Merge branch '3.0' into fix/TD-18445

cmake/rust-bindings_CMakeLists.txt.in

@@ -1,12 +0,0 @@
-
-# rust-bindings
-ExternalProject_Add(rust-bindings
-        GIT_REPOSITORY https://github.com/songtianyi/tdengine-rust-bindings.git
-        GIT_TAG 7ed7a97
-        SOURCE_DIR "${TD_SOURCE_DIR}/examples/rust"
-        BINARY_DIR "${TD_SOURCE_DIR}/examples/rust"
-        CONFIGURE_COMMAND "" 
-        BUILD_COMMAND ""  
-        INSTALL_COMMAND ""
-        TEST_COMMAND ""
-        )

contrib/CMakeLists.txt

@@ -105,11 +105,6 @@ if(${BUILD_WITH_SQLITE})
     cat("${TD_SUPPORT_DIR}/sqlite_CMakeLists.txt.in" ${CONTRIB_TMP_FILE})
 endif(${BUILD_WITH_SQLITE})
 
-# rust-bindings
-if(${RUST_BINDINGS})
-    cat("${TD_SUPPORT_DIR}/rust-bindings_CMakeLists.txt.in" ${CONTRIB_TMP_FILE})
-endif(${RUST_BINDINGS})
-
 # lucene
 if(${BUILD_WITH_LUCENE})
     cat("${TD_SUPPORT_DIR}/lucene_CMakeLists.txt.in" ${CONTRIB_TMP_FILE})

docs/examples/c/tmq_example.c

@@ -0,0 +1,275 @@
+/*
+ * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
+ *
+ * This program is free software: you can use, redistribute, and/or modify
+ * it under the terms of the GNU Affero General Public License, version 3
+ * or later ("AGPL"), as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <assert.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+#include "taos.h"
+
+static int  running = 1;
+static char dbName[64] = "tmqdb";
+static char stbName[64] = "stb";
+static char topicName[64] = "topicname";
+
+static int32_t msg_process(TAOS_RES* msg) {
+  char    buf[1024];
+  int32_t rows = 0;
+
+  const char* topicName = tmq_get_topic_name(msg);
+  const char* dbName = tmq_get_db_name(msg);
+  int32_t     vgroupId = tmq_get_vgroup_id(msg);
+
+  printf("topic: %s\n", topicName);
+  printf("db: %s\n", dbName);
+  printf("vgroup id: %d\n", vgroupId);
+
+  while (1) {
+    TAOS_ROW row = taos_fetch_row(msg);
+    if (row == NULL) break;
+
+    TAOS_FIELD* fields = taos_fetch_fields(msg);
+    int32_t     numOfFields = taos_field_count(msg);
+    int32_t*    length = taos_fetch_lengths(msg);
+    int32_t     precision = taos_result_precision(msg);
+    rows++;
+    taos_print_row(buf, row, fields, numOfFields);
+    printf("row content: %s\n", buf);
+  }
+
+  return rows;
+}
+
+static int32_t init_env() {
+  TAOS* pConn = taos_connect("localhost", "root", "taosdata", NULL, 0);
+  if (pConn == NULL) {
+    return -1;
+  }
+
+  TAOS_RES* pRes;
+  // drop database if exists
+  printf("create database\n");
+  pRes = taos_query(pConn, "drop database if exists tmqdb");
+  if (taos_errno(pRes) != 0) {
+    printf("error in drop tmqdb, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  // create database
+  pRes = taos_query(pConn, "create database tmqdb");
+  if (taos_errno(pRes) != 0) {
+    printf("error in create tmqdb, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  // create super table
+  printf("create super table\n");
+  pRes = taos_query(
+      pConn, "create table tmqdb.stb (ts timestamp, c1 int, c2 float, c3 varchar(16)) tags(t1 int, t3 varchar(16))");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create super table stb, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  // create sub tables
+  printf("create sub tables\n");
+  pRes = taos_query(pConn, "create table tmqdb.ctb0 using tmqdb.stb tags(0, 'subtable0')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create super table ctb0, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "create table tmqdb.ctb1 using tmqdb.stb tags(1, 'subtable1')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create super table ctb1, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "create table tmqdb.ctb2 using tmqdb.stb tags(2, 'subtable2')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create super table ctb2, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "create table tmqdb.ctb3 using tmqdb.stb tags(3, 'subtable3')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create super table ctb3, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  // insert data
+  printf("insert data into sub tables\n");
+  pRes = taos_query(pConn, "insert into tmqdb.ctb0 values(now, 0, 0, 'a0')(now+1s, 0, 0, 'a00')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to insert into ctb0, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "insert into tmqdb.ctb1 values(now, 1, 1, 'a1')(now+1s, 11, 11, 'a11')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to insert into ctb0, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "insert into tmqdb.ctb2 values(now, 2, 2, 'a1')(now+1s, 22, 22, 'a22')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to insert into ctb0, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "insert into tmqdb.ctb3 values(now, 3, 3, 'a1')(now+1s, 33, 33, 'a33')");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to insert into ctb0, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  taos_close(pConn);
+  return 0;
+}
+
+int32_t create_topic() {
+  printf("create topic\n");
+  TAOS_RES* pRes;
+  TAOS*     pConn = taos_connect("localhost", "root", "taosdata", NULL, 0);
+  if (pConn == NULL) {
+    return -1;
+  }
+
+  pRes = taos_query(pConn, "use tmqdb");
+  if (taos_errno(pRes) != 0) {
+    printf("error in use tmqdb, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  pRes = taos_query(pConn, "create topic topicname as select ts, c1, c2, c3, tbname from tmqdb.stb where c1 > 1");
+  if (taos_errno(pRes) != 0) {
+    printf("failed to create topic topicname, reason:%s\n", taos_errstr(pRes));
+    return -1;
+  }
+  taos_free_result(pRes);
+
+  taos_close(pConn);
+  return 0;
+}
+
+void tmq_commit_cb_print(tmq_t* tmq, int32_t code, void* param) {
+  printf("tmq_commit_cb_print() code: %d, tmq: %p, param: %p\n", code, tmq, param);
+}
+
+tmq_t* build_consumer() {
+  tmq_conf_res_t code;
+  tmq_conf_t*    conf = tmq_conf_new();
+  code = tmq_conf_set(conf, "enable.auto.commit", "true");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "auto.commit.interval.ms", "1000");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "group.id", "cgrpName");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "client.id", "user defined name");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "td.connect.user", "root");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "td.connect.pass", "taosdata");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "auto.offset.reset", "earliest");
+  if (TMQ_CONF_OK != code) return NULL;
+  code = tmq_conf_set(conf, "experimental.snapshot.enable", "false");
+  if (TMQ_CONF_OK != code) return NULL;
+
+  tmq_conf_set_auto_commit_cb(conf, tmq_commit_cb_print, NULL);
+
+  tmq_t* tmq = tmq_consumer_new(conf, NULL, 0);
+  tmq_conf_destroy(conf);
+  return tmq;
+}
+
+tmq_list_t* build_topic_list() {
+  tmq_list_t* topicList = tmq_list_new();
+  int32_t     code = tmq_list_append(topicList, "topicname");
+  if (code) {
+    return NULL;
+  }
+  return topicList;
+}
+
+void basic_consume_loop(tmq_t* tmq) {
+  int32_t totalRows = 0;
+  int32_t msgCnt = 0;
+  int32_t timeout = 5000;
+  while (running) {
+    TAOS_RES* tmqmsg = tmq_consumer_poll(tmq, timeout);
+    if (tmqmsg) {
+      msgCnt++;
+      totalRows += msg_process(tmqmsg);
+      taos_free_result(tmqmsg);
+    } else {
+      break;
+    }
+  }
+
+  fprintf(stderr, "%d msg consumed, include %d rows\n", msgCnt, totalRows);
+}
+
+int main(int argc, char* argv[]) {
+  int32_t code;
+
+  if (init_env() < 0) {
+    return -1;
+  }
+
+  if (create_topic() < 0) {
+    return -1;
+  }
+
+  tmq_t* tmq = build_consumer();
+  if (NULL == tmq) {
+    fprintf(stderr, "%% build_consumer() fail!\n");
+    return -1;
+  }
+
+  tmq_list_t* topic_list = build_topic_list();
+  if (NULL == topic_list) {
+    return -1;
+  }
+
+  if ((code = tmq_subscribe(tmq, topic_list))) {
+    fprintf(stderr, "%% Failed to tmq_subscribe(): %s\n", tmq_err2str(code));
+  }
+  tmq_list_destroy(topic_list);
+
+  basic_consume_loop(tmq);
+
+  code = tmq_consumer_close(tmq);
+  if (code) {
+    fprintf(stderr, "%% Failed to close consumer: %s\n", tmq_err2str(code));
+  } else {
+    fprintf(stderr, "%% Consumer closed\n");
+  }
+
+  return 0;
+}

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

@@ -724,7 +724,6 @@ consumer.close();
 
 </TabItem>
 
-
 <TabItem value="Go" label="Go">
 
 ```go
@@ -769,6 +768,7 @@ consumer.Unsubscribe();
 // 关闭消费
 consumer.Close();
 ```
+
 </TabItem>
 
 </Tabs>
@@ -809,7 +809,7 @@ SHOW SUBSCRIPTIONS;
 <Tabs defaultValue="java" groupId="lang">
 
 <TabItem label="C" value="c">
-  <CDemo>
+  <CDemo />
 </TabItem>
 
 <TabItem label="Java" value="java">

docs/zh/07-develop/_sub_c.mdx

@@ -1,3 +1,3 @@
 ```c
-{{#include docs/examples/c/tmq-example.c}}
+{{#include docs/examples/c/tmq_example.c}}
 ```

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

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

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

@@ -1,4 +0,0 @@
----
-sidebar_label: taosX
-title: 使用 taosX 在集群间复制数据
----

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

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

examples/c/stream_demo.c

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

examples/rust/.gitignore

@@ -0,0 +1,2 @@
+/target
+Cargo.lock

examples/rust/Cargo.toml

@@ -0,0 +1,18 @@
+[package]
+name = "rust"
+version = "0.1.0"
+edition = "2021"
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]
+taos = "*"
+
+[dev-dependencies]
+chrono = "0.4"
+itertools = "0.10.3"
+pretty_env_logger = "0.4.0"
+serde = { version = "1", features = ["derive"] }
+serde_json = "1"
+tokio = { version = "1", features = ["full"] }
+anyhow = "1"

examples/rust/examples/bind-tags.rs

@@ -0,0 +1,80 @@
+use anyhow::Result;
+use serde::Deserialize;
+use taos::*;
+
+#[tokio::main]
+async fn main() -> Result<()> {
+    let taos = TaosBuilder::from_dsn("taos://")?.build()?;
+    taos.exec_many([
+        "drop database if exists test",
+        "create database test keep 36500",
+        "use test",
+        "create table tb1 (ts timestamp, c1 bool, c2 tinyint, c3 smallint, c4 int, c5 bigint,
+            c6 tinyint unsigned, c7 smallint unsigned, c8 int unsigned, c9 bigint unsigned,
+            c10 float, c11 double, c12 varchar(100), c13 nchar(100)) tags(t1 varchar(100))",
+    ])
+    .await?;
+    let mut stmt = Stmt::init(&taos)?;
+    stmt.prepare(
+        "insert into ? using tb1 tags(?) values(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)",
+    )?;
+    stmt.set_tbname("d0")?;
+    stmt.set_tags(&[Value::VarChar("涛思".to_string())])?;
+
+    let params = vec![
+        ColumnView::from_millis_timestamp(vec![164000000000]),
+        ColumnView::from_bools(vec![true]),
+        ColumnView::from_tiny_ints(vec![i8::MAX]),
+        ColumnView::from_small_ints(vec![i16::MAX]),
+        ColumnView::from_ints(vec![i32::MAX]),
+        ColumnView::from_big_ints(vec![i64::MAX]),
+        ColumnView::from_unsigned_tiny_ints(vec![u8::MAX]),
+        ColumnView::from_unsigned_small_ints(vec![u16::MAX]),
+        ColumnView::from_unsigned_ints(vec![u32::MAX]),
+        ColumnView::from_unsigned_big_ints(vec![u64::MAX]),
+        ColumnView::from_floats(vec![f32::MAX]),
+        ColumnView::from_doubles(vec![f64::MAX]),
+        ColumnView::from_varchar(vec!["ABC"]),
+        ColumnView::from_nchar(vec!["涛思数据"]),
+    ];
+    let rows = stmt.bind(&params)?.add_batch()?.execute()?;
+    assert_eq!(rows, 1);
+
+    #[derive(Debug, Deserialize)]
+    #[allow(dead_code)]
+    struct Row {
+        ts: String,
+        c1: bool,
+        c2: i8,
+        c3: i16,
+        c4: i32,
+        c5: i64,
+        c6: u8,
+        c7: u16,
+        c8: u32,
+        c9: u64,
+        c10: Option<f32>,
+        c11: f64,
+        c12: String,
+        c13: String,
+        t1: serde_json::Value,
+    }
+
+    let rows: Vec<Row> = taos
+        .query("select * from tb1")
+        .await?
+        .deserialize()
+        .try_collect()
+        .await?;
+    let row = &rows[0];
+    dbg!(&row);
+    assert_eq!(row.c5, i64::MAX);
+    assert_eq!(row.c8, u32::MAX);
+    assert_eq!(row.c9, u64::MAX);
+    assert_eq!(row.c10.unwrap(), f32::MAX);
+    // assert_eq!(row.c11, f64::MAX);
+    assert_eq!(row.c12, "ABC");
+    assert_eq!(row.c13, "涛思数据");
+
+    Ok(())
+}

examples/rust/examples/bind.rs

@@ -0,0 +1,74 @@
+use anyhow::Result;
+use serde::Deserialize;
+use taos::*;
+
+#[tokio::main]
+async fn main() -> Result<()> {
+    let taos = TaosBuilder::from_dsn("taos://")?.build()?;
+    taos.exec_many([
+        "drop database if exists test_bindable",
+        "create database test_bindable keep 36500",
+        "use test_bindable",
+        "create table tb1 (ts timestamp, c1 bool, c2 tinyint, c3 smallint, c4 int, c5 bigint,
+            c6 tinyint unsigned, c7 smallint unsigned, c8 int unsigned, c9 bigint unsigned,
+            c10 float, c11 double, c12 varchar(100), c13 nchar(100))",
+    ])
+    .await?;
+    let mut stmt = Stmt::init(&taos)?;
+    stmt.prepare("insert into tb1 values(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)")?;
+    let params = vec![
+        ColumnView::from_millis_timestamp(vec![0]),
+        ColumnView::from_bools(vec![true]),
+        ColumnView::from_tiny_ints(vec![i8::MAX]),
+        ColumnView::from_small_ints(vec![i16::MAX]),
+        ColumnView::from_ints(vec![i32::MAX]),
+        ColumnView::from_big_ints(vec![i64::MAX]),
+        ColumnView::from_unsigned_tiny_ints(vec![u8::MAX]),
+        ColumnView::from_unsigned_small_ints(vec![u16::MAX]),
+        ColumnView::from_unsigned_ints(vec![u32::MAX]),
+        ColumnView::from_unsigned_big_ints(vec![u64::MAX]),
+        ColumnView::from_floats(vec![f32::MAX]),
+        ColumnView::from_doubles(vec![f64::MAX]),
+        ColumnView::from_varchar(vec!["ABC"]),
+        ColumnView::from_nchar(vec!["涛思数据"]),
+    ];
+    let rows = stmt.bind(&params)?.add_batch()?.execute()?;
+    assert_eq!(rows, 1);
+
+    #[derive(Debug, Deserialize)]
+    #[allow(dead_code)]
+    struct Row {
+        ts: String,
+        c1: bool,
+        c2: i8,
+        c3: i16,
+        c4: i32,
+        c5: i64,
+        c6: u8,
+        c7: u16,
+        c8: u32,
+        c9: u64,
+        c10: Option<f32>,
+        c11: f64,
+        c12: String,
+        c13: String,
+    }
+
+    let rows: Vec<Row> = taos
+        .query("select * from tb1")
+        .await?
+        .deserialize()
+        .try_collect()
+        .await?;
+    let row = &rows[0];
+    dbg!(&row);
+    assert_eq!(row.c5, i64::MAX);
+    assert_eq!(row.c8, u32::MAX);
+    assert_eq!(row.c9, u64::MAX);
+    assert_eq!(row.c10.unwrap(), f32::MAX);
+    // assert_eq!(row.c11, f64::MAX);
+    assert_eq!(row.c12, "ABC");
+    assert_eq!(row.c13, "涛思数据");
+
+    Ok(())
+}

examples/rust/examples/query.rs

@@ -0,0 +1,106 @@
+use std::time::Duration;
+
+use chrono::{DateTime, Local};
+use taos::*;
+
+#[tokio::main]
+async fn main() -> anyhow::Result<()> {
+    let dsn = "taos://";
+
+    let opts = PoolBuilder::new()
+        .max_size(5000) // max connections
+        .max_lifetime(Some(Duration::from_secs(60 * 60))) // lifetime of each connection
+        .min_idle(Some(1000)) // minimal idle connections
+        .connection_timeout(Duration::from_secs(2));
+
+    let pool = TaosBuilder::from_dsn(dsn)?.with_pool_builder(opts)?;
+
+    let taos = pool.get()?;
+
+    let db = "query";
+
+    // prepare database
+    taos.exec_many([
+        format!("DROP DATABASE IF EXISTS `{db}`"),
+        format!("CREATE DATABASE `{db}`"),
+        format!("USE `{db}`"),
+    ])
+    .await?;
+
+    let inserted = taos.exec_many([
+        // create super table
+        "CREATE TABLE `meters` (`ts` TIMESTAMP, `current` FLOAT, `voltage` INT, `phase` FLOAT) TAGS (`groupid` INT, `location` BINARY(16))",
+        // create child table
+        "CREATE TABLE `d0` USING `meters` TAGS(0, 'Los Angles')",
+        // insert into child table
+        "INSERT INTO `d0` values(now - 10s, 10, 116, 0.32)",
+        // insert with NULL values
+        "INSERT INTO `d0` values(now - 8s, NULL, NULL, NULL)",
+        // insert and automatically create table with tags if not exists
+        "INSERT INTO `d1` USING `meters` TAGS(1, 'San Francisco') values(now - 9s, 10.1, 119, 0.33)",
+        // insert many records in a single sql
+        "INSERT INTO `d1` values (now-8s, 10, 120, 0.33) (now - 6s, 10, 119, 0.34) (now - 4s, 11.2, 118, 0.322)",
+    ]).await?;
+
+    assert_eq!(inserted, 6);
+    loop {
+        let count: usize = taos
+            .query_one("select count(*) from `meters`")
+            .await?
+            .unwrap_or_default();
+
+        if count >= 6 {
+            break;
+        } else {
+            println!("waiting for data");
+        }
+    }
+
+    let mut result = taos.query("select tbname, * from `meters`").await?;
+
+    for field in result.fields() {
+        println!("got field: {}", field.name());
+    }
+
+    // Query option 1, use rows stream.
+    let mut rows = result.rows();
+    let mut nrows = 0;
+    while let Some(row) = rows.try_next().await? {
+        for (col, (name, value)) in row.enumerate() {
+            println!(
+                "[{}] got value in col {} (named `{:>8}`): {}",
+                nrows, col, name, value
+            );
+        }
+        nrows += 1;
+    }
+
+    // Query options 2, use deserialization with serde.
+    #[derive(Debug, serde::Deserialize)]
+    #[allow(dead_code)]
+    struct Record {
+        tbname: String,
+        // deserialize timestamp to chrono::DateTime<Local>
+        ts: DateTime<Local>,
+        // float to f32
+        current: Option<f32>,
+        // int to i32
+        voltage: Option<i32>,
+        phase: Option<f32>,
+        groupid: i32,
+        // binary/varchar to String
+        location: String,
+    }
+
+    let records: Vec<Record> = taos
+        .query("select tbname, * from `meters`")
+        .await?
+        .deserialize()
+        .try_collect()
+        .await?;
+
+    dbg!(result.summary());
+    assert_eq!(records.len(), 6);
+    dbg!(records);
+    Ok(())
+}

examples/rust/examples/subscribe.rs

@@ -0,0 +1,103 @@
+use std::time::Duration;
+
+use chrono::{DateTime, Local};
+use taos::*;
+
+// Query options 2, use deserialization with serde.
+#[derive(Debug, serde::Deserialize)]
+#[allow(dead_code)]
+struct Record {
+    // deserialize timestamp to chrono::DateTime<Local>
+    ts: DateTime<Local>,
+    // float to f32
+    current: Option<f32>,
+    // int to i32
+    voltage: Option<i32>,
+    phase: Option<f32>,
+}
+
+async fn prepare(taos: Taos) -> anyhow::Result<()> {
+    let inserted = taos.exec_many([
+        // create child table
+        "CREATE TABLE `d0` USING `meters` TAGS(0, 'Los Angles')",
+        // insert into child table
+        "INSERT INTO `d0` values(now - 10s, 10, 116, 0.32)",
+        // insert with NULL values
+        "INSERT INTO `d0` values(now - 8s, NULL, NULL, NULL)",
+        // insert and automatically create table with tags if not exists
+        "INSERT INTO `d1` USING `meters` TAGS(1, 'San Francisco') values(now - 9s, 10.1, 119, 0.33)",
+        // insert many records in a single sql
+        "INSERT INTO `d1` values (now-8s, 10, 120, 0.33) (now - 6s, 10, 119, 0.34) (now - 4s, 11.2, 118, 0.322)",
+    ]).await?;
+    assert_eq!(inserted, 6);
+    Ok(())
+}
+
+#[tokio::main]
+async fn main() -> anyhow::Result<()> {
+    // std::env::set_var("RUST_LOG", "debug");
+    pretty_env_logger::init();
+    let dsn = "taos://localhost:6030";
+    let builder = TaosBuilder::from_dsn(dsn)?;
+
+    let taos = builder.build()?;
+    let db = "tmq";
+
+    // prepare database
+    taos.exec_many([
+        "DROP TOPIC IF EXISTS tmq_meters".to_string(),
+        format!("DROP DATABASE IF EXISTS `{db}`"),
+        format!("CREATE DATABASE `{db}`"),
+        format!("USE `{db}`"),
+        // create super table
+        "CREATE TABLE `meters` (`ts` TIMESTAMP, `current` FLOAT, `voltage` INT, `phase` FLOAT) TAGS (`groupid` INT, `location` BINARY(16))".to_string(),
+        // create topic for subscription
+        format!("CREATE TOPIC tmq_meters with META AS DATABASE {db}")
+    ])
+    .await?;
+
+    let task = tokio::spawn(prepare(taos));
+
+    tokio::time::sleep(Duration::from_secs(1)).await;
+
+    // subscribe
+    let tmq = TmqBuilder::from_dsn("taos://localhost:6030/?group.id=test")?;
+
+    let mut consumer = tmq.build()?;
+    consumer.subscribe(["tmq_meters"]).await?;
+
+    {
+        let mut stream = consumer.stream();
+
+        while let Some((offset, message)) = stream.try_next().await? {
+            // get information from offset
+
+            // the topic
+            let topic = offset.topic();
+            // the vgroup id, like partition id in kafka.
+            let vgroup_id = offset.vgroup_id();
+            println!("* in vgroup id {vgroup_id} of topic {topic}\n");
+
+            if let Some(data) = message.into_data() {
+                while let Some(block) = data.fetch_raw_block().await? {
+                    // one block for one table, get table name if needed
+                    let name = block.table_name();
+                    let records: Vec<Record> = block.deserialize().try_collect()?;
+                    println!(
+                        "** table: {}, got {} records: {:#?}\n",
+                        name.unwrap(),
+                        records.len(),
+                        records
+                    );
+                }
+            }
+            consumer.commit(offset).await?;
+        }
+    }
+
+    consumer.unsubscribe().await;
+
+    task.await??;
+
+    Ok(())
+}

examples/rust/src/main.rs

@@ -0,0 +1,3 @@
+fn main() {
+    println!("Hello, world!");
+}

include/common/tcommon.h

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

include/libs/stream/tstream.h

@@ -53,6 +53,7 @@ enum {
   TASK_SCHED_STATUS__WAITING,
   TASK_SCHED_STATUS__ACTIVE,
   TASK_SCHED_STATUS__FAILED,
+  TASK_SCHED_STATUS__DROPPING,
 };
 
 enum {
@@ -127,6 +128,10 @@ typedef struct {
   int8_t type;
 } SStreamCheckpoint;
 
+typedef struct {
+  int8_t type;
+} SStreamTaskDestroy;
+
 typedef struct {
   int8_t       type;
   SSDataBlock* pBlock;
@@ -211,7 +216,6 @@ typedef struct {
   void*     vnode;
   FTbSink*  tbSinkFunc;
   STSchema* pTSchema;
-  SHashObj* pHash;  // groupId to tbuid
 } STaskSinkTb;
 
 typedef void FSmaSink(void* vnode, int64_t smaId, const SArray* data);

include/libs/sync/sync.h

@@ -30,7 +30,10 @@ extern bool gRaftDetailLog;
 #define SYNC_SPEED_UP_HB_TIMER       400
 #define SYNC_SPEED_UP_AFTER_MS       (1000 * 20)
 #define SYNC_SLOW_DOWN_RANGE         100
-#define SYNC_MAX_READ_RANGE    10
+#define SYNC_MAX_READ_RANGE          2
+#define SYNC_MAX_PROGRESS_WAIT_MS    4000
+#define SYNC_MAX_START_TIME_RANGE_MS (1000 * 20)
+#define SYNC_MAX_RECV_TIME_RANGE_MS  1000
 
 #define SYNC_MAX_BATCH_SIZE 1
 #define SYNC_INDEX_BEGIN    0

include/libs/sync/syncTools.h

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

include/util/taoserror.h

@@ -291,6 +291,7 @@ int32_t* taosGetErrno();
 #define TSDB_CODE_MND_STREAM_NOT_EXIST          TAOS_DEF_ERROR_CODE(0, 0x03F1)
 #define TSDB_CODE_MND_INVALID_STREAM_OPTION     TAOS_DEF_ERROR_CODE(0, 0x03F2)
 #define TSDB_CODE_MND_STREAM_MUST_BE_DELETED    TAOS_DEF_ERROR_CODE(0, 0x03F3)
+#define TSDB_CODE_MND_STREAM_TASK_DROPPED       TAOS_DEF_ERROR_CODE(0, 0x03F4)
 
 // mnode-sma
 #define TSDB_CODE_MND_SMA_ALREADY_EXIST         TAOS_DEF_ERROR_CODE(0, 0x0480)

source/common/src/tdatablock.c

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

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

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

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

@@ -1145,7 +1145,7 @@ static int32_t mndAddSuperTableTag(const SStbObj *pOld, SStbObj *pNew, SArray *p
   return 0;
 }
 
-int32_t mndCheckColAndTagModifiable(SMnode *pMnode, const char *stbname, int64_t suid, col_id_t colId) {
+static int32_t mndCheckAlterColForTopic(SMnode *pMnode, const char *stbFullName, int64_t suid, col_id_t colId) {
   SSdb *pSdb = pMnode->pSdb;
   void *pIter = NULL;
   while (1) {
@@ -1154,7 +1154,7 @@ int32_t mndCheckColAndTagModifiable(SMnode *pMnode, const char *stbname, int64_t
     if (pIter == NULL) break;
 
     mDebug("topic:%s, check tag and column modifiable, stb:%s suid:%" PRId64 " colId:%d, subType:%d sql:%s",
-           pTopic->name, stbname, suid, colId, pTopic->subType, pTopic->sql);
+           pTopic->name, stbFullName, suid, colId, pTopic->subType, pTopic->sql);
     if (pTopic->subType != TOPIC_SUB_TYPE__COLUMN) {
       sdbRelease(pSdb, pTopic);
       continue;
@@ -1192,20 +1192,66 @@ int32_t mndCheckColAndTagModifiable(SMnode *pMnode, const char *stbname, int64_t
     sdbRelease(pSdb, pTopic);
     nodesDestroyNode(pAst);
   }
+  return 0;
+}
 
+static int32_t mndCheckAlterColForStream(SMnode *pMnode, const char *stbFullName, int64_t suid, col_id_t colId) {
+  SSdb *pSdb = pMnode->pSdb;
+  void *pIter = NULL;
+  while (1) {
+    SStreamObj *pStream = NULL;
+    pIter = sdbFetch(pSdb, SDB_STREAM, pIter, (void **)&pStream);
+    if (pIter == NULL) break;
+
+    SNode *pAst = NULL;
+    if (nodesStringToNode(pStream->ast, &pAst) != 0) {
+      ASSERT(0);
+      return -1;
+    }
+
+    SNodeList *pNodeList = NULL;
+    nodesCollectColumns((SSelectStmt *)pAst, SQL_CLAUSE_FROM, NULL, COLLECT_COL_TYPE_ALL, &pNodeList);
+    SNode *pNode = NULL;
+    FOREACH(pNode, pNodeList) {
+      SColumnNode *pCol = (SColumnNode *)pNode;
+
+      if (pCol->tableId != suid) {
+        mDebug("stream:%s, check colId:%d passed", pStream->name, pCol->colId);
+        goto NEXT;
+      }
+      if (pCol->colId > 0 && pCol->colId == colId) {
+        sdbRelease(pSdb, pStream);
+        nodesDestroyNode(pAst);
+        terrno = TSDB_CODE_MND_STREAM_MUST_BE_DELETED;
+        mError("stream:%s, check colId:%d conflicted", pStream->name, pCol->colId);
+        return -1;
+      }
+      mDebug("stream:%s, check colId:%d passed", pStream->name, pCol->colId);
+    }
+
+  NEXT:
+    sdbRelease(pSdb, pStream);
+    nodesDestroyNode(pAst);
+  }
+  return 0;
+}
+
+static int32_t mndCheckAlterColForTSma(SMnode *pMnode, const char *stbFullName, int64_t suid, col_id_t colId) {
+  SSdb *pSdb = pMnode->pSdb;
+  void *pIter = NULL;
   while (1) {
     SSmaObj *pSma = NULL;
     pIter = sdbFetch(pSdb, SDB_SMA, pIter, (void **)&pSma);
     if (pIter == NULL) break;
 
-    mDebug("tsma:%s, check tag and column modifiable, stb:%s suid:%" PRId64 " colId:%d, sql:%s", pSma->name, stbname,
+    mDebug("tsma:%s, check tag and column modifiable, stb:%s suid:%" PRId64 " colId:%d, sql:%s", pSma->name,
-           suid, colId, pSma->sql);
+           stbFullName, suid, colId, pSma->sql);
 
     SNode *pAst = NULL;
     if (nodesStringToNode(pSma->ast, &pAst) != 0) {
       terrno = TSDB_CODE_SDB_INVALID_DATA_CONTENT;
       mError("tsma:%s, check tag and column modifiable, stb:%s suid:%" PRId64 " colId:%d failed since parse AST err",
-             pSma->name, stbname, suid, colId);
+             pSma->name, stbFullName, suid, colId);
       return -1;
     }
 
@@ -1218,7 +1264,7 @@ int32_t mndCheckColAndTagModifiable(SMnode *pMnode, const char *stbname, int64_t
 
       if ((pCol->tableId != suid) && (pSma->stbUid != suid)) {
         mDebug("tsma:%s, check colId:%d passed", pSma->name, pCol->colId);
-        goto NEXT2;
+        goto NEXT;
       }
       if ((pCol->colId) > 0 && (pCol->colId == colId)) {
         sdbRelease(pSdb, pSma);
@@ -1230,11 +1276,24 @@ int32_t mndCheckColAndTagModifiable(SMnode *pMnode, const char *stbname, int64_t
       mDebug("tsma:%s, check colId:%d passed", pSma->name, pCol->colId);
     }
 
-  NEXT2:
+  NEXT:
     sdbRelease(pSdb, pSma);
     nodesDestroyNode(pAst);
   }
+  return 0;
+}
 
+int32_t mndCheckColAndTagModifiable(SMnode *pMnode, const char *stbFullName, int64_t suid, col_id_t colId) {
+  if (mndCheckAlterColForTopic(pMnode, stbFullName, suid, colId) < 0) {
+    return -1;
+  }
+  if (mndCheckAlterColForStream(pMnode, stbFullName, suid, colId) < 0) {
+    return -1;
+  }
+
+  if (mndCheckAlterColForTSma(pMnode, stbFullName, suid, colId) < 0) {
+    return -1;
+  }
   return 0;
 }
 
@@ -1930,6 +1989,98 @@ _OVER:
   return code;
 }
 
+static int32_t mndCheckDropStbForTopic(SMnode *pMnode, const char *stbFullName, int64_t suid) {
+  SSdb *pSdb = pMnode->pSdb;
+  void *pIter = NULL;
+  while (1) {
+    SMqTopicObj *pTopic = NULL;
+    pIter = sdbFetch(pSdb, SDB_TOPIC, pIter, (void **)&pTopic);
+    if (pIter == NULL) break;
+
+    if (pTopic->subType == TOPIC_SUB_TYPE__TABLE) {
+      if (pTopic->stbUid == suid) {
+        sdbRelease(pSdb, pTopic);
+        return -1;
+      }
+    }
+
+    if (pTopic->subType != TOPIC_SUB_TYPE__COLUMN) {
+      sdbRelease(pSdb, pTopic);
+      continue;
+    }
+
+    SNode *pAst = NULL;
+    if (nodesStringToNode(pTopic->ast, &pAst) != 0) {
+      ASSERT(0);
+      return -1;
+    }
+
+    SNodeList *pNodeList = NULL;
+    nodesCollectColumns((SSelectStmt *)pAst, SQL_CLAUSE_FROM, NULL, COLLECT_COL_TYPE_ALL, &pNodeList);
+    SNode *pNode = NULL;
+    FOREACH(pNode, pNodeList) {
+      SColumnNode *pCol = (SColumnNode *)pNode;
+
+      if (pCol->tableId == suid) {
+        sdbRelease(pSdb, pTopic);
+        nodesDestroyNode(pAst);
+        return -1;
+      } else {
+        goto NEXT;
+      }
+    }
+  NEXT:
+    sdbRelease(pSdb, pTopic);
+    nodesDestroyNode(pAst);
+  }
+  return 0;
+}
+
+static int32_t mndCheckDropStbForStream(SMnode *pMnode, const char *stbFullName, int64_t suid) {
+  SSdb *pSdb = pMnode->pSdb;
+  void *pIter = NULL;
+  while (1) {
+    SStreamObj *pStream = NULL;
+    pIter = sdbFetch(pSdb, SDB_STREAM, pIter, (void **)&pStream);
+    if (pIter == NULL) break;
+
+    if (pStream->smaId != 0) {
+      sdbRelease(pSdb, pStream);
+      continue;
+    }
+
+    if (pStream->targetStbUid == suid) {
+      sdbRelease(pSdb, pStream);
+      return -1;
+    }
+
+    SNode *pAst = NULL;
+    if (nodesStringToNode(pStream->ast, &pAst) != 0) {
+      ASSERT(0);
+      return -1;
+    }
+
+    SNodeList *pNodeList = NULL;
+    nodesCollectColumns((SSelectStmt *)pAst, SQL_CLAUSE_FROM, NULL, COLLECT_COL_TYPE_ALL, &pNodeList);
+    SNode *pNode = NULL;
+    FOREACH(pNode, pNodeList) {
+      SColumnNode *pCol = (SColumnNode *)pNode;
+
+      if (pCol->tableId == suid) {
+        sdbRelease(pSdb, pStream);
+        nodesDestroyNode(pAst);
+        return -1;
+      } else {
+        goto NEXT;
+      }
+    }
+  NEXT:
+    sdbRelease(pSdb, pStream);
+    nodesDestroyNode(pAst);
+  }
+  return 0;
+}
+
 static int32_t mndProcessDropStbReq(SRpcMsg *pReq) {
   SMnode      *pMnode = pReq->info.node;
   int32_t      code = -1;
@@ -1971,6 +2122,16 @@ static int32_t mndProcessDropStbReq(SRpcMsg *pReq) {
     goto _OVER;
   }
 
+  if (mndCheckDropStbForTopic(pMnode, dropReq.name, pStb->uid) < 0) {
+    terrno = TSDB_CODE_MND_TOPIC_MUST_BE_DELETED;
+    goto _OVER;
+  }
+
+  if (mndCheckDropStbForStream(pMnode, dropReq.name, pStb->uid) < 0) {
+    terrno = TSDB_CODE_MND_STREAM_MUST_BE_DELETED;
+    goto _OVER;
+  }
+
   code = mndDropStb(pMnode, pReq, pDb, pStb);
   if (code == 0) code = TSDB_CODE_ACTION_IN_PROGRESS;
 

source/dnode/vnode/inc/vnode.h

@@ -91,8 +91,9 @@ typedef struct SMetaEntry  SMetaEntry;
 void        metaReaderInit(SMetaReader *pReader, SMeta *pMeta, int32_t flags);
 void        metaReaderClear(SMetaReader *pReader);
 int32_t     metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid);
+int32_t     metaGetTableTags(SMeta *pMeta, uint64_t suid, SArray *uidList, SHashObj *tags);
 int32_t     metaReadNext(SMetaReader *pReader);
-const void *metaGetTableTagVal(SMetaEntry *pEntry, int16_t type, STagVal *tagVal);
+const void *metaGetTableTagVal(void *tag, int16_t type, STagVal *tagVal);
 int         metaGetTableNameByUid(void *meta, uint64_t uid, char *tbName);
 
 typedef struct SMetaFltParam {

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

@@ -87,7 +87,7 @@ int metaOpen(SVnode *pVnode, SMeta **ppMeta) {
   }
 
   // open pCtbIdx
-  ret = tdbTbOpen("ctb.idx", sizeof(SCtbIdxKey), 0, ctbIdxKeyCmpr, pMeta->pEnv, &pMeta->pCtbIdx);
+  ret = tdbTbOpen("ctb.idx", sizeof(SCtbIdxKey), -1, ctbIdxKeyCmpr, pMeta->pEnv, &pMeta->pCtbIdx);
   if (ret < 0) {
     metaError("vgId:%d, failed to open meta child table index since %s", TD_VID(pVnode), tstrerror(terrno));
     goto _err;

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

@@ -53,6 +53,80 @@ _err:
   return -1;
 }
 
+// int metaGetTableEntryByUidTest(void* meta, SArray *uidList) {
+//
+//   SArray* readerList = taosArrayInit(taosArrayGetSize(uidList), sizeof(SMetaReader));
+//   SArray* uidVersion = taosArrayInit(taosArrayGetSize(uidList), sizeof(STbDbKey));
+//   SMeta  *pMeta = meta;
+//   int64_t version;
+//   SHashObj *uHash = taosHashInit(32, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), false, HASH_NO_LOCK);
+//
+//   int64_t stt1 = taosGetTimestampUs();
+//   for(int i = 0; i < taosArrayGetSize(uidList); i++) {
+//     void* ppVal = NULL;
+//     int vlen = 0;
+//     uint64_t *  uid = taosArrayGet(uidList, i);
+//     // query uid.idx
+//     if (tdbTbGet(pMeta->pUidIdx, uid, sizeof(*uid), &ppVal, &vlen) < 0) {
+//       continue;
+//     }
+//     version = *(int64_t *)ppVal;
+//
+//     STbDbKey tbDbKey = {.version = version, .uid = *uid};
+//     taosArrayPush(uidVersion, &tbDbKey);
+//     taosHashPut(uHash, uid, sizeof(int64_t), ppVal, sizeof(int64_t));
+//   }
+//   int64_t stt2 = taosGetTimestampUs();
+//   qDebug("metaGetTableEntryByUidTest1 rows:%d, cost:%ld us", taosArrayGetSize(uidList), stt2-stt1);
+//
+//   TBC        *pCur = NULL;
+//   tdbTbcOpen(pMeta->pTbDb, &pCur, NULL);
+//   tdbTbcMoveToFirst(pCur);
+//   void *pKey = NULL;
+//   int   kLen = 0;
+//
+//   while(1){
+//     SMetaReader pReader = {0};
+//     int32_t ret = tdbTbcNext(pCur, &pKey, &kLen, &pReader.pBuf, &pReader.szBuf);
+//     if (ret < 0) break;
+//     STbDbKey *tmp = (STbDbKey*)pKey;
+//     int64_t *ver = (int64_t*)taosHashGet(uHash, &tmp->uid, sizeof(int64_t));
+//     if(ver == NULL || *ver != tmp->version) continue;
+//     taosArrayPush(readerList, &pReader);
+//   }
+//   tdbTbcClose(pCur);
+//
+//   taosArrayClear(readerList);
+//   int64_t stt3 = taosGetTimestampUs();
+//   qDebug("metaGetTableEntryByUidTest2 rows:%d, cost:%ld us", taosArrayGetSize(uidList), stt3-stt2);
+//   for(int i = 0; i < taosArrayGetSize(uidVersion); i++) {
+//     SMetaReader pReader = {0};
+//
+//     STbDbKey *tbDbKey = taosArrayGet(uidVersion, i);
+//     // query table.db
+//     if (tdbTbGet(pMeta->pTbDb, tbDbKey, sizeof(STbDbKey), &pReader.pBuf, &pReader.szBuf) < 0) {
+//       continue;
+//     }
+//     taosArrayPush(readerList, &pReader);
+//   }
+//   int64_t stt4 = taosGetTimestampUs();
+//   qDebug("metaGetTableEntryByUidTest3 rows:%d, cost:%ld us", taosArrayGetSize(uidList), stt4-stt3);
+//
+//   for(int i = 0; i < taosArrayGetSize(readerList); i++){
+//     SMetaReader* pReader  = taosArrayGet(readerList, i);
+//     metaReaderInit(pReader, meta, 0);
+//     // decode the entry
+//     tDecoderInit(&pReader->coder, pReader->pBuf, pReader->szBuf);
+//
+//     if (metaDecodeEntry(&pReader->coder, &pReader->me) < 0) {
+//     }
+//     metaReaderClear(pReader);
+//   }
+//   int64_t stt5 = taosGetTimestampUs();
+//   qDebug("metaGetTableEntryByUidTest4 rows:%d, cost:%ld us", taosArrayGetSize(readerList), stt5-stt4);
+//   return 0;
+// }
+
 int metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid) {
   SMeta  *pMeta = pReader->pMeta;
   int64_t version;
@@ -749,9 +823,8 @@ SArray *metaGetSmaTbUids(SMeta *pMeta) {
 
 #endif
 
-const void *metaGetTableTagVal(SMetaEntry *pEntry, int16_t type, STagVal *val) {
+const void *metaGetTableTagVal(void *pTag, int16_t type, STagVal *val) {
-  ASSERT(pEntry->type == TSDB_CHILD_TABLE);
+  STag *tag = (STag *)pTag;
-  STag *tag = (STag *)pEntry->ctbEntry.pTags;
   if (type == TSDB_DATA_TYPE_JSON) {
     return tag;
   }
@@ -853,6 +926,9 @@ int32_t metaFilterTableIds(SMeta *pMeta, SMetaFltParam *param, SArray *pUids) {
         break;
       }
     }
+    if (p->suid != pKey->suid) {
+      break;
+    }
     first = false;
     if (p != NULL) {
       int32_t cmp = (*param->filterFunc)(p->data, pKey->data, pKey->type);
@@ -889,6 +965,38 @@ END:
   return ret;
 }
 
+int32_t metaGetTableTags(SMeta *pMeta, uint64_t suid, SArray *uidList, SHashObj *tags) {
+  SMCtbCursor *pCur = metaOpenCtbCursor(pMeta, suid);
+
+  SHashObj *uHash = NULL;
+  size_t    len = taosArrayGetSize(uidList);  // len > 0 means there already have uids
+  if (len > 0) {
+    uHash = taosHashInit(32, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), false, HASH_NO_LOCK);
+    for (int i = 0; i < len; i++) {
+      int64_t *uid = taosArrayGet(uidList, i);
+      taosHashPut(uHash, uid, sizeof(int64_t), &i, sizeof(i));
+    }
+  }
+  while (1) {
+    tb_uid_t id = metaCtbCursorNext(pCur);
+    if (id == 0) {
+      break;
+    }
+
+    if (len > 0 && taosHashGet(uHash, &id, sizeof(int64_t)) == NULL) {
+      continue;
+    } else if (len == 0) {
+      taosArrayPush(uidList, &id);
+    }
+
+    taosHashPut(tags, &id, sizeof(int64_t), pCur->pVal, pCur->vLen);
+  }
+
+  taosHashCleanup(uHash);
+  metaCloseCtbCursor(pCur);
+  return TSDB_CODE_SUCCESS;
+}
+
 int32_t metaGetInfo(SMeta *pMeta, int64_t uid, SMetaInfo *pInfo) {
   int32_t code = 0;
   void   *pData = NULL;

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

@@ -181,10 +181,28 @@ int metaDelJsonVarFromIdx(SMeta *pMeta, const SMetaEntry *pCtbEntry, const SSche
 
 int metaCreateSTable(SMeta *pMeta, int64_t version, SVCreateStbReq *pReq) {
   SMetaEntry  me = {0};
+  int         kLen = 0;
+  int         vLen = 0;
+  const void *pKey = NULL;
+  const void *pVal = NULL;
+  void       *pBuf = NULL;
+  int32_t     szBuf = 0;
+  void       *p = NULL;
 
   // validate req
-  if (tdbTbGet(pMeta->pNameIdx, pReq->name, strlen(pReq->name), NULL, NULL) == 0) {
+  void *pData = NULL;
+  int   nData = 0;
+  if (tdbTbGet(pMeta->pNameIdx, pReq->name, strlen(pReq->name) + 1, &pData, &nData) == 0) {
+    tb_uid_t uid = *(tb_uid_t *)pData;
+    tdbFree(pData);
+    SMetaInfo info;
+    metaGetInfo(pMeta, uid, &info);
+    if (info.uid == info.suid) {
       return 0;
+    } else {
+      terrno = TSDB_CODE_TDB_TABLE_ALREADY_EXIST;
+      return -1;
+    }
   }
 
   // set structs
@@ -865,6 +883,9 @@ static int metaUpdateTableTagVal(SMeta *pMeta, int64_t version, SVAlterTbReq *pA
     metaUpdateTagIdx(pMeta, &ctbEntry);
   }
 
+  SCtbIdxKey ctbIdxKey = {.suid = ctbEntry.ctbEntry.suid, .uid = uid};
+  tdbTbUpsert(pMeta->pCtbIdx, &ctbIdxKey, sizeof(ctbIdxKey), ctbEntry.ctbEntry.pTags, ((STag*)(ctbEntry.ctbEntry.pTags))->len, &pMeta->txn);
+
   tDecoderClear(&dc1);
   tDecoderClear(&dc2);
   if (ctbEntry.ctbEntry.pTags) taosMemoryFree((void *)ctbEntry.ctbEntry.pTags);
@@ -1069,7 +1090,8 @@ static int metaUpdateTtlIdx(SMeta *pMeta, const SMetaEntry *pME) {
 
 static int metaUpdateCtbIdx(SMeta *pMeta, const SMetaEntry *pME) {
   SCtbIdxKey ctbIdxKey = {.suid = pME->ctbEntry.suid, .uid = pME->uid};
-  return tdbTbInsert(pMeta->pCtbIdx, &ctbIdxKey, sizeof(ctbIdxKey), NULL, 0, &pMeta->txn);
+
+  return tdbTbInsert(pMeta->pCtbIdx, &ctbIdxKey, sizeof(ctbIdxKey), pME->ctbEntry.pTags, ((STag*)(pME->ctbEntry.pTags))->len, &pMeta->txn);
 }
 
 int metaCreateTagIdxKey(tb_uid_t suid, int32_t cid, const void *pTagData, int32_t nTagData, int8_t type, tb_uid_t uid,

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

@@ -628,8 +628,6 @@ int32_t tqProcessVgChangeReq(STQ* pTq, int64_t version, char* msg, int32_t msgLe
 }
 
 int32_t tqExpandTask(STQ* pTq, SStreamTask* pTask) {
-  int32_t code = 0;
-
   if (pTask->taskLevel == TASK_LEVEL__AGG) {
     ASSERT(taosArrayGetSize(pTask->childEpInfo) != 0);
   }
@@ -640,8 +638,7 @@ int32_t tqExpandTask(STQ* pTq, SStreamTask* pTask) {
   pTask->outputQueue = streamQueueOpen();
 
   if (pTask->inputQueue == NULL || pTask->outputQueue == NULL) {
-    code = -1;
+    return -1;
-    goto FAIL;
   }
 
   pTask->inputStatus = TASK_INPUT_STATUS__NORMAL;
@@ -686,14 +683,9 @@ int32_t tqExpandTask(STQ* pTq, SStreamTask* pTask) {
 
   streamSetupTrigger(pTask);
 
-  tqInfo("deploy stream task on vg %d, task id %d, child id %d", TD_VID(pTq->pVnode), pTask->taskId,
+  tqInfo("expand stream task on vg %d, task id %d, child id %d", TD_VID(pTq->pVnode), pTask->taskId,
          pTask->selfChildId);
-
+  return 0;
-FAIL:
-  if (pTask->inputQueue) streamQueueClose(pTask->inputQueue);
-  if (pTask->outputQueue) streamQueueClose(pTask->outputQueue);
-  // TODO free executor
-  return code;
 }
 
 int32_t tqProcessTaskDeployReq(STQ* pTq, int64_t version, char* msg, int32_t msgLen) {

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

@@ -231,11 +231,12 @@ void tqTableSink(SStreamTask* pTask, void* vnode, int64_t ver, void* data) {
 
   ASSERT(pTask->tbSink.pTSchema);
   deleteReq.deleteReqs = taosArrayInit(0, sizeof(SSingleDeleteReq));
-  SSubmitReq* pReq = tqBlockToSubmit(pVnode, pRes, pTask->tbSink.pTSchema, true, pTask->tbSink.stbUid,
+  SSubmitReq* submitReq = tqBlockToSubmit(pVnode, pRes, pTask->tbSink.pTSchema, true, pTask->tbSink.stbUid,
                                           pTask->tbSink.stbFullName, &deleteReq);
 
   tqDebug("vgId:%d, task %d convert blocks over, put into write-queue", TD_VID(pVnode), pTask->taskId);
 
+  if (taosArrayGetSize(deleteReq.deleteReqs) != 0) {
     int32_t code;
     int32_t len;
     tEncodeSize(tEncodeSBatchDeleteReq, &deleteReq, len, code);
@@ -244,21 +245,21 @@ void tqTableSink(SStreamTask* pTask, void* vnode, int64_t ver, void* data) {
       ASSERT(0);
     }
     SEncoder encoder;
-  void*    buf = rpcMallocCont(len + sizeof(SMsgHead));
+    void*    serializedDeleteReq = rpcMallocCont(len + sizeof(SMsgHead));
-  void*    abuf = POINTER_SHIFT(buf, sizeof(SMsgHead));
+    void*    abuf = POINTER_SHIFT(serializedDeleteReq, sizeof(SMsgHead));
     tEncoderInit(&encoder, abuf, len);
     tEncodeSBatchDeleteReq(&encoder, &deleteReq);
     tEncoderClear(&encoder);
 
-  ((SMsgHead*)buf)->vgId = pVnode->config.vgId;
+    ((SMsgHead*)serializedDeleteReq)->vgId = pVnode->config.vgId;
 
-  if (taosArrayGetSize(deleteReq.deleteReqs) != 0) {
     SRpcMsg msg = {
         .msgType = TDMT_VND_BATCH_DEL,
-        .pCont = buf,
+        .pCont = serializedDeleteReq,
         .contLen = len + sizeof(SMsgHead),
     };
     if (tmsgPutToQueue(&pVnode->msgCb, WRITE_QUEUE, &msg) != 0) {
+      rpcFreeCont(serializedDeleteReq);
       tqDebug("failed to put into write-queue since %s", terrstr());
     }
   }
@@ -268,11 +269,12 @@ void tqTableSink(SStreamTask* pTask, void* vnode, int64_t ver, void* data) {
   // build write msg
   SRpcMsg msg = {
       .msgType = TDMT_VND_SUBMIT,
-      .pCont = pReq,
+      .pCont = submitReq,
-      .contLen = ntohl(pReq->length),
+      .contLen = ntohl(submitReq->length),
   };
 
   if (tmsgPutToQueue(&pVnode->msgCb, WRITE_QUEUE, &msg) != 0) {
+    rpcFreeCont(submitReq);
     tqDebug("failed to put into write-queue since %s", terrstr());
   }
 }

source/libs/executor/src/executil.c

@@ -221,7 +221,7 @@ EDealRes doTranslateTagExpr(SNode** pNode, void* pContext) {
 
     STagVal tagVal = {0};
     tagVal.cid = pSColumnNode->colId;
-    const char* p = metaGetTableTagVal(&mr->me, pSColumnNode->node.resType.type, &tagVal);
+    const char* p = metaGetTableTagVal(mr->me.ctbEntry.pTags, pSColumnNode->node.resType.type, &tagVal);
     if (p == NULL) {
       res->node.resType.type = TSDB_DATA_TYPE_NULL;
     } else if (pSColumnNode->node.resType.type == TSDB_DATA_TYPE_JSON) {
@@ -298,6 +298,209 @@ int32_t isQualifiedTable(STableKeyInfo* info, SNode* pTagCond, void* metaHandle,
   return TSDB_CODE_SUCCESS;
 }
 
+typedef struct tagFilterAssist{
+  SHashObj *colHash;
+  int32_t   index;
+  SArray   *cInfoList;
+}tagFilterAssist;
+
+static EDealRes getColumn(SNode** pNode, void* pContext) {
+  SColumnNode* pSColumnNode = NULL;
+  if (QUERY_NODE_COLUMN == nodeType((*pNode))) {
+    pSColumnNode = *(SColumnNode**)pNode;
+  }else if(QUERY_NODE_FUNCTION == nodeType((*pNode))){
+    SFunctionNode* pFuncNode = *(SFunctionNode**)(pNode);
+    if (pFuncNode->funcType == FUNCTION_TYPE_TBNAME) {
+      pSColumnNode = (SColumnNode*)nodesMakeNode(QUERY_NODE_COLUMN);
+      if (NULL == pSColumnNode) {
+        return DEAL_RES_ERROR;
+      }
+      pSColumnNode->colId = -1;
+      pSColumnNode->colType = COLUMN_TYPE_TBNAME;
+      pSColumnNode->node.resType.type = TSDB_DATA_TYPE_VARCHAR;
+      pSColumnNode->node.resType.bytes = TSDB_TABLE_FNAME_LEN - 1 + VARSTR_HEADER_SIZE;
+      nodesDestroyNode(*pNode);
+      *pNode = (SNode*)pSColumnNode;
+    }else{
+      return DEAL_RES_CONTINUE;
+    }
+  }else{
+    return DEAL_RES_CONTINUE;
+  }
+
+  tagFilterAssist *pData = (tagFilterAssist *)pContext;
+  void *data = taosHashGet(pData->colHash, &pSColumnNode->colId, sizeof(pSColumnNode->colId));
+  if(!data){
+    taosHashPut(pData->colHash, &pSColumnNode->colId, sizeof(pSColumnNode->colId), pNode, sizeof((*pNode)));
+    pSColumnNode->slotId = pData->index++;
+    SColumnInfo cInfo = {.colId = pSColumnNode->colId, .type = pSColumnNode->node.resType.type, .bytes = pSColumnNode->node.resType.bytes};
+#if TAG_FILTER_DEBUG
+    qDebug("tagfilter build column info, slotId:%d, colId:%d, type:%d", pSColumnNode->slotId, cInfo.colId, cInfo.type);
+#endif
+    taosArrayPush(pData->cInfoList, &cInfo);
+  }else{
+    SColumnNode* col = *(SColumnNode**)data;
+    pSColumnNode->slotId = col->slotId;
+  }
+
+  return DEAL_RES_CONTINUE;
+}
+
+static int32_t createResultData(SDataType* pType, int32_t numOfRows, SScalarParam* pParam) {
+  SColumnInfoData* pColumnData = taosMemoryCalloc(1, sizeof(SColumnInfoData));
+  if (pColumnData == NULL) {
+    terrno = TSDB_CODE_OUT_OF_MEMORY;
+    return terrno;
+  }
+
+  pColumnData->info.type      = pType->type;
+  pColumnData->info.bytes     = pType->bytes;
+  pColumnData->info.scale     = pType->scale;
+  pColumnData->info.precision = pType->precision;
+
+  int32_t code = colInfoDataEnsureCapacity(pColumnData, numOfRows);
+  if (code != TSDB_CODE_SUCCESS) {
+    terrno = TSDB_CODE_OUT_OF_MEMORY;
+    taosMemoryFree(pColumnData);
+    return terrno;
+  }
+
+  pParam->columnData = pColumnData;
+  pParam->colAlloced = true;
+  return TSDB_CODE_SUCCESS;
+}
+
+static SColumnInfoData* getColInfoResult(void* metaHandle, uint64_t suid, SArray* uidList, SNode* pTagCond){
+  int32_t code = TSDB_CODE_SUCCESS;
+  SArray* pBlockList = NULL;
+  SSDataBlock* pResBlock = NULL;
+  SHashObj * tags = NULL;
+  SScalarParam output = {0};
+
+  tagFilterAssist ctx = {0};
+  ctx.colHash = taosHashInit(4, taosGetDefaultHashFunction(TSDB_DATA_TYPE_SMALLINT), false, HASH_NO_LOCK);
+  if(ctx.colHash == NULL){
+    terrno = TSDB_CODE_OUT_OF_MEMORY;
+    goto end;
+  }
+  ctx.index = 0;
+  ctx.cInfoList = taosArrayInit(4, sizeof(SColumnInfo));
+  if(ctx.cInfoList == NULL){
+    terrno = TSDB_CODE_OUT_OF_MEMORY;
+    goto end;
+  }
+
+  nodesRewriteExprPostOrder(&pTagCond, getColumn, (void *)&ctx);
+
+  pResBlock = createDataBlock();
+  if (pResBlock == NULL) {
+    terrno = TSDB_CODE_OUT_OF_MEMORY;
+    goto end;
+  }
+
+  for (int32_t i = 0; i < taosArrayGetSize(ctx.cInfoList); ++i) {
+    SColumnInfoData colInfo = {{0}, 0};
+    colInfo.info = *(SColumnInfo*)taosArrayGet(ctx.cInfoList, i);
+    blockDataAppendColInfo(pResBlock, &colInfo);
+  }
+
+//  int64_t stt = taosGetTimestampUs();
+  tags = taosHashInit(32, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), false, HASH_NO_LOCK);
+  code = metaGetTableTags(metaHandle, suid, uidList, tags);
+  if (code != TSDB_CODE_SUCCESS) {
+    terrno = code;
+    goto end;
+  }
+
+  int32_t rows = taosArrayGetSize(uidList);
+  if(rows == 0){
+    goto end;
+  }
+//  int64_t stt1 = taosGetTimestampUs();
+//  qDebug("generate tag meta rows:%d, cost:%ld us", rows, stt1-stt);
+
+  code = blockDataEnsureCapacity(pResBlock, rows);
+  if (code != TSDB_CODE_SUCCESS) {
+    terrno = code;
+    goto end;
+  }
+
+//  int64_t st = taosGetTimestampUs();
+  for (int32_t i = 0; i < rows; i++) {
+    int64_t* uid = taosArrayGet(uidList, i);
+    void* tag = taosHashGet(tags, uid, sizeof(int64_t));
+    ASSERT(tag);
+    for(int32_t j = 0; j < taosArrayGetSize(pResBlock->pDataBlock); j++){
+      SColumnInfoData* pColInfo = (SColumnInfoData*)taosArrayGet(pResBlock->pDataBlock, j);
+
+      if(pColInfo->info.colId == -1){     // tbname
+        char str[TSDB_TABLE_FNAME_LEN + VARSTR_HEADER_SIZE] = {0};
+        metaGetTableNameByUid(metaHandle, *uid, str);
+        colDataAppend(pColInfo, i, str, false);
+#if TAG_FILTER_DEBUG
+        qDebug("tagfilter uid:%ld, tbname:%s", *uid, str+2);
+#endif
+      }else{
+        STagVal tagVal = {0};
+        tagVal.cid = pColInfo->info.colId;
+        const char* p = metaGetTableTagVal(tag, pColInfo->info.type, &tagVal);
+
+        if (p == NULL || (pColInfo->info.type == TSDB_DATA_TYPE_JSON && ((STag*)p)->nTag == 0)){
+          colDataAppend(pColInfo, i, p, true);
+        } else if (pColInfo->info.type == TSDB_DATA_TYPE_JSON) {
+          colDataAppend(pColInfo, i, p, false);
+        } else if (IS_VAR_DATA_TYPE(pColInfo->info.type)) {
+          char *tmp = taosMemoryCalloc(tagVal.nData + VARSTR_HEADER_SIZE + 1, 1);
+          varDataSetLen(tmp, tagVal.nData);
+          memcpy(tmp + VARSTR_HEADER_SIZE, tagVal.pData, tagVal.nData);
+          colDataAppend(pColInfo, i, tmp, false);
+#if TAG_FILTER_DEBUG
+          qDebug("tagfilter varch:%s", tmp+2);
+#endif
+          taosMemoryFree(tmp);
+        } else {
+          colDataAppend(pColInfo, i, (const char*)&tagVal.i64, false);
+#if TAG_FILTER_DEBUG
+          if(pColInfo->info.type == TSDB_DATA_TYPE_INT){
+            qDebug("tagfilter int:%d", *(int*)(&tagVal.i64));
+          }else if(pColInfo->info.type == TSDB_DATA_TYPE_DOUBLE){
+            qDebug("tagfilter double:%f", *(double *)(&tagVal.i64));
+          }
+#endif
+        }
+      }
+    }
+  }
+  pResBlock->info.rows = rows;
+
+//  int64_t st1 = taosGetTimestampUs();
+//  qDebug("generate tag block rows:%d, cost:%ld us", rows, st1-st);
+
+  pBlockList = taosArrayInit(2, POINTER_BYTES);
+  taosArrayPush(pBlockList, &pResBlock);
+
+  SDataType type = {.type = TSDB_DATA_TYPE_BOOL, .bytes = sizeof(bool)};
+  code = createResultData(&type, rows, &output);
+  if (code != TSDB_CODE_SUCCESS) {
+    goto end;
+  }
+
+  code = scalarCalculate(pTagCond, pBlockList, &output);
+  if(code != TSDB_CODE_SUCCESS){
+    terrno = code;
+  }
+//  int64_t st2 = taosGetTimestampUs();
+//  qDebug("calculate tag block rows:%d, cost:%ld us", rows, st2-st1);
+
+end:
+  taosHashCleanup(tags);
+  taosHashCleanup(ctx.colHash);
+  taosArrayDestroy(ctx.cInfoList);
+  blockDataDestroy(pResBlock);
+  taosArrayDestroy(pBlockList);
+  return output.columnData;
+}
+
 int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode, SNode* pTagCond, SNode* pTagIndexCond,
                      STableListInfo* pListInfo) {
   int32_t code = TSDB_CODE_SUCCESS;
@@ -308,62 +511,67 @@ int32_t getTableList(void* metaHandle, void* pVnode, SScanPhysiNode* pScanNode,
   }
 
   uint64_t tableUid = pScanNode->uid;
-
   pListInfo->suid = pScanNode->suid;
+  SArray* res = taosArrayInit(8, sizeof(uint64_t));
 
   if (pScanNode->tableType == TSDB_SUPER_TABLE) {
     if (pTagIndexCond) {
       SIndexMetaArg metaArg = {
           .metaEx = metaHandle, .idx = tsdbGetIdx(metaHandle), .ivtIdx = tsdbGetIvtIdx(metaHandle), .suid = tableUid};
 
-      SArray*       res = taosArrayInit(8, sizeof(uint64_t));
+//      int64_t stt = taosGetTimestampUs();
       SIdxFltStatus status = SFLT_NOT_INDEX;
       code = doFilterTag(pTagIndexCond, &metaArg, res, &status);
       if (code != 0 || status == SFLT_NOT_INDEX) {
         qError("failed to get tableIds from index, reason:%s, suid:%" PRIu64, tstrerror(code), tableUid);
-        //        code = TSDB_CODE_INDEX_REBUILDING;
+        code = TDB_CODE_SUCCESS;
-        code = vnodeGetAllTableList(pVnode, tableUid, pListInfo->pTableList);
+      }
-      } else {
+
-        qDebug("success to get tableIds, size:%d, suid:%" PRIu64, (int)taosArrayGetSize(res), tableUid);
+//      int64_t stt1 = taosGetTimestampUs();
+//      qDebug("generate table list, cost:%ld us", stt1-stt);
+    }else if(!pTagCond){
+      vnodeGetCtbIdList(pVnode, pScanNode->suid, res);
+    }
+  } else {  // Create one table group.
+    taosArrayPush(res, &tableUid);
+  }
+
+  if (pTagCond) {
+    SColumnInfoData* pColInfoData = getColInfoResult(metaHandle, pListInfo->suid, res, pTagCond);
+    if(terrno != TDB_CODE_SUCCESS){
+      colDataDestroy(pColInfoData);
+      taosMemoryFreeClear(pColInfoData);
+      taosArrayDestroy(res);
+      return terrno;
+    }
+
+    int32_t i = 0;
+    int32_t j = 0;
+    int32_t len = taosArrayGetSize(res);
+    while (i < taosArrayGetSize(res) && j < len && pColInfoData) {
+      void* var = POINTER_SHIFT(pColInfoData->pData, j * pColInfoData->info.bytes);
+
+      int64_t* uid = taosArrayGet(res, i);
+      qDebug("tagfilter get uid:%ld, res:%d", *uid, *(bool*)var);
+      if (*(bool*)var == false) {
+        taosArrayRemove(res, i);
+        j++;
+        continue;
+      }
+      i++;
+      j++;
+    }
+    colDataDestroy(pColInfoData);
+    taosMemoryFreeClear(pColInfoData);
   }
 
   for (int i = 0; i < taosArrayGetSize(res); i++) {
     STableKeyInfo info = {.uid = *(uint64_t*)taosArrayGet(res, i), .groupId = 0};
     taosArrayPush(pListInfo->pTableList, &info);
+    qDebug("tagfilter get uid:%ld", info.uid);
   }
+
   taosArrayDestroy(res);
-    } else {
-      code = vnodeGetAllTableList(pVnode, tableUid, pListInfo->pTableList);
-    }
-
-    if (code != TSDB_CODE_SUCCESS) {
-      qError("failed to get tableIds, reason:%s, suid:%" PRIu64, tstrerror(code), tableUid);
-      terrno = code;
-      return code;
-    }
-  } else {  // Create one table group.
-    STableKeyInfo info = {.uid = tableUid, .groupId = 0};
-    taosArrayPush(pListInfo->pTableList, &info);
-  }
-
-  if (pTagCond) {
-    int32_t i = 0;
-    while (i < taosArrayGetSize(pListInfo->pTableList)) {
-      STableKeyInfo* info = taosArrayGet(pListInfo->pTableList, i);
-
-      bool qualified = true;
-      code = isQualifiedTable(info, pTagCond, metaHandle, &qualified);
-      if (code != TSDB_CODE_SUCCESS) {
-        return code;
-      }
-
-      if (!qualified) {
-        taosArrayRemove(pListInfo->pTableList, i);
-        continue;
-      }
-      i++;
-    }
-  }
 
   pListInfo->pGroupList = taosArrayInit(4, POINTER_BYTES);
   if (pListInfo->pGroupList == NULL) {

source/libs/executor/src/scanoperator.c

@@ -440,7 +440,7 @@ int32_t addTagPseudoColumnData(SReadHandle* pHandle, SExprInfo* pPseudoExpr, int
     } else {  // these are tags
       STagVal tagVal = {0};
       tagVal.cid = pExpr->base.pParam[0].pCol->colId;
-      const char* p = metaGetTableTagVal(&mr.me, pColInfoData->info.type, &tagVal);
+      const char* p = metaGetTableTagVal(mr.me.ctbEntry.pTags, pColInfoData->info.type, &tagVal);
 
       char* data = NULL;
       if (pColInfoData->info.type != TSDB_DATA_TYPE_JSON && p != NULL) {
@@ -2506,7 +2506,7 @@ static SSDataBlock* doTagScan(SOperatorInfo* pOperator) {
       } else {  // it is a tag value
         STagVal val = {0};
         val.cid = pExprInfo[j].base.pParam[0].pCol->colId;
-        const char* p = metaGetTableTagVal(&mr.me, pDst->info.type, &val);
+        const char* p = metaGetTableTagVal(mr.me.ctbEntry.pTags, pDst->info.type, &val);
 
         char* data = NULL;
         if (pDst->info.type != TSDB_DATA_TYPE_JSON && p != NULL) {

source/libs/scalar/src/scalar.c

@@ -292,6 +292,9 @@ int32_t sclInitParam(SNode* node, SScalarParam *param, SScalarCtx *ctx, int32_t
       }
 
       SColumnInfoData *columnData = (SColumnInfoData *)taosArrayGet(block->pDataBlock, ref->slotId);
+#if TAG_FILTER_DEBUG
+      qDebug("tagfilter column info, slotId:%d, colId:%d, type:%d", ref->slotId, columnData->info.colId, columnData->info.type);
+#endif
       param->numOfRows = block->info.rows;
       param->columnData = columnData;
       break;

source/libs/stream/inc/streamInc.h

@@ -32,7 +32,6 @@ typedef struct {
 
 static SStreamGlobalEnv streamEnv;
 
-int32_t streamExec(SStreamTask* pTask);
 int32_t streamPipelineExec(SStreamTask* pTask, int32_t batchNum, bool dispatch);
 
 int32_t streamDispatch(SStreamTask* pTask);

source/libs/stream/src/stream.c

@@ -185,7 +185,9 @@ int32_t streamProcessDispatchReq(SStreamTask* pTask, SStreamDispatchReq* pReq, S
   tFreeStreamDispatchReq(pReq);
 
   if (exec) {
-    streamTryExec(pTask);
+    if (streamTryExec(pTask) < 0) {
+      return -1;
+    }
 
     if (pTask->outputType == TASK_OUTPUT__FIXED_DISPATCH || pTask->outputType == TASK_OUTPUT__SHUFFLE_DISPATCH) {
       streamDispatch(pTask);
@@ -221,7 +223,9 @@ int32_t streamProcessDispatchRsp(SStreamTask* pTask, SStreamDispatchRsp* pRsp) {
 }
 
 int32_t streamProcessRunReq(SStreamTask* pTask) {
-  streamTryExec(pTask);
+  if (streamTryExec(pTask) < 0) {
+    return -1;
+  }
 
   if (pTask->outputType == TASK_OUTPUT__FIXED_DISPATCH || pTask->outputType == TASK_OUTPUT__SHUFFLE_DISPATCH) {
     streamDispatch(pTask);

source/libs/stream/src/streamMeta.c

@@ -15,6 +15,7 @@
 
 #include "executor.h"
 #include "tstream.h"
+#include "ttimer.h"
 
 SStreamMeta* streamMetaOpen(const char* path, void* ahandle, FTaskExpand expandFunc) {
   SStreamMeta* pMeta = taosMemoryCalloc(1, sizeof(SStreamMeta));
@@ -99,16 +100,19 @@ int32_t streamMetaAddSerializedTask(SStreamMeta* pMeta, int64_t startVer, char*
     goto FAIL;
   }
 
-  taosHashPut(pMeta->pTasks, &pTask->taskId, sizeof(int32_t), &pTask, sizeof(void*));
+  if (taosHashPut(pMeta->pTasks, &pTask->taskId, sizeof(int32_t), &pTask, sizeof(void*)) < 0) {
+    goto FAIL;
+  }
 
   if (tdbTbUpsert(pMeta->pTaskDb, &pTask->taskId, sizeof(int32_t), msg, msgLen, &pMeta->txn) < 0) {
+    taosHashRemove(pMeta->pTasks, &pTask->taskId, sizeof(int32_t));
     ASSERT(0);
-    return -1;
+    goto FAIL;
   }
   return 0;
 
 FAIL:
-  if (pTask) taosMemoryFree(pTask);
+  if (pTask) tFreeSStreamTask(pTask);
   return -1;
 }
 
@@ -158,11 +162,28 @@ int32_t streamMetaRemoveTask(SStreamMeta* pMeta, int32_t taskId) {
     SStreamTask* pTask = *ppTask;
     taosHashRemove(pMeta->pTasks, &taskId, sizeof(int32_t));
     atomic_store_8(&pTask->taskStatus, TASK_STATUS__DROPPING);
-  }
 
     if (tdbTbDelete(pMeta->pTaskDb, &taskId, sizeof(int32_t), &pMeta->txn) < 0) {
       /*return -1;*/
     }
+
+    if (pTask->triggerParam != 0) {
+      taosTmrStop(pTask->timer);
+    }
+
+    while (1) {
+      int8_t schedStatus =
+          atomic_val_compare_exchange_8(&pTask->schedStatus, TASK_SCHED_STATUS__INACTIVE, TASK_SCHED_STATUS__DROPPING);
+      if (schedStatus == TASK_SCHED_STATUS__INACTIVE) {
+        tFreeSStreamTask(pTask);
+        break;
+      } else if (schedStatus == TASK_SCHED_STATUS__DROPPING) {
+        break;
+      }
+      taosMsleep(10);
+    }
+  }
+
   return 0;
 }
 

source/libs/stream/src/streamQueue.c

@@ -13,7 +13,7 @@
  * along with this program. If not, see <http://www.gnu.org/licenses/>.
  */
 
-#include "tstream.h"
+#include "streamInc.h"
 
 SStreamQueue* streamQueueOpen() {
   SStreamQueue* pQueue = taosMemoryCalloc(1, sizeof(SStreamQueue));
@@ -36,9 +36,12 @@ void streamQueueClose(SStreamQueue* queue) {
   while (1) {
     void* qItem = streamQueueNextItem(queue);
     if (qItem) {
-      taosFreeQitem(qItem);
+      streamFreeQitem(qItem);
     } else {
-      return;
+      break;
     }
   }
+  taosFreeQall(queue->qall);
+  taosCloseQueue(queue->queue);
+  taosMemoryFree(queue);
 }

source/libs/stream/src/streamTask.c

@@ -152,9 +152,17 @@ int32_t tDecodeSStreamTask(SDecoder* pDecoder, SStreamTask* pTask) {
 }
 
 void tFreeSStreamTask(SStreamTask* pTask) {
-  streamQueueClose(pTask->inputQueue);
+  if (pTask->inputQueue) streamQueueClose(pTask->inputQueue);
-  streamQueueClose(pTask->outputQueue);
+  if (pTask->outputQueue) streamQueueClose(pTask->outputQueue);
   if (pTask->exec.qmsg) taosMemoryFree(pTask->exec.qmsg);
   if (pTask->exec.executor) qDestroyTask(pTask->exec.executor);
+  taosArrayDestroy(pTask->childEpInfo);
+  if (pTask->outputType == TASK_OUTPUT__TABLE) {
+    tDeleteSSchemaWrapper(pTask->tbSink.pSchemaWrapper);
+    taosMemoryFree(pTask->tbSink.pTSchema);
+  }
+  if (pTask->outputType == TASK_OUTPUT__SHUFFLE_DISPATCH) {
+    taosArrayDestroy(pTask->shuffleDispatcher.dbInfo.pVgroupInfos);
+  }
   taosMemoryFree(pTask);
 }

source/libs/sync/inc/syncIndexMgr.h

@@ -31,6 +31,10 @@ typedef struct SSyncIndexMgr {
   SRaftId (*replicas)[TSDB_MAX_REPLICA];
   SyncIndex index[TSDB_MAX_REPLICA];
   SyncTerm  privateTerm[TSDB_MAX_REPLICA];  // for advanced function
+
+  int64_t startTimeArr[TSDB_MAX_REPLICA];
+  int64_t recvTimeArr[TSDB_MAX_REPLICA];
+
   int32_t    replicaNum;
   SSyncNode *pSyncNode;
 } SSyncIndexMgr;
@@ -44,6 +48,11 @@ SyncIndex      syncIndexMgrGetIndex(SSyncIndexMgr *pSyncIndexMgr, const SRaftId
 cJSON         *syncIndexMgr2Json(SSyncIndexMgr *pSyncIndexMgr);
 char          *syncIndexMgr2Str(SSyncIndexMgr *pSyncIndexMgr);
 
+void    syncIndexMgrSetStartTime(SSyncIndexMgr *pSyncIndexMgr, const SRaftId *pRaftId, int64_t startTime);
+int64_t syncIndexMgrGetStartTime(SSyncIndexMgr *pSyncIndexMgr, const SRaftId *pRaftId);
+void    syncIndexMgrSetRecvTime(SSyncIndexMgr *pSyncIndexMgr, const SRaftId *pRaftId, int64_t recvTime);
+int64_t syncIndexMgrGetRecvTime(SSyncIndexMgr *pSyncIndexMgr, const SRaftId *pRaftId);
+
 // void     syncIndexMgrSetTerm(SSyncIndexMgr *pSyncIndexMgr, const SRaftId *pRaftId, SyncTerm term);
 // SyncTerm syncIndexMgrGetTerm(SSyncIndexMgr *pSyncIndexMgr, const SRaftId *pRaftId);
 

source/libs/sync/inc/syncInt.h

@@ -269,6 +269,8 @@ int32_t syncNodeLeaderTransfer(SSyncNode* pSyncNode);
 int32_t syncNodeLeaderTransferTo(SSyncNode* pSyncNode, SNodeInfo newLeader);
 int32_t syncDoLeaderTransfer(SSyncNode* ths, SRpcMsg* pRpcMsg, SSyncRaftEntry* pEntry);
 
+int32_t syncNodeDynamicQuorum(const SSyncNode* pSyncNode);
+
 // trace log
 void syncLogSendRequestVote(SSyncNode* pSyncNode, const SyncRequestVote* pMsg, const char* s);
 void syncLogRecvRequestVote(SSyncNode* pSyncNode, const SyncRequestVote* pMsg, const char* s);

source/libs/sync/inc/syncReplication.h

@@ -55,6 +55,8 @@ int32_t syncNodeAppendEntriesPeers(SSyncNode* pSyncNode);
 int32_t syncNodeAppendEntriesPeersSnapshot(SSyncNode* pSyncNode);
 int32_t syncNodeAppendEntriesPeersSnapshot2(SSyncNode* pSyncNode);
 
+int32_t syncNodeAppendEntriesOnePeer(SSyncNode* pSyncNode, SRaftId* pDestId, SyncIndex nextIndex);
+
 int32_t syncNodeReplicate(SSyncNode* pSyncNode, bool isTimer);
 int32_t syncNodeAppendEntries(SSyncNode* pSyncNode, const SRaftId* destRaftId, const SyncAppendEntries* pMsg);
 int32_t syncNodeAppendEntriesBatch(SSyncNode* pSyncNode, const SRaftId* destRaftId, const SyncAppendEntriesBatch* pMsg);

source/libs/sync/src/syncAppendEntries.c

@@ -148,6 +148,7 @@ int32_t syncNodeOnAppendEntriesCb(SSyncNode* ths, SyncAppendEntries* pMsg) {
     pReply->term = ths->pRaftStore->currentTerm;
     pReply->success = false;
     pReply->matchIndex = SYNC_INDEX_INVALID;
+    pReply->startTime = ths->startTime;
 
     // msg event log
     syncLogSendAppendEntriesReply(ths, pReply, "");
@@ -290,6 +291,8 @@ int32_t syncNodeOnAppendEntriesCb(SSyncNode* ths, SyncAppendEntries* pMsg) {
       pReply->matchIndex = pMsg->prevLogIndex;
     }
 
+    pReply->startTime = ths->startTime;
+
     // msg event log
     syncLogSendAppendEntriesReply(ths, pReply, "");
 
@@ -603,6 +606,7 @@ int32_t syncNodeOnAppendEntriesSnapshot2Cb(SSyncNode* ths, SyncAppendEntriesBatc
       pReply->privateTerm = ths->pNewNodeReceiver->privateTerm;
       pReply->success = true;
       pReply->matchIndex = matchIndex;
+      pReply->startTime = ths->startTime;
 
       // msg event log
       syncLogSendAppendEntriesReply(ths, pReply, "");
@@ -651,6 +655,7 @@ int32_t syncNodeOnAppendEntriesSnapshot2Cb(SSyncNode* ths, SyncAppendEntriesBatc
       pReply->privateTerm = ths->pNewNodeReceiver->privateTerm;
       pReply->success = false;
       pReply->matchIndex = ths->commitIndex;
+      pReply->startTime = ths->startTime;
 
       // msg event log
       syncLogSendAppendEntriesReply(ths, pReply, "");
@@ -729,6 +734,7 @@ int32_t syncNodeOnAppendEntriesSnapshot2Cb(SSyncNode* ths, SyncAppendEntriesBatc
       pReply->privateTerm = ths->pNewNodeReceiver->privateTerm;
       pReply->success = true;
       pReply->matchIndex = hasAppendEntries ? pMsg->prevLogIndex + pMsg->dataCount : pMsg->prevLogIndex;
+      pReply->startTime = ths->startTime;
 
       // msg event log
       syncLogSendAppendEntriesReply(ths, pReply, "");
@@ -874,6 +880,7 @@ int32_t syncNodeOnAppendEntriesSnapshotCb(SSyncNode* ths, SyncAppendEntries* pMs
       pReply->privateTerm = ths->pNewNodeReceiver->privateTerm;
       pReply->success = true;
       pReply->matchIndex = matchIndex;
+      pReply->startTime = ths->startTime;
 
       // msg event log
       syncLogSendAppendEntriesReply(ths, pReply, "");
@@ -919,6 +926,7 @@ int32_t syncNodeOnAppendEntriesSnapshotCb(SSyncNode* ths, SyncAppendEntries* pMs
       pReply->privateTerm = ths->pNewNodeReceiver->privateTerm;
       pReply->success = false;
       pReply->matchIndex = SYNC_INDEX_INVALID;
+      pReply->startTime = ths->startTime;
 
       // msg event log
       syncLogSendAppendEntriesReply(ths, pReply, "");
@@ -984,6 +992,7 @@ int32_t syncNodeOnAppendEntriesSnapshotCb(SSyncNode* ths, SyncAppendEntries* pMs
       pReply->privateTerm = ths->pNewNodeReceiver->privateTerm;
       pReply->success = true;
       pReply->matchIndex = hasAppendEntries ? pMsg->prevLogIndex + 1 : pMsg->prevLogIndex;
+      pReply->startTime = ths->startTime;
 
       // msg event log
       syncLogSendAppendEntriesReply(ths, pReply, "");

source/libs/sync/src/syncAppendEntriesReply.c

@@ -64,6 +64,10 @@ int32_t syncNodeOnAppendEntriesReplyCb(SSyncNode* ths, SyncAppendEntriesReply* p
 
   ASSERT(pMsg->term == ths->pRaftStore->currentTerm);
 
+  // update time
+  syncIndexMgrSetStartTime(ths->pNextIndex, &(pMsg->srcId), pMsg->startTime);
+  syncIndexMgrSetRecvTime(ths->pNextIndex, &(pMsg->srcId), taosGetTimestampMs());
+
   SyncIndex beforeNextIndex = syncIndexMgrGetIndex(ths->pNextIndex, &(pMsg->srcId));
   SyncIndex beforeMatchIndex = syncIndexMgrGetIndex(ths->pMatchIndex, &(pMsg->srcId));
 
@@ -170,6 +174,10 @@ int32_t syncNodeOnAppendEntriesReplySnapshot2Cb(SSyncNode* ths, SyncAppendEntrie
 
   ASSERT(pMsg->term == ths->pRaftStore->currentTerm);
 
+  // update time
+  syncIndexMgrSetStartTime(ths->pNextIndex, &(pMsg->srcId), pMsg->startTime);
+  syncIndexMgrSetRecvTime(ths->pNextIndex, &(pMsg->srcId), taosGetTimestampMs());
+
   SyncIndex beforeNextIndex = syncIndexMgrGetIndex(ths->pNextIndex, &(pMsg->srcId));
   SyncIndex beforeMatchIndex = syncIndexMgrGetIndex(ths->pMatchIndex, &(pMsg->srcId));
 
@@ -330,6 +338,10 @@ int32_t syncNodeOnAppendEntriesReplySnapshotCb(SSyncNode* ths, SyncAppendEntries
 
   ASSERT(pMsg->term == ths->pRaftStore->currentTerm);
 
+  // update time
+  syncIndexMgrSetStartTime(ths->pNextIndex, &(pMsg->srcId), pMsg->startTime);
+  syncIndexMgrSetRecvTime(ths->pNextIndex, &(pMsg->srcId), taosGetTimestampMs());
+
   SyncIndex beforeNextIndex = syncIndexMgrGetIndex(ths->pNextIndex, &(pMsg->srcId));
   SyncIndex beforeMatchIndex = syncIndexMgrGetIndex(ths->pMatchIndex, &(pMsg->srcId));
 

source/libs/sync/src/syncCommit.c

@@ -133,6 +133,63 @@ bool syncAgreeIndex(SSyncNode* pSyncNode, SRaftId* pRaftId, SyncIndex index) {
   return false;
 }
 
+static inline int64_t syncNodeAbs64(int64_t a, int64_t b) {
+  ASSERT(a >= 0);
+  ASSERT(b >= 0);
+
+  int64_t c = a > b ? a - b : b - a;
+  return c;
+}
+
+int32_t syncNodeDynamicQuorum(const SSyncNode* pSyncNode) {
+  int32_t quorum = 1;  // self
+
+  int64_t timeNow = taosGetTimestampMs();
+  for (int i = 0; i < pSyncNode->peersNum; ++i) {
+    int64_t peerStartTime = syncIndexMgrGetStartTime(pSyncNode->pNextIndex, &(pSyncNode->peersId)[i]);
+    int64_t peerRecvTime = syncIndexMgrGetRecvTime(pSyncNode->pNextIndex, &(pSyncNode->peersId)[i]);
+
+    int64_t recvTimeDiff = syncNodeAbs64(peerRecvTime, timeNow);
+    int64_t startTimeDiff = syncNodeAbs64(peerStartTime, pSyncNode->startTime);
+
+    int32_t addQuorum = 0;
+
+    if (recvTimeDiff < SYNC_MAX_RECV_TIME_RANGE_MS) {
+      addQuorum = 1;
+    } else {
+      addQuorum = 0;
+    }
+
+    if (startTimeDiff > SYNC_MAX_START_TIME_RANGE_MS) {
+      addQuorum = 0;
+    }
+
+    quorum += addQuorum;
+  }
+
+  ASSERT(quorum <= pSyncNode->replicaNum);
+
+  if (quorum < pSyncNode->quorum) {
+    quorum = pSyncNode->quorum;
+  }
+
+  return quorum;
+}
+
+bool syncAgree(SSyncNode* pSyncNode, SyncIndex index) {
+  int agreeCount = 0;
+  for (int i = 0; i < pSyncNode->replicaNum; ++i) {
+    if (syncAgreeIndex(pSyncNode, &(pSyncNode->replicasId[i]), index)) {
+      ++agreeCount;
+    }
+    if (agreeCount >= syncNodeDynamicQuorum(pSyncNode)) {
+      return true;
+    }
+  }
+  return false;
+}
+
+/*
 bool syncAgree(SSyncNode* pSyncNode, SyncIndex index) {
   int agreeCount = 0;
   for (int i = 0; i < pSyncNode->replicaNum; ++i) {
@@ -145,3 +202,4 @@ bool syncAgree(SSyncNode* pSyncNode, SyncIndex index) {
   }
   return false;
 }
+*/

source/libs/sync/src/syncIndexMgr.c

@@ -47,6 +47,13 @@ void syncIndexMgrDestroy(SSyncIndexMgr *pSyncIndexMgr) {
 void syncIndexMgrClear(SSyncIndexMgr *pSyncIndexMgr) {
   memset(pSyncIndexMgr->index, 0, sizeof(pSyncIndexMgr->index));
   memset(pSyncIndexMgr->privateTerm, 0, sizeof(pSyncIndexMgr->privateTerm));
+
+  // int64_t timeNow = taosGetMonotonicMs();
+  for (int i = 0; i < pSyncIndexMgr->replicaNum; ++i) {
+    pSyncIndexMgr->startTimeArr[i] = 0;
+    pSyncIndexMgr->recvTimeArr[i] = 0;
+  }
+
   /*
   for (int i = 0; i < pSyncIndexMgr->replicaNum; ++i) {
     pSyncIndexMgr->index[i] = 0;
@@ -68,7 +75,8 @@ void syncIndexMgrSetIndex(SSyncIndexMgr *pSyncIndexMgr, const SRaftId *pRaftId,
   char     host[128];
   uint16_t port;
   syncUtilU642Addr(pRaftId->addr, host, sizeof(host), &port);
-  sError("vgId:%d, index mgr set for %s:%d, index:%" PRId64 " error", pSyncIndexMgr->pSyncNode->vgId, host, port, index);
+  sError("vgId:%d, index mgr set for %s:%d, index:%" PRId64 " error", pSyncIndexMgr->pSyncNode->vgId, host, port,
+         index);
 }
 
 SyncIndex syncIndexMgrGetIndex(SSyncIndexMgr *pSyncIndexMgr, const SRaftId *pRaftId) {
@@ -130,6 +138,60 @@ char *syncIndexMgr2Str(SSyncIndexMgr *pSyncIndexMgr) {
   return serialized;
 }
 
+void syncIndexMgrSetStartTime(SSyncIndexMgr *pSyncIndexMgr, const SRaftId *pRaftId, int64_t startTime) {
+  for (int i = 0; i < pSyncIndexMgr->replicaNum; ++i) {
+    if (syncUtilSameId(&((*(pSyncIndexMgr->replicas))[i]), pRaftId)) {
+      (pSyncIndexMgr->startTimeArr)[i] = startTime;
+      return;
+    }
+  }
+
+  // maybe config change
+  // ASSERT(0);
+  char     host[128];
+  uint16_t port;
+  syncUtilU642Addr(pRaftId->addr, host, sizeof(host), &port);
+  sError("vgId:%d, index mgr set for %s:%d, start-time:%" PRId64 " error", pSyncIndexMgr->pSyncNode->vgId, host, port,
+         startTime);
+}
+
+int64_t syncIndexMgrGetStartTime(SSyncIndexMgr *pSyncIndexMgr, const SRaftId *pRaftId) {
+  for (int i = 0; i < pSyncIndexMgr->replicaNum; ++i) {
+    if (syncUtilSameId(&((*(pSyncIndexMgr->replicas))[i]), pRaftId)) {
+      int64_t startTime = (pSyncIndexMgr->startTimeArr)[i];
+      return startTime;
+    }
+  }
+  ASSERT(0);
+}
+
+void syncIndexMgrSetRecvTime(SSyncIndexMgr *pSyncIndexMgr, const SRaftId *pRaftId, int64_t recvTime) {
+  for (int i = 0; i < pSyncIndexMgr->replicaNum; ++i) {
+    if (syncUtilSameId(&((*(pSyncIndexMgr->replicas))[i]), pRaftId)) {
+      (pSyncIndexMgr->recvTimeArr)[i] = recvTime;
+      return;
+    }
+  }
+
+  // maybe config change
+  // ASSERT(0);
+  char     host[128];
+  uint16_t port;
+  syncUtilU642Addr(pRaftId->addr, host, sizeof(host), &port);
+  sError("vgId:%d, index mgr set for %s:%d, recv-time:%" PRId64 " error", pSyncIndexMgr->pSyncNode->vgId, host, port,
+         recvTime);
+}
+
+int64_t syncIndexMgrGetRecvTime(SSyncIndexMgr *pSyncIndexMgr, const SRaftId *pRaftId) {
+  for (int i = 0; i < pSyncIndexMgr->replicaNum; ++i) {
+    if (syncUtilSameId(&((*(pSyncIndexMgr->replicas))[i]), pRaftId)) {
+      int64_t recvTime = (pSyncIndexMgr->recvTimeArr)[i];
+      return recvTime;
+    }
+  }
+  ASSERT(0);
+}
+
 // for debug -------------------
 void syncIndexMgrPrint(SSyncIndexMgr *pObj) {
   char *serialized = syncIndexMgr2Str(pObj);

source/libs/sync/src/syncMain.c

@@ -1682,13 +1682,13 @@ inline void syncNodeEventLog(const SSyncNode* pSyncNode, char* str) {
                ", sby:%d, "
                "stgy:%d, bch:%d, "
                "r-num:%d, "
-               "lcfg:%" PRId64 ", chging:%d, rsto:%d, elt:%" PRId64 ", hb:%" PRId64 ", %s",
+               "lcfg:%" PRId64 ", chging:%d, rsto:%d, dquorum:%d, elt:%" PRId64 ", hb:%" PRId64 ", %s",
                pSyncNode->vgId, syncUtilState2String(pSyncNode->state), str, pSyncNode->pRaftStore->currentTerm,
                pSyncNode->commitIndex, logBeginIndex, logLastIndex, snapshot.lastApplyIndex, snapshot.lastApplyTerm,
                pSyncNode->pRaftCfg->isStandBy, pSyncNode->pRaftCfg->snapshotStrategy, pSyncNode->pRaftCfg->batchSize,
                pSyncNode->replicaNum, pSyncNode->pRaftCfg->lastConfigIndex, pSyncNode->changing,
-               pSyncNode->restoreFinish, pSyncNode->electTimerLogicClockUser, pSyncNode->heartbeatTimerLogicClockUser,
+               pSyncNode->restoreFinish, syncNodeDynamicQuorum(pSyncNode), pSyncNode->electTimerLogicClockUser,
-               printStr);
+               pSyncNode->heartbeatTimerLogicClockUser, printStr);
     } else {
       snprintf(logBuf, sizeof(logBuf), "%s", str);
     }
@@ -1706,12 +1706,13 @@ inline void syncNodeEventLog(const SSyncNode* pSyncNode, char* str) {
                ", sby:%d, "
                "stgy:%d, bch:%d, "
                "r-num:%d, "
-               "lcfg:%" PRId64 ", chging:%d, rsto:%d, %s",
+               "lcfg:%" PRId64 ", chging:%d, rsto:%d, dquorum:%d, elt:%" PRId64 ", hb:%" PRId64 ", %s",
                pSyncNode->vgId, syncUtilState2String(pSyncNode->state), str, pSyncNode->pRaftStore->currentTerm,
                pSyncNode->commitIndex, logBeginIndex, logLastIndex, snapshot.lastApplyIndex, snapshot.lastApplyTerm,
                pSyncNode->pRaftCfg->isStandBy, pSyncNode->pRaftCfg->snapshotStrategy, pSyncNode->pRaftCfg->batchSize,
                pSyncNode->replicaNum, pSyncNode->pRaftCfg->lastConfigIndex, pSyncNode->changing,
-               pSyncNode->restoreFinish, printStr);
+               pSyncNode->restoreFinish, syncNodeDynamicQuorum(pSyncNode), pSyncNode->electTimerLogicClockUser,
+               pSyncNode->heartbeatTimerLogicClockUser, printStr);
     } else {
       snprintf(s, len, "%s", str);
     }

source/libs/sync/src/syncMessage.c

@@ -1947,6 +1947,8 @@ cJSON* syncAppendEntriesReply2Json(const SyncAppendEntriesReply* pMsg) {
     cJSON_AddNumberToObject(pRoot, "success", pMsg->success);
     snprintf(u64buf, sizeof(u64buf), "%" PRId64, pMsg->matchIndex);
     cJSON_AddStringToObject(pRoot, "matchIndex", u64buf);
+    snprintf(u64buf, sizeof(u64buf), "%" PRId64, pMsg->startTime);
+    cJSON_AddStringToObject(pRoot, "startTime", u64buf);
   }
 
   cJSON* pJson = cJSON_CreateObject();

source/libs/sync/src/syncReplication.c

@@ -116,6 +116,120 @@ int32_t syncNodeAppendEntriesPeers(SSyncNode* pSyncNode) {
   return ret;
 }
 
+int32_t syncNodeAppendEntriesOnePeer(SSyncNode* pSyncNode, SRaftId* pDestId, SyncIndex nextIndex) {
+  int32_t ret = 0;
+
+  // pre index, pre term
+  SyncIndex preLogIndex = syncNodeGetPreIndex(pSyncNode, nextIndex);
+  SyncTerm  preLogTerm = syncNodeGetPreTerm(pSyncNode, nextIndex);
+  if (preLogTerm == SYNC_TERM_INVALID) {
+    SyncIndex newNextIndex = syncNodeGetLastIndex(pSyncNode) + 1;
+    // SyncIndex newNextIndex = nextIndex + 1;
+
+    syncIndexMgrSetIndex(pSyncNode->pNextIndex, pDestId, newNextIndex);
+    syncIndexMgrSetIndex(pSyncNode->pMatchIndex, pDestId, SYNC_INDEX_INVALID);
+    sError("vgId:%d, sync get pre term error, nextIndex:%" PRId64 ", update next-index:%" PRId64
+           ", match-index:%d, raftid:%" PRId64,
+           pSyncNode->vgId, nextIndex, newNextIndex, SYNC_INDEX_INVALID, pDestId->addr);
+    return -1;
+  }
+
+  // entry pointer array
+  SSyncRaftEntry* entryPArr[SYNC_MAX_BATCH_SIZE];
+  memset(entryPArr, 0, sizeof(entryPArr));
+
+  // get entry batch
+  int32_t   getCount = 0;
+  SyncIndex getEntryIndex = nextIndex;
+  for (int32_t i = 0; i < pSyncNode->pRaftCfg->batchSize; ++i) {
+    SSyncRaftEntry* pEntry = NULL;
+    int32_t         code = pSyncNode->pLogStore->syncLogGetEntry(pSyncNode->pLogStore, getEntryIndex, &pEntry);
+    if (code == 0) {
+      ASSERT(pEntry != NULL);
+      entryPArr[i] = pEntry;
+      getCount++;
+      getEntryIndex++;
+
+    } else {
+      break;
+    }
+  }
+
+  // event log
+  do {
+    char     logBuf[128];
+    char     host[64];
+    uint16_t port;
+    syncUtilU642Addr(pDestId->addr, host, sizeof(host), &port);
+    snprintf(logBuf, sizeof(logBuf), "build batch:%d for %s:%d", getCount, host, port);
+    syncNodeEventLog(pSyncNode, logBuf);
+  } while (0);
+
+  // build msg
+  SyncAppendEntriesBatch* pMsg = syncAppendEntriesBatchBuild(entryPArr, getCount, pSyncNode->vgId);
+  ASSERT(pMsg != NULL);
+
+  // free entries
+  for (int32_t i = 0; i < pSyncNode->pRaftCfg->batchSize; ++i) {
+    SSyncRaftEntry* pEntry = entryPArr[i];
+    if (pEntry != NULL) {
+      syncEntryDestory(pEntry);
+      entryPArr[i] = NULL;
+    }
+  }
+
+  // prepare msg
+  pMsg->srcId = pSyncNode->myRaftId;
+  pMsg->destId = *pDestId;
+  pMsg->term = pSyncNode->pRaftStore->currentTerm;
+  pMsg->prevLogIndex = preLogIndex;
+  pMsg->prevLogTerm = preLogTerm;
+  pMsg->commitIndex = pSyncNode->commitIndex;
+  pMsg->privateTerm = 0;
+  pMsg->dataCount = getCount;
+
+  // send msg
+  syncNodeAppendEntriesBatch(pSyncNode, pDestId, pMsg);
+
+  // speed up
+  if (pMsg->dataCount > 0 && pSyncNode->commitIndex - pMsg->prevLogIndex > SYNC_SLOW_DOWN_RANGE) {
+    ret = 1;
+
+#if 0
+      do {
+        char     logBuf[128];
+        char     host[64];
+        uint16_t port;
+        syncUtilU642Addr(pDestId->addr, host, sizeof(host), &port);
+        snprintf(logBuf, sizeof(logBuf), "maybe speed up for %s:%d, pre-index:%ld", host, port, pMsg->prevLogIndex);
+        syncNodeEventLog(pSyncNode, logBuf);
+      } while (0);
+#endif
+  }
+
+  syncAppendEntriesBatchDestroy(pMsg);
+
+  return ret;
+}
+
+int32_t syncNodeAppendEntriesPeersSnapshot2(SSyncNode* pSyncNode) {
+  if (pSyncNode->state != TAOS_SYNC_STATE_LEADER) {
+    return -1;
+  }
+
+  int32_t ret = 0;
+  for (int i = 0; i < pSyncNode->peersNum; ++i) {
+    SRaftId* pDestId = &(pSyncNode->peersId[i]);
+
+    // next index
+    SyncIndex nextIndex = syncIndexMgrGetIndex(pSyncNode->pNextIndex, pDestId);
+    ret = syncNodeAppendEntriesOnePeer(pSyncNode, pDestId, nextIndex);
+  }
+
+  return ret;
+}
+
+#if 0
 int32_t syncNodeAppendEntriesPeersSnapshot2(SSyncNode* pSyncNode) {
   if (pSyncNode->state != TAOS_SYNC_STATE_LEADER) {
     return -1;
@@ -221,6 +335,7 @@ int32_t syncNodeAppendEntriesPeersSnapshot2(SSyncNode* pSyncNode) {
 
   return ret;
 }
+#endif
 
 int32_t syncNodeAppendEntriesPeersSnapshot(SSyncNode* pSyncNode) {
   ASSERT(pSyncNode->state == TAOS_SYNC_STATE_LEADER);

source/libs/sync/test/syncAppendEntriesReplyTest.cpp

@@ -24,6 +24,7 @@ SyncAppendEntriesReply *createMsg() {
   pMsg->matchIndex = 77;
   pMsg->term = 33;
   pMsg->privateTerm = 44;
+  pMsg->startTime = taosGetTimestampMs();
   return pMsg;
 }
 
@@ -89,6 +90,8 @@ void test5() {
 }
 
 int main() {
+  gRaftDetailLog = true;
+
   tsAsyncLog = 0;
   sDebugFlag = DEBUG_TRACE + DEBUG_SCREEN + DEBUG_FILE;
   logTest();

source/util/src/tarray.c

@@ -386,6 +386,7 @@ void* taosArrayDestroy(SArray* pArray) {
 }
 
 void taosArrayDestroyP(SArray* pArray, FDelete fp) {
+  if(!pArray) return;
   for (int32_t i = 0; i < pArray->size; i++) {
     fp(*(void**)TARRAY_GET_ELEM(pArray, i));
   }

source/util/src/terror.c

@@ -293,6 +293,7 @@ TAOS_DEFINE_ERROR(TSDB_CODE_MND_CGROUP_USED,              "Consumer group being
 TAOS_DEFINE_ERROR(TSDB_CODE_MND_STREAM_ALREADY_EXIST,     "Stream already exists")
 TAOS_DEFINE_ERROR(TSDB_CODE_MND_STREAM_NOT_EXIST,         "Stream not exist")
 TAOS_DEFINE_ERROR(TSDB_CODE_MND_INVALID_STREAM_OPTION,    "Invalid stream option")
+TAOS_DEFINE_ERROR(TSDB_CODE_MND_STREAM_MUST_BE_DELETED,   "Stream must be dropped first")
 
 // mnode-sma
 TAOS_DEFINE_ERROR(TSDB_CODE_MND_SMA_ALREADY_EXIST,        "SMA already exists")

tests/script/tsim/db/basic2.sim

@@ -3,6 +3,21 @@ system sh/deploy.sh -n dnode1 -i 1
 system sh/exec.sh -n dnode1 -s start
 sql connect
 
+print =============== conflict stb
+sql create database db vgroups 1;
+sql use db;
+sql create table stb (ts timestamp, i int) tags (j int);
+sql_error create table stb using stb tags (1);
+sql_error create table stb (ts timestamp, i int);
+
+sql create table ctb (ts timestamp, i int);
+sql_error create table ctb (ts timestamp, i int) tags (j int);
+
+sql create table ntb (ts timestamp, i int);
+sql_error create table ntb (ts timestamp, i int) tags (j int);
+
+sql drop database db
+
 print =============== create database d1
 sql create database d1
 sql use d1

tests/system-test/2-query/json_tag.py

@@ -197,7 +197,7 @@ class TDTestCase:
 
         # test where with json tag
         tdSql.query(f"select * from {dbname}.jsons1_1 where jtag is not null")
-        tdSql.query(f"select * from {dbname}.jsons1 where jtag='{{\"tag1\":11,\"tag2\":\"\"}}'")
+        tdSql.error(f"select * from {dbname}.jsons1 where jtag='{{\"tag1\":11,\"tag2\":\"\"}}'")
         tdSql.error(f"select * from {dbname}.jsons1 where jtag->'tag1'={{}}")
 
         # test json error