define TK_LINEAR               252
+#define TK_PREV                 252
-#define TK_NEXT                 253
+#define TK_LINEAR               253
-#define TK_HAVING               254
+#define TK_NEXT                 254
-#define TK_RANGE                255
+#define TK_HAVING               255
-#define TK_EVERY                256
+#define TK_RANGE                256
-#define TK_ORDER                257
+#define TK_EVERY                257
-#define TK_SLIMIT               258
+#define TK_ORDER                258
-#define TK_SOFFSET              259
+#define TK_SLIMIT               259
-#define TK_LIMIT                260
+#define TK_SOFFSET              260
-#define TK_OFFSET               261
+#define TK_LIMIT                261
-#define TK_ASC                  262
+#define TK_OFFSET               262
-#define TK_NULLS                263
+#define TK_ASC                  263
-#define TK_ABORT                264
+#define TK_NULLS                264
-#define TK_AFTER                265
+#define TK_ABORT                265
-#define TK_ATTACH               266
+#define TK_AFTER                266
-#define TK_BEFORE               267
+#define TK_ATTACH               267
-#define TK_BEGIN                268
+#define TK_BEFORE               268
-#define TK_BITAND               269
+#define TK_BEGIN                269
-#define TK_BITNOT               270
+#define TK_BITAND               270
-#define TK_BITOR                271
+#define TK_BITNOT               271
-#define TK_BLOCKS               272
+#define TK_BITOR                272
-#define TK_CHANGE               273
+#define TK_BLOCKS               273
-#define TK_COMMA                274
+#define TK_CHANGE               274
-#define TK_COMPACT              275
+#define TK_COMMA                275
-#define TK_CONCAT               276
+#define TK_COMPACT              276
-#define TK_CONFLICT             277
+#define TK_CONCAT               277
-#define TK_COPY                 278
+#define TK_CONFLICT             278
-#define TK_DEFERRED             279
+#define TK_COPY                 279
-#define TK_DELIMITERS           280
+#define TK_DEFERRED             280
-#define TK_DETACH               281
+#define TK_DELIMITERS           281
-#define TK_DIVIDE               282
+#define TK_DETACH               282
-#define TK_DOT                  283
+#define TK_DIVIDE               283
-#define TK_EACH                 284
+#define TK_DOT                  284
-#define TK_FAIL                 285
+#define TK_EACH                 285
-#define TK_FILE                 286
+#define TK_FAIL                 286
-#define TK_FOR                  287
+#define TK_FILE                 287
-#define TK_GLOB                 288
+#define TK_FOR                  288
-#define TK_ID                   289
+#define TK_GLOB                 289
-#define TK_IMMEDIATE            290
+#define TK_ID                   290
-#define TK_IMPORT               291
+#define TK_IMMEDIATE            291
-#define TK_INITIALLY            292
+#define TK_IMPORT               292
-#define TK_INSTEAD              293
+#define TK_INITIALLY            293
-#define TK_ISNULL               294
+#define TK_INSTEAD              294
-#define TK_KEY                  295
+#define TK_ISNULL               295
-#define TK_MODULES              296
+#define TK_KEY                  296
-#define TK_NK_BITNOT            297
+#define TK_MODULES              297
-#define TK_NK_SEMI              298
+#define TK_NK_BITNOT            298
-#define TK_NOTNULL              299
+#define TK_NK_SEMI              299
-#define TK_OF                   300
+#define TK_NOTNULL              300
-#define TK_PLUS                 301
+#define TK_OF                   301
-#define TK_PRIVILEGE            302
+#define TK_PLUS                 302
-#define TK_RAISE                303
+#define TK_PRIVILEGE            303
-#define TK_REPLACE              304
+#define TK_RAISE                304
-#define TK_RESTRICT             305
+#define TK_REPLACE              305
-#define TK_ROW                  306
+#define TK_RESTRICT             306
-#define TK_SEMI                 307
+#define TK_ROW                  307
-#define TK_STAR                 308
+#define TK_SEMI                 308
-#define TK_STATEMENT            309
+#define TK_STAR                 309
-#define TK_STRING               310
+#define TK_STATEMENT            310
-#define TK_TIMES                311
+#define TK_STRING               311
-#define TK_UPDATE               312
+#define TK_TIMES                312
-#define TK_VALUES               313
+#define TK_UPDATE               313
-#define TK_VARIABLE             314
+#define TK_VALUES               314
-#define TK_VIEW                 315
+#define TK_VARIABLE             315
-#define TK_WAL                  316
+#define TK_VIEW                 316
+#define TK_WAL                  317
 
 #define TK_NK_SPACE   300
 #define TK_NK_COMMENT 301

include/common/ttypes.h

@@ -290,6 +290,7 @@ typedef struct {
   (IS_SIGNED_NUMERIC_TYPE(_t) || (_t) == (TSDB_DATA_TYPE_BOOL) || (_t) == (TSDB_DATA_TYPE_TIMESTAMP))
 #define IS_CONVERT_AS_UNSIGNED(_t) (IS_UNSIGNED_NUMERIC_TYPE(_t) || (_t) == (TSDB_DATA_TYPE_BOOL))
 
+#if 0
 // TODO remove this function
 static FORCE_INLINE bool isNull(const void *val, int32_t type) {
   switch (type) {
@@ -325,6 +326,7 @@ static FORCE_INLINE bool isNull(const void *val, int32_t type) {
       return false;
   };
 }
+#endif
 
 typedef struct tDataTypeDescriptor {
   int16_t type;

include/dnode/snode/snode.h

@@ -68,8 +68,8 @@ int32_t sndGetLoad(SSnode *pSnode, SSnodeLoad *pLoad);
  * @param pMsg The request message
  * @param pRsp The response message
  */
-int32_t sndProcessUMsg(SSnode *pSnode, SRpcMsg *pMsg);
+int32_t sndProcessWriteMsg(SSnode *pSnode, SRpcMsg *pMsg, SRpcMsg *pRsp);
-int32_t sndProcessSMsg(SSnode *pSnode, SRpcMsg *pMsg);
+int32_t sndProcessStreamMsg(SSnode *pSnode, SRpcMsg *pMsg);
 
 #ifdef __cplusplus
 }

include/libs/catalog/catalog.h

@@ -203,6 +203,14 @@ int32_t catalogUpdateTableMeta(SCatalog* pCatalog, STableMetaRsp* rspMsg);
 
 int32_t catalogUpdateTableMeta(SCatalog* pCatalog, STableMetaRsp* rspMsg);
 
+int32_t catalogGetCachedTableMeta(SCatalog* pCtg, SRequestConnInfo* pConn, const SName* pTableName, STableMeta** pTableMeta);
+
+int32_t catalogGetCachedSTableMeta(SCatalog* pCtg, SRequestConnInfo* pConn, const SName* pTableName,
+                            STableMeta** pTableMeta);
+
+int32_t catalogGetCachedTableHashVgroup(SCatalog* pCtg, SRequestConnInfo* pConn, const SName* pTableName,
+                                  SVgroupInfo* pVgroup, bool* exists);
+
 /**
  * Force refresh DB's local cached vgroup info.
  * @param pCtg (input, got with catalogGetHandle)

include/libs/executor/executor.h

@@ -89,6 +89,16 @@ qTaskInfo_t qCreateQueueExecTaskInfo(void* msg, SReadHandle* readers, int32_t* n
  */
 int32_t qSetMultiStreamInput(qTaskInfo_t tinfo, const void* pBlocks, size_t numOfBlocks, int32_t type);
 
+/**
+ * Set block for sma
+ * @param tinfo
+ * @param pBlocks
+ * @param numOfInputBlock
+ * @param type
+ * @return
+ */
+int32_t qSetSMAInput(qTaskInfo_t tinfo, const void* pBlocks, size_t numOfBlocks, int32_t type);
+
 /**
  * Update the table id list, add or remove.
  *
@@ -169,6 +179,7 @@ int32_t qSerializeTaskStatus(qTaskInfo_t tinfo, char** pOutput, int32_t* len);
 
 int32_t qDeserializeTaskStatus(qTaskInfo_t tinfo, const char* pInput, int32_t len);
 
+STimeWindow getAlignQueryTimeWindow(SInterval* pInterval, int32_t precision, int64_t key);
 /**
  * return the scan info, in the form of tuple of two items, including table uid and current timestamp
  * @param tinfo
@@ -200,9 +211,11 @@ int32_t qExtractStreamScanner(qTaskInfo_t tinfo, void** scanner);
 
 int32_t qStreamInput(qTaskInfo_t tinfo, void* pItem);
 
-int32_t qStreamPrepareRecover(qTaskInfo_t tinfo, int64_t startVer, int64_t endVer);
+int32_t qStreamSetParamForRecover(qTaskInfo_t tinfo);
-
+int32_t qStreamSourceRecoverStep1(qTaskInfo_t tinfo, int64_t ver);
-STimeWindow getAlignQueryTimeWindow(SInterval* pInterval, int32_t precision, int64_t key);
+int32_t qStreamSourceRecoverStep2(qTaskInfo_t tinfo, int64_t ver);
+int32_t qStreamRecoverFinish(qTaskInfo_t tinfo);
+int32_t qStreamRestoreParam(qTaskInfo_t tinfo);
 
 #ifdef __cplusplus
 }

include/libs/nodes/cmdnodes.h

@@ -274,6 +274,7 @@ typedef struct SShowTableDistributedStmt {
 typedef struct SShowDnodeVariablesStmt {
   ENodeType type;
   SNode*    pDnodeId;
+  SNode*    pLikePattern;
 } SShowDnodeVariablesStmt;
 
 typedef struct SShowVnodesStmt {
@@ -373,6 +374,7 @@ typedef struct SStreamOptions {
   int8_t    triggerType;
   SNode*    pDelay;
   SNode*    pWatermark;
+  int8_t    fillHistory;
   int8_t    ignoreExpired;
 } SStreamOptions;
 

include/libs/nodes/plannodes.h

@@ -567,6 +567,7 @@ typedef struct SSubplan {
   SDataSinkNode* pDataSink;     // data of the subplan flow into the datasink
   SNode*         pTagCond;
   SNode*         pTagIndexCond;
+  bool           showRewrite;
 } SSubplan;
 
 typedef enum EExplainMode { EXPLAIN_MODE_DISABLE = 1, EXPLAIN_MODE_STATIC, EXPLAIN_MODE_ANALYZE } EExplainMode;
@@ -585,8 +586,6 @@ typedef struct SQueryPlan {
   SExplainInfo explainInfo;
 } SQueryPlan;
 
-void nodesWalkPhysiPlan(SNode* pNode, FNodeWalker walker, void* pContext);
-
 const char* dataOrderStr(EDataOrderLevel order);
 
 #ifdef __cplusplus

include/libs/nodes/querynodes.h

@@ -347,11 +347,6 @@ typedef struct SInsertStmt {
   uint8_t    precision;
 } SInsertStmt;
 
-typedef enum {
-  PAYLOAD_TYPE_KV = 0,
-  PAYLOAD_TYPE_RAW = 1,
-} EPayloadType;
-
 typedef struct SVgDataBlocks {
   SVgroupInfo vg;
   int32_t     numOfTables;  // number of tables in current submit block
@@ -363,7 +358,6 @@ typedef struct SVnodeModifOpStmt {
   ENodeType   nodeType;
   ENodeType   sqlNodeType;
   SArray*     pDataBlocks;  // data block for each vgroup, SArray<SVgDataBlocks*>.
-  uint8_t     payloadType;  // EPayloadType. 0: K-V payload for non-prepare insert, 1: rawPayload for prepare insert
   uint32_t    insertType;   // insert data from [file|sql statement| bound statement]
   const char* sql;          // current sql statement position
 } SVnodeModifOpStmt;

include/libs/planner/planner.h

@@ -61,7 +61,6 @@ int32_t qStringToSubplan(const char* pStr, SSubplan** pSubplan);
 int32_t qSubPlanToMsg(const SSubplan* pSubplan, char** pStr, int32_t* pLen);
 int32_t qMsgToSubplan(const char* pStr, int32_t len, SSubplan** pSubplan);
 
-char*       qQueryPlanToString(const SQueryPlan* pPlan);
 SQueryPlan* qStringToQueryPlan(const char* pStr);
 
 void qDestroyQueryPlan(SQueryPlan* pPlan);

include/libs/stream/streamState.h

@@ -99,6 +99,8 @@ int32_t streamStateSeekLast(SStreamState* pState, SStreamStateCur* pCur);
 int32_t streamStateCurNext(SStreamState* pState, SStreamStateCur* pCur);
 int32_t streamStateCurPrev(SStreamState* pState, SStreamStateCur* pCur);
 
+// char* streamStateSessionDump(SStreamState* pState);
+
 #ifdef __cplusplus
 }
 #endif

include/libs/stream/tstream.h

@@ -47,7 +47,9 @@ enum {
   TASK_STATUS__FAIL,
   TASK_STATUS__STOP,
   TASK_STATUS__RECOVER_DOWNSTREAM,
-  TASK_STATUS__RECOVER_SELF,
+  TASK_STATUS__RECOVER_PREPARE,
+  TASK_STATUS__RECOVER1,
+  TASK_STATUS__RECOVER2,
 };
 
 enum {
@@ -329,6 +331,9 @@ typedef struct SStreamTask {
   // state backend
   SStreamState* pState;
 
+  // do not serialize
+  int32_t recoverWaitingChild;
+
 } SStreamTask;
 
 int32_t tEncodeStreamEpInfo(SEncoder* pEncoder, const SStreamChildEpInfo* pInfo);
@@ -435,6 +440,20 @@ typedef struct {
   int32_t rspToTaskId;
 } SStreamRetrieveRsp;
 
+typedef struct {
+  int64_t streamId;
+  int32_t taskId;
+} SStreamRecoverStep1Req, SStreamRecoverStep2Req;
+
+typedef struct {
+  int64_t streamId;
+  int32_t taskId;
+  int32_t childId;
+} SStreamRecoverFinishReq;
+
+int32_t tEncodeSStreamRecoverFinishReq(SEncoder* pEncoder, const SStreamRecoverFinishReq* pReq);
+int32_t tDecodeSStreamRecoverFinishReq(SDecoder* pDecoder, SStreamRecoverFinishReq* pReq);
+
 #if 0
 typedef struct {
   int64_t streamId;
@@ -521,8 +540,29 @@ int32_t streamProcessRetrieveRsp(SStreamTask* pTask, SStreamRetrieveRsp* pRsp);
 int32_t streamTryExec(SStreamTask* pTask);
 int32_t streamSchedExec(SStreamTask* pTask);
 
-typedef int32_t FTaskExpand(void* ahandle, SStreamTask* pTask);
+int32_t streamScanExec(SStreamTask* pTask, int32_t batchSz);
 
+// recover and fill history
+// common
+int32_t streamSetParamForRecover(SStreamTask* pTask);
+int32_t streamRestoreParam(SStreamTask* pTask);
+int32_t streamSetStatusNormal(SStreamTask* pTask);
+// source level
+int32_t streamSourceRecoverPrepareStep1(SStreamTask* pTask, int64_t ver);
+int32_t streamBuildSourceRecover1Req(SStreamTask* pTask, SStreamRecoverStep1Req* pReq);
+int32_t streamSourceRecoverScanStep1(SStreamTask* pTask);
+int32_t streamBuildSourceRecover2Req(SStreamTask* pTask, SStreamRecoverStep2Req* pReq);
+int32_t streamSourceRecoverScanStep2(SStreamTask* pTask, int64_t ver);
+int32_t streamDispatchRecoverFinishReq(SStreamTask* pTask);
+// agg level
+int32_t streamAggRecoverPrepare(SStreamTask* pTask);
+// int32_t streamAggChildrenRecoverFinish(SStreamTask* pTask);
+int32_t streamProcessRecoverFinishReq(SStreamTask* pTask, int32_t childId);
+
+// expand and deploy
+typedef int32_t FTaskExpand(void* ahandle, SStreamTask* pTask, int64_t ver);
+
+// meta
 typedef struct SStreamMeta {
   char*        path;
   TDB*         db;
@@ -533,12 +573,13 @@ typedef struct SStreamMeta {
   void*        ahandle;
   TXN          txn;
   FTaskExpand* expandFunc;
+  int32_t      vgId;
 } SStreamMeta;
 
-SStreamMeta* streamMetaOpen(const char* path, void* ahandle, FTaskExpand expandFunc);
+SStreamMeta* streamMetaOpen(const char* path, void* ahandle, FTaskExpand expandFunc, int32_t vgId);
 void         streamMetaClose(SStreamMeta* streamMeta);
 
-// int32_t      streamMetaAddTask(SStreamMeta* pMeta, SStreamTask* pTask);
+int32_t      streamMetaAddTask(SStreamMeta* pMeta, int64_t ver, SStreamTask* pTask);
 int32_t      streamMetaAddSerializedTask(SStreamMeta* pMeta, int64_t startVer, char* msg, int32_t msgLen);
 int32_t      streamMetaRemoveTask(SStreamMeta* pMeta, int32_t taskId);
 SStreamTask* streamMetaGetTask(SStreamMeta* pMeta, int32_t taskId);

include/libs/sync/sync.h

@@ -38,7 +38,7 @@ extern bool gRaftDetailLog;
 #define SYNC_DEL_WAL_MS              (1000 * 60)
 #define SYNC_ADD_QUORUM_COUNT        3
 #define SYNC_MNODE_LOG_RETENTION     10000
-#define SYNC_VNODE_LOG_RETENTION     500
+#define SYNC_VNODE_LOG_RETENTION     100
 
 #define SYNC_APPEND_ENTRIES_TIMEOUT_MS 10000
 

include/libs/wal/wal.h

@@ -33,16 +33,16 @@ extern "C" {
 #define wTrace(...) { if (wDebugFlag & DEBUG_TRACE) { taosPrintLog("WAL ",       DEBUG_TRACE, wDebugFlag, __VA_ARGS__); }}
 // clang-format on
 
-#define WAL_PROTO_VER     0
+#define WAL_PROTO_VER        0
-#define WAL_NOSUFFIX_LEN  20
+#define WAL_NOSUFFIX_LEN     20
-#define WAL_SUFFIX_AT     (WAL_NOSUFFIX_LEN + 1)
+#define WAL_SUFFIX_AT        (WAL_NOSUFFIX_LEN + 1)
-#define WAL_LOG_SUFFIX    "log"
+#define WAL_LOG_SUFFIX       "log"
-#define WAL_INDEX_SUFFIX  "idx"
+#define WAL_INDEX_SUFFIX     "idx"
-#define WAL_REFRESH_MS    1000
+#define WAL_REFRESH_MS       1000
-#define WAL_PATH_LEN      (TSDB_FILENAME_LEN + 12)
+#define WAL_PATH_LEN         (TSDB_FILENAME_LEN + 12)
-#define WAL_FILE_LEN      (WAL_PATH_LEN + 32)
+#define WAL_FILE_LEN         (WAL_PATH_LEN + 32)
-#define WAL_MAGIC         0xFAFBFCFDF4F3F2F1ULL
+#define WAL_MAGIC            0xFAFBFCFDF4F3F2F1ULL
-#define WAL_SCAN_BUF_SIZE (1024 * 1024 * 3)
+#define WAL_SCAN_BUF_SIZE    (1024 * 1024 * 3)
 #define WAL_RECOV_SIZE_LIMIT (100 * WAL_SCAN_BUF_SIZE)
 
 typedef enum {
@@ -204,7 +204,6 @@ SWalRef *walRefCommittedVer(SWal *);
 SWalRef *walOpenRef(SWal *);
 void     walCloseRef(SWal *pWal, int64_t refId);
 int32_t  walRefVer(SWalRef *, int64_t ver);
-int32_t  walPreRefVer(SWalRef *pRef, int64_t ver);
 void     walUnrefVer(SWalRef *);
 
 // helper function for raft

include/util/tdef.h

@@ -234,9 +234,9 @@ typedef enum ELogicConditionType {
  *  - Secondly, if all cols are VarDataT type except primary key, we need 4 bits to store the offset, thus
  *    the final value is 65531-(4096-1)*4 = 49151.
  */
-#define TSDB_MAX_BYTES_PER_ROW  49151
+#define TSDB_MAX_BYTES_PER_ROW 49151
-#define TSDB_MAX_TAGS_LEN       16384
+#define TSDB_MAX_TAGS_LEN      16384
-#define TSDB_MAX_TAGS           128
+#define TSDB_MAX_TAGS          128
 
 #define TSDB_MAX_COL_TAG_NUM  (TSDB_MAX_COLUMNS + TSDB_MAX_TAGS)
 #define TSDB_MAX_JSON_TAG_LEN 16384
@@ -370,13 +370,13 @@ typedef enum ELogicConditionType {
 
 #define TSDB_DB_MIN_WAL_RETENTION_PERIOD -1
 #define TSDB_REP_DEF_DB_WAL_RET_PERIOD   0
-#define TSDB_REPS_DEF_DB_WAL_RET_PERIOD  (24 * 60 * 60 * 4)
+#define TSDB_REPS_DEF_DB_WAL_RET_PERIOD  0
 #define TSDB_DB_MIN_WAL_RETENTION_SIZE   -1
 #define TSDB_REP_DEF_DB_WAL_RET_SIZE     0
-#define TSDB_REPS_DEF_DB_WAL_RET_SIZE    -1
+#define TSDB_REPS_DEF_DB_WAL_RET_SIZE    0
 #define TSDB_DB_MIN_WAL_ROLL_PERIOD      0
 #define TSDB_REP_DEF_DB_WAL_ROLL_PERIOD  0
-#define TSDB_REPS_DEF_DB_WAL_ROLL_PERIOD (24 * 60 * 60 * 1)
+#define TSDB_REPS_DEF_DB_WAL_ROLL_PERIOD 0
 #define TSDB_DB_MIN_WAL_SEGMENT_SIZE     0
 #define TSDB_DEFAULT_DB_WAL_SEGMENT_SIZE 0
 
@@ -409,7 +409,7 @@ typedef enum ELogicConditionType {
 #define TSDB_MAX_RPC_THREADS 10
 #endif
 
-#define TSDB_QUERY_TYPE_NON_TYPE      0x00u  // none type
+#define TSDB_QUERY_TYPE_NON_TYPE 0x00u  // none type
 
 #define TSDB_META_COMPACT_RATIO 0  // disable tsdb meta compact by default
 
@@ -486,7 +486,7 @@ enum {
 
 #define MAX_NUM_STR_SIZE 40
 
-#define MAX_META_MSG_IN_BATCH 1048576
+#define MAX_META_MSG_IN_BATCH   1048576
 #define MAX_META_BATCH_RSP_SIZE (1 * 1048576 * 1024)
 
 #ifdef __cplusplus

include/util/tfunctional.h

@@ -1,56 +0,0 @@
-/*
- * 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/>.
- */
-#ifndef _TD_UTIL_FUNCTIONAL_H_
-#define _TD_UTIL_FUNCTIONAL_H_
-
-#include "os.h"
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-// TODO: hard to use, trying to rewrite it using va_list
-
-typedef void* (*GenericVaFunc)(void* args[]);
-typedef int32_t (*I32VaFunc)(void* args[]);
-typedef void (*VoidVaFunc)(void* args[]);
-
-typedef struct GenericSavedFunc {
-  GenericVaFunc func;
-  void*         args[];
-} tGenericSavedFunc;
-
-typedef struct I32SavedFunc {
-  I32VaFunc func;
-  void*     args[];
-} tI32SavedFunc;
-
-typedef struct VoidSavedFunc {
-  VoidVaFunc func;
-  void*      args[];
-} tVoidSavedFunc;
-
-tGenericSavedFunc* genericSavedFuncInit(GenericVaFunc func, int32_t numOfArgs);
-tI32SavedFunc*     i32SavedFuncInit(I32VaFunc func, int32_t numOfArgs);
-tVoidSavedFunc*    voidSavedFuncInit(VoidVaFunc func, int32_t numOfArgs);
-void*              genericInvoke(tGenericSavedFunc* const pSavedFunc);
-int32_t            i32Invoke(tI32SavedFunc* const pSavedFunc);
-void               voidInvoke(tVoidSavedFunc* const pSavedFunc);
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif /*_TD_UTIL_FUNCTIONAL_H_*/

include/util/tskiplist.h

@@ -19,7 +19,6 @@
 #include "os.h"
 #include "taos.h"
 #include "tarray.h"
-#include "tfunctional.h"
 
 #ifdef __cplusplus
 extern "C" {
@@ -67,7 +66,6 @@ typedef struct SSkipList {
   uint32_t           size;
   SSkipListNode     *pHead;  // point to the first element
   SSkipListNode     *pTail;  // point to the last element
-  tGenericSavedFunc *insertHandleFn;
 } SSkipList;
 
 typedef struct SSkipListIterator {

packaging/cfg/taos.cfg

@@ -5,42 +5,50 @@
 #                                                      #
 ########################################################
 
-# The end point of the first dnode in the cluster to be connected to when `taosd` or `taos` is started
+######### 0. Client only configurations #############
+
+# The interval for TDengine CLI to send heartbeat to mnode
+# shellActivityTimer        3
+
+
+############### 1. Cluster End point ############################
+
+# The end point of the first dnode in the cluster to be connected to when this dnode or a TDengine CLI `taos` is started
 # firstEp                   hostname:6030
 
-# The end point of the second dnode to be connected to if the firstEp is not available when `taosd` or `taos` is started
+# The end point of the second dnode to be connected to if the firstEp is not available
 # secondEp
 
-# The FQDN of the host where `taosd` will be started. It can be IP address
+
+############### 2. Configuration Parameters of current dnode #####
+
+# The FQDN of the host on which this dnode will be started. It can be IP address
 # fqdn                      hostname
 
-# The port for external access after `taosd` is started     
+# The port for external access after this dnode is started     
 # serverPort                6030
 
 # The maximum number of connections a dnode can accept
-# maxShellConns         5000
+# maxShellConns             5000
 
-# The directory for writing log files
+# The directory for writing log files, if you are using Windows platform please change to Windows path
 # logDir                    /var/log/taos
 
-# All data files are stored in this directory
+# All data files are stored in this directory, if you are using Windows platform please change to Windows path
 # dataDir                   /var/lib/taos
 
-# temporary file's directory
+# temporary file's directory, if you are using Windows platform please change to Windows path
 # tempDir                   /tmp/
 
 # Switch for allowing TDengine to collect and report service usage information
 # telemetryReporting        1
 
-# The maximum number of vnodes supported by dnode
+# The maximum number of vnodes supported by this dnode
 # supportVnodes             0
 
-# The interval of dnode reporting status to mnode
+# The interval of this dnode reporting status to mnode, [1..10] seconds
 # statusInterval            1
 
-# The interval for TDengine CLI to send heartbeat to mnode
-# shellActivityTimer        3
-
 # The minimum sliding window time, milli-second
 # minSlidingTime            10
 
@@ -66,13 +74,14 @@
 
 # system time zone
 # timezone              Asia/Shanghai (CST, +0800)
+
 # system time zone (for windows 10)
 # timezone                  UTC-8
 
 # system locale
 # locale                    en_US.UTF-8
 
-# default system charset
+# system charset
 # charset                   UTF-8
 
 # stop writing logs when the disk size of the log folder is less than this value
@@ -81,7 +90,7 @@
 # stop writing temporary files when the disk size of the tmp folder is less than this value
 # minimalTmpDirGB           1.0
 
-# if disk free space is less than this value, taosd service exit directly within startup process
+# if free disk space is less than this value, this dnode will fail to start
 # minimalDataDirGB          2.0
 
 # enable/disable system monitor 
@@ -91,14 +100,18 @@
 # max number of lines per log filters
 # numOfLogLines             10000000
 
-# enable/disable async log
+# write log in async way: 1 - async, 0 - sync
 # asyncLog                  1
 
-# time of keeping log files, days
+# time period of keeping log files, in days
 # logKeepDays               0
 
-# The following parameters are used for debug purpose only.
+
-# debugFlag 8 bits mask: FILE-SCREEN-UNUSED-HeartBeat-DUMP-TRACE_WARN-ERROR
+############ 3. Debug Flag and levels #############################################
+
+# The following parameters are used for debug purpose only by this dnode.
+# debugFlag is a 8 bits mask: FILE-SCREEN-UNUSED-HeartBeat-DUMP-TRACE_WARN-ERROR
+# Available debug levels are: 
 # 131: output warning and error 
 # 135: output debug, warning and error
 # 143: output trace, debug, warning and error to log

packaging/docker/DockerfileCloud

@@ -0,0 +1,26 @@
+FROM ubuntu:18.04
+
+WORKDIR /root
+
+ARG pkgFile
+ARG dirName
+ARG cpuType
+RUN echo ${pkgFile} && echo ${dirName}
+
+COPY ${pkgFile} /root/
+ENV TINI_VERSION v0.19.0
+ADD https://github.com/krallin/tini/releases/download/${TINI_VERSION}/tini-${cpuType} /tini
+ENV DEBIAN_FRONTEND=noninteractive
+WORKDIR /root/
+RUN tar -zxf ${pkgFile} && cd /root/${dirName}/ && /bin/bash install.sh -e no && cd /root && rm /root/${pkgFile} && rm -rf /root/${dirName} && apt-get update && apt-get install -y locales tzdata netcat && locale-gen en_US.UTF-8 && apt-get clean && rm -rf /var/lib/apt/lists/ && chmod +x /tini
+
+ENV LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/lib" \
+    LC_CTYPE=en_US.UTF-8 \
+    LANG=en_US.UTF-8 \
+    LC_ALL=en_US.UTF-8
+COPY ./run.sh /usr/bin/
+COPY ./bin/* /usr/bin/
+
+ENTRYPOINT ["/tini", "--", "/usr/bin/entrypoint.sh"]
+CMD ["bash", "-c", "/usr/bin/run.sh"]
+VOLUME [ "/var/lib/taos", "/var/log/taos", "/corefile" ]

packaging/docker/dockerbuild.sh

@@ -19,8 +19,9 @@ passWord=""
 pkgFile=""
 verType="stable"
 dockerLatest="n"
+cloudBuild="n"
 
-while getopts "hc:n:p:f:V:a:b:" arg
+while getopts "hc:n:p:f:V:a:b:d:" arg
 do
   case $arg in
     c)
@@ -47,6 +48,10 @@ do
       #echo "verType=$OPTARG"
       verType=$(echo $OPTARG)
       ;;
+    d)
+      #echo "cloudBuild=$OPTARG"
+      cloudBuild=$(echo $OPTARG)
+      ;;
     a)
       #echo "dockerLatest=$OPTARG"
       dockerLatest=$(echo $OPTARG)
@@ -58,6 +63,7 @@ do
       echo "                      -V [stable | beta] "
       echo "                      -f [pkg file] "
       echo "                      -a [y | n ]   "
+      echo "                      -d [cloud build ] "
       exit 0
       ;;
     ?) #unknow option 
@@ -83,6 +89,9 @@ else
   echo "unknow verType, nor stabel or beta"
   exit 1
 fi
+if [ "$cloudBuild" == "y" ]; then
+  dockername=cloud-${dockername}
+fi
 
 
 echo "cpuType=${cpuType} version=${version} pkgFile=${pkgFile} verType=${verType} "
@@ -90,10 +99,15 @@ echo "$(pwd)"
 echo "====NOTES: ${pkgFile} must be in the same directory as dockerbuild.sh===="
 
 scriptDir=$(dirname $(readlink -f $0))
-comunityArchiveDir=/nas/TDengine/v$version/community   # community version’package directory
 communityDir=${scriptDir}/../../../community
 DockerfilePath=${communityDir}/packaging/docker/
-Dockerfile=${communityDir}/packaging/docker/Dockerfile
+if [ "$cloudBuild" == "y" ]; then
+  comunityArchiveDir=/nas/TDengine/v$version/cloud
+  Dockerfile=${communityDir}/packaging/docker/DockerfileCloud
+else
+  comunityArchiveDir=/nas/TDengine/v$version/community
+  Dockerfile=${communityDir}/packaging/docker/Dockerfile
+fi
 cd ${scriptDir}
 cp -f ${comunityArchiveDir}/${pkgFile}  .
 
@@ -111,42 +125,15 @@ else
 fi
 
 docker build --rm -f "${Dockerfile}"  --network=host -t tdengine/tdengine-${dockername}:${version} "." --build-arg pkgFile=${pkgFile} --build-arg dirName=${dirName} --build-arg cpuType=${cpuTypeAlias}
-docker login -u tdengine -p ${passWord}  #replace the docker registry username and password
+if [ "$cloudBuild" != "y" ]; then
-docker push tdengine/tdengine-${dockername}:${version}
+  docker login -u tdengine -p ${passWord}  #replace the docker registry username and password
-
+  docker push tdengine/tdengine-${dockername}:${version}
-if [ -n "$(docker ps -aq)" ] ;then 
-  echo "delete docker process"
-  docker stop $(docker ps -aq)
-  docker rm $(docker ps -aq)
 fi
 
-if [  -n "$(pidof taosd)" ] ;then
-   echo "kill taosd "
-   kill -9 $(pidof taosd)
-fi
-
-if [ -n "$(pidof power)" ]  ;then
-  echo "kill power "
-  kill -9 $(pidof power)
-fi
-
-
-echo ">>>>>>>>>>>>> check whether  tdengine/tdengine-${dockername}:${version}  has been published"
-docker run -d --name doctest -p 6030-6049:6030-6049 -p 6030-6049:6030-6049/udp  tdengine/tdengine-${dockername}:${version}
-sleep 2
-curl -u root:taosdata -d 'show variables;'  127.0.0.1:6041/rest/sql > temp1.data
-data_version=$( cat temp1.data |jq .data| jq '.[]' |grep "version" -A 2 -B 1  | jq ".[1]")
-echo "${data_version}" 
-if [ "${data_version}" == "\"${version}\"" ] ; then
-    echo  "docker version is right "
-else
-    echo  "docker version is wrong "
-    exit 1
-fi
-rm -rf  temp1.data
-
 # set this version to latest version 
-if  [ ${dockerLatest} == 'y' ]  ;then
+if  [ "$cloudBuild" != "y" ] && [ ${dockerLatest} == 'y' ]  ;then
   docker tag tdengine/tdengine-${dockername}:${version} tdengine/tdengine-${dockername}:latest
   docker push tdengine/tdengine-${dockername}:latest
 fi
+
+rm -f ${pkgFile}

packaging/docker/run.sh

@@ -0,0 +1,36 @@
+#!/bin/bash
+TAOS_RUN_TAOSBENCHMARK_TEST_ONCE=0
+while ((1))
+do
+    # echo "outer loop: $a"
+    sleep 10
+    output=`taos -k`
+    status=${output:0:1}
+    # echo $output
+    # echo $status
+    if [ "$status"x = "0"x ]
+    then
+        taosd &
+    fi
+    # echo "$status"x "$TAOS_RUN_TAOSBENCHMARK_TEST"x "$TAOS_RUN_TAOSBENCHMARK_TEST_ONCE"x
+    if [ "$status"x = "2"x ] && [ "$TAOS_RUN_TAOSBENCHMARK_TEST"x = "1"x ] && [ "$TAOS_RUN_TAOSBENCHMARK_TEST_ONCE"x = "0"x ]
+    then
+        TAOS_RUN_TAOSBENCHMARK_TEST_ONCE=1
+        # result=`taos -s "show databases;" | grep " test "`
+        # if [ "${result:0:5}"x != " test"x ]
+        # then
+        #     taosBenchmark -y -t 1000 -n 1000 -S 900000
+        # fi
+        taos -s "select stable_name from information_schema.ins_stables where db_name = 'test';"|grep -q -w meters
+        if [ $? -ne 0 ]; then
+            taosBenchmark -y -t 1000 -n 1000 -S 900000
+	    taos -s "create user admin_user pass 'NDS65R6t' sysinfo 0;"
+	    taos -s "GRANT ALL on test.* to admin_user;"
+        fi
+    fi
+    # check taosadapter
+    nc -z localhost 6041
+    if [ $? -ne 0 ]; then
+	taosadapter &
+    fi
+done

packaging/release.sh

@@ -17,7 +17,7 @@ set -e
 #             -H [ false | true]
 
 # set parameters by default value
-verMode=edge    # [cluster, edge]
+verMode=edge    # [cluster, edge, cloud]
 verType=stable  # [stable, beta]
 cpuType=x64     # [aarch32 | aarch64 | x64 | x86 | mips64 ...]
 osType=Linux    # [Linux | Kylin | Alpine | Raspberrypi | Darwin | Windows | Ningsi60 | Ningsi80 |...]
@@ -169,7 +169,7 @@ build_time=$(date +"%F %R")
 # get commint id from git
 gitinfo=$(git rev-parse --verify HEAD)
 
-if [[ "$verMode" == "cluster" ]]; then
+if [[ "$verMode" == "cluster" ]] || [[ "$verMode" == "cloud" ]]; then
   enterprise_dir="${top_dir}/../enterprise"
   cd ${enterprise_dir}
   gitinfoOfInternal=$(git rev-parse --verify HEAD)
@@ -205,7 +205,7 @@ else
   BUILD_HTTP=false
 fi
 
-if [[ "$verMode" == "cluster" ]]; then
+if [[ "$verMode" == "cluster" ]] || [[ "$verMode" == "cloud" ]]; then
   BUILD_HTTP=internal
 fi
 
@@ -217,10 +217,12 @@ fi
 
 # check support cpu type
 if [[ "$cpuType" == "x64" ]] || [[ "$cpuType" == "aarch64" ]] || [[ "$cpuType" == "aarch32" ]] || [[ "$cpuType" == "arm64" ]] || [[ "$cpuType" == "arm32" ]] || [[ "$cpuType" == "mips64" ]]; then
-  if [ "$verMode" != "cluster" ]; then
+  if [ "$verMode" == "edge" ]; then
     # community-version compile
     cmake ../ -DCPUTYPE=${cpuType} -DWEBSOCKET=true -DOSTYPE=${osType} -DSOMODE=${soMode} -DDBNAME=${dbName} -DVERTYPE=${verType} -DVERDATE="${build_time}" -DGITINFO=${gitinfo} -DGITINFOI=${gitinfoOfInternal} -DVERNUMBER=${verNumber} -DVERCOMPATIBLE=${verNumberComp} -DPAGMODE=${pagMode} -DBUILD_HTTP=${BUILD_HTTP} -DBUILD_TOOLS=${BUILD_TOOLS} ${allocator_macro}
-  else
+  elif [ "$verMode" == "cloud" ]; then
+    cmake ../../ -DCPUTYPE=${cpuType} -DWEBSOCKET=true -DBUILD_CLOUD=true -DOSTYPE=${osType} -DSOMODE=${soMode} -DDBNAME=${dbName} -DVERTYPE=${verType} -DVERDATE="${build_time}" -DGITINFO=${gitinfo} -DGITINFOI=${gitinfoOfInternal} -DVERNUMBER=${verNumber} -DVERCOMPATIBLE=${verNumberComp} -DBUILD_HTTP=${BUILD_HTTP} -DBUILD_TOOLS=${BUILD_TOOLS} ${allocator_macro}
+  elif [ "$verMode" == "cluster" ]; then
     if [[ "$dbName" != "taos" ]]; then
       replace_enterprise_$dbName
     fi
@@ -244,7 +246,7 @@ cd ${curr_dir}
 
 # 3. Call the corresponding script for packaging
 if [ "$osType" != "Darwin" ]; then
-  if [[ "$verMode" != "cluster" ]] && [[ "$pagMode" == "full" ]] && [[ "$cpuType" == "x64" ]] && [[ "$dbName" == "taos" ]]; then
+  if [[ "$verMode" != "cluster" ]] && [[ "$verMode" != "cloud" ]] && [[ "$pagMode" == "full" ]] && [[ "$cpuType" == "x64" ]] && [[ "$dbName" == "taos" ]]; then
     ret='0'
     command -v dpkg >/dev/null 2>&1 || { ret='1'; }
     if [ "$ret" -eq 0 ]; then

packaging/tools/make_install.bat

@@ -47,6 +47,15 @@ copy %binary_dir%\\build\\bin\\udfd.exe %tagert_dir% > nul
 if exist %binary_dir%\\build\\bin\\taosBenchmark.exe (
     copy %binary_dir%\\build\\bin\\taosBenchmark.exe %tagert_dir% > nul
 )
+if exist %binary_dir%\\build\\lib\\taosws.dll.lib (
+    copy %binary_dir%\\build\\lib\\taosws.dll.lib %tagert_dir%\\driver  > nul
+)
+if exist %binary_dir%\\build\\lib\\taosws.dll (
+    copy %binary_dir%\\build\\lib\\taosws.dll %tagert_dir%\\driver  > nul
+)
+if exist %binary_dir%\\build\\bin\\taosdump.exe (
+    copy %binary_dir%\\build\\bin\\taosdump.exe %tagert_dir% > nul
+)
 if exist %binary_dir%\\build\\bin\\taosadapter.exe (
     copy %binary_dir%\\build\\bin\\taosadapter.exe %tagert_dir% > nul
 )
@@ -54,4 +63,7 @@ if exist %binary_dir%\\build\\bin\\taosadapter.exe (
 mshta vbscript:createobject("shell.application").shellexecute("%~s0",":hasAdmin","","runas",1)(window.close)&& echo To start/stop TDengine with administrator privileges: sc start/stop taosd &goto :eof
 :hasAdmin
 copy /y C:\\TDengine\\driver\\taos.dll C:\\Windows\\System32 > nul
+if exist C:\\TDengine\\driver\\taosws.dll (
+    copy /y C:\\TDengine\\driver\\taosws.dll C:\\Windows\\System32 > nul
+)
 sc query "taosd" >nul || sc create "taosd" binPath= "C:\\TDengine\\taosd.exe --win_service" start= DEMAND

packaging/tools/makeclient.sh

@@ -39,6 +39,8 @@ release_dir="${top_dir}/release"
 
 if [ "$verMode" == "cluster" ]; then
   install_dir="${release_dir}/${productName}-enterprise-client-${version}"
+elif [ "$verMode" == "cloud" ]; then
+  install_dir="${release_dir}/${productName}-cloud-client-${version}"
 else
   install_dir="${release_dir}/${productName}-client-${version}"
 fi
@@ -138,6 +140,10 @@ if [ "$verMode" == "cluster" ]; then
   sed 's/verMode=edge/verMode=cluster/g' ${install_dir}/install_client.sh >>install_client_temp.sh
   mv install_client_temp.sh ${install_dir}/install_client.sh
 fi
+if [ "$verMode" == "cloud" ]; then
+  sed 's/verMode=edge/verMode=cloud/g' ${install_dir}/install_client.sh >>install_client_temp.sh
+  mv install_client_temp.sh ${install_dir}/install_client.sh
+fi
 
 if [ "$pagMode" == "lite" ]; then
   sed 's/pagMode=full/pagMode=lite/g' ${install_dir}/install_client.sh >>install_client_temp.sh
@@ -161,7 +167,7 @@ if [[ $productName == "TDengine" ]]; then
     mkdir -p ${install_dir}/examples/taosbenchmark-json && cp ${examples_dir}/../tools/taos-tools/example/* ${install_dir}/examples/taosbenchmark-json
   fi
 
-  if [ "$verMode" == "cluster" ]; then
+  if [ "$verMode" == "cluster" ] || [ "$verMode" == "cloud" ]; then
       # Copy connector
       connector_dir="${code_dir}/connector"
       mkdir -p ${install_dir}/connector

packaging/tools/makepkg.sh

@@ -39,6 +39,8 @@ release_dir="${top_dir}/release"
 #package_name='linux'
 if [ "$verMode" == "cluster" ]; then
   install_dir="${release_dir}/${productName}-enterprise-server-${version}"
+elif [ "$verMode" == "cloud" ]; then
+  install_dir="${release_dir}/${productName}-cloud-server-${version}"
 else
   install_dir="${release_dir}/${productName}-server-${version}"
 fi
@@ -217,7 +219,10 @@ fi
 if [ "$verMode" == "cluster" ]; then
   sed 's/verMode=edge/verMode=cluster/g' ${install_dir}/bin/remove.sh >>remove_temp.sh
   mv remove_temp.sh ${install_dir}/bin/remove.sh
-
+fi
+if [ "$verMode" == "cloud" ]; then
+  sed 's/verMode=edge/verMode=cloud/g' ${install_dir}/bin/remove.sh >>remove_temp.sh
+  mv remove_temp.sh ${install_dir}/bin/remove.sh
 fi
 
 cd ${install_dir}
@@ -234,6 +239,10 @@ if [ "$verMode" == "cluster" ]; then
   sed 's/verMode=edge/verMode=cluster/g' ${install_dir}/install.sh >>install_temp.sh
   mv install_temp.sh ${install_dir}/install.sh
 fi
+if [ "$verMode" == "cloud" ]; then
+  sed 's/verMode=edge/verMode=cloud/g' ${install_dir}/install.sh >>install_temp.sh
+  mv install_temp.sh ${install_dir}/install.sh
+fi
 if [ "$pagMode" == "lite" ]; then
   sed 's/pagMode=full/pagMode=lite/g' ${install_dir}/install.sh >>install_temp.sh
   mv install_temp.sh ${install_dir}/install.sh
@@ -288,7 +297,7 @@ mkdir -p ${install_dir}/driver && cp ${lib_files} ${install_dir}/driver && echo
 [ -f ${wslib_files} ] && cp ${wslib_files} ${install_dir}/driver || :
 
 # Copy connector
-if [ "$verMode" == "cluster" ]; then
+if [ "$verMode" == "cluster" ] || [ "$verMode" == "cloud" ]; then
     connector_dir="${code_dir}/connector"
     mkdir -p ${install_dir}/connector
     if [[ "$pagMode" != "lite" ]] && [[ "$cpuType" != "aarch32" ]]; then

source/client/inc/clientInt.h

@@ -46,7 +46,6 @@ extern "C" {
 
 #define ERROR_MSG_BUF_DEFAULT_SIZE 512
 #define HEARTBEAT_INTERVAL         1500  // ms
-#define SYNC_ON_TOP_OF_ASYNC       1
 
 enum {
   RES_TYPE__QUERY = 1,
@@ -271,9 +270,8 @@ void    doFreeReqResultInfo(SReqResultInfo* pResInfo);
 int32_t transferTableNameList(const char* tbList, int32_t acctId, char* dbName, SArray** pReq);
 void    syncCatalogFn(SMetaData* pResult, void* param, int32_t code);
 
-SRequestObj* execQuery(uint64_t connId, const char* sql, int sqlLen, bool validateOnly);
+TAOS_RES* taosQueryImpl(TAOS* taos, const char* sql, bool validateOnly);
-TAOS_RES*    taosQueryImpl(TAOS* taos, const char* sql, bool validateOnly);
+void      taosAsyncQueryImpl(uint64_t connId, const char* sql, __taos_async_fn_t fp, void* param, bool validateOnly);
-void         taosAsyncQueryImpl(uint64_t connId, const char* sql, __taos_async_fn_t fp, void* param, bool validateOnly);
 
 int32_t getVersion1BlockMetaSize(const char* p, int32_t numOfCols);
 
@@ -349,8 +347,6 @@ void processMsgFromServer(void* parent, SRpcMsg* pMsg, SEpSet* pEpSet);
 STscObj* taos_connect_internal(const char* ip, const char* user, const char* pass, const char* auth, const char* db,
                                uint16_t port, int connType);
 
-SRequestObj* launchQuery(uint64_t connId, const char* sql, int sqlLen, bool validateOnly, bool inRetry);
-
 int32_t parseSql(SRequestObj* pRequest, bool topicQuery, SQuery** pQuery, SStmtCallback* pStmtCb);
 
 int32_t getPlan(SRequestObj* pRequest, SQuery* pQuery, SQueryPlan** pPlan, SArray* pNodeList);

source/client/inc/clientLog.h

@@ -30,7 +30,7 @@ extern "C" {
 #define tscDebug(...)  do { if (cDebugFlag & DEBUG_DEBUG) { taosPrintLog("TSC ", DEBUG_DEBUG, cDebugFlag, __VA_ARGS__); }} while(0)
 #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)
+//#define tscPerf(...)   do { if (cDebugFlag & DEBUG_INFO)  { taosPrintLog("TSC ", DEBUG_INFO, cDebugFlag, __VA_ARGS__); }} while(0)
 // clang-format on
 
 #ifdef __cplusplus

source/client/src/clientEnv.c

@@ -77,19 +77,19 @@ static void deregisterRequest(SRequestObj *pRequest) {
            pRequest->self, pTscObj->id, pRequest->requestId, duration / 1000.0, num, currentInst);
 
   if (QUERY_NODE_VNODE_MODIF_STMT == pRequest->stmtType) {
-    tscPerf("insert duration %" PRId64 "us: syntax:%" PRId64 "us, ctg:%" PRId64 "us, semantic:%" PRId64
+//    tscPerf("insert duration %" PRId64 "us: syntax:%" PRId64 "us, ctg:%" PRId64 "us, semantic:%" PRId64
-            "us, exec:%" PRId64 "us",
+//            "us, exec:%" PRId64 "us",
-            duration, pRequest->metric.syntaxEnd - pRequest->metric.syntaxStart,
+//            duration, pRequest->metric.syntaxEnd - pRequest->metric.syntaxStart,
-            pRequest->metric.ctgEnd - pRequest->metric.ctgStart, pRequest->metric.semanticEnd - pRequest->metric.ctgEnd,
+//            pRequest->metric.ctgEnd - pRequest->metric.ctgStart, pRequest->metric.semanticEnd - pRequest->metric.ctgEnd,
-            pRequest->metric.execEnd - pRequest->metric.semanticEnd);
+//            pRequest->metric.execEnd - pRequest->metric.semanticEnd);
     atomic_add_fetch_64((int64_t *)&pActivity->insertElapsedTime, duration);
   } else if (QUERY_NODE_SELECT_STMT == pRequest->stmtType) {
-    tscPerf("select duration %" PRId64 "us: syntax:%" PRId64 "us, ctg:%" PRId64 "us, semantic:%" PRId64
+//    tscPerf("select duration %" PRId64 "us: syntax:%" PRId64 "us, ctg:%" PRId64 "us, semantic:%" PRId64
-            "us, planner:%" PRId64 "us, exec:%" PRId64 "us, reqId:0x%" PRIx64,
+//            "us, planner:%" PRId64 "us, exec:%" PRId64 "us, reqId:0x%" PRIx64,
-            duration, pRequest->metric.syntaxEnd - pRequest->metric.syntaxStart,
+//            duration, pRequest->metric.syntaxEnd - pRequest->metric.syntaxStart,
-            pRequest->metric.ctgEnd - pRequest->metric.ctgStart, pRequest->metric.semanticEnd - pRequest->metric.ctgEnd,
+//            pRequest->metric.ctgEnd - pRequest->metric.ctgStart, pRequest->metric.semanticEnd - pRequest->metric.ctgEnd,
-            pRequest->metric.planEnd - pRequest->metric.semanticEnd,
+//            pRequest->metric.planEnd - pRequest->metric.semanticEnd,
-            pRequest->metric.resultReady - pRequest->metric.planEnd, pRequest->requestId);
+//            pRequest->metric.resultReady - pRequest->metric.planEnd, pRequest->requestId);
 
     atomic_add_fetch_64((int64_t *)&pActivity->queryElapsedTime, duration);
   }

source/client/src/clientImpl.c

@@ -1013,28 +1013,6 @@ SRequestObj* launchQueryImpl(SRequestObj* pRequest, SQuery* pQuery, bool keepQue
   return pRequest;
 }
 
-SRequestObj* launchQuery(uint64_t connId, const char* sql, int sqlLen, bool validateOnly, bool inRetry) {
-  SRequestObj* pRequest = NULL;
-  SQuery*      pQuery = NULL;
-
-  int32_t code = buildRequest(connId, sql, sqlLen, NULL, validateOnly, &pRequest);
-  if (code != TSDB_CODE_SUCCESS) {
-    terrno = code;
-    return NULL;
-  }
-
-  code = parseSql(pRequest, false, &pQuery, NULL);
-  if (code != TSDB_CODE_SUCCESS) {
-    pRequest->code = code;
-    return pRequest;
-  }
-
-  pRequest->inRetry = inRetry;
-  pRequest->stableQuery = pQuery->stableQuery;
-
-  return launchQueryImpl(pRequest, pQuery, false, NULL);
-}
-
 static int32_t asyncExecSchQuery(SRequestObj* pRequest, SQuery* pQuery, SMetaData* pResultMeta,
                                  SSqlCallbackWrapper* pWrapper) {
   pRequest->type = pQuery->msgType;
@@ -1197,32 +1175,6 @@ int32_t removeMeta(STscObj* pTscObj, SArray* tbList) {
   return TSDB_CODE_SUCCESS;
 }
 
-//  todo remove it soon
-SRequestObj* execQuery(uint64_t connId, const char* sql, int sqlLen, bool validateOnly) {
-  SRequestObj* pRequest = NULL;
-  int32_t      retryNum = 0;
-  int32_t      code = 0;
-  bool         inRetry = false;
-
-  do {
-    destroyRequest(pRequest);
-    pRequest = launchQuery(connId, sql, sqlLen, validateOnly, inRetry);
-    if (pRequest == NULL || TSDB_CODE_SUCCESS == pRequest->code || !NEED_CLIENT_HANDLE_ERROR(pRequest->code)) {
-      break;
-    }
-
-    code = refreshMeta(pRequest->pTscObj, pRequest);
-    if (code) {
-      pRequest->code = code;
-      break;
-    }
-
-    inRetry = true;
-  } while (retryNum++ < REQUEST_TOTAL_EXEC_TIMES);
-
-  return pRequest;
-}
-
 int initEpSetFromCfg(const char* firstEp, const char* secondEp, SCorEpSet* pEpSet) {
   pEpSet->version = 0;
 
@@ -2291,7 +2243,6 @@ TAOS_RES* taosQueryImpl(TAOS* taos, const char* sql, bool validateOnly) {
     return NULL;
   }
 
-#if SYNC_ON_TOP_OF_ASYNC
   SSyncQueryParam* param = taosMemoryCalloc(1, sizeof(SSyncQueryParam));
   tsem_init(&param->sem, 0, 0);
 
@@ -2301,15 +2252,4 @@ TAOS_RES* taosQueryImpl(TAOS* taos, const char* sql, bool validateOnly) {
     param->pRequest->syncQuery = true;
   }
   return param->pRequest;
-#else
-  size_t sqlLen = strlen(sql);
-  if (sqlLen > (size_t)TSDB_MAX_ALLOWED_SQL_LEN) {
-    tscError("sql string exceeds max length:%d", TSDB_MAX_ALLOWED_SQL_LEN);
-    terrno = TSDB_CODE_TSC_EXCEED_SQL_LIMIT;
-    return NULL;
-  }
-
-  TAOS_RES* pRes = execQuery(*(int64_t*)taos, sql, sqlLen, validateOnly);
-  return pRes;
-#endif
 }

source/client/src/clientMain.c

@@ -261,12 +261,7 @@ TAOS_ROW taos_fetch_row(TAOS_RES *res) {
       return NULL;
     }
 
-#if SYNC_ON_TOP_OF_ASYNC
     return doAsyncFetchRows(pRequest, true, true);
-#else
-    return doFetchRows(pRequest, true, true);
-#endif
-
   } else if (TD_RES_TMQ(res) || TD_RES_TMQ_METADATA(res)) {
     SMqRspObj      *msg = ((SMqRspObj *)res);
     SReqResultInfo *pResultInfo;
@@ -549,11 +544,7 @@ int taos_fetch_block_s(TAOS_RES *res, int *numOfRows, TAOS_ROW *rows) {
       return 0;
     }
 
-#if SYNC_ON_TOP_OF_ASYNC
     doAsyncFetchRows(pRequest, false, true);
-#else
-    doFetchRows(pRequest, true, true);
-#endif
 
     // TODO refactor
     SReqResultInfo *pResultInfo = &pRequest->body.resInfo;
@@ -601,11 +592,7 @@ int taos_fetch_raw_block(TAOS_RES *res, int *numOfRows, void **pData) {
     return 0;
   }
 
-#if SYNC_ON_TOP_OF_ASYNC
   doAsyncFetchRows(pRequest, false, false);
-#else
-  doFetchRows(pRequest, false, false);
-#endif
 
   SReqResultInfo *pResultInfo = &pRequest->body.resInfo;
 
@@ -989,7 +976,7 @@ const void *taos_get_raw_block(TAOS_RES *res) {
   return pRequest->body.resInfo.pData;
 }
 
-int taos_get_db_route_info(TAOS* taos, const char* db, TAOS_DB_ROUTE_INFO* dbInfo) {
+int taos_get_db_route_info(TAOS *taos, const char *db, TAOS_DB_ROUTE_INFO *dbInfo) {
   if (NULL == taos) {
     terrno = TSDB_CODE_TSC_DISCONNECTED;
     return terrno;
@@ -1001,16 +988,16 @@ int taos_get_db_route_info(TAOS* taos, const char* db, TAOS_DB_ROUTE_INFO* dbInf
     return terrno;
   }
 
-  int64_t       connId = *(int64_t *)taos;
+  int64_t      connId = *(int64_t *)taos;
-  SRequestObj  *pRequest = NULL;
+  SRequestObj *pRequest = NULL;
-  char *sql = "taos_get_db_route_info";
+  char        *sql = "taos_get_db_route_info";
-  int32_t code = buildRequest(connId, sql, strlen(sql), NULL, false, &pRequest);
+  int32_t      code = buildRequest(connId, sql, strlen(sql), NULL, false, &pRequest);
   if (code != TSDB_CODE_SUCCESS) {
     terrno = code;
     return terrno;
   }
 
-  STscObj *pTscObj = pRequest->pTscObj;
+  STscObj  *pTscObj = pRequest->pTscObj;
   SCatalog *pCtg = NULL;
   code = catalogGetHandle(pTscObj->pAppInfo->clusterId, &pCtg);
   if (code != TSDB_CODE_SUCCESS) {
@@ -1038,7 +1025,7 @@ _return:
   return code;
 }
 
-int taos_get_table_vgId(TAOS* taos, const char* db, const char* table, int* vgId) {
+int taos_get_table_vgId(TAOS *taos, const char *db, const char *table, int *vgId) {
   if (NULL == taos) {
     terrno = TSDB_CODE_TSC_DISCONNECTED;
     return terrno;
@@ -1050,15 +1037,15 @@ int taos_get_table_vgId(TAOS* taos, const char* db, const char* table, int* vgId
     return terrno;
   }
 
-  int64_t       connId = *(int64_t *)taos;
+  int64_t      connId = *(int64_t *)taos;
-  SRequestObj  *pRequest = NULL;
+  SRequestObj *pRequest = NULL;
-  char *sql = "taos_get_table_vgId";
+  char        *sql = "taos_get_table_vgId";
-  int32_t code = buildRequest(connId, sql, strlen(sql), NULL, false, &pRequest);
+  int32_t      code = buildRequest(connId, sql, strlen(sql), NULL, false, &pRequest);
   if (code != TSDB_CODE_SUCCESS) {
     return terrno;
   }
 
-  STscObj *pTscObj = pRequest->pTscObj;
+  STscObj  *pTscObj = pRequest->pTscObj;
   SCatalog *pCtg = NULL;
   code = catalogGetHandle(pTscObj->pAppInfo->clusterId, &pCtg);
   if (code != TSDB_CODE_SUCCESS) {

source/client/src/clientRawBlockWrite.c

@@ -1369,7 +1369,6 @@ int taos_write_raw_block(TAOS* taos, int rows, char* pData, const char* tbname)
     goto end;
   }
   SVnodeModifOpStmt* nodeStmt = (SVnodeModifOpStmt*)(pQuery->pRoot);
-  nodeStmt->payloadType = PAYLOAD_TYPE_KV;
   nodeStmt->pDataBlocks = taosArrayInit(1, POINTER_BYTES);
 
   SVgDataBlocks* dst = taosMemoryCalloc(1, sizeof(SVgDataBlocks));
@@ -1625,7 +1624,6 @@ static int32_t tmqWriteRawDataImpl(TAOS* taos, void* data, int32_t dataLen) {
     goto end;
   }
   SVnodeModifOpStmt* nodeStmt = (SVnodeModifOpStmt*)(pQuery->pRoot);
-  nodeStmt->payloadType = PAYLOAD_TYPE_KV;
 
   int32_t numOfVg = taosHashGetSize(pVgHash);
   nodeStmt->pDataBlocks = taosArrayInit(numOfVg, POINTER_BYTES);
@@ -1929,7 +1927,6 @@ static int32_t tmqWriteRawMetaDataImpl(TAOS* taos, void* data, int32_t dataLen)
     goto end;
   }
   SVnodeModifOpStmt* nodeStmt = (SVnodeModifOpStmt*)(pQuery->pRoot);
-  nodeStmt->payloadType = PAYLOAD_TYPE_KV;
 
   int32_t numOfVg = taosHashGetSize(pVgHash);
   nodeStmt->pDataBlocks = taosArrayInit(numOfVg, POINTER_BYTES);

source/client/src/clientSml.c

@@ -1531,7 +1531,6 @@ static SSmlHandle *smlBuildSmlInfo(STscObj *pTscObj, SRequestObj *request, SMLPr
     uError("SML:0x%" PRIx64 " create info->pQuery->pRoot error", info->id);
     goto cleanup;
   }
-  ((SVnodeModifOpStmt *)(info->pQuery->pRoot))->payloadType = PAYLOAD_TYPE_KV;
 
   if (pTscObj) {
     info->taos = pTscObj;

source/client/src/clientTmq.c

@@ -1619,7 +1619,7 @@ void* tmqHandleAllRsp(tmq_t* tmq, int64_t timeout, bool pollIfReset) {
       taosGetQitem(tmq->qall, (void**)&rspWrapper);
 
       if (rspWrapper == NULL) {
-        tscDebug("consumer %" PRId64 " mqueue empty", tmq->consumerId);
+        /*tscDebug("consumer %" PRId64 " mqueue empty", tmq->consumerId);*/
         return NULL;
       }
     }
@@ -1732,6 +1732,7 @@ TAOS_RES* tmq_consumer_poll(tmq_t* tmq, int64_t timeout) {
 
   // in no topic status, delayed task also need to be processed
   if (atomic_load_8(&tmq->status) == TMQ_CONSUMER_STATUS__INIT) {
+    tscDebug("consumer:%" PRId64 ", poll return since consumer status is init", tmq->consumerId);
     return NULL;
   }
 

source/common/src/systable.c

@@ -230,8 +230,8 @@ static const SSysDbTableSchema transSchema[] = {
 };
 
 static const SSysDbTableSchema configSchema[] = {
-    {.name = "name", .bytes = TSDB_CONFIG_OPTION_LEN + VARSTR_HEADER_SIZE, .type = TSDB_DATA_TYPE_VARCHAR, .sysInfo = true},
+    {.name = "name", .bytes = TSDB_CONFIG_OPTION_LEN + VARSTR_HEADER_SIZE, .type = TSDB_DATA_TYPE_VARCHAR, .sysInfo = false},
-    {.name = "value", .bytes = TSDB_CONFIG_VALUE_LEN + VARSTR_HEADER_SIZE, .type = TSDB_DATA_TYPE_VARCHAR, .sysInfo = true},
+    {.name = "value", .bytes = TSDB_CONFIG_VALUE_LEN + VARSTR_HEADER_SIZE, .type = TSDB_DATA_TYPE_VARCHAR, .sysInfo = false},
 };
 
 static const SSysDbTableSchema variablesSchema[] = {
@@ -270,7 +270,7 @@ static const SSysTableMeta infosMeta[] = {
     {TSDB_INS_TABLE_MNODES, mnodesSchema, tListLen(mnodesSchema), true},
     {TSDB_INS_TABLE_MODULES, modulesSchema, tListLen(modulesSchema), true},
     {TSDB_INS_TABLE_QNODES, qnodesSchema, tListLen(qnodesSchema), true},
-    //    {TSDB_INS_TABLE_SNODES, snodesSchema, tListLen(snodesSchema)},
+    {TSDB_INS_TABLE_SNODES, snodesSchema, tListLen(snodesSchema)},
     {TSDB_INS_TABLE_CLUSTER, clusterSchema, tListLen(clusterSchema), true},
     {TSDB_INS_TABLE_DATABASES, userDBSchema, tListLen(userDBSchema), false},
     {TSDB_INS_TABLE_FUNCTIONS, userFuncSchema, tListLen(userFuncSchema), false},
@@ -282,7 +282,7 @@ static const SSysTableMeta infosMeta[] = {
     {TSDB_INS_TABLE_USERS, userUsersSchema, tListLen(userUsersSchema), false},
     {TSDB_INS_TABLE_LICENCES, grantsSchema, tListLen(grantsSchema), true},
     {TSDB_INS_TABLE_VGROUPS, vgroupsSchema, tListLen(vgroupsSchema), true},
-    {TSDB_INS_TABLE_CONFIGS, configSchema, tListLen(configSchema), true},
+    {TSDB_INS_TABLE_CONFIGS, configSchema, tListLen(configSchema), false},
     {TSDB_INS_TABLE_DNODE_VARIABLES, variablesSchema, tListLen(variablesSchema), true},
     {TSDB_INS_TABLE_TOPICS, topicSchema, tListLen(topicSchema), false},
     {TSDB_INS_TABLE_SUBSCRIPTIONS, subscriptionSchema, tListLen(subscriptionSchema), false},

source/common/src/tdatablock.c

@@ -1707,8 +1707,6 @@ void* tDecodeDataBlock(const void* buf, SSDataBlock* pBlock) {
 
     if (IS_VAR_DATA_TYPE(data.info.type)) {
       buf = taosDecodeBinary(buf, (void**)&data.varmeta.offset, pBlock->info.rows * sizeof(int32_t));
-      data.varmeta.length = pBlock->info.rows * sizeof(int32_t);
-      data.varmeta.allocLen = data.varmeta.length;
     } else {
       buf = taosDecodeBinary(buf, (void**)&data.nullbitmap, BitmapLen(pBlock->info.rows));
     }
@@ -1716,6 +1714,10 @@ void* tDecodeDataBlock(const void* buf, SSDataBlock* pBlock) {
     int32_t len = 0;
     buf = taosDecodeFixedI32(buf, &len);
     buf = taosDecodeBinary(buf, (void**)&data.pData, len);
+    if (IS_VAR_DATA_TYPE(data.info.type)) {
+      data.varmeta.length = len;
+      data.varmeta.allocLen = len;
+    }
     taosArrayPush(pBlock->pDataBlock, &data);
   }
   return (void*)buf;

source/common/src/tglobal.c

@@ -55,8 +55,8 @@ int32_t tsNumOfVnodeSyncThreads = 2;
 int32_t tsNumOfVnodeRsmaThreads = 2;
 int32_t tsNumOfQnodeQueryThreads = 4;
 int32_t tsNumOfQnodeFetchThreads = 1;
-int32_t tsNumOfSnodeSharedThreads = 2;
+int32_t tsNumOfSnodeStreamThreads = 4;
-int32_t tsNumOfSnodeUniqueThreads = 2;
+int32_t tsNumOfSnodeWriteThreads = 1;
 
 // monitor
 bool     tsEnableMonitor = true;
@@ -82,6 +82,7 @@ bool tsSmlDataFormat = false;
 // query
 int32_t tsQueryPolicy = 1;
 int32_t tsQueryRspPolicy = 0;
+bool    tsEnableQueryHb = false;
 int32_t tsQuerySmaOptimize = 0;
 int32_t tsQueryRsmaTolerance = 1000;  // the tolerance time (ms) to judge from which level to query rsma data.
 bool    tsQueryPlannerTrace = false;
@@ -132,7 +133,7 @@ int32_t  tsDiskCfgNum = 0;
 SDiskCfg tsDiskCfg[TFS_MAX_DISKS] = {0};
 
 // stream scheduler
-bool tsSchedStreamToSnode = true;
+bool tsDeployOnSnode = true;
 
 /*
  * minimum scale for whole system, millisecond by default
@@ -284,6 +285,7 @@ static int32_t taosAddClientCfg(SConfig *pCfg) {
   if (cfgAddInt32(pCfg, "compressMsgSize", tsCompressMsgSize, -1, 100000000, 1) != 0) return -1;
   if (cfgAddInt32(pCfg, "compressColData", tsCompressColData, -1, 100000000, 1) != 0) return -1;
   if (cfgAddInt32(pCfg, "queryPolicy", tsQueryPolicy, 1, 4, 1) != 0) return -1;
+  if (cfgAddBool(pCfg, "enableQueryHb", tsEnableQueryHb, false) != 0) return -1;
   if (cfgAddInt32(pCfg, "querySmaOptimize", tsQuerySmaOptimize, 0, 1, 1) != 0) return -1;
   if (cfgAddBool(pCfg, "queryPlannerTrace", tsQueryPlannerTrace, true) != 0) return -1;
   if (cfgAddInt32(pCfg, "queryNodeChunkSize", tsQueryNodeChunkSize, 1024, 128 * 1024, true) != 0) return -1;
@@ -388,13 +390,13 @@ static int32_t taosAddServerCfg(SConfig *pCfg) {
   //  tsNumOfQnodeFetchThreads = TMAX(tsNumOfQnodeFetchThreads, 4);
   //  if (cfgAddInt32(pCfg, "numOfQnodeFetchThreads", tsNumOfQnodeFetchThreads, 1, 1024, 0) != 0) return -1;
 
-  tsNumOfSnodeSharedThreads = tsNumOfCores / 4;
+  tsNumOfSnodeStreamThreads = tsNumOfCores / 4;
-  tsNumOfSnodeSharedThreads = TRANGE(tsNumOfSnodeSharedThreads, 2, 4);
+  tsNumOfSnodeStreamThreads = TRANGE(tsNumOfSnodeStreamThreads, 2, 4);
-  if (cfgAddInt32(pCfg, "numOfSnodeSharedThreads", tsNumOfSnodeSharedThreads, 2, 1024, 0) != 0) return -1;
+  if (cfgAddInt32(pCfg, "numOfSnodeSharedThreads", tsNumOfSnodeStreamThreads, 2, 1024, 0) != 0) return -1;
 
-  tsNumOfSnodeUniqueThreads = tsNumOfCores / 4;
+  tsNumOfSnodeWriteThreads = tsNumOfCores / 4;
-  tsNumOfSnodeUniqueThreads = TRANGE(tsNumOfSnodeUniqueThreads, 2, 4);
+  tsNumOfSnodeWriteThreads = TRANGE(tsNumOfSnodeWriteThreads, 2, 4);
-  if (cfgAddInt32(pCfg, "numOfSnodeUniqueThreads", tsNumOfSnodeUniqueThreads, 2, 1024, 0) != 0) return -1;
+  if (cfgAddInt32(pCfg, "numOfSnodeUniqueThreads", tsNumOfSnodeWriteThreads, 2, 1024, 0) != 0) return -1;
 
   tsRpcQueueMemoryAllowed = tsTotalMemoryKB * 1024 * 0.1;
   tsRpcQueueMemoryAllowed = TRANGE(tsRpcQueueMemoryAllowed, TSDB_MAX_MSG_SIZE * 10LL, TSDB_MAX_MSG_SIZE * 10000LL);
@@ -540,17 +542,17 @@ static int32_t taosUpdateServerCfg(SConfig *pCfg) {
 
   pItem = cfgGetItem(tsCfg, "numOfSnodeSharedThreads");
   if (pItem != NULL && pItem->stype == CFG_STYPE_DEFAULT) {
-    tsNumOfSnodeSharedThreads = numOfCores / 4;
+    tsNumOfSnodeStreamThreads = numOfCores / 4;
-    tsNumOfSnodeSharedThreads = TRANGE(tsNumOfSnodeSharedThreads, 2, 4);
+    tsNumOfSnodeStreamThreads = TRANGE(tsNumOfSnodeStreamThreads, 2, 4);
-    pItem->i32 = tsNumOfSnodeSharedThreads;
+    pItem->i32 = tsNumOfSnodeStreamThreads;
     pItem->stype = stype;
   }
 
   pItem = cfgGetItem(tsCfg, "numOfSnodeUniqueThreads");
   if (pItem != NULL && pItem->stype == CFG_STYPE_DEFAULT) {
-    tsNumOfSnodeUniqueThreads = numOfCores / 4;
+    tsNumOfSnodeWriteThreads = numOfCores / 4;
-    tsNumOfSnodeUniqueThreads = TRANGE(tsNumOfSnodeUniqueThreads, 2, 4);
+    tsNumOfSnodeWriteThreads = TRANGE(tsNumOfSnodeWriteThreads, 2, 4);
-    pItem->i32 = tsNumOfSnodeUniqueThreads;
+    pItem->i32 = tsNumOfSnodeWriteThreads;
     pItem->stype = stype;
   }
 
@@ -644,6 +646,7 @@ static int32_t taosSetClientCfg(SConfig *pCfg) {
   tsCompressColData = cfgGetItem(pCfg, "compressColData")->i32;
   tsNumOfTaskQueueThreads = cfgGetItem(pCfg, "numOfTaskQueueThreads")->i32;
   tsQueryPolicy = cfgGetItem(pCfg, "queryPolicy")->i32;
+  tsEnableQueryHb = cfgGetItem(pCfg, "enableQueryHb")->bval;
   tsQuerySmaOptimize = cfgGetItem(pCfg, "querySmaOptimize")->i32;
   tsQueryPlannerTrace = cfgGetItem(pCfg, "queryPlannerTrace")->bval;
   tsQueryNodeChunkSize = cfgGetItem(pCfg, "queryNodeChunkSize")->i32;
@@ -693,8 +696,8 @@ static int32_t taosSetServerCfg(SConfig *pCfg) {
   tsNumOfVnodeRsmaThreads = cfgGetItem(pCfg, "numOfVnodeRsmaThreads")->i32;
   tsNumOfQnodeQueryThreads = cfgGetItem(pCfg, "numOfQnodeQueryThreads")->i32;
   //  tsNumOfQnodeFetchThreads = cfgGetItem(pCfg, "numOfQnodeFetchThreads")->i32;
-  tsNumOfSnodeSharedThreads = cfgGetItem(pCfg, "numOfSnodeSharedThreads")->i32;
+  tsNumOfSnodeStreamThreads = cfgGetItem(pCfg, "numOfSnodeSharedThreads")->i32;
-  tsNumOfSnodeUniqueThreads = cfgGetItem(pCfg, "numOfSnodeUniqueThreads")->i32;
+  tsNumOfSnodeWriteThreads = cfgGetItem(pCfg, "numOfSnodeUniqueThreads")->i32;
   tsRpcQueueMemoryAllowed = cfgGetItem(pCfg, "rpcQueueMemoryAllowed")->i64;
 
   tsEnableMonitor = cfgGetItem(pCfg, "monitor")->bval;
@@ -780,6 +783,8 @@ int32_t taosSetCfg(SConfig *pCfg, char *name) {
       if (strcasecmp("enableCoreFile", name) == 0) {
         bool enableCore = cfgGetItem(pCfg, "enableCoreFile")->bval;
         taosSetCoreDump(enableCore);
+      } else if (strcasecmp("enableQueryHb", name) == 0) {
+        tsEnableQueryHb = cfgGetItem(pCfg, "enableQueryHb")->bval;
       }
       break;
     }
@@ -941,9 +946,9 @@ int32_t taosSetCfg(SConfig *pCfg, char *name) {
                 tsNumOfQnodeFetchThreads = cfgGetItem(pCfg, "numOfQnodeFetchThreads")->i32;
         */
       } else if (strcasecmp("numOfSnodeSharedThreads", name) == 0) {
-        tsNumOfSnodeSharedThreads = cfgGetItem(pCfg, "numOfSnodeSharedThreads")->i32;
+        tsNumOfSnodeStreamThreads = cfgGetItem(pCfg, "numOfSnodeSharedThreads")->i32;
       } else if (strcasecmp("numOfSnodeUniqueThreads", name) == 0) {
-        tsNumOfSnodeUniqueThreads = cfgGetItem(pCfg, "numOfSnodeUniqueThreads")->i32;
+        tsNumOfSnodeWriteThreads = cfgGetItem(pCfg, "numOfSnodeUniqueThreads")->i32;
       } else if (strcasecmp("numOfLogLines", name) == 0) {
         tsNumOfLogLines = cfgGetItem(pCfg, "numOfLogLines")->i32;
       }

source/common/src/tmsg.c

@@ -4853,6 +4853,7 @@ int32_t tSerializeSCMCreateStreamReq(void *buf, int32_t bufLen, const SCMCreateS
   if (tEncodeCStr(&encoder, pReq->sourceDB) < 0) return -1;
   if (tEncodeCStr(&encoder, pReq->targetStbFullName) < 0) return -1;
   if (tEncodeI8(&encoder, pReq->igExists) < 0) return -1;
+  if (tEncodeI8(&encoder, pReq->fillHistory) < 0) return -1;
   if (tEncodeI32(&encoder, sqlLen) < 0) return -1;
   if (tEncodeI32(&encoder, astLen) < 0) return -1;
   if (tEncodeI8(&encoder, pReq->triggerType) < 0) return -1;
@@ -4889,6 +4890,7 @@ int32_t tDeserializeSCMCreateStreamReq(void *buf, int32_t bufLen, SCMCreateStrea
   if (tDecodeCStrTo(&decoder, pReq->sourceDB) < 0) return -1;
   if (tDecodeCStrTo(&decoder, pReq->targetStbFullName) < 0) return -1;
   if (tDecodeI8(&decoder, &pReq->igExists) < 0) return -1;
+  if (tDecodeI8(&decoder, &pReq->fillHistory) < 0) return -1;
   if (tDecodeI32(&decoder, &sqlLen) < 0) return -1;
   if (tDecodeI32(&decoder, &astLen) < 0) return -1;
   if (tDecodeI8(&decoder, &pReq->triggerType) < 0) return -1;
@@ -4994,6 +4996,7 @@ int32_t tDeserializeSMRecoverStreamReq(void *buf, int32_t bufLen, SMRecoverStrea
 }
 
 void tFreeSCMCreateStreamReq(SCMCreateStreamReq *pReq) {
+  taosArrayDestroy(pReq->pTags);
   taosMemoryFreeClear(pReq->sql);
   taosMemoryFreeClear(pReq->ast);
 }
@@ -5158,7 +5161,7 @@ int tDecodeSVCreateTbReq(SDecoder *pCoder, SVCreateTbReq *pReq) {
       char  name[TSDB_COL_NAME_LEN] = {0};
       char *tmp = NULL;
       if (tDecodeCStr(pCoder, &tmp) < 0) return -1;
-      strcpy(name, tmp);
+      strncpy(name, tmp, TSDB_COL_NAME_LEN - 1);
       taosArrayPush(pReq->ctb.tagName, name);
     }
   } else if (pReq->type == TSDB_NORMAL_TABLE) {
@@ -5986,7 +5989,11 @@ int32_t tDecodeSMqDataRsp(SDecoder *pDecoder, SMqDataRsp *pRsp) {
       if (pRsp->withSchema) {
         SSchemaWrapper *pSW = (SSchemaWrapper *)taosMemoryCalloc(1, sizeof(SSchemaWrapper));
         if (pSW == NULL) return -1;
-        if (tDecodeSSchemaWrapper(pDecoder, pSW) < 0) return -1;
+        if (tDecodeSSchemaWrapper(pDecoder, pSW) < 0) {
+          taosMemoryFree(pSW);
+          return -1;
+        }
+
         taosArrayPush(pRsp->blockSchema, &pSW);
       }
 
@@ -6067,7 +6074,10 @@ int32_t tDecodeSTaosxRsp(SDecoder *pDecoder, STaosxRsp *pRsp) {
       if (pRsp->withSchema) {
         SSchemaWrapper *pSW = (SSchemaWrapper *)taosMemoryCalloc(1, sizeof(SSchemaWrapper));
         if (pSW == NULL) return -1;
-        if (tDecodeSSchemaWrapper(pDecoder, pSW) < 0) return -1;
+        if (tDecodeSSchemaWrapper(pDecoder, pSW) < 0) {
+          taosMemoryFree(pSW);
+          return -1;
+        }
         taosArrayPush(pRsp->blockSchema, &pSW);
       }
 

source/common/src/ttime.c

@@ -507,29 +507,29 @@ int64_t convertTimeFromPrecisionToUnit(int64_t time, int32_t fromPrecision, char
   double  tmp = time;
   switch (toUnit) {
     case 's': {
-      tmp /= (NANOSECOND_PER_SEC / factors[fromPrecision]);  // the result of division is an integer
       time /= (NANOSECOND_PER_SEC / factors[fromPrecision]);
+      tmp = (double)time;
       break;
     }
     case 'm':
-      tmp /= (NANOSECOND_PER_MINUTE / factors[fromPrecision]);  // the result of division is an integer
       time /= (NANOSECOND_PER_MINUTE / factors[fromPrecision]);
+      tmp = (double)time;
       break;
     case 'h':
-      tmp /= (NANOSECOND_PER_HOUR / factors[fromPrecision]);  // the result of division is an integer
       time /= (NANOSECOND_PER_HOUR / factors[fromPrecision]);
+      tmp = (double)time;
       break;
     case 'd':
-      tmp /= (NANOSECOND_PER_DAY / factors[fromPrecision]);  // the result of division is an integer
       time /= (NANOSECOND_PER_DAY / factors[fromPrecision]);
+      tmp = (double)time;
       break;
     case 'w':
-      tmp /= (NANOSECOND_PER_WEEK / factors[fromPrecision]);  // the result of division is an integer
       time /= (NANOSECOND_PER_WEEK / factors[fromPrecision]);
+      tmp = (double)time;
       break;
     case 'a':
-      tmp /= (NANOSECOND_PER_MSEC / factors[fromPrecision]);  // the result of division is an integer
       time /= (NANOSECOND_PER_MSEC / factors[fromPrecision]);
+      tmp = (double)time;
       break;
     case 'u':
       // the result of (NANOSECOND_PER_USEC/(double)factors[fromPrecision]) maybe a double
@@ -540,13 +540,13 @@ int64_t convertTimeFromPrecisionToUnit(int64_t time, int32_t fromPrecision, char
           break;
         }
         case TSDB_TIME_PRECISION_MICRO: {
-          tmp /= 1;
           time /= 1;
+          tmp = (double)time;
           break;
         }
         case TSDB_TIME_PRECISION_NANO: {
-          tmp /= 1000;
           time /= 1000;
+          tmp = (double)time;
           break;
         }
       }

source/common/src/ttszip.c

@@ -187,7 +187,9 @@ void* tsBufDestroy(STSBuf* pTSBuf) {
 
   if (pTSBuf->autoDelete) {
     //    ("tsBuf %p destroyed, delete tmp file:%s", pTSBuf, pTSBuf->path);
-    taosRemoveFile(pTSBuf->path);
+    if (taosRemoveFile(pTSBuf->path) != 0) {
+      // tscError("tsBuf %p destroyed, failed to remove tmp file:%s", pTSBuf, pTSBuf->path);
+    }
   } else {
     //    tscDebug("tsBuf %p destroyed, tmp file:%s, remains", pTSBuf, pTSBuf->path);
   }

source/dnode/mgmt/mgmt_dnode/src/dmHandle.c

@@ -332,7 +332,6 @@ SArray *dmGetMsgHandles() {
 
   // Requests handled by MNODE
   if (dmSetMgmtHandle(pArray, TDMT_MND_GRANT, dmPutNodeMsgToMgmtQueue, 0) == NULL) goto _OVER;
-  // if (dmSetMgmtHandle(pArray, TDMT_MND_GRANT_RSP, dmPutNodeMsgToMgmtQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_AUTH_RSP, dmPutNodeMsgToMgmtQueue, 0) == NULL) goto _OVER;
 
   code = 0;

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

@@ -124,6 +124,7 @@ SArray *mmGetMsgHandles() {
   if (dmSetMgmtHandle(pArray, TDMT_MND_VGROUP_LIST, mmPutMsgToReadQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_REDISTRIBUTE_VGROUP, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_MERGE_VGROUP, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_MND_SPLIT_VGROUP, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_BALANCE_VGROUP, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_CREATE_FUNC, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_RETRIEVE_FUNC, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;

source/dnode/mgmt/mgmt_snode/inc/smInt.h

@@ -30,9 +30,9 @@ typedef struct SSnodeMgmt {
   SMsgCb        msgCb;
   const char   *path;
   const char   *name;
-  int8_t        uniqueWorkerInUse;
+  int8_t        writeWorkerInUse;
-  SArray       *uniqueWorkers;  // SArray<SMultiWorker*>
+  SArray       *writeWroker;  // SArray<SMultiWorker*>
-  SSingleWorker sharedWorker;
+  SSingleWorker streamWorker;
 } SSnodeMgmt;
 
 // smHandle.c
@@ -43,10 +43,11 @@ int32_t smProcessDropReq(const SMgmtInputOpt *pInput, SRpcMsg *pMsg);
 // smWorker.c
 int32_t smStartWorker(SSnodeMgmt *pMgmt);
 void    smStopWorker(SSnodeMgmt *pMgmt);
+int32_t smPutMsgToQueue(SSnodeMgmt *pMgmt, EQueueType qtype, SRpcMsg *pMsg);
 int32_t smPutNodeMsgToMgmtQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg);
-int32_t smPutNodeMsgToUniqueQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg);
+int32_t smPutNodeMsgToWriteQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg);
-int32_t smPutNodeMsgToSharedQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg);
+int32_t smPutNodeMsgToStreamQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg);
-int32_t smPutNodeMsgToExecQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg);
+void    sndEnqueueStreamDispatch(SSnode *pSnode, SRpcMsg *pMsg);
 
 #ifdef __cplusplus
 }

source/dnode/mgmt/mgmt_snode/src/smHandle.c

@@ -69,13 +69,13 @@ SArray *smGetMsgHandles() {
 
   if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_DEPLOY, smPutNodeMsgToMgmtQueue, 1) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_DROP, smPutNodeMsgToMgmtQueue, 1) == NULL) goto _OVER;
-  if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_RUN, smPutNodeMsgToSharedQueue, 1) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_RUN, smPutNodeMsgToStreamQueue, 1) == NULL) goto _OVER;
-  if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_DISPATCH, smPutNodeMsgToSharedQueue, 1) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_DISPATCH, smPutNodeMsgToStreamQueue, 1) == NULL) goto _OVER;
-  if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_DISPATCH_RSP, smPutNodeMsgToSharedQueue, 1) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_DISPATCH_RSP, smPutNodeMsgToStreamQueue, 1) == NULL) goto _OVER;
-  if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_RECOVER, smPutNodeMsgToSharedQueue, 1) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_RECOVER, smPutNodeMsgToStreamQueue, 1) == NULL) goto _OVER;
-  if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_RECOVER_RSP, smPutNodeMsgToSharedQueue, 1) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_STREAM_TASK_RECOVER_RSP, smPutNodeMsgToStreamQueue, 1) == NULL) goto _OVER;
-  if (dmSetMgmtHandle(pArray, TDMT_STREAM_RETRIEVE, smPutNodeMsgToSharedQueue, 1) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_STREAM_RETRIEVE, smPutNodeMsgToStreamQueue, 1) == NULL) goto _OVER;
-  if (dmSetMgmtHandle(pArray, TDMT_STREAM_RETRIEVE_RSP, smPutNodeMsgToSharedQueue, 1) == NULL) goto _OVER;
+  if (dmSetMgmtHandle(pArray, TDMT_STREAM_RETRIEVE_RSP, smPutNodeMsgToStreamQueue, 1) == NULL) goto _OVER;
 
   code = 0;
 _OVER:

source/dnode/mgmt/mgmt_snode/src/smInt.c

@@ -45,6 +45,7 @@ int32_t smOpen(SMgmtInputOpt *pInput, SMgmtOutputOpt *pOutput) {
   pMgmt->name = pInput->name;
   pMgmt->msgCb = pInput->msgCb;
   pMgmt->msgCb.mgmt = pMgmt;
+  pMgmt->msgCb.putToQueueFp = (PutToQueueFp)smPutMsgToQueue;
 
   SSnodeOpt option = {0};
   smInitOption(pMgmt, &option);

source/dnode/mgmt/mgmt_snode/src/smWorker.c

@@ -26,18 +26,24 @@ static inline void smSendRsp(SRpcMsg *pMsg, int32_t code) {
   tmsgSendRsp(&rsp);
 }
 
-static void smProcessUniqueQueue(SQueueInfo *pInfo, STaosQall *qall, int32_t numOfMsgs) {
+static void smProcessWriteQueue(SQueueInfo *pInfo, STaosQall *qall, int32_t numOfMsgs) {
   SSnodeMgmt *pMgmt = pInfo->ahandle;
 
   for (int32_t i = 0; i < numOfMsgs; i++) {
     SRpcMsg *pMsg = NULL;
     taosGetQitem(qall, (void **)&pMsg);
+    const STraceId *trace = &pMsg->info.traceId;
 
-    dTrace("msg:%p, get from snode-unique queue", pMsg);
+    dTrace("msg:%p, get from snode-write queue", pMsg);
-    if (sndProcessUMsg(pMgmt->pSnode, pMsg) < 0) {
+    int32_t code = sndProcessWriteMsg(pMgmt->pSnode, pMsg, NULL);
-      ASSERT(0);
+    if (code < 0) {
+      dGError("snd, msg:%p failed to process write since %s", pMsg, terrstr(code));
+      if (pMsg->info.handle != NULL) {
+        tmsgSendRsp(pMsg);
+      }
+    } else {
+      smSendRsp(pMsg, 0);
     }
-    smSendRsp(pMsg, 0);
 
     dTrace("msg:%p, is freed", pMsg);
     rpcFreeCont(pMsg->pCont);
@@ -45,13 +51,15 @@ static void smProcessUniqueQueue(SQueueInfo *pInfo, STaosQall *qall, int32_t num
   }
 }
 
-static void smProcessSharedQueue(SQueueInfo *pInfo, SRpcMsg *pMsg) {
+static void smProcessStreamQueue(SQueueInfo *pInfo, SRpcMsg *pMsg) {
-  SSnodeMgmt *pMgmt = pInfo->ahandle;
+  SSnodeMgmt     *pMgmt = pInfo->ahandle;
+  const STraceId *trace = &pMsg->info.traceId;
 
-  dTrace("msg:%p, get from snode-shared queue", pMsg);
+  dTrace("msg:%p, get from snode-stream queue", pMsg);
-  if (sndProcessSMsg(pMgmt->pSnode, pMsg) < 0) {
+  int32_t code = sndProcessStreamMsg(pMgmt->pSnode, pMsg);
+  if (code < 0) {
+    dGError("snd, msg:%p failed to process stream since %s", pMsg, terrstr(code));
     smSendRsp(pMsg, terrno);
-    ASSERT(0);
   }
 
   dTrace("msg:%p, is freed", pMsg);
@@ -60,44 +68,44 @@ static void smProcessSharedQueue(SQueueInfo *pInfo, SRpcMsg *pMsg) {
 }
 
 int32_t smStartWorker(SSnodeMgmt *pMgmt) {
-  pMgmt->uniqueWorkers = taosArrayInit(0, sizeof(SMultiWorker *));
+  pMgmt->writeWroker = taosArrayInit(0, sizeof(SMultiWorker *));
-  if (pMgmt->uniqueWorkers == NULL) {
+  if (pMgmt->writeWroker == NULL) {
     terrno = TSDB_CODE_OUT_OF_MEMORY;
     return -1;
   }
 
-  for (int32_t i = 0; i < tsNumOfSnodeUniqueThreads; i++) {
+  for (int32_t i = 0; i < tsNumOfSnodeWriteThreads; i++) {
-    SMultiWorker *pUniqueWorker = taosMemoryMalloc(sizeof(SMultiWorker));
+    SMultiWorker *pWriteWorker = taosMemoryMalloc(sizeof(SMultiWorker));
-    if (pUniqueWorker == NULL) {
+    if (pWriteWorker == NULL) {
       terrno = TSDB_CODE_OUT_OF_MEMORY;
       return -1;
     }
 
     SMultiWorkerCfg cfg = {
         .max = 1,
-        .name = "snode-unique",
+        .name = "snode-write",
-        .fp = smProcessUniqueQueue,
+        .fp = smProcessWriteQueue,
         .param = pMgmt,
     };
-    if (tMultiWorkerInit(pUniqueWorker, &cfg) != 0) {
+    if (tMultiWorkerInit(pWriteWorker, &cfg) != 0) {
       dError("failed to start snode-unique worker since %s", terrstr());
       return -1;
     }
-    if (taosArrayPush(pMgmt->uniqueWorkers, &pUniqueWorker) == NULL) {
+    if (taosArrayPush(pMgmt->writeWroker, &pWriteWorker) == NULL) {
       terrno = TSDB_CODE_OUT_OF_MEMORY;
       return -1;
     }
   }
 
   SSingleWorkerCfg cfg = {
-      .min = tsNumOfSnodeSharedThreads,
+      .min = tsNumOfSnodeStreamThreads,
-      .max = tsNumOfSnodeSharedThreads,
+      .max = tsNumOfSnodeStreamThreads,
-      .name = "snode-shared",
+      .name = "snode-stream",
-      .fp = (FItem)smProcessSharedQueue,
+      .fp = (FItem)smProcessStreamQueue,
       .param = pMgmt,
   };
 
-  if (tSingleWorkerInit(&pMgmt->sharedWorker, &cfg)) {
+  if (tSingleWorkerInit(&pMgmt->streamWorker, &cfg)) {
     dError("failed to start snode shared-worker since %s", terrstr());
     return -1;
   }
@@ -107,30 +115,50 @@ int32_t smStartWorker(SSnodeMgmt *pMgmt) {
 }
 
 void smStopWorker(SSnodeMgmt *pMgmt) {
-  for (int32_t i = 0; i < taosArrayGetSize(pMgmt->uniqueWorkers); i++) {
+  for (int32_t i = 0; i < taosArrayGetSize(pMgmt->writeWroker); i++) {
-    SMultiWorker *pWorker = taosArrayGetP(pMgmt->uniqueWorkers, i);
+    SMultiWorker *pWorker = taosArrayGetP(pMgmt->writeWroker, i);
     tMultiWorkerCleanup(pWorker);
   }
-  taosArrayDestroy(pMgmt->uniqueWorkers);
+  taosArrayDestroy(pMgmt->writeWroker);
-  tSingleWorkerCleanup(&pMgmt->sharedWorker);
+  tSingleWorkerCleanup(&pMgmt->streamWorker);
   dDebug("snode workers are closed");
 }
 
-static FORCE_INLINE int32_t smGetSWIdFromMsg(SRpcMsg *pMsg) {
+int32_t smPutMsgToQueue(SSnodeMgmt *pMgmt, EQueueType qtype, SRpcMsg *pRpc) {
-  SMsgHead *pHead = pMsg->pCont;
+  SRpcMsg *pMsg = taosAllocateQitem(sizeof(SRpcMsg), RPC_QITEM);
-  pHead->vgId = htonl(pHead->vgId);
+  if (pMsg == NULL) {
-  return pHead->vgId % tsNumOfSnodeUniqueThreads;
+    rpcFreeCont(pRpc->pCont);
-}
+    pRpc->pCont = NULL;
+    return -1;
+  }
 
-static FORCE_INLINE int32_t smGetSWTypeFromMsg(SRpcMsg *pMsg) {
+  SSnode *pSnode = pMgmt->pSnode;
-  /*SMsgHead *pHead = pMsg->pCont;*/
+  if (pSnode == NULL) {
-  /*pHead->workerType = htonl(pHead->workerType);*/
+    dError("snode: msg:%p failed to put into vnode queue since %s, type:%s qtype:%d", pMsg, terrstr(),
-  /*return pHead->workerType;*/
+           TMSG_INFO(pMsg->msgType), qtype);
+    return -1;
+  }
+
+  SMsgHead *pHead = pRpc->pCont;
+  pHead->contLen = htonl(pHead->contLen);
+  pHead->vgId = SNODE_HANDLE;
+  memcpy(pMsg, pRpc, sizeof(SRpcMsg));
+
+  switch (qtype) {
+    case STREAM_QUEUE:
+      smPutNodeMsgToStreamQueue(pMgmt, pMsg);
+      break;
+    case WRITE_QUEUE:
+      smPutNodeMsgToWriteQueue(pMgmt, pMsg);
+      break;
+    default:
+      ASSERT(0);
+  }
   return 0;
 }
 
 int32_t smPutNodeMsgToMgmtQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg) {
-  SMultiWorker *pWorker = taosArrayGetP(pMgmt->uniqueWorkers, 0);
+  SMultiWorker *pWorker = taosArrayGetP(pMgmt->writeWroker, 0);
   if (pWorker == NULL) {
     terrno = TSDB_CODE_INVALID_MSG;
     return -1;
@@ -141,9 +169,8 @@ int32_t smPutNodeMsgToMgmtQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg) {
   return 0;
 }
 
-int32_t smPutNodeMsgToUniqueQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg) {
+int32_t smPutNodeMsgToWriteQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg) {
-  int32_t       index = smGetSWIdFromMsg(pMsg);
+  SMultiWorker *pWorker = taosArrayGetP(pMgmt->writeWroker, 0);
-  SMultiWorker *pWorker = taosArrayGetP(pMgmt->uniqueWorkers, index);
   if (pWorker == NULL) {
     terrno = TSDB_CODE_INVALID_MSG;
     return -1;
@@ -154,19 +181,14 @@ int32_t smPutNodeMsgToUniqueQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg) {
   return 0;
 }
 
-int32_t smPutNodeMsgToSharedQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg) {
+int32_t smPutNodeMsgToStreamQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg) {
-  SSingleWorker *pWorker = &pMgmt->sharedWorker;
+  SSingleWorker *pWorker = &pMgmt->streamWorker;
 
   dTrace("msg:%p, put into worker %s", pMsg, pWorker->name);
-  taosWriteQitem(pWorker->queue, pMsg);
+  if (pMsg->msgType == TDMT_STREAM_TASK_DISPATCH) {
-  return 0;
+    sndEnqueueStreamDispatch(pMgmt->pSnode, pMsg);
-}
-
-int32_t smPutNodeMsgToExecQueue(SSnodeMgmt *pMgmt, SRpcMsg *pMsg) {
-  int32_t workerType = smGetSWTypeFromMsg(pMsg);
-  if (workerType == SND_WORKER_TYPE__SHARED) {
-    return smPutNodeMsgToSharedQueue(pMgmt, pMsg);
   } else {
-    return smPutNodeMsgToUniqueQueue(pMgmt, pMsg);
+    taosWriteQitem(pWorker->queue, pMsg);
   }
+  return 0;
 }

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

@@ -307,6 +307,10 @@ int32_t vmGetQueueSize(SVnodeMgmt *pMgmt, int32_t vgId, EQueueType qtype) {
     }
     vmReleaseVnode(pMgmt, pVnode);
   }
+  if (size < 0) {
+    dError("vgId:%d, can't get size from queue since %s, qtype:%d", vgId, terrstr(), qtype);
+    size = 0;
+  }
   return size;
 }
 

source/dnode/mgmt/test/sut/inc/server.h

@@ -20,6 +20,7 @@ class TestServer {
  public:
   bool Start();
   void Stop();
+bool runnning;
   
  private:
   TdThread threadId;

source/dnode/mgmt/test/sut/src/server.cpp

@@ -16,8 +16,18 @@
 #include "sut.h"
 
 void* serverLoop(void* param) {
-  dmInit();
+  TestServer* server = (TestServer*)param;
-  dmRun();
+  server->runnning = false;
+
+  if (dmInit() != 0) {
+    return NULL;
+  }
+
+  server->runnning = true;
+  if (dmRun() != 0) {
+    return NULL;
+  }
+
   dmCleanup();
   return NULL;
 }
@@ -26,10 +36,10 @@ bool TestServer::Start() {
   TdThreadAttr thAttr;
   taosThreadAttrInit(&thAttr);
   taosThreadAttrSetDetachState(&thAttr, PTHREAD_CREATE_JOINABLE);
-  taosThreadCreate(&threadId, &thAttr, serverLoop, NULL);
+  taosThreadCreate(&threadId, &thAttr, serverLoop, this);
   taosThreadAttrDestroy(&thAttr);
   taosMsleep(2100);
-  return true;
+  return runnning;
 }
 
 void TestServer::Stop() {

source/dnode/mgmt/test/sut/src/sut.cpp

@@ -53,7 +53,10 @@ void Testbase::Init(const char* path, int16_t port) {
   taosMkDir(path);
   InitLog(TD_TMP_DIR_PATH "td");
 
-  server.Start();
+  if (!server.Start()) {
+    printf("failed to start server, exit\n");
+    exit(0);
+  };
   client.Init("root", "taosdata");
   showRsp = NULL;
 }

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

@@ -80,8 +80,8 @@ typedef struct {
 } SProfileMgmt;
 
 typedef struct {
-  SRWLatch lock;
+  TdThreadMutex lock;
-  char     email[TSDB_FQDN_LEN];
+  char           email[TSDB_FQDN_LEN];
 } STelemMgmt;
 
 typedef struct {

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

@@ -1,50 +0,0 @@
-/*
- * 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/>.
- */
-
-#ifndef _TD_MND_OFFSET_H_
-#define _TD_MND_OFFSET_H_
-
-#include "mndInt.h"
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-int32_t mndInitOffset(SMnode *pMnode);
-void    mndCleanupOffset(SMnode *pMnode);
-
-SMqOffsetObj *mndAcquireOffset(SMnode *pMnode, const char *key);
-void          mndReleaseOffset(SMnode *pMnode, SMqOffsetObj *pOffset);
-
-SSdbRaw *mndOffsetActionEncode(SMqOffsetObj *pOffset);
-SSdbRow *mndOffsetActionDecode(SSdbRaw *pRaw);
-
-int32_t mndCreateOffsets(STrans *pTrans, const char *cgroup, const char *topicName, const SArray *vgs);
-
-static FORCE_INLINE int32_t mndMakePartitionKey(char *key, const char *cgroup, const char *topicName, int32_t vgId) {
-  return snprintf(key, TSDB_PARTITION_KEY_LEN, "%d:%s:%s", vgId, cgroup, topicName);
-}
-
-int32_t mndDropOffsetByDB(SMnode *pMnode, STrans *pTrans, SDbObj *pDb);
-int32_t mndDropOffsetByTopic(SMnode *pMnode, STrans *pTrans, const char *topic);
-int32_t mndDropOffsetBySubKey(SMnode *pMnode, STrans *pTrans, const char *subKey);
-
-bool mndOffsetFromTopic(SMqOffsetObj *pOffset, const char *topic);
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif /*_TD_MND_OFFSET_H_*/

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

@@ -31,6 +31,10 @@ void             mndReleaseSubscribe(SMnode *pMnode, SMqSubscribeObj *pSub);
 
 int32_t mndMakeSubscribeKey(char *key, const char *cgroup, const char *topicName);
 
+static FORCE_INLINE int32_t mndMakePartitionKey(char *key, const char *cgroup, const char *topicName, int32_t vgId) {
+  return snprintf(key, TSDB_PARTITION_KEY_LEN, "%d:%s:%s", vgId, cgroup, topicName);
+}
+
 int32_t mndDropSubByDB(SMnode *pMnode, STrans *pTrans, SDbObj *pDb);
 int32_t mndDropSubByTopic(SMnode *pMnode, STrans *pTrans, const char *topic);
 int32_t mndSetDropSubCommitLogs(SMnode *pMnode, STrans *pTrans, SMqSubscribeObj *pSub);

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

@@ -18,7 +18,6 @@
 #include "mndDb.h"
 #include "mndDnode.h"
 #include "mndMnode.h"
-#include "mndOffset.h"
 #include "mndPrivilege.h"
 #include "mndShow.h"
 #include "mndStb.h"
@@ -408,12 +407,6 @@ static int32_t mndProcessAskEpReq(SRpcMsg *pMsg) {
             .offset = -1,
         };
 
-        // 2.2.2 fetch vg offset
-        SMqOffsetObj *pOffsetObj = mndAcquireOffset(pMnode, offsetKey);
-        if (pOffsetObj != NULL) {
-          vgEp.offset = atomic_load_64(&pOffsetObj->offset);
-          mndReleaseOffset(pMnode, pOffsetObj);
-        }
         taosArrayPush(topicEp.vgs, &vgEp);
       }
       taosArrayPush(rsp.topics, &topicEp);

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

@@ -17,7 +17,6 @@
 #include "mndDb.h"
 #include "mndCluster.h"
 #include "mndDnode.h"
-#include "mndOffset.h"
 #include "mndPrivilege.h"
 #include "mndShow.h"
 #include "mndSma.h"
@@ -641,12 +640,8 @@ static int32_t mndSetDbCfgFromAlterDbReq(SDbObj *pDb, SAlterDbReq *pAlter) {
   }
 
   if (pAlter->pageSize > 0 && pAlter->pageSize != pDb->cfg.pageSize) {
-#if 1
-    terrno = TSDB_CODE_OPS_NOT_SUPPORT;
-#else
     pDb->cfg.pageSize = pAlter->pageSize;
     terrno = 0;
-#endif
   }
 
   if (pAlter->daysPerFile > 0 && pAlter->daysPerFile != pDb->cfg.daysPerFile) {

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

@@ -776,11 +776,6 @@ static int32_t mndProcessDropDnodeReq(SRpcMsg *pReq) {
     }
   }
 
-  if (numOfVnodes > 0) {
-    terrno = TSDB_CODE_OPS_NOT_SUPPORT;
-    goto _OVER;
-  }
-
   code = mndDropDnode(pMnode, pReq, pDnode, pMObj, pQObj, pSObj, numOfVnodes);
   if (code == 0) code = TSDB_CODE_ACTION_IN_PROGRESS;
 

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

@@ -23,7 +23,6 @@
 #include "mndGrant.h"
 #include "mndInfoSchema.h"
 #include "mndMnode.h"
-#include "mndOffset.h"
 #include "mndPerfSchema.h"
 #include "mndPrivilege.h"
 #include "mndProfile.h"
@@ -305,7 +304,6 @@ static int32_t mndInitSteps(SMnode *pMnode) {
   if (mndAllocStep(pMnode, "mnode-topic", mndInitTopic, mndCleanupTopic) != 0) return -1;
   if (mndAllocStep(pMnode, "mnode-consumer", mndInitConsumer, mndCleanupConsumer) != 0) return -1;
   if (mndAllocStep(pMnode, "mnode-subscribe", mndInitSubscribe, mndCleanupSubscribe) != 0) return -1;
-  if (mndAllocStep(pMnode, "mnode-offset", mndInitOffset, mndCleanupOffset) != 0) return -1;
   if (mndAllocStep(pMnode, "mnode-vgroup", mndInitVgroup, mndCleanupVgroup) != 0) return -1;
   if (mndAllocStep(pMnode, "mnode-stb", mndInitStb, mndCleanupStb) != 0) return -1;
   if (mndAllocStep(pMnode, "mnode-sma", mndInitSma, mndCleanupSma) != 0) return -1;
@@ -424,7 +422,7 @@ SMnode *mndOpen(const char *path, const SMnodeOpt *pOption) {
     return NULL;
   }
 
-  mInfo("mnode open successfully ");
+  mInfo("mnode open successfully");
   return pMnode;
 }
 

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

@@ -1,370 +0,0 @@
-/*
- * 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/>.
- */
-
-#define _DEFAULT_SOURCE
-#include "mndOffset.h"
-#include "mndDb.h"
-#include "mndDnode.h"
-#include "mndMnode.h"
-#include "mndPrivilege.h"
-#include "mndShow.h"
-#include "mndStb.h"
-#include "mndTopic.h"
-#include "mndTrans.h"
-#include "mndUser.h"
-#include "mndVgroup.h"
-#include "tname.h"
-
-#define MND_OFFSET_VER_NUMBER   1
-#define MND_OFFSET_RESERVE_SIZE 64
-
-static int32_t mndOffsetActionInsert(SSdb *pSdb, SMqOffsetObj *pOffset);
-static int32_t mndOffsetActionDelete(SSdb *pSdb, SMqOffsetObj *pOffset);
-static int32_t mndOffsetActionUpdate(SSdb *pSdb, SMqOffsetObj *pOffset, SMqOffsetObj *pNewOffset);
-static int32_t mndProcessCommitOffsetReq(SRpcMsg *pReq);
-
-int32_t mndInitOffset(SMnode *pMnode) {
-  SSdbTable table = {
-      .sdbType = SDB_OFFSET,
-      .keyType = SDB_KEY_BINARY,
-      .encodeFp = (SdbEncodeFp)mndOffsetActionEncode,
-      .decodeFp = (SdbDecodeFp)mndOffsetActionDecode,
-      .insertFp = (SdbInsertFp)mndOffsetActionInsert,
-      .updateFp = (SdbUpdateFp)mndOffsetActionUpdate,
-      .deleteFp = (SdbDeleteFp)mndOffsetActionDelete,
-  };
-
-  mndSetMsgHandle(pMnode, TDMT_MND_MQ_COMMIT_OFFSET, mndProcessCommitOffsetReq);
-
-  return sdbSetTable(pMnode->pSdb, table);
-}
-
-void mndCleanupOffset(SMnode *pMnode) {}
-
-bool mndOffsetFromTopic(SMqOffsetObj *pOffset, const char *topic) {
-  int32_t i = 0;
-  while (pOffset->key[i] != ':') i++;
-  while (pOffset->key[i] != ':') i++;
-  if (strcmp(&pOffset->key[i + 1], topic) == 0) return true;
-  return false;
-}
-
-bool mndOffsetFromSubKey(SMqOffsetObj *pOffset, const char *subKey) {
-  int32_t i = 0;
-  while (pOffset->key[i] != ':') i++;
-  if (strcmp(&pOffset->key[i + 1], subKey) == 0) return true;
-  return false;
-}
-SSdbRaw *mndOffsetActionEncode(SMqOffsetObj *pOffset) {
-  terrno = TSDB_CODE_OUT_OF_MEMORY;
-  void   *buf = NULL;
-  int32_t tlen = tEncodeSMqOffsetObj(NULL, pOffset);
-  int32_t size = sizeof(int32_t) + tlen + MND_OFFSET_RESERVE_SIZE;
-
-  SSdbRaw *pRaw = sdbAllocRaw(SDB_OFFSET, MND_OFFSET_VER_NUMBER, size);
-  if (pRaw == NULL) goto OFFSET_ENCODE_OVER;
-
-  buf = taosMemoryMalloc(tlen);
-  if (buf == NULL) goto OFFSET_ENCODE_OVER;
-
-  void *abuf = buf;
-  tEncodeSMqOffsetObj(&abuf, pOffset);
-
-  int32_t dataPos = 0;
-  SDB_SET_INT32(pRaw, dataPos, tlen, OFFSET_ENCODE_OVER);
-  SDB_SET_BINARY(pRaw, dataPos, buf, tlen, OFFSET_ENCODE_OVER);
-  SDB_SET_RESERVE(pRaw, dataPos, MND_OFFSET_RESERVE_SIZE, OFFSET_ENCODE_OVER);
-  SDB_SET_DATALEN(pRaw, dataPos, OFFSET_ENCODE_OVER);
-
-  terrno = TSDB_CODE_SUCCESS;
-
-OFFSET_ENCODE_OVER:
-  taosMemoryFreeClear(buf);
-  if (terrno != TSDB_CODE_SUCCESS) {
-    mError("offset:%s, failed to encode to raw:%p since %s", pOffset->key, pRaw, terrstr());
-    sdbFreeRaw(pRaw);
-    return NULL;
-  }
-
-  mTrace("offset:%s, encode to raw:%p, row:%p", pOffset->key, pRaw, pOffset);
-  return pRaw;
-}
-
-SSdbRow *mndOffsetActionDecode(SSdbRaw *pRaw) {
-  terrno = TSDB_CODE_OUT_OF_MEMORY;
-  void *buf = NULL;
-
-  int8_t sver = 0;
-  if (sdbGetRawSoftVer(pRaw, &sver) != 0) goto OFFSET_DECODE_OVER;
-
-  if (sver != MND_OFFSET_VER_NUMBER) {
-    terrno = TSDB_CODE_SDB_INVALID_DATA_VER;
-    goto OFFSET_DECODE_OVER;
-  }
-
-  int32_t  size = sizeof(SMqOffsetObj);
-  SSdbRow *pRow = sdbAllocRow(size);
-  if (pRow == NULL) goto OFFSET_DECODE_OVER;
-
-  SMqOffsetObj *pOffset = sdbGetRowObj(pRow);
-  if (pOffset == NULL) goto OFFSET_DECODE_OVER;
-
-  int32_t dataPos = 0;
-  int32_t tlen;
-  SDB_GET_INT32(pRaw, dataPos, &tlen, OFFSET_DECODE_OVER);
-  buf = taosMemoryMalloc(tlen + 1);
-  if (buf == NULL) goto OFFSET_DECODE_OVER;
-  SDB_GET_BINARY(pRaw, dataPos, buf, tlen, OFFSET_DECODE_OVER);
-  SDB_GET_RESERVE(pRaw, dataPos, MND_OFFSET_RESERVE_SIZE, OFFSET_DECODE_OVER);
-
-  if (tDecodeSMqOffsetObj(buf, pOffset) == NULL) {
-    goto OFFSET_DECODE_OVER;
-  }
-
-  terrno = TSDB_CODE_SUCCESS;
-
-OFFSET_DECODE_OVER:
-  taosMemoryFreeClear(buf);
-  if (terrno != TSDB_CODE_SUCCESS) {
-    mError("offset:%s, failed to decode from raw:%p since %s", pOffset->key, pRaw, terrstr());
-    taosMemoryFreeClear(pRow);
-    return NULL;
-  }
-
-  mTrace("offset:%s, decode from raw:%p, row:%p", pOffset->key, pRaw, pOffset);
-  return pRow;
-}
-
-int32_t mndCreateOffsets(STrans *pTrans, const char *cgroup, const char *topicName, const SArray *vgs) {
-  int32_t sz = taosArrayGetSize(vgs);
-  for (int32_t i = 0; i < sz; i++) {
-    int32_t      vgId = *(int32_t *)taosArrayGet(vgs, i);
-    SMqOffsetObj offsetObj = {0};
-    if (mndMakePartitionKey(offsetObj.key, cgroup, topicName, vgId) < 0) {
-      return -1;
-    }
-    // TODO assign db
-    offsetObj.offset = -1;
-    SSdbRaw *pOffsetRaw = mndOffsetActionEncode(&offsetObj);
-    if (pOffsetRaw == NULL) {
-      return -1;
-    }
-    (void)sdbSetRawStatus(pOffsetRaw, SDB_STATUS_READY);
-    // commit log or redo log?
-    if (mndTransAppendRedolog(pTrans, pOffsetRaw) < 0) {
-      return -1;
-    }
-  }
-  return 0;
-}
-
-static int32_t mndProcessCommitOffsetReq(SRpcMsg *pMsg) {
-  char key[TSDB_PARTITION_KEY_LEN];
-
-  SMnode              *pMnode = pMsg->info.node;
-  char                *msgStr = pMsg->pCont;
-  SMqCMCommitOffsetReq commitOffsetReq;
-  SDecoder             decoder;
-  tDecoderInit(&decoder, msgStr, pMsg->contLen);
-
-  tDecodeSMqCMCommitOffsetReq(&decoder, &commitOffsetReq);
-
-  STrans *pTrans = mndTransCreate(pMnode, TRN_POLICY_ROLLBACK, TRN_CONFLICT_NOTHING, pMsg, "commit-offset");
-  if (pTrans == NULL) {
-    terrno = TSDB_CODE_OUT_OF_MEMORY;
-    tDecoderClear(&decoder);
-    return -1;
-  }
-
-  for (int32_t i = 0; i < commitOffsetReq.num; i++) {
-    SMqOffset *pOffset = &commitOffsetReq.offsets[i];
-    mInfo("commit offset %" PRId64 " to vgId:%d of consumer group %s on topic %s", pOffset->offset, pOffset->vgId,
-          pOffset->cgroup, pOffset->topicName);
-    if (mndMakePartitionKey(key, pOffset->cgroup, pOffset->topicName, pOffset->vgId) < 0) {
-      mError("submit offset to topic %s failed", pOffset->topicName);
-      return -1;
-    }
-    bool          create = false;
-    SMqOffsetObj *pOffsetObj = mndAcquireOffset(pMnode, key);
-    if (pOffsetObj == NULL) {
-      SMqTopicObj *pTopic = mndAcquireTopic(pMnode, pOffset->topicName);
-      if (pTopic == NULL) {
-        terrno = TSDB_CODE_MND_TOPIC_NOT_EXIST;
-        mError("submit offset to topic %s failed since %s", pOffset->topicName, terrstr());
-        continue;
-      }
-      pOffsetObj = taosMemoryMalloc(sizeof(SMqOffsetObj));
-      pOffsetObj->dbUid = pTopic->dbUid;
-      mndReleaseTopic(pMnode, pTopic);
-      memcpy(pOffsetObj->key, key, TSDB_PARTITION_KEY_LEN);
-      create = true;
-    }
-    pOffsetObj->offset = pOffset->offset;
-    SSdbRaw *pOffsetRaw = mndOffsetActionEncode(pOffsetObj);
-    (void)sdbSetRawStatus(pOffsetRaw, SDB_STATUS_READY);
-    mndTransAppendCommitlog(pTrans, pOffsetRaw);
-    if (create) {
-      taosMemoryFree(pOffsetObj);
-    } else {
-      mndReleaseOffset(pMnode, pOffsetObj);
-    }
-  }
-
-  tDecoderClear(&decoder);
-
-  if (mndTransPrepare(pMnode, pTrans) != 0) {
-    mError("mq-commit-offset-trans:%d, failed to prepare since %s", pTrans->id, terrstr());
-    mndTransDrop(pTrans);
-    return -1;
-  }
-
-  mndTransDrop(pTrans);
-  return TSDB_CODE_ACTION_IN_PROGRESS;
-}
-
-static int32_t mndOffsetActionInsert(SSdb *pSdb, SMqOffsetObj *pOffset) {
-  mTrace("offset:%s, perform insert action", pOffset->key);
-  return 0;
-}
-
-static int32_t mndOffsetActionDelete(SSdb *pSdb, SMqOffsetObj *pOffset) {
-  mTrace("offset:%s, perform delete action", pOffset->key);
-  return 0;
-}
-
-static int32_t mndOffsetActionUpdate(SSdb *pSdb, SMqOffsetObj *pOldOffset, SMqOffsetObj *pNewOffset) {
-  mTrace("offset:%s, perform update action", pOldOffset->key);
-  atomic_store_64(&pOldOffset->offset, pNewOffset->offset);
-  return 0;
-}
-
-SMqOffsetObj *mndAcquireOffset(SMnode *pMnode, const char *key) {
-  SSdb         *pSdb = pMnode->pSdb;
-  SMqOffsetObj *pOffset = sdbAcquire(pSdb, SDB_OFFSET, key);
-  if (pOffset == NULL && terrno == TSDB_CODE_SDB_OBJ_NOT_THERE) {
-    terrno = TSDB_CODE_MND_OFFSET_NOT_EXIST;
-  }
-  return pOffset;
-}
-
-void mndReleaseOffset(SMnode *pMnode, SMqOffsetObj *pOffset) {
-  SSdb *pSdb = pMnode->pSdb;
-  sdbRelease(pSdb, pOffset);
-}
-
-static void mndCancelGetNextOffset(SMnode *pMnode, void *pIter) {
-  SSdb *pSdb = pMnode->pSdb;
-  sdbCancelFetch(pSdb, pIter);
-}
-
-static int32_t mndSetDropOffsetCommitLogs(SMnode *pMnode, STrans *pTrans, SMqOffsetObj *pOffset) {
-  SSdbRaw *pCommitRaw = mndOffsetActionEncode(pOffset);
-  if (pCommitRaw == NULL) return -1;
-  if (mndTransAppendCommitlog(pTrans, pCommitRaw) != 0) return -1;
-  if (sdbSetRawStatus(pCommitRaw, SDB_STATUS_DROPPED) != 0) return -1;
-  return 0;
-}
-
-static int32_t mndSetDropOffsetRedoLogs(SMnode *pMnode, STrans *pTrans, SMqOffsetObj *pOffset) {
-  SSdbRaw *pRedoRaw = mndOffsetActionEncode(pOffset);
-  if (pRedoRaw == NULL) return -1;
-  if (mndTransAppendRedolog(pTrans, pRedoRaw) != 0) return -1;
-  if (sdbSetRawStatus(pRedoRaw, SDB_STATUS_DROPPED) != 0) return -1;
-  return 0;
-}
-
-int32_t mndDropOffsetByDB(SMnode *pMnode, STrans *pTrans, SDbObj *pDb) {
-  int32_t code = 0;
-  SSdb   *pSdb = pMnode->pSdb;
-
-  void         *pIter = NULL;
-  SMqOffsetObj *pOffset = NULL;
-  while (1) {
-    pIter = sdbFetch(pSdb, SDB_OFFSET, pIter, (void **)&pOffset);
-    if (pIter == NULL) break;
-
-    if (pOffset->dbUid != pDb->uid) {
-      sdbRelease(pSdb, pOffset);
-      continue;
-    }
-
-    if (mndSetDropOffsetCommitLogs(pMnode, pTrans, pOffset) < 0) {
-      sdbRelease(pSdb, pOffset);
-      sdbCancelFetch(pSdb, pIter);
-      code = -1;
-      break;
-    }
-
-    sdbRelease(pSdb, pOffset);
-  }
-
-  return code;
-}
-
-int32_t mndDropOffsetByTopic(SMnode *pMnode, STrans *pTrans, const char *topic) {
-  int32_t code = -1;
-  SSdb   *pSdb = pMnode->pSdb;
-
-  void         *pIter = NULL;
-  SMqOffsetObj *pOffset = NULL;
-  while (1) {
-    pIter = sdbFetch(pSdb, SDB_OFFSET, pIter, (void **)&pOffset);
-    if (pIter == NULL) break;
-
-    if (!mndOffsetFromTopic(pOffset, topic)) {
-      sdbRelease(pSdb, pOffset);
-      continue;
-    }
-
-    if (mndSetDropOffsetCommitLogs(pMnode, pTrans, pOffset) < 0) {
-      sdbRelease(pSdb, pOffset);
-      goto END;
-    }
-
-    sdbRelease(pSdb, pOffset);
-  }
-
-  code = 0;
-END:
-  return code;
-}
-
-int32_t mndDropOffsetBySubKey(SMnode *pMnode, STrans *pTrans, const char *subKey) {
-  int32_t code = -1;
-  SSdb   *pSdb = pMnode->pSdb;
-
-  void         *pIter = NULL;
-  SMqOffsetObj *pOffset = NULL;
-  while (1) {
-    pIter = sdbFetch(pSdb, SDB_OFFSET, pIter, (void **)&pOffset);
-    if (pIter == NULL) break;
-
-    if (!mndOffsetFromSubKey(pOffset, subKey)) {
-      sdbRelease(pSdb, pOffset);
-      continue;
-    }
-
-    if (mndSetDropOffsetCommitLogs(pMnode, pTrans, pOffset) < 0) {
-      sdbRelease(pSdb, pOffset);
-      goto END;
-    }
-
-    sdbRelease(pSdb, pOffset);
-  }
-
-  code = 0;
-END:
-  return code;
-}

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

@@ -18,7 +18,6 @@
 #include "mndDb.h"
 #include "mndDnode.h"
 #include "mndMnode.h"
-#include "mndOffset.h"
 #include "mndShow.h"
 #include "mndSnode.h"
 #include "mndStb.h"
@@ -33,7 +32,7 @@
 #include "tname.h"
 #include "tuuid.h"
 
-extern bool tsSchedStreamToSnode;
+extern bool tsDeployOnSnode;
 
 static int32_t mndAddTaskToTaskSet(SArray* pArray, SStreamTask* pTask) {
   int32_t childId = taosArrayGetSize(pArray);
@@ -191,7 +190,7 @@ int32_t mndAssignTaskToSnode(SMnode* pMnode, SStreamTask* pTask, SSubplan* plan,
   pTask->nodeId = SNODE_HANDLE;
   pTask->epSet = mndAcquireEpFromSnode(pMnode, pSnode);
 
-  plan->execNode.nodeId = 0;
+  plan->execNode.nodeId = SNODE_HANDLE;
   plan->execNode.epSet = pTask->epSet;
 
   if (qSubPlanToString(plan, &pTask->exec.qmsg, &msgLen) < 0) {
@@ -374,7 +373,7 @@ int32_t mndScheduleStream(SMnode* pMnode, SStreamObj* pStream) {
         return -1;
       }
 
-      if (tsSchedStreamToSnode) {
+      if (tsDeployOnSnode) {
         SSnodeObj* pSnode = mndSchedFetchOneSnode(pMnode);
         if (pSnode == NULL) {
           SVgObj* pVgroup = mndSchedFetchOneVg(pMnode, pStream->sourceDbUid);

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

@@ -273,9 +273,6 @@ _OVER:
 }
 
 static int32_t mndProcessCreateSnodeReq(SRpcMsg *pReq) {
-#if 1
-  return TSDB_CODE_OPS_NOT_SUPPORT;
-#else
   SMnode          *pMnode = pReq->info.node;
   int32_t          code = -1;
   SSnodeObj       *pObj = NULL;
@@ -318,7 +315,6 @@ _OVER:
   mndReleaseSnode(pMnode, pObj);
   mndReleaseDnode(pMnode, pDnode);
   return code;
-#endif
 }
 
 static int32_t mndSetDropSnodeRedoLogs(STrans *pTrans, SSnodeObj *pObj) {
@@ -390,9 +386,6 @@ _OVER:
 }
 
 static int32_t mndProcessDropSnodeReq(SRpcMsg *pReq) {
-#if 1
-  return TSDB_CODE_OPS_NOT_SUPPORT;
-#else
   SMnode *pMnode = pReq->info.node;
   int32_t code = -1;
   SSnodeObj *pObj = NULL;
@@ -429,7 +422,6 @@ _OVER:
 
   mndReleaseSnode(pMnode, pObj);
   return code;
-#endif
 }
 
 static int32_t mndRetrieveSnodes(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pBlock, int32_t rows) {

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

@@ -19,7 +19,6 @@
 #include "mndDb.h"
 #include "mndDnode.h"
 #include "mndMnode.h"
-#include "mndOffset.h"
 #include "mndScheduler.h"
 #include "mndShow.h"
 #include "mndStb.h"
@@ -77,7 +76,6 @@ int32_t mndInitSubscribe(SMnode *pMnode) {
   mndSetMsgHandle(pMnode, TDMT_VND_MQ_VG_CHANGE_RSP, mndTransProcessRsp);
   mndSetMsgHandle(pMnode, TDMT_VND_MQ_VG_DELETE_RSP, mndTransProcessRsp);
   mndSetMsgHandle(pMnode, TDMT_MND_MQ_DO_REBALANCE, mndProcessRebalanceReq);
-  mndSetMsgHandle(pMnode, TDMT_MND_MQ_DO_REBALANCE, mndProcessRebalanceReq);
   mndSetMsgHandle(pMnode, TDMT_MND_MQ_DROP_CGROUP, mndProcessDropCgroupReq);
   mndSetMsgHandle(pMnode, TDMT_MND_MQ_DROP_CGROUP_RSP, mndTransProcessRsp);
 
@@ -684,12 +682,6 @@ static int32_t mndProcessDropCgroupReq(SRpcMsg *pReq) {
 
   mInfo("trans:%d, used to drop cgroup:%s on topic %s", pTrans->id, dropReq.cgroup, dropReq.topic);
 
-  if (mndDropOffsetBySubKey(pMnode, pTrans, pSub->key) < 0) {
-    mndReleaseSubscribe(pMnode, pSub);
-    mndTransDrop(pTrans);
-    return -1;
-  }
-
   if (mndSetDropSubCommitLogs(pMnode, pTrans, pSub) < 0) {
     mError("cgroup %s on topic:%s, failed to drop since %s", dropReq.cgroup, dropReq.topic, terrstr());
     mndReleaseSubscribe(pMnode, pSub);

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

@@ -127,8 +127,10 @@ static int32_t mndProcessTelemTimer(SRpcMsg* pReq) {
   STelemMgmt* pMgmt = &pMnode->telemMgmt;
   if (!tsEnableTelem) return 0;
 
-  taosWLockLatch(&pMgmt->lock);
+  taosThreadMutexLock(&pMgmt->lock);
   char* pCont = mndBuildTelemetryReport(pMnode);
+  taosThreadMutexUnlock(&pMgmt->lock);
+
   if (pCont != NULL) {
     if (taosSendHttpReport(tsTelemServer, tsTelemPort, pCont, strlen(pCont), HTTP_FLAT) != 0) {
       mError("failed to send telemetry report");
@@ -137,18 +139,20 @@ static int32_t mndProcessTelemTimer(SRpcMsg* pReq) {
     }
     taosMemoryFree(pCont);
   }
-  taosWUnLockLatch(&pMgmt->lock);
   return 0;
 }
 
 int32_t mndInitTelem(SMnode* pMnode) {
   STelemMgmt* pMgmt = &pMnode->telemMgmt;
 
-  taosInitRWLatch(&pMgmt->lock);
+  taosThreadMutexInit(&pMgmt->lock, NULL);
   taosGetEmail(pMgmt->email, sizeof(pMgmt->email));
   mndSetMsgHandle(pMnode, TDMT_MND_TELEM_TIMER, mndProcessTelemTimer);
 
   return 0;
 }
 
-void mndCleanupTelem(SMnode* pMnode) {}
+void mndCleanupTelem(SMnode* pMnode) {
+  STelemMgmt* pMgmt = &pMnode->telemMgmt;
+  taosThreadMutexDestroy(&pMgmt->lock);
+}

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

@@ -18,7 +18,6 @@
 #include "mndDb.h"
 #include "mndDnode.h"
 #include "mndMnode.h"
-#include "mndOffset.h"
 #include "mndPrivilege.h"
 #include "mndShow.h"
 #include "mndStb.h"

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

@@ -918,13 +918,19 @@ static void mndTransSendRpcRsp(SMnode *pMnode, STrans *pTrans) {
       sendRsp = true;
     }
   } else {
-    if (pTrans->stage == TRN_STAGE_REDO_ACTION && pTrans->failedTimes > 6) {
+    if (pTrans->stage == TRN_STAGE_REDO_ACTION && ((code == TSDB_CODE_APP_NOT_READY && pTrans->failedTimes > 60) ||
+                                                   (code != TSDB_CODE_APP_NOT_READY && pTrans->failedTimes > 6))) {
       if (code == 0) code = TSDB_CODE_MND_TRANS_UNKNOW_ERROR;
       sendRsp = true;
     }
   }
 
-  if (!sendRsp) return;
+  if (!sendRsp) {
+    return;
+  } else {
+    mInfo("trans:%d, send rsp, stage:%s failedTimes:%d code:0x%x", pTrans->id, mndTransStr(pTrans->stage),
+          pTrans->failedTimes, code);
+  }
 
   int32_t size = taosArrayGetSize(pTrans->pRpcArray);
   if (size <= 0) return;

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

@@ -60,7 +60,7 @@ int32_t mndInitVgroup(SMnode *pMnode) {
   mndSetMsgHandle(pMnode, TDMT_VND_COMPACT_RSP, mndTransProcessRsp);
 
   mndSetMsgHandle(pMnode, TDMT_MND_REDISTRIBUTE_VGROUP, mndProcessRedistributeVgroupMsg);
-  mndSetMsgHandle(pMnode, TDMT_MND_MERGE_VGROUP, mndProcessSplitVgroupMsg);
+  mndSetMsgHandle(pMnode, TDMT_MND_SPLIT_VGROUP, mndProcessSplitVgroupMsg);
   mndSetMsgHandle(pMnode, TDMT_MND_BALANCE_VGROUP, mndProcessBalanceVgroupMsg);
 
   mndAddShowRetrieveHandle(pMnode, TSDB_MGMT_TABLE_VGROUP, mndRetrieveVgroups);
@@ -260,6 +260,12 @@ void *mndBuildCreateVnodeReq(SMnode *pMnode, SDnodeObj *pDnode, SDbObj *pDb, SVg
     return NULL;
   }
 
+  mInfo("vgId:%d, build create vnode req, replica:%d selfIndex:%d strict:%d", createReq.vgId, createReq.replica,
+        createReq.selfIndex, createReq.strict);
+  for (int32_t i = 0; i < createReq.replica; ++i) {
+    mInfo("vgId:%d, replica:%d ep:%s:%u", createReq.vgId, i, createReq.replicas[i].fqdn, createReq.replicas[i].port);
+  }
+
   int32_t contLen = tSerializeSCreateVnodeReq(NULL, 0, &createReq);
   if (contLen < 0) {
     terrno = TSDB_CODE_OUT_OF_MEMORY;
@@ -293,6 +299,7 @@ static void *mndBuildAlterVnodeConfigReq(SMnode *pMnode, SDbObj *pDb, SVgObj *pV
   alterReq.strict = pDb->cfg.strict;
   alterReq.cacheLast = pDb->cfg.cacheLast;
 
+  mInfo("vgId:%d, build alter vnode config req", pVgroup->vgId);
   int32_t contLen = tSerializeSAlterVnodeConfigReq(NULL, 0, &alterReq);
   if (contLen < 0) {
     terrno = TSDB_CODE_OUT_OF_MEMORY;
@@ -342,7 +349,7 @@ static void *mndBuildAlterVnodeReplicaReq(SMnode *pMnode, SDbObj *pDb, SVgObj *p
     }
   }
   alterReq.replica = pVgroup->replica;
-  mInfo("vgId:%d, start to alter vnode, replica:%d selfIndex:%d strict:%d", alterReq.vgId, alterReq.replica,
+  mInfo("vgId:%d, build alter vnode req, replica:%d selfIndex:%d strict:%d", alterReq.vgId, alterReq.replica,
         alterReq.selfIndex, alterReq.strict);
   for (int32_t i = 0; i < alterReq.replica; ++i) {
     mInfo("vgId:%d, replica:%d ep:%s:%u", alterReq.vgId, i, alterReq.replicas[i].fqdn, alterReq.replicas[i].port);
@@ -377,6 +384,7 @@ void *mndBuildDropVnodeReq(SMnode *pMnode, SDnodeObj *pDnode, SDbObj *pDb, SVgOb
   memcpy(dropReq.db, pDb->name, TSDB_DB_FNAME_LEN);
   dropReq.dbUid = pDb->uid;
 
+  mInfo("vgId:%d, build drop vnode req", dropReq.vgId);
   int32_t contLen = tSerializeSDropVnodeReq(NULL, 0, &dropReq);
   if (contLen < 0) {
     terrno = TSDB_CODE_OUT_OF_MEMORY;
@@ -609,35 +617,6 @@ SEpSet mndGetVgroupEpset(SMnode *pMnode, const SVgObj *pVgroup) {
   return epset;
 }
 
-static bool mndGetVgroupMaxReplicaFp(SMnode *pMnode, void *pObj, void *p1, void *p2, void *p3) {
-  SVgObj  *pVgroup = pObj;
-  int64_t  uid = *(int64_t *)p1;
-  int8_t  *pReplica = p2;
-  int32_t *pNumOfVgroups = p3;
-
-  if (pVgroup->dbUid == uid) {
-    *pReplica = TMAX(*pReplica, pVgroup->replica);
-    (*pNumOfVgroups)++;
-  }
-
-  return true;
-}
-
-static int32_t mndGetVgroupMaxReplica(SMnode *pMnode, char *dbName, int8_t *pReplica, int32_t *pNumOfVgroups) {
-  SSdb   *pSdb = pMnode->pSdb;
-  SDbObj *pDb = mndAcquireDb(pMnode, dbName);
-  if (pDb == NULL) {
-    terrno = TSDB_CODE_MND_DB_NOT_SELECTED;
-    return -1;
-  }
-
-  *pReplica = 1;
-  *pNumOfVgroups = 0;
-  sdbTraverse(pSdb, SDB_VGROUP, mndGetVgroupMaxReplicaFp, &pDb->uid, pReplica, pNumOfVgroups);
-  mndReleaseDb(pMnode, pDb);
-  return 0;
-}
-
 static int32_t mndRetrieveVgroups(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pBlock, int32_t rows) {
   SMnode *pMnode = pReq->info.node;
   SSdb   *pSdb = pMnode->pSdb;
@@ -1104,7 +1083,9 @@ int32_t mndSetMoveVgroupInfoToTrans(SMnode *pMnode, STrans *pTrans, SDbObj *pDb,
   mInfo("vgId:%d, will add 1 vnodes", pVgroup->vgId);
   if (mndAddVnodeToVgroup(pMnode, &newVg, pArray) != 0) return -1;
   if (mndAddCreateVnodeAction(pMnode, pTrans, pDb, &newVg, &newVg.vnodeGid[newVg.replica - 1]) != 0) return -1;
-  if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, &newVg, -1) != 0) return -1;
+  for (int32_t i = 0; i < newVg.replica - 1; ++i) {
+    if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, &newVg, newVg.vnodeGid[i].dnodeId) != 0) return -1;
+  }
   if (mndAddAlterVnodeConfirmAction(pMnode, pTrans, pDb, &newVg) != 0) return -1;
 
   mInfo("vgId:%d, will remove 1 vnodes", pVgroup->vgId);
@@ -1112,8 +1093,10 @@ int32_t mndSetMoveVgroupInfoToTrans(SMnode *pMnode, STrans *pTrans, SDbObj *pDb,
   SVnodeGid del = newVg.vnodeGid[vnIndex];
   newVg.vnodeGid[vnIndex] = newVg.vnodeGid[newVg.replica];
   memset(&newVg.vnodeGid[newVg.replica], 0, sizeof(SVnodeGid));
-  if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, &newVg, -1) != 0) return -1;
   if (mndAddDropVnodeAction(pMnode, pTrans, pDb, &newVg, &del, true) != 0) return -1;
+  for (int32_t i = 0; i < newVg.replica; ++i) {
+    if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, &newVg, newVg.vnodeGid[i].dnodeId) != 0) return -1;
+  }
   if (mndAddAlterVnodeConfirmAction(pMnode, pTrans, pDb, &newVg) != 0) return -1;
 
   {
@@ -1186,7 +1169,9 @@ static int32_t mndAddIncVgroupReplicaToTrans(SMnode *pMnode, STrans *pTrans, SDb
   pGid->syncState = TAOS_SYNC_STATE_ERROR;
 
   if (mndAddCreateVnodeAction(pMnode, pTrans, pDb, pVgroup, pGid) != 0) return -1;
-  if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, pVgroup, -1) != 0) return -1;
+  for (int32_t i = 0; i < pVgroup->replica - 1; ++i) {
+    if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, pVgroup, pVgroup->vnodeGid[i].dnodeId) != 0) return -1;
+  }
   if (mndAddAlterVnodeConfirmAction(pMnode, pTrans, pDb, pVgroup) != 0) return -1;
 
   return 0;
@@ -1212,8 +1197,10 @@ static int32_t mndAddDecVgroupReplicaFromTrans(SMnode *pMnode, STrans *pTrans, S
   memcpy(pGid, &pVgroup->vnodeGid[pVgroup->replica], sizeof(SVnodeGid));
   memset(&pVgroup->vnodeGid[pVgroup->replica], 0, sizeof(SVnodeGid));
 
-  if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, pVgroup, -1) != 0) return -1;
   if (mndAddDropVnodeAction(pMnode, pTrans, pDb, pVgroup, &delGid, true) != 0) return -1;
+  for (int32_t i = 0; i < pVgroup->replica; ++i) {
+    if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, pVgroup, pVgroup->vnodeGid[i].dnodeId) != 0) return -1;
+  }
   if (mndAddAlterVnodeConfirmAction(pMnode, pTrans, pDb, pVgroup) != 0) return -1;
 
   return 0;
@@ -1334,9 +1321,6 @@ _OVER:
 }
 
 static int32_t mndProcessRedistributeVgroupMsg(SRpcMsg *pReq) {
-#if 1
-  return TSDB_CODE_OPS_NOT_SUPPORT;
-#else
   SMnode    *pMnode = pReq->info.node;
   SDnodeObj *pNew1 = NULL;
   SDnodeObj *pNew2 = NULL;
@@ -1530,7 +1514,6 @@ _OVER:
   mndReleaseDb(pMnode, pDb);
 
   return code;
-#endif
 }
 
 static int32_t mndCheckDnodeMemory(SMnode *pMnode, SDbObj *pOldDb, SDbObj *pNewDb, SVgObj *pOldVgroup,
@@ -1658,35 +1641,81 @@ static int32_t mndSplitVgroup(SMnode *pMnode, SRpcMsg *pReq, SDbObj *pDb, SVgObj
   if (newVg1.replica == 1) {
     if (mndAddVnodeToVgroup(pMnode, &newVg1, pArray) != 0) goto _OVER;
     if (mndAddCreateVnodeAction(pMnode, pTrans, pDb, &newVg1, &newVg1.vnodeGid[1]) != 0) goto _OVER;
-    if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, &newVg1, -1) != 0) goto _OVER;
+    if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, &newVg1, newVg1.vnodeGid[0].dnodeId) != 0) goto _OVER;
     if (mndAddAlterVnodeConfirmAction(pMnode, pTrans, pDb, &newVg1) != 0) goto _OVER;
   } else if (newVg1.replica == 3) {
     SVnodeGid del1 = {0};
     if (mndRemoveVnodeFromVgroup(pMnode, &newVg1, pArray, &del1) != 0) goto _OVER;
-    if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, &newVg1, -1) != 0) goto _OVER;
+    if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, &newVg1, newVg1.vnodeGid[0].dnodeId) != 0) goto _OVER;
+    if (mndAddAlterVnodeReplicaAction(pMnode, pTrans, pDb, &newVg1, newVg1.vnodeGid[1].dnodeId) != 0) goto _OVER;
     if (mndAddDropVnodeAction(pMnode, pTrans, pDb, &newVg1, &del1, true) != 0) goto _OVER;
     if (mndAddAlterVnodeConfirmAction(pMnode, pTrans, pDb, &newVg1) != 0) goto _OVER;
   } else {
     goto _OVER;
   }
 
+  mInfo("vgId:%d, vgroup info after adjust replica, replica:%d hashBegin:%u hashEnd:%u vnode:0 dnode:%d", newVg1.vgId,
+        newVg1.replica, newVg1.hashBegin, newVg1.hashEnd, newVg1.vnodeGid[0].dnodeId);
+  for (int32_t i = 0; i < newVg1.replica; ++i) {
+    mInfo("vgId:%d, vnode:%d dnode:%d", newVg1.vgId, i, newVg1.vnodeGid[i].dnodeId);
+  }
+
   SVgObj newVg2 = {0};
-  memcpy(&newVg1, &newVg2, sizeof(SVgObj));
+  memcpy(&newVg2, &newVg1, sizeof(SVgObj));
   newVg1.replica = 1;
-  newVg1.hashEnd = (newVg1.hashBegin + newVg1.hashEnd) / 2;
+  newVg1.hashEnd = newVg1.hashBegin / 2 + newVg1.hashEnd / 2;
   memset(&newVg1.vnodeGid[1], 0, sizeof(SVnodeGid));
 
   newVg2.replica = 1;
   newVg2.hashBegin = newVg1.hashEnd + 1;
   memcpy(&newVg2.vnodeGid[0], &newVg2.vnodeGid[1], sizeof(SVnodeGid));
-  memset(&newVg1.vnodeGid[1], 0, sizeof(SVnodeGid));
+  memset(&newVg2.vnodeGid[1], 0, sizeof(SVnodeGid));
+
+  mInfo("vgId:%d, vgroup info after adjust hash, replica:%d hashBegin:%u hashEnd:%u vnode:0 dnode:%d", newVg1.vgId,
+        newVg1.replica, newVg1.hashBegin, newVg1.hashEnd, newVg1.vnodeGid[0].dnodeId);
+  mInfo("vgId:%d, vgroup info after adjust hash, replica:%d hashBegin:%u hashEnd:%u vnode:0 dnode:%d", newVg2.vgId,
+        newVg2.replica, newVg2.hashBegin, newVg2.hashEnd, newVg2.vnodeGid[0].dnodeId);
 
   if (mndAddAlterVnodeHashRangeAction(pMnode, pTrans, pDb, &newVg1) != 0) goto _OVER;
   if (mndAddAlterVnodeHashRangeAction(pMnode, pTrans, pDb, &newVg2) != 0) goto _OVER;
 
-  // adjust vgroup
+#if 0
-  if (mndBuildAlterVgroupAction(pMnode, pTrans, pDb, pDb, &newVg1, pArray) != 0) goto _OVER;
+  // adjust vgroup replica
-  if (mndBuildAlterVgroupAction(pMnode, pTrans, pDb, pDb, &newVg2, pArray) != 0) goto _OVER;
+  if (pDb->cfg.replications != newVg1.replica) {
+    if (mndBuildAlterVgroupAction(pMnode, pTrans, pDb, pDb, &newVg1, pArray) != 0) goto _OVER;
+  }
+  if (pDb->cfg.replications != newVg2.replica) {
+    if (mndBuildAlterVgroupAction(pMnode, pTrans, pDb, pDb, &newVg2, pArray) != 0) goto _OVER;
+  }
+#endif
+
+  {
+    pRaw = mndVgroupActionEncode(&newVg1);
+    if (pRaw == NULL || mndTransAppendCommitlog(pTrans, pRaw) != 0) goto _OVER;
+    (void)sdbSetRawStatus(pRaw, SDB_STATUS_READY);
+    pRaw = NULL;
+  }
+
+  {
+    pRaw = mndVgroupActionEncode(&newVg2);
+    if (pRaw == NULL || mndTransAppendCommitlog(pTrans, pRaw) != 0) goto _OVER;
+    (void)sdbSetRawStatus(pRaw, SDB_STATUS_READY);
+    pRaw = NULL;
+  }
+
+  mInfo("vgId:%d, vgroup info after adjust hash, replica:%d hashBegin:%u hashEnd:%u vnode:0 dnode:%d", newVg1.vgId,
+        newVg1.replica, newVg1.hashBegin, newVg1.hashEnd, newVg1.vnodeGid[0].dnodeId);
+  for (int32_t i = 0; i < newVg1.replica; ++i) {
+    mInfo("vgId:%d, vnode:%d dnode:%d", newVg1.vgId, i, newVg1.vnodeGid[i].dnodeId);
+  }
+  mInfo("vgId:%d, vgroup info after adjust hash, replica:%d hashBegin:%u hashEnd:%u vnode:0 dnode:%d", newVg2.vgId,
+        newVg2.replica, newVg2.hashBegin, newVg2.hashEnd, newVg2.vnodeGid[0].dnodeId);
+  for (int32_t i = 0; i < newVg1.replica; ++i) {
+    mInfo("vgId:%d, vnode:%d dnode:%d", newVg2.vgId, i, newVg2.vnodeGid[i].dnodeId);
+  }
+
+  if (mndTransPrepare(pMnode, pTrans) != 0) goto _OVER;
+  code = 0;
 
 _OVER:
   mndTransDrop(pTrans);
@@ -1862,13 +1891,10 @@ _OVER:
 }
 
 static int32_t mndProcessBalanceVgroupMsg(SRpcMsg *pReq) {
-#if 1
-  return TSDB_CODE_OPS_NOT_SUPPORT;
-#else
   SMnode *pMnode = pReq->info.node;
   int32_t code = -1;
   SArray *pArray = NULL;
-  void *pIter = NULL;
+  void   *pIter = NULL;
   int64_t curMs = taosGetTimestampMs();
 
   SBalanceVgroupReq req = {0};
@@ -1913,7 +1939,6 @@ _OVER:
 
   taosArrayDestroy(pArray);
   return code;
-#endif
 }
 
 bool mndVgroupInDb(SVgObj *pVgroup, int64_t dbUid) { return !pVgroup->isTsma && pVgroup->dbUid == dbUid; }

source/dnode/mnode/impl/test/sdb/sdbTest.cpp

@@ -367,6 +367,80 @@ bool sdbTraverseFail(SMnode *pMnode, SStrObj *pObj, int32_t *p1, int32_t *p2, in
   return false;
 }
 
+#ifndef WINDOWS
+
+TEST_F(MndTestSdb, 00_API) {
+  SMnode  mnode = {0};
+  SSdbOpt opt = {0};
+  opt.pMnode = &mnode;
+  opt.path = TD_TMP_DIR_PATH "mnode_test_sdb";
+  taosRemoveDir(opt.path);
+  SSdb *pSdb = sdbInit(&opt);
+
+  SSdbTable table = {.sdbType = SDB_USER, .keyType = SDB_KEY_BINARY};
+  sdbSetTable(pSdb, table);
+
+  // sdbRow.c
+  SSdbRow *pRow1 = sdbAllocRow(-128);
+  ASSERT_EQ(pRow1 == NULL, 1);
+
+  void *pRow2 = sdbGetRowObj(NULL);
+  ASSERT_EQ(pRow2 == NULL, 1);
+
+  //sdbRaw.c
+  SStrObj  strObj;
+  SSdbRaw *pRaw1 = NULL;
+  strSetDefault(&strObj, 1);
+  
+  pRaw1 = strEncode(&strObj);
+  int32_t id = sdbGetIdFromRaw(pSdb, pRaw1);
+  ASSERT_EQ(id, -2);
+
+  SSdbRaw *pRaw2 = sdbAllocRaw(SDB_USER, 1, -128);
+  ASSERT_EQ(pRaw2 == NULL, 1);
+
+  ASSERT_EQ(sdbSetRawInt8(NULL, 0, 0), -1);
+  ASSERT_EQ(sdbSetRawInt8(pRaw1, -128, 0), -1);
+  ASSERT_EQ(sdbSetRawInt32(NULL, 0, 0), -1);
+  ASSERT_EQ(sdbSetRawInt32(pRaw1, -128, 0), -1);
+  ASSERT_EQ(sdbSetRawInt16(NULL, 0, 0), -1);
+  ASSERT_EQ(sdbSetRawInt16(pRaw1, -128, 0), -1);
+  ASSERT_EQ(sdbSetRawInt64(NULL, 0, 0), -1);
+  ASSERT_EQ(sdbSetRawInt64(pRaw1, -128, 0), -1);
+  ASSERT_EQ(sdbSetRawBinary(NULL, 0, "12", 3), -1);
+  ASSERT_EQ(sdbSetRawBinary(pRaw1, 9028, "12", 3), -1);
+  ASSERT_EQ(sdbSetRawDataLen(NULL, 0), -1);
+  ASSERT_EQ(sdbSetRawDataLen(pRaw1, 9000), -1);
+  ASSERT_EQ(sdbSetRawStatus(NULL, SDB_STATUS_READY), -1);
+  ASSERT_EQ(sdbSetRawStatus(pRaw1, SDB_STATUS_INIT), -1);
+
+  ASSERT_EQ(sdbGetRawInt8(NULL, 0, 0), -1);
+  ASSERT_EQ(sdbGetRawInt8(pRaw1, 9000, 0), -1);
+  ASSERT_EQ(sdbGetRawInt32(NULL, 0, 0), -1);
+  ASSERT_EQ(sdbGetRawInt32(pRaw1, 9000, 0), -1);
+  ASSERT_EQ(sdbGetRawInt16(NULL, 0, 0), -1);
+  ASSERT_EQ(sdbGetRawInt16(pRaw1, 9000, 0), -1);
+  ASSERT_EQ(sdbGetRawInt64(NULL, 0, 0), -1);
+  ASSERT_EQ(sdbGetRawInt64(pRaw1, 9000, 0), -1);
+  ASSERT_EQ(sdbGetRawBinary(NULL, 0, 0, 4096), -1);
+  ASSERT_EQ(sdbGetRawBinary(pRaw1, 9000, 0, 112), -1);
+  ASSERT_EQ(sdbGetRawSoftVer(NULL, 0), -1);
+  ASSERT_EQ(sdbGetRawTotalSize(NULL), -1);
+
+  // sdbHash.c
+  EXPECT_STREQ(sdbTableName((ESdbType)100), "undefine");
+  EXPECT_STREQ(sdbStatusName((ESdbStatus)100), "undefine");
+  ASSERT_EQ(sdbGetTableVer(pSdb, (ESdbType)100), -1);
+
+  SSdbRaw *pRaw3 = sdbAllocRaw((ESdbType)-12, 1, 128);
+  ASSERT_NE(sdbWriteWithoutFree(pSdb, pRaw3), 0);
+  pSdb->hashObjs[1] = NULL;
+  SSdbRaw *pRaw4 = sdbAllocRaw((ESdbType)1, 1, 128);
+  ASSERT_NE(sdbWriteWithoutFree(pSdb, pRaw4), 0);
+}
+
+#endif
+
 TEST_F(MndTestSdb, 01_Write_Str) {
   void    *pIter = NULL;
   int32_t  num = 0;
@@ -885,8 +959,8 @@ TEST_F(MndTestSdb, 01_Read_Str) {
     ASSERT_EQ(sdbWrite(pSdb, pRaw), 0);
     pI32Obj = (SI32Obj *)sdbAcquire(pSdb, SDB_VGROUP, &key);
     ASSERT_EQ(pI32Obj, nullptr);
-    int32_t code = terrno;
+    // int32_t code = terrno;
-    ASSERT_EQ(code, TSDB_CODE_SDB_OBJ_DROPPING);
+    // ASSERT_EQ(code, TSDB_CODE_SDB_OBJ_DROPPING);
   }
 
   {
@@ -899,8 +973,8 @@ TEST_F(MndTestSdb, 01_Read_Str) {
     ASSERT_EQ(sdbWrite(pSdb, pRaw), 0);
     pI32Obj = (SI32Obj *)sdbAcquire(pSdb, SDB_VGROUP, &key);
     ASSERT_EQ(pI32Obj, nullptr);
-    int32_t code = terrno;
+    // int32_t code = terrno;
-    ASSERT_EQ(code, TSDB_CODE_SDB_OBJ_CREATING);
+    // ASSERT_EQ(code, TSDB_CODE_SDB_OBJ_CREATING);
   }
 
   {

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

@@ -654,15 +654,12 @@ int32_t sdbStopWrite(SSdb *pSdb, SSdbIter *pIter, bool isApply, int64_t index, i
   }
 
   if (config > 0) {
-    ASSERT(pSdb->commitConfig == config);
     pSdb->commitConfig = config;
   }
   if (term > 0) {
-    ASSERT(pSdb->commitTerm == term);
     pSdb->commitTerm = term;
   }
   if (index > 0) {
-    ASSERT(pSdb->commitIndex == index);
     pSdb->commitIndex = index;
   }
 

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

@@ -241,9 +241,7 @@ int32_t sdbWriteWithoutFree(SSdb *pSdb, SSdbRaw *pRaw) {
 
   SdbDecodeFp decodeFp = pSdb->decodeFps[pRaw->type];
   SSdbRow    *pRow = (*decodeFp)(pRaw);
-  if (pRow == NULL) {
+  if (pRow == NULL) return terrno;
-    return terrno;
-  }
 
   pRow->type = pRaw->type;
 

source/dnode/qnode/src/qnode.c

@@ -90,7 +90,7 @@ int32_t qndProcessQueryMsg(SQnode *pQnode, int64_t ts, SRpcMsg *pMsg) {
       code = qWorkerProcessFetchMsg(pQnode, pQnode->pQuery, pMsg, ts);
       break;
     case TDMT_SCH_CANCEL_TASK:
-      code = qWorkerProcessCancelMsg(pQnode, pQnode->pQuery, pMsg, ts);
+      //code = qWorkerProcessCancelMsg(pQnode, pQnode->pQuery, pMsg, ts);
       break;
     case TDMT_SCH_DROP_TASK:
       code = qWorkerProcessDropMsg(pQnode, pQnode->pQuery, pMsg, ts);

source/dnode/snode/inc/sndInt.h

@@ -31,7 +31,9 @@ extern "C" {
 #endif
 
 typedef struct SSnode {
-  SMsgCb msgCb;
+  char*        path;
+  SStreamMeta* pMeta;
+  SMsgCb       msgCb;
 } SSnode;
 
 #if 0

source/dnode/snode/src/snode.c

@@ -15,197 +15,197 @@
 
 #include "executor.h"
 #include "sndInt.h"
+#include "tstream.h"
 #include "tuuid.h"
-/*SSnode *sndOpen(const char *path, const SSnodeOpt *pOption) { return NULL; }*/
+
-/*void    sndClose(SSnode *pSnode) {}*/
+void sndEnqueueStreamDispatch(SSnode *pSnode, SRpcMsg *pMsg) {
-int32_t sndProcessUMsg(SSnode *pSnode, SRpcMsg *pMsg) { return 0; }
+  char   *msgStr = pMsg->pCont;
-int32_t sndProcessSMsg(SSnode *pSnode, SRpcMsg *pMsg) { return 0; }
+  char   *msgBody = POINTER_SHIFT(msgStr, sizeof(SMsgHead));
+  int32_t msgLen = pMsg->contLen - sizeof(SMsgHead);
+  int32_t code = 0;
+
+  SStreamDispatchReq req;
+  SDecoder           decoder;
+  tDecoderInit(&decoder, msgBody, msgLen);
+  if (tDecodeStreamDispatchReq(&decoder, &req) < 0) {
+    code = TSDB_CODE_MSG_DECODE_ERROR;
+    tDecoderClear(&decoder);
+    goto FAIL;
+  }
+  tDecoderClear(&decoder);
+
+  int32_t taskId = req.taskId;
+
+  SStreamTask *pTask = streamMetaGetTask(pSnode->pMeta, taskId);
+  if (pTask) {
+    SRpcMsg rsp = {
+        .info = pMsg->info,
+        .code = 0,
+    };
+    streamProcessDispatchReq(pTask, &req, &rsp, false);
+    rpcFreeCont(pMsg->pCont);
+    taosFreeQitem(pMsg);
+    return;
+  }
+
+FAIL:
+  if (pMsg->info.handle == NULL) return;
+  SRpcMsg rsp = {
+      .code = code,
+      .info = pMsg->info,
+  };
+  tmsgSendRsp(&rsp);
+  rpcFreeCont(pMsg->pCont);
+  taosFreeQitem(pMsg);
+}
+
+int32_t sndExpandTask(SSnode *pSnode, SStreamTask *pTask, int64_t ver) {
+  ASSERT(pTask->taskLevel == TASK_LEVEL__AGG);
+  ASSERT(taosArrayGetSize(pTask->childEpInfo) != 0);
+
+  pTask->schedStatus = TASK_SCHED_STATUS__INACTIVE;
+  pTask->inputQueue = streamQueueOpen();
+  pTask->outputQueue = streamQueueOpen();
+
+  if (pTask->inputQueue == NULL || pTask->outputQueue == NULL) {
+    return -1;
+  }
+
+  pTask->inputStatus = TASK_INPUT_STATUS__NORMAL;
+  pTask->outputStatus = TASK_OUTPUT_STATUS__NORMAL;
+
+  pTask->pMsgCb = &pSnode->msgCb;
+
+  pTask->startVer = ver;
+
+  pTask->pState = streamStateOpen(pSnode->path, pTask, false, -1, -1);
+  if (pTask->pState == NULL) {
+    return -1;
+  }
+
+  SReadHandle mgHandle = {
+      .vnode = NULL,
+      .numOfVgroups = (int32_t)taosArrayGetSize(pTask->childEpInfo),
+      .pStateBackend = pTask->pState,
+  };
+  pTask->exec.executor = qCreateStreamExecTaskInfo(pTask->exec.qmsg, &mgHandle);
+  ASSERT(pTask->exec.executor);
+
+  return 0;
+}
 
 SSnode *sndOpen(const char *path, const SSnodeOpt *pOption) {
   SSnode *pSnode = taosMemoryCalloc(1, sizeof(SSnode));
   if (pSnode == NULL) {
+    terrno = TSDB_CODE_OUT_OF_MEMORY;
     return NULL;
   }
+  pSnode->path = strdup(path);
+  if (pSnode->path == NULL) {
+    terrno = TSDB_CODE_OUT_OF_MEMORY;
+    goto FAIL;
+  }
   pSnode->msgCb = pOption->msgCb;
-#if 0
+
-  pSnode->pMeta = sndMetaNew();
+  pSnode->pMeta = streamMetaOpen(path, pSnode, (FTaskExpand *)sndExpandTask, SNODE_HANDLE);
   if (pSnode->pMeta == NULL) {
-    taosMemoryFree(pSnode);
+    terrno = TSDB_CODE_OUT_OF_MEMORY;
-    return NULL;
+    goto FAIL;
   }
-#endif
+
   return pSnode;
+
+FAIL:
+  taosMemoryFree(pSnode->path);
+  taosMemoryFree(pSnode);
+  return NULL;
 }
 
 void sndClose(SSnode *pSnode) {
-  /*sndMetaDelete(pSnode->pMeta);*/
+  streamMetaClose(pSnode->pMeta);
+  taosMemoryFree(pSnode->path);
   taosMemoryFree(pSnode);
 }
 
 int32_t sndGetLoad(SSnode *pSnode, SSnodeLoad *pLoad) { return 0; }
 
-#if 0
+int32_t sndProcessTaskDeployReq(SSnode *pSnode, char *msg, int32_t msgLen) {
-SStreamMeta *sndMetaNew() {
+  int32_t code;
-  SStreamMeta *pMeta = taosMemoryCalloc(1, sizeof(SStreamMeta));
-  if (pMeta == NULL) {
-    return NULL;
-  }
-  pMeta->pHash = taosHashInit(64, taosGetDefaultHashFunction(TSDB_DATA_TYPE_INT), true, HASH_NO_LOCK);
-  if (pMeta->pHash == NULL) {
-    taosMemoryFree(pMeta);
-    return NULL;
-  }
-  return pMeta;
-}
-
-void sndMetaDelete(SStreamMeta *pMeta) {
-  taosHashCleanup(pMeta->pHash);
-  taosMemoryFree(pMeta);
-}
-
-int32_t sndMetaDeployTask(SStreamMeta *pMeta, SStreamTask *pTask) {
-  pTask->exec.executor = qCreateStreamExecTaskInfo(pTask->exec.qmsg, NULL);
-  return taosHashPut(pMeta->pHash, &pTask->taskId, sizeof(int32_t), pTask, sizeof(void *));
-}
-
-SStreamTask *sndMetaGetTask(SStreamMeta *pMeta, int32_t taskId) {
-  return taosHashGet(pMeta->pHash, &taskId, sizeof(int32_t));
-}
-
-int32_t sndMetaRemoveTask(SStreamMeta *pMeta, int32_t taskId) {
-  SStreamTask *pTask = taosHashGet(pMeta->pHash, &taskId, sizeof(int32_t));
-  if (pTask == NULL) {
-    return -1;
-  }
-  taosMemoryFree(pTask->exec.qmsg);
-  // TODO:free executor
-  taosMemoryFree(pTask);
-  return taosHashRemove(pMeta->pHash, &taskId, sizeof(int32_t));
-}
-
-static int32_t sndProcessTaskDeployReq(SSnode *pNode, SRpcMsg *pMsg) {
-  SStreamMeta *pMeta = pNode->pMeta;
-  char        *msg = POINTER_SHIFT(pMsg->pCont, sizeof(SMsgHead));
-  int32_t      msgLen = pMsg->contLen - sizeof(SMsgHead);
 
+  // 1.deserialize msg and build task
   SStreamTask *pTask = taosMemoryCalloc(1, sizeof(SStreamTask));
   if (pTask == NULL) {
     return -1;
   }
   SDecoder decoder;
   tDecoderInit(&decoder, (uint8_t *)msg, msgLen);
-  if (tDecodeSStreamTask(&decoder, pTask) < 0) {
+  code = tDecodeSStreamTask(&decoder, pTask);
-    ASSERT(0);
+  if (code < 0) {
+    tDecoderClear(&decoder);
+    taosMemoryFree(pTask);
+    return -1;
   }
   tDecoderClear(&decoder);
 
-  pTask->execStatus = TASK_EXEC_STATUS__IDLE;
+  ASSERT(pTask->taskLevel == TASK_LEVEL__AGG);
 
-  pTask->inputQueue = streamQueueOpen();
+  // 2.save task
-  pTask->outputQueue = streamQueueOpen();
+  code = streamMetaAddTask(pSnode->pMeta, -1, pTask);
-  pTask->inputStatus = TASK_INPUT_STATUS__NORMAL;
+  if (code < 0) {
-  pTask->outputStatus = TASK_INPUT_STATUS__NORMAL;
+    return -1;
+  }
 
-  if (pTask->inputQueue == NULL || pTask->outputQueue == NULL) goto FAIL;
+  // 3.go through recover steps to fill history
-
+  if (pTask->fillHistory) {
-  pTask->pMsgCb = &pNode->msgCb;
+    streamSetParamForRecover(pTask);
-
+    streamAggRecoverPrepare(pTask);
-  pTask->exec.executor = qCreateStreamExecTaskInfo(pTask->exec.qmsg, NULL);
+  }
-  ASSERT(pTask->exec.executor);
-
-  streamSetupTrigger(pTask);
-
-  qInfo("deploy stream: stream id %" PRId64 " task id %d child id %d on snode", pTask->streamId, pTask->taskId,
-        pTask->selfChildId);
-
-  taosHashPut(pMeta->pHash, &pTask->taskId, sizeof(int32_t), &pTask, sizeof(void *));
 
   return 0;
-
-FAIL:
-  if (pTask->inputQueue) streamQueueClose(pTask->inputQueue);
-  if (pTask->outputQueue) streamQueueClose(pTask->outputQueue);
-  if (pTask) taosMemoryFree(pTask);
-  return -1;
 }
 
-static int32_t sndProcessTaskRunReq(SSnode *pNode, SRpcMsg *pMsg) {
+int32_t sndProcessTaskDropReq(SSnode *pSnode, char *msg, int32_t msgLen) {
-  SStreamMeta       *pMeta = pNode->pMeta;
+  SVDropStreamTaskReq *pReq = (SVDropStreamTaskReq *)msg;
+  return streamMetaRemoveTask(pSnode->pMeta, pReq->taskId);
+}
+
+int32_t sndProcessTaskRunReq(SSnode *pSnode, SRpcMsg *pMsg) {
   SStreamTaskRunReq *pReq = pMsg->pCont;
   int32_t            taskId = pReq->taskId;
-  SStreamTask       *pTask = *(SStreamTask **)taosHashGet(pMeta->pHash, &taskId, sizeof(int32_t));
+  SStreamTask       *pTask = streamMetaGetTask(pSnode->pMeta, taskId);
-  streamProcessRunReq(pTask);
+  if (pTask) {
-  return 0;
+    streamProcessRunReq(pTask);
+    return 0;
+  } else {
+    return -1;
+  }
 }
 
-static int32_t sndProcessTaskDispatchReq(SSnode *pNode, SRpcMsg *pMsg) {
+int32_t sndProcessTaskDispatchReq(SSnode *pSnode, SRpcMsg *pMsg, bool exec) {
-  SStreamMeta *pMeta = pNode->pMeta;
+  char              *msgStr = pMsg->pCont;
-
+  char              *msgBody = POINTER_SHIFT(msgStr, sizeof(SMsgHead));
-  char   *msgStr = pMsg->pCont;
+  int32_t            msgLen = pMsg->contLen - sizeof(SMsgHead);
-  char   *msgBody = POINTER_SHIFT(msgStr, sizeof(SMsgHead));
-  int32_t msgLen = pMsg->contLen - sizeof(SMsgHead);
-
   SStreamDispatchReq req;
   SDecoder           decoder;
-  tDecoderInit(&decoder, msgBody, msgLen);
+  tDecoderInit(&decoder, (uint8_t *)msgBody, msgLen);
   tDecodeStreamDispatchReq(&decoder, &req);
-  int32_t      taskId = req.taskId;
+  int32_t taskId = req.taskId;
-  SStreamTask *pTask = *(SStreamTask **)taosHashGet(pMeta->pHash, &taskId, sizeof(int32_t));
-  SRpcMsg      rsp = {
-           .info = pMsg->info,
-           .code = 0,
-  };
-  streamProcessDispatchReq(pTask, &req, &rsp, true);
-  return 0;
-}
 
-static int32_t sndProcessTaskRecoverReq(SSnode *pNode, SRpcMsg *pMsg) {
+  SStreamTask *pTask = streamMetaGetTask(pSnode->pMeta, taskId);
-  SStreamMeta *pMeta = pNode->pMeta;
+  if (pTask) {
-
+    SRpcMsg rsp = {
-  SStreamTaskRecoverReq *pReq = pMsg->pCont;
+        .info = pMsg->info,
-  int32_t                taskId = pReq->taskId;
+        .code = 0,
-  SStreamTask           *pTask = *(SStreamTask **)taosHashGet(pMeta->pHash, &taskId, sizeof(int32_t));
+    };
-  streamProcessRecoverReq(pTask, pReq, pMsg);
+    streamProcessDispatchReq(pTask, &req, &rsp, exec);
-  return 0;
+    return 0;
-}
+  } else {
-
+    return -1;
-static int32_t sndProcessTaskDispatchRsp(SSnode *pNode, SRpcMsg *pMsg) {
-  SStreamMeta *pMeta = pNode->pMeta;
-
-  SStreamDispatchRsp *pRsp = POINTER_SHIFT(pMsg->pCont, sizeof(SMsgHead));
-  int32_t             taskId = pRsp->taskId;
-  SStreamTask        *pTask = *(SStreamTask **)taosHashGet(pMeta->pHash, &taskId, sizeof(int32_t));
-  streamProcessDispatchRsp(pTask, pRsp);
-  return 0;
-}
-
-static int32_t sndProcessTaskRecoverRsp(SSnode *pNode, SRpcMsg *pMsg) {
-  SStreamMeta *pMeta = pNode->pMeta;
-
-  SStreamTaskRecoverRsp *pRsp = pMsg->pCont;
-  int32_t                taskId = pRsp->rspTaskId;
-  SStreamTask           *pTask = *(SStreamTask **)taosHashGet(pMeta->pHash, &taskId, sizeof(int32_t));
-  streamProcessRecoverRsp(pTask, pRsp);
-  return 0;
-}
-
-static int32_t sndProcessTaskDropReq(SSnode *pNode, SRpcMsg *pMsg) {
-  SStreamMeta *pMeta = pNode->pMeta;
-
-  char                *msg = pMsg->pCont;
-  int32_t              msgLen = pMsg->contLen;
-  SVDropStreamTaskReq *pReq = (SVDropStreamTaskReq *)msg;
-  int32_t              code = taosHashRemove(pMeta->pHash, &pReq->taskId, sizeof(int32_t));
-  ASSERT(code == 0);
-  if (code == 0) {
-    // sendrsp
   }
-  return code;
+  return 0;
 }
 
-static int32_t sndProcessTaskRetrieveReq(SSnode *pNode, SRpcMsg *pMsg) {
+int32_t sndProcessTaskRetrieveReq(SSnode *pSnode, SRpcMsg *pMsg) {
-  SStreamMeta *pMeta = pNode->pMeta;
-
   char              *msgStr = pMsg->pCont;
   char              *msgBody = POINTER_SHIFT(msgStr, sizeof(SMsgHead));
   int32_t            msgLen = pMsg->contLen - sizeof(SMsgHead);
@@ -213,53 +213,64 @@ static int32_t sndProcessTaskRetrieveReq(SSnode *pNode, SRpcMsg *pMsg) {
   SDecoder           decoder;
   tDecoderInit(&decoder, msgBody, msgLen);
   tDecodeStreamRetrieveReq(&decoder, &req);
+  tDecoderClear(&decoder);
   int32_t      taskId = req.dstTaskId;
-  SStreamTask *pTask = *(SStreamTask **)taosHashGet(pMeta->pHash, &taskId, sizeof(int32_t));
+  SStreamTask *pTask = streamMetaGetTask(pSnode->pMeta, taskId);
-  if (atomic_load_8(&pTask->taskStatus) != TASK_STATUS__NORMAL) {
+  if (pTask) {
+    SRpcMsg rsp = {
+        .info = pMsg->info,
+        .code = 0,
+    };
+    streamProcessRetrieveReq(pTask, &req, &rsp);
+    tDeleteStreamRetrieveReq(&req);
     return 0;
+  } else {
+    return -1;
+  }
+}
+
+int32_t sndProcessTaskDispatchRsp(SSnode *pSnode, SRpcMsg *pMsg) {
+  SStreamDispatchRsp *pRsp = POINTER_SHIFT(pMsg->pCont, sizeof(SMsgHead));
+  int32_t             taskId = pRsp->taskId;
+  SStreamTask        *pTask = streamMetaGetTask(pSnode->pMeta, taskId);
+  if (pTask) {
+    streamProcessDispatchRsp(pTask, pRsp);
+    return 0;
+  } else {
+    return -1;
   }
-  SRpcMsg rsp = {
-      .info = pMsg->info,
-      .code = 0,
-  };
-  streamProcessRetrieveReq(pTask, &req, &rsp);
   return 0;
 }
 
-static int32_t sndProcessTaskRetrieveRsp(SSnode *pNode, SRpcMsg *pMsg) {
+int32_t sndProcessTaskRetrieveRsp(SSnode *pSnode, SRpcMsg *pMsg) {
   //
   return 0;
 }
 
-int32_t sndProcessUMsg(SSnode *pSnode, SRpcMsg *pMsg) {
+int32_t sndProcessWriteMsg(SSnode *pSnode, SRpcMsg *pMsg, SRpcMsg *pRsp) {
-  // stream deploy
+  void   *pReq = POINTER_SHIFT(pMsg->pCont, sizeof(SMsgHead));
-  // stream stop/resume
+  int32_t len = pMsg->contLen - sizeof(SMsgHead);
-  // operator exec
   switch (pMsg->msgType) {
     case TDMT_STREAM_TASK_DEPLOY:
-      return sndProcessTaskDeployReq(pSnode, pMsg);
+      return sndProcessTaskDeployReq(pSnode, pReq, len);
     case TDMT_STREAM_TASK_DROP:
-      return sndProcessTaskDropReq(pSnode, pMsg);
+      return sndProcessTaskDropReq(pSnode, pReq, len);
     default:
       ASSERT(0);
   }
   return 0;
 }
 
-int32_t sndProcessSMsg(SSnode *pSnode, SRpcMsg *pMsg) {
+int32_t sndProcessStreamMsg(SSnode *pSnode, SRpcMsg *pMsg) {
   switch (pMsg->msgType) {
     case TDMT_STREAM_TASK_RUN:
       return sndProcessTaskRunReq(pSnode, pMsg);
     case TDMT_STREAM_TASK_DISPATCH:
-      return sndProcessTaskDispatchReq(pSnode, pMsg);
+      return sndProcessTaskDispatchReq(pSnode, pMsg, true);
-    case TDMT_STREAM_TASK_RECOVER:
-      return sndProcessTaskRecoverReq(pSnode, pMsg);
     case TDMT_STREAM_RETRIEVE:
-      return sndProcessTaskRecoverReq(pSnode, pMsg);
+      return sndProcessTaskRetrieveReq(pSnode, pMsg);
     case TDMT_STREAM_TASK_DISPATCH_RSP:
       return sndProcessTaskDispatchRsp(pSnode, pMsg);
-    case TDMT_STREAM_TASK_RECOVER_RSP:
-      return sndProcessTaskRecoverRsp(pSnode, pMsg);
     case TDMT_STREAM_RETRIEVE_RSP:
       return sndProcessTaskRetrieveRsp(pSnode, pMsg);
     default:
@@ -267,4 +278,3 @@ int32_t sndProcessSMsg(SSnode *pSnode, SRpcMsg *pMsg) {
   }
   return 0;
 }
-#endif

source/dnode/vnode/inc/vnode.h

@@ -152,22 +152,26 @@ typedef struct STsdbReader STsdbReader;
 #define CACHESCAN_RETRIEVE_LAST_ROW    0x4
 #define CACHESCAN_RETRIEVE_LAST        0x8
 
-int32_t  tsdbSetTableId(STsdbReader *pReader, int64_t uid);
+int32_t tsdbSetTableId(STsdbReader *pReader, int64_t uid);
-int32_t  tsdbReaderOpen(SVnode *pVnode, SQueryTableDataCond *pCond, SArray *pTableList, STsdbReader **ppReader,
+int32_t tsdbReaderOpen(SVnode *pVnode, SQueryTableDataCond *pCond, SArray *pTableList, STsdbReader **ppReader,
-                        const char *idstr);
+                       const char *idstr);
+
 void     tsdbReaderClose(STsdbReader *pReader);
 bool     tsdbNextDataBlock(STsdbReader *pReader);
+bool     tsdbTableNextDataBlock(STsdbReader *pReader, int64_t uid);
 void     tsdbRetrieveDataBlockInfo(STsdbReader *pReader, SDataBlockInfo *pDataBlockInfo);
 int32_t  tsdbRetrieveDatablockSMA(STsdbReader *pReader, SColumnDataAgg ***pBlockStatis, bool *allHave);
 SArray  *tsdbRetrieveDataBlock(STsdbReader *pTsdbReadHandle, SArray *pColumnIdList);
 int32_t  tsdbReaderReset(STsdbReader *pReader, SQueryTableDataCond *pCond);
 int32_t  tsdbGetFileBlocksDistInfo(STsdbReader *pReader, STableBlockDistInfo *pTableBlockInfo);
 int64_t  tsdbGetNumOfRowsInMemTable(STsdbReader *pHandle);
+bool     tsdbIsAscendingOrder(STsdbReader *pReader);
 void    *tsdbGetIdx(SMeta *pMeta);
 void    *tsdbGetIvtIdx(SMeta *pMeta);
 uint64_t getReaderMaxVersion(STsdbReader *pReader);
 
-int32_t tsdbCacherowsReaderOpen(void *pVnode, int32_t type, SArray *pTableIdList, int32_t numOfCols, void **pReader);
+int32_t tsdbCacherowsReaderOpen(void *pVnode, int32_t type, SArray *pTableIdList, int32_t numOfCols, uint64_t suid,
+                                void **pReader);
 int32_t tsdbRetrieveCacheRows(void *pReader, SSDataBlock *pResBlock, const int32_t *slotIds, SArray *pTableUids);
 void   *tsdbCacherowsReaderClose(void *pReader);
 int32_t tsdbGetTableSchema(SVnode *pVnode, int64_t uid, STSchema **pSchema, int64_t *suid);

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

@@ -155,7 +155,7 @@ int64_t tqFetchLog(STQ* pTq, STqHandle* pHandle, int64_t* fetchOffset, SWalCkHea
 
 // tqExec
 int32_t tqTaosxScanLog(STQ* pTq, STqHandle* pHandle, SSubmitReq* pReq, STaosxRsp* pRsp);
-int32_t tqAddBlockDataToRsp(const SSDataBlock* pBlock, SMqDataRsp* pRsp, int32_t numOfCols);
+int32_t tqAddBlockDataToRsp(const SSDataBlock* pBlock, SMqDataRsp* pRsp, int32_t numOfCols, int8_t precision);
 int32_t tqSendDataRsp(STQ* pTq, const SRpcMsg* pMsg, const SMqPollReq* pReq, const SMqDataRsp* pRsp);
 int32_t tqPushDataRsp(STQ* pTq, STqPushEntry* pPushEntry);
 
@@ -181,15 +181,15 @@ int32_t         tqOffsetDelete(STqOffsetStore* pStore, const char* subscribeKey)
 int32_t         tqOffsetCommitFile(STqOffsetStore* pStore);
 
 // tqSink
-void tqSinkToTableMerge(SStreamTask* pTask, void* vnode, int64_t ver, void* data);
+// void tqSinkToTableMerge(SStreamTask* pTask, void* vnode, int64_t ver, void* data);
 void tqSinkToTablePipeline(SStreamTask* pTask, void* vnode, int64_t ver, void* data);
 
 // tqOffset
-char*   tqOffsetBuildFName(const char* path, int32_t ver);
+char*   tqOffsetBuildFName(const char* path, int32_t fVer);
 int32_t tqOffsetRestoreFromFile(STqOffsetStore* pStore, const char* fname);
 
 // tqStream
-int32_t tqExpandTask(STQ* pTq, SStreamTask* pTask);
+int32_t tqExpandTask(STQ* pTq, SStreamTask* pTask, int64_t ver);
 
 #ifdef __cplusplus
 }

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

@@ -298,29 +298,6 @@ int32_t tsdbMerge(STsdb *pTsdb);
 #define TSDB_CACHE_LAST_ROW(c) (((c).cacheLast & 1) > 0)
 #define TSDB_CACHE_LAST(c)     (((c).cacheLast & 2) > 0)
 
-// tsdbCache ==============================================================================================
-typedef struct {
-  TSKEY   ts;
-  SColVal colVal;
-} SLastCol;
-
-int32_t tsdbOpenCache(STsdb *pTsdb);
-void    tsdbCloseCache(STsdb *pTsdb);
-int32_t tsdbCacheInsertLast(SLRUCache *pCache, tb_uid_t uid, STSRow *row, STsdb *pTsdb);
-int32_t tsdbCacheInsertLastrow(SLRUCache *pCache, STsdb *pTsdb, tb_uid_t uid, STSRow *row, bool dup);
-int32_t tsdbCacheGetLastH(SLRUCache *pCache, tb_uid_t uid, STsdb *pTsdb, LRUHandle **h);
-int32_t tsdbCacheGetLastrowH(SLRUCache *pCache, tb_uid_t uid, STsdb *pTsdb, LRUHandle **h);
-int32_t tsdbCacheRelease(SLRUCache *pCache, LRUHandle *h);
-
-int32_t tsdbCacheDeleteLastrow(SLRUCache *pCache, tb_uid_t uid, TSKEY eKey);
-int32_t tsdbCacheDeleteLast(SLRUCache *pCache, tb_uid_t uid, TSKEY eKey);
-int32_t tsdbCacheDelete(SLRUCache *pCache, tb_uid_t uid, TSKEY eKey);
-
-void   tsdbCacheSetCapacity(SVnode *pVnode, size_t capacity);
-size_t tsdbCacheGetCapacity(SVnode *pVnode);
-
-int32_t tsdbCacheLastArray2Row(SArray *pLastArray, STSRow **ppRow, STSchema *pSchema);
-
 // tsdbDiskData ==============================================================================================
 int32_t tDiskDataBuilderCreate(SDiskDataBuilder **ppBuilder);
 void   *tDiskDataBuilderDestroy(SDiskDataBuilder *pBuilder);
@@ -729,6 +706,45 @@ void               resetLastBlockLoadInfo(SSttBlockLoadInfo *pLoadInfo);
 void               getLastBlockLoadInfo(SSttBlockLoadInfo *pLoadInfo, int64_t *blocks, double *el);
 void              *destroyLastBlockLoadInfo(SSttBlockLoadInfo *pLoadInfo);
 
+// tsdbCache ==============================================================================================
+typedef struct SCacheRowsReader {
+  SVnode            *pVnode;
+  STSchema          *pSchema;
+  uint64_t           uid;
+  uint64_t           suid;
+  char             **transferBuf;  // todo remove it soon
+  int32_t            numOfCols;
+  int32_t            type;
+  int32_t            tableIndex;  // currently returned result tables
+  SArray            *pTableList;  // table id list
+  SSttBlockLoadInfo *pLoadInfo;
+  STsdbReadSnap     *pReadSnap;
+  SDataFReader      *pDataFReader;
+  SDataFReader      *pDataFReaderLast;
+} SCacheRowsReader;
+
+typedef struct {
+  TSKEY   ts;
+  SColVal colVal;
+} SLastCol;
+
+int32_t tsdbOpenCache(STsdb *pTsdb);
+void    tsdbCloseCache(STsdb *pTsdb);
+int32_t tsdbCacheInsertLast(SLRUCache *pCache, tb_uid_t uid, STSRow *row, STsdb *pTsdb);
+int32_t tsdbCacheInsertLastrow(SLRUCache *pCache, STsdb *pTsdb, tb_uid_t uid, STSRow *row, bool dup);
+int32_t tsdbCacheGetLastH(SLRUCache *pCache, tb_uid_t uid, SCacheRowsReader *pr, LRUHandle **h);
+int32_t tsdbCacheGetLastrowH(SLRUCache *pCache, tb_uid_t uid, SCacheRowsReader *pr, LRUHandle **h);
+int32_t tsdbCacheRelease(SLRUCache *pCache, LRUHandle *h);
+
+int32_t tsdbCacheDeleteLastrow(SLRUCache *pCache, tb_uid_t uid, TSKEY eKey);
+int32_t tsdbCacheDeleteLast(SLRUCache *pCache, tb_uid_t uid, TSKEY eKey);
+int32_t tsdbCacheDelete(SLRUCache *pCache, tb_uid_t uid, TSKEY eKey);
+
+void   tsdbCacheSetCapacity(SVnode *pVnode, size_t capacity);
+size_t tsdbCacheGetCapacity(SVnode *pVnode);
+
+int32_t tsdbCacheLastArray2Row(SArray *pLastArray, STSRow **ppRow, STSchema *pSchema);
+
 // ========== inline functions ==========
 static FORCE_INLINE int32_t tsdbKeyCmprFn(const void *p1, const void *p2) {
   TSDBKEY *pKey1 = (TSDBKEY *)p1;

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

@@ -1062,7 +1062,7 @@ int32_t metaFilterCreateTime(SMeta *pMeta, SMetaFltParam *param, SArray *pUids)
   if (tdbTbcMoveTo(pCursor->pCur, &ctimeKey, sizeof(ctimeKey), &cmp) < 0) {
     goto END;
   }
-  bool    first = true;
+
   int32_t valid = 0;
   while (1) {
     void   *entryKey = NULL;
@@ -1074,7 +1074,13 @@ int32_t metaFilterCreateTime(SMeta *pMeta, SMetaFltParam *param, SArray *pUids)
 
     int32_t cmp = (*param->filterFunc)((void *)&p->ctime, (void *)&pCtimeKey->ctime, param->type);
     if (cmp == 0) taosArrayPush(pUids, &p->uid);
-    if (cmp == -1) break;
+
+    if (param->reverse == false) {
+      if (cmp == -1) break;
+    } else if (param->reverse) {
+      if (cmp == 1) break;
+    }
+
     valid = param->reverse ? tdbTbcMoveToPrev(pCursor->pCur) : tdbTbcMoveToNext(pCursor->pCur);
     if (valid < 0) break;
   }

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

@@ -572,8 +572,12 @@ static int metaBuildCtimeIdxKey(SCtimeIdxKey *ctimeKey, const SMetaEntry *pME) {
 }
 
 static int metaBuildNColIdxKey(SNcolIdxKey *ncolKey, const SMetaEntry *pME) {
-  ncolKey->ncol = pME->ntbEntry.schemaRow.nCols;
+  if (pME->type == TSDB_NORMAL_TABLE) {
-  ncolKey->uid = pME->uid;
+    ncolKey->ncol = pME->ntbEntry.schemaRow.nCols;
+    ncolKey->uid = pME->uid;
+  } else {
+    return -1;
+  }
   return 0;
 }
 
@@ -777,9 +781,13 @@ static int metaAlterTableColumn(SMeta *pMeta, int64_t version, SVAlterTbReq *pAl
     terrno = TSDB_CODE_VND_INVALID_TABLE_ACTION;
     goto _err;
   }
-
   // search the column to add/drop/update
   pSchema = &entry.ntbEntry.schemaRow;
+
+  // save old entry
+  SMetaEntry oldEntry = {.type = TSDB_NORMAL_TABLE, .uid = entry.uid};
+  oldEntry.ntbEntry.schemaRow.nCols = pSchema->nCols;
+
   int32_t iCol = 0;
   for (;;) {
     pColumn = NULL;
@@ -872,6 +880,9 @@ static int metaAlterTableColumn(SMeta *pMeta, int64_t version, SVAlterTbReq *pAl
 
   entry.version = version;
 
+  metaDeleteNcolIdx(pMeta, &oldEntry);
+  metaUpdateNcolIdx(pMeta, &entry);
+
   // do actual write
   metaWLock(pMeta);
 

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

@@ -15,11 +15,11 @@
 
 #include "sma.h"
 
-#define RSMA_QTASKEXEC_SMOOTH_SIZE (100)                                                 // cnt
+#define RSMA_QTASKEXEC_SMOOTH_SIZE (100)     // cnt
-#define RSMA_SUBMIT_BATCH_SIZE     (1024)                                                // cnt
+#define RSMA_SUBMIT_BATCH_SIZE     (1024)    // cnt
-#define RSMA_FETCH_DELAY_MAX       (120000)                                              // ms
+#define RSMA_FETCH_DELAY_MAX       (120000)  // ms
-#define RSMA_FETCH_ACTIVE_MAX      (1000)                                                // ms
+#define RSMA_FETCH_ACTIVE_MAX      (1000)    // ms
-#define RSMA_FETCH_INTERVAL        (5000)                                                // ms
+#define RSMA_FETCH_INTERVAL        (5000)    // ms
 
 SSmaMgmt smaMgmt = {
     .inited = 0,
@@ -839,7 +839,7 @@ static int32_t tdExecuteRSmaImpl(SSma *pSma, const void *pMsg, int32_t msgSize,
     tdRsmaPrintSubmitReq(pSma, pReq);
   }
 #endif
-  if (qSetMultiStreamInput(qTaskInfo, pMsg, msgSize, inputType) < 0) {
+  if (qSetSMAInput(qTaskInfo, pMsg, msgSize, inputType) < 0) {
     smaError("vgId:%d, rsma %" PRIi8 " qSetStreamInput failed since %s", SMA_VID(pSma), level, tstrerror(terrno));
     return TSDB_CODE_FAILED;
   }
@@ -1404,7 +1404,7 @@ static int32_t tdRSmaFetchAllResult(SSma *pSma, SRSmaInfo *pInfo) {
 
       pItem->nScanned = 0;
 
-      if ((terrno = qSetMultiStreamInput(taskInfo, &dataBlock, 1, STREAM_INPUT__DATA_BLOCK)) < 0) {
+      if ((terrno = qSetSMAInput(taskInfo, &dataBlock, 1, STREAM_INPUT__DATA_BLOCK)) < 0) {
         goto _err;
       }
       if (tdRSmaExecAndSubmitResult(pSma, taskInfo, pItem, pInfo->pTSchema, pInfo->suid) < 0) {