Merge branch '3.0' into feature/stream

2022-07-19 19:46:36 +08:00 · 2022-07-19 19:46:36 +08:00 · 414673db57
parent 6362f9e435 88f13276f8
commit 414673db57
122 changed files with 2090 additions and 937 deletions
--- a/docs/en/14-reference/03-connector/java.mdx
+++ b/docs/en/14-reference/03-connector/java.mdx
@ -202,6 +202,10 @@ The configuration parameters in the URL are as follows.
 - batchfetch: true: pull the result set in batch when executing the query; false: pull the result set row by row. The default value is false. batchfetch uses HTTP for data transfer. The JDBC REST connection supports bulk data pulling function in taos-jdbcdriver-2.0.38 and TDengine 2.4.0.12 and later versions. taos-jdbcdriver and TDengine transfer data via WebSocket connection. Compared with HTTP, WebSocket enables JDBC REST connection to support large data volume querying and improve query performance.
 - charset: specify the charset to parse the string, this parameter is valid only when set batchfetch to true.
 - batchErrorIgnore: true: when executing executeBatch of Statement, if one SQL execution fails in the middle, continue to execute the following SQL. false: no longer execute any statement after the failed SQL. The default value is: false.
+- httpConnectTimeout: REST connection timeout in milliseconds, the default value is 5000 ms.
+- httpSocketTimeout: socket timeout in milliseconds, the default value is 5000 ms. It only takes effect when batchfetch is false.
+- messageWaitTimeout: message transmission timeout in milliseconds, the default value is 3000 ms. It only takes effect when batchfetch is true.
+- useSSL: connecting Securely Using SSL. true: using SSL conneciton, false: not using SSL connection.

 **Note**: Some configuration items (e.g., locale, timezone) do not work in the REST connection.

@ -257,14 +261,18 @@ In the above example, a connection is established to `taosdemo.com`, port is 603

 The configuration parameters in properties are as follows.

- TSDBDriver.PROPERTY_KEY_USER: Login TDengine user name, default value 'root'.
+- TSDBDriver.PROPERTY_KEY_USER: login TDengine user name, default value 'root'.
 - TSDBDriver.PROPERTY_KEY_PASSWORD: user login password, default value 'taosdata'.
 - TSDBDriver.PROPERTY_KEY_BATCH_LOAD: true: pull the result set in batch when executing query; false: pull the result set row by row. The default value is: false.
 - TSDBDriver.PROPERTY_KEY_BATCH_ERROR_IGNORE: true: when executing executeBatch of Statement, if there is a SQL execution failure in the middle, continue to execute the following sq. false: no longer execute any statement after the failed SQL. The default value is: false.
- TSDBDriver.PROPERTY_KEY_CONFIG_DIR: Only works when using JDBC native connection. Client configuration file directory path, default value `/etc/taos` on Linux OS, default value `C:/TDengine/cfg` on Windows OS.
+- TSDBDriver.PROPERTY_KEY_CONFIG_DIR: only works when using JDBC native connection. Client configuration file directory path, default value `/etc/taos` on Linux OS, default value `C:/TDengine/cfg` on Windows OS.
 - TSDBDriver.PROPERTY_KEY_CHARSET: In the character set used by the client, the default value is the system character set.
 - TSDBDriver.PROPERTY_KEY_LOCALE: this only takes effect when using JDBC native connection. Client language environment, the default value is system current locale.
 - TSDBDriver.PROPERTY_KEY_TIME_ZONE: only takes effect when using JDBC native connection. In the time zone used by the client, the default value is the system's current time zone.
+- TSDBDriver.HTTP_CONNECT_TIMEOUT: REST connection timeout in milliseconds, the default value is 5000 ms. It only takes effect when using JDBC REST connection.
+- TSDBDriver.HTTP_SOCKET_TIMEOUT: socket timeout in milliseconds, the default value is 5000 ms. It only takes effect when using JDBC REST connection and batchfetch is false.
+- TSDBDriver.PROPERTY_KEY_MESSAGE_WAIT_TIMEOUT: message transmission timeout in milliseconds, the default value is 3000 ms. It only takes effect when using JDBC REST connection and batchfetch is true.
+- TSDBDriver.PROPERTY_KEY_USE_SSL: connecting Securely Using SSL. true: using SSL conneciton, false: not using SSL connection. It only takes effect when using using JDBC REST connection.
  For JDBC native connections, you can specify other parameters, such as log level, SQL length, etc., by specifying URL and Properties. For more detailed configuration, please refer to [Client Configuration](/reference/config/#Client-Only).

 ### Priority of configuration parameters
@ -812,11 +820,12 @@ Please refer to: [JDBC example](https://github.com/taosdata/TDengine/tree/develo

 ## Recent update logs

-| taos-jdbcdriver version |                major changes                 |
-| :---------------------: | :------------------------------------------: |
-|         2.0.38          | JDBC REST connections add bulk pull function |
-|         2.0.37          |         Added support for json tags          |
-|         2.0.36          |      Add support for schemaless writing      |
+| taos-jdbcdriver version |                 major changes                  |
+| :---------------------: | :--------------------------------------------: |
+|     2.0.39 - 2.0.40     | Add REST connection/request timeout parameters |
+|         2.0.38          |  JDBC REST connections add bulk pull function  |
+|         2.0.37          |               Support json tags                |
+|         2.0.36          |           Support schemaless writing           |

 ## Frequently Asked Questions

--- a/docs/zh/14-reference/03-connector/java.mdx
+++ b/docs/zh/14-reference/03-connector/java.mdx
@ -201,6 +201,10 @@ url 中的配置参数如下：
 - batchfetch: true：在执行查询时批量拉取结果集；false：逐行拉取结果集。默认值为：false。逐行拉取结果集使用 HTTP 方式进行数据传输。从 taos-jdbcdriver-2.0.38 和 TDengine 2.4.0.12 版本开始，JDBC REST 连接增加批量拉取数据功能。taos-jdbcdriver 与 TDengine 之间通过 WebSocket 连接进行数据传输。相较于 HTTP，WebSocket 可以使 JDBC REST 连接支持大数据量查询，并提升查询性能。
 - charset: 当开启批量拉取数据时，指定解析字符串数据的字符集。
 - batchErrorIgnore：true：在执行 Statement 的 executeBatch 时，如果中间有一条 SQL 执行失败，继续执行下面的 SQL 了。false：不再执行失败 SQL 后的任何语句。默认值为：false。
+- httpConnectTimeout: 连接超时时间，单位 ms， 默认值为 5000。
+- httpSocketTimeout: socket 超时时间，单位 ms，默认值为 5000。仅在 batchfetch 设置为 false 时生效。
+- messageWaitTimeout: 消息超时时间, 单位 ms， 默认值为 3000。 仅在 batchfetch 设置为 true 时生效。
+- useSSL: 连接中是否使用 SSL。

 **注意**：部分配置项（比如：locale、timezone）在 REST 连接中不生效。

@ -264,7 +268,11 @@ properties 中的配置参数如下：
 - TSDBDriver.PROPERTY_KEY_CHARSET：客户端使用的字符集，默认值为系统字符集。
 - TSDBDriver.PROPERTY_KEY_LOCALE：仅在使用 JDBC 原生连接时生效。 客户端语言环境，默认值系统当前 locale。
 - TSDBDriver.PROPERTY_KEY_TIME_ZONE：仅在使用 JDBC 原生连接时生效。 客户端使用的时区，默认值为系统当前时区。
- 此外对 JDBC 原生连接，通过指定 URL 和 Properties 还可以指定其他参数，比如日志级别、SQL 长度等。更多详细配置请参考[客户端配置](/reference/config/#仅客户端适用)。
+- TSDBDriver.HTTP_CONNECT_TIMEOUT: 连接超时时间，单位 ms， 默认值为 5000。仅在 REST 连接时生效。
+- TSDBDriver.HTTP_SOCKET_TIMEOUT: socket 超时时间，单位 ms，默认值为 5000。仅在 REST 连接且 batchfetch 设置为 false 时生效。
+- TSDBDriver.PROPERTY_KEY_MESSAGE_WAIT_TIMEOUT: 消息超时时间, 单位 ms， 默认值为 3000。 仅在 REST 连接且 batchfetch 设置为 true 时生效。
+- TSDBDriver.PROPERTY_KEY_USE_SSL: 连接中是否使用 SSL。仅在 REST 连接时生效。
+  此外对 JDBC 原生连接，通过指定 URL 和 Properties 还可以指定其他参数，比如日志级别、SQL 长度等。更多详细配置请参考[客户端配置](/reference/config/#仅客户端适用)。

 ### 配置参数的优先级

@ -809,6 +817,7 @@ Query OK, 1 row(s) in set (0.000141s)

 | taos-jdbcdriver 版本 |            主要变化            |
 | :------------------: | :----------------------------: |
+|   2.0.39 - 2.0.40    |  增加 REST 连接/请求 超时设置  |
 |        2.0.38        | JDBC REST 连接增加批量拉取功能 |
 |        2.0.37        |      增加对 json tag 支持      |
 |        2.0.36        |   增加对 schemaless 写入支持   |
--- a/examples/c/tmq.c
+++ b/examples/c/tmq.c
@ -28,8 +28,9 @@ static void msg_process(TAOS_RES* msg) {
  printf("db: %s\n", tmq_get_db_name(msg));
  printf("vg: %d\n", tmq_get_vgroup_id(msg));
  if (tmq_get_res_type(msg) == TMQ_RES_TABLE_META) {
-    tmq_raw_data* raw = tmq_get_raw_meta(msg);
-    if (raw) {
+    tmq_raw_data raw = {0};
+    int32_t code = tmq_get_raw_meta(msg, &raw);
+    if (code == 0) {
      TAOS* pConn = taos_connect("192.168.1.86", "root", "taosdata", NULL, 0);
      if (pConn == NULL) {
        return;
@ -53,7 +54,6 @@ static void msg_process(TAOS_RES* msg) {
      printf("write raw data: %s\n", tmq_err2str(ret));
      taos_close(pConn);
    }
-    tmq_free_raw_meta(raw);
    char* result = tmq_get_json_meta(msg);
    if (result) {
      printf("meta result: %s\n", result);
--- a/include/client/taos.h
+++ b/include/client/taos.h
@ -259,13 +259,17 @@ enum tmq_res_t {
  TMQ_RES_TABLE_META = 2,
 };

+typedef struct tmq_raw_data{
+  void*    raw_meta;
+  uint32_t raw_meta_len;
+  uint16_t raw_meta_type;
+} tmq_raw_data;
+
 typedef enum tmq_res_t tmq_res_t;
-typedef struct tmq_raw_data tmq_raw_data;

 DLL_EXPORT tmq_res_t     tmq_get_res_type(TAOS_RES *res);
-DLL_EXPORT tmq_raw_data *tmq_get_raw_meta(TAOS_RES *res);
-DLL_EXPORT int32_t       taos_write_raw_meta(TAOS *taos, tmq_raw_data *raw_meta);
-DLL_EXPORT void          tmq_free_raw_meta(tmq_raw_data *rawMeta);
+DLL_EXPORT int32_t       tmq_get_raw_meta(TAOS_RES *res, tmq_raw_data *raw_meta);
+DLL_EXPORT int32_t       taos_write_raw_meta(TAOS *taos, tmq_raw_data raw_meta);
 DLL_EXPORT char         *tmq_get_json_meta(TAOS_RES *res);   // Returning null means error. Returned result need to be freed by tmq_free_json_meta
 DLL_EXPORT void          tmq_free_json_meta(char* jsonMeta);
 DLL_EXPORT const char   *tmq_get_topic_name(TAOS_RES *res);
--- a/include/libs/executor/executor.h
+++ b/include/libs/executor/executor.h
@ -40,6 +40,7 @@ typedef struct SReadHandle {
  bool    initMetaReader;
  bool    initTableReader;
  bool    initTqReader;
+  int32_t numOfVgroups;
 } SReadHandle;

 // in queue mode, data streams are seperated by msg
@ -63,7 +64,7 @@ qTaskInfo_t qCreateStreamExecTaskInfo(void* msg, SReadHandle* readers);
 * @param SReadHandle
 * @return
 */
-qTaskInfo_t qCreateQueueExecTaskInfo(void* msg, SReadHandle* readers);
+qTaskInfo_t qCreateQueueExecTaskInfo(void* msg, SReadHandle* readers, int32_t* numOfCols);

 /**
 * Set the input data block for the stream scan.
--- a/include/libs/qcom/query.h
+++ b/include/libs/qcom/query.h
@ -162,9 +162,12 @@ typedef struct SRequestConnInfo {
  SEpSet   mgmtEps;
 } SRequestConnInfo;

+typedef void (*__freeFunc)(void *param);
+
 typedef struct SMsgSendInfo {
  __async_send_cb_fn_t fp;      // async callback function
  STargetInfo          target;  // for update epset
+  __freeFunc           paramFreeFp;
  void*                param;
  uint64_t             requestId;
  uint64_t             requestObjRefId;
@ -188,6 +191,8 @@ int32_t cleanupTaskQueue();
 */
 int32_t taosAsyncExec(__async_exec_fn_t execFn, void* execParam, int32_t* code);

+void destroySendMsgInfo(SMsgSendInfo* pMsgBody);
+
 int32_t asyncSendMsgToServerExt(void* pTransporter, SEpSet* epSet, int64_t* pTransporterId, const SMsgSendInfo* pInfo,
                                bool persistHandle, void* ctx);

--- a/include/libs/stream/tstream.h
+++ b/include/libs/stream/tstream.h
@ -269,6 +269,7 @@ typedef struct SStreamTask {
  int64_t startVer;
  int64_t checkpointVer;
  int64_t processedVer;
+  int32_t numOfVgroups;

  // children info
  SArray* childEpInfo;  // SArray<SStreamChildEpInfo*>
--- a/include/libs/sync/sync.h
+++ b/include/libs/sync/sync.h
@ -26,7 +26,7 @@ extern "C" {

 extern bool gRaftDetailLog;

-#define SYNC_RESP_TTL_MS 5000
+#define SYNC_RESP_TTL_MS 10000

 #define SYNC_MAX_BATCH_SIZE 500
 #define SYNC_INDEX_BEGIN 0
--- a/packaging/deb/makedeb.sh
+++ b/packaging/deb/makedeb.sh
@ -73,7 +73,7 @@ cp ${compile_dir}/../include/client/taos.h          ${pkg_dir}${install_home_pat
 cp ${compile_dir}/../include/common/taosdef.h       ${pkg_dir}${install_home_path}/include
 cp ${compile_dir}/../include/util/taoserror.h       ${pkg_dir}${install_home_path}/include
 cp ${compile_dir}/../include/libs/function/taosudf.h       ${pkg_dir}${install_home_path}/include
-cp ${compile_dir}/../src/inc/taosws.h               ${pkg_dir}${install_home_path}/include ||:
+cp ${compile_dir}/build/include/taosws.h            ${pkg_dir}${install_home_path}/include ||:
 cp -r ${top_dir}/examples/*                         ${pkg_dir}${install_home_path}/examples
 #cp -r ${top_dir}/src/connector/python               ${pkg_dir}${install_home_path}/connector
 #cp -r ${top_dir}/src/connector/go                   ${pkg_dir}${install_home_path}/connector
--- a/packaging/rpm/tdengine.spec
+++ b/packaging/rpm/tdengine.spec
@ -80,7 +80,7 @@ cp %{_compiledir}/../include/client/taos.h          %{buildroot}%{homepath}/incl
 cp %{_compiledir}/../include/common/taosdef.h       %{buildroot}%{homepath}/include
 cp %{_compiledir}/../include/util/taoserror.h       %{buildroot}%{homepath}/include
 cp %{_compiledir}/../include/libs/function/taosudf.h       %{buildroot}%{homepath}/include
-cp %{_compiledir}/../src/inc/taosws.h               %{buildroot}%{homepath}/include ||:
+cp %{_compiledir}/build/include/taosws.h            %{buildroot}%{homepath}/include ||:
 #cp -r %{_compiledir}/../src/connector/python        %{buildroot}%{homepath}/connector
 #cp -r %{_compiledir}/../src/connector/go            %{buildroot}%{homepath}/connector
 #cp -r %{_compiledir}/../src/connector/nodejs        %{buildroot}%{homepath}/connector
--- a/packaging/tools/install.sh
+++ b/packaging/tools/install.sh
@ -229,13 +229,13 @@ function install_lib() {
  ${csudo}ln -s ${install_main_dir}/driver/libtaos.* ${lib_link_dir}/libtaos.so.1
  ${csudo}ln -s ${lib_link_dir}/libtaos.so.1 ${lib_link_dir}/libtaos.so

-  ${csudo}ln -s ${lib_link_dir}/libtaosws.so ${lib_link_dir}/libtaosws.so || :
+  [ -f ${install_main_dir}/driver/libtaosws.so ] && ${csudo}ln -s ${install_main_dir}/driver/libtaosws.so ${lib_link_dir}/libtaosws.so || :

  if [[ -d ${lib64_link_dir} && ! -e ${lib64_link_dir}/libtaos.so ]]; then
    ${csudo}ln -s ${install_main_dir}/driver/libtaos.* ${lib64_link_dir}/libtaos.so.1 || :
    ${csudo}ln -s ${lib64_link_dir}/libtaos.so.1 ${lib64_link_dir}/libtaos.so || :

-    ${csudo}ln -s ${lib64_link_dir}/libtaosws.so ${lib64_link_dir}/libtaosws.so || :
+    ${csudo}ln -s ${install_main_dir}/libtaosws.so ${lib64_link_dir}/libtaosws.so || :
  fi

  ${csudo}ldconfig
--- a/packaging/tools/install_client.sh
+++ b/packaging/tools/install_client.sh
@ -116,6 +116,7 @@ function install_bin() {

 function clean_lib() {
    sudo rm -f /usr/lib/libtaos.* || :
+    sudo rm -f /usr/lib/libtaosws.* || :
    sudo rm -rf ${lib_dir} || :
 }

@ -123,6 +124,9 @@ function install_lib() {
    # Remove links
    ${csudo}rm -f ${lib_link_dir}/libtaos.*         || :
    ${csudo}rm -f ${lib64_link_dir}/libtaos.*       || :
+
+    ${csudo}rm -f ${lib_link_dir}/libtaosws.*         || :
+    ${csudo}rm -f ${lib64_link_dir}/libtaosws.*       || :
    #${csudo}rm -rf ${v15_java_app_dir}              || :

    ${csudo}cp -rf ${script_dir}/driver/* ${install_main_dir}/driver && ${csudo}chmod 777 ${install_main_dir}/driver/*
@ -131,13 +135,19 @@ function install_lib() {
        ${csudo}ln -s ${install_main_dir}/driver/libtaos.* ${lib_link_dir}/libtaos.so.1
        ${csudo}ln -s ${lib_link_dir}/libtaos.so.1 ${lib_link_dir}/libtaos.so

+        [ -f ${install_main_dir}/driver/libtaosws.so ] && ${csudo}ln -s ${install_main_dir}/driver/libtaosws.so ${lib_link_dir}/libtaosws.so
+
        if [ -d "${lib64_link_dir}" ]; then
            ${csudo}ln -s ${install_main_dir}/driver/libtaos.* ${lib64_link_dir}/libtaos.so.1       || :
            ${csudo}ln -s ${lib64_link_dir}/libtaos.so.1 ${lib64_link_dir}/libtaos.so               || :
+
+            [ -f ${install_main_dir}/driver/libtaosws.so ] && ${csudo}ln -s ${install_main_dir}/driver/libtaosws.so ${lib64_link_dir}/libtaosws.so       || :
        fi
    else
        ${csudo}ln -s ${install_main_dir}/driver/libtaos.* ${lib_link_dir}/libtaos.1.dylib
        ${csudo}ln -s ${lib_link_dir}/libtaos.1.dylib ${lib_link_dir}/libtaos.dylib
+
+        [ -f ${install_main_dir}/driver/libtaosws.dylib ] && ${csudo}ln -s ${install_main_dir}/driver/libtaosws.dylib ${lib_link_dir}/libtaosws.dylib
    fi

    if [ "$osType" != "Darwin" ]; then
@ -154,6 +164,8 @@ function install_header() {
    ${csudo}ln -s ${install_main_dir}/include/taosdef.h ${inc_link_dir}/taosdef.h
    ${csudo}ln -s ${install_main_dir}/include/taoserror.h ${inc_link_dir}/taoserror.h
    ${csudo}ln -s ${install_main_dir}/include/taosudf.h ${inc_link_dir}/taosudf.h    
+
+    [ -f ${install_main_dir}/include/taosws.h ] && ${csudo}ln -s ${install_main_dir}/include/taosws.h ${inc_link_dir}/taos.h
 }

 function install_jemalloc() {
--- a/packaging/tools/makeclient.sh
+++ b/packaging/tools/makeclient.sh
@ -57,12 +57,16 @@ if [ "$osType" != "Darwin" ]; then
        ${script_dir}/get_client.sh"
  fi
  lib_files="${build_dir}/lib/libtaos.so.${version}"
+  wslib_files="${build_dir}/lib/libtaosws.so"
 else
  bin_files="${build_dir}/bin/${clientName} ${script_dir}/remove_client.sh"
  lib_files="${build_dir}/lib/libtaos.${version}.dylib"
+  wslib_files="${build_dir}/lib/libtaosws.dylib"
 fi

 header_files="${code_dir}/include/client/taos.h ${code_dir}/include/common/taosdef.h ${code_dir}/include/util/taoserror.h ${code_dir}/include/libs/function/taosudf.h"
+wsheader_files="${build_dir}/include/taosws.h"
+
 if [ "$dbName" != "taos" ]; then
  cfg_dir="${top_dir}/../enterprise/packaging/cfg"
 else
@ -74,6 +78,8 @@ install_files="${script_dir}/install_client.sh"
 # make directories.
 mkdir -p ${install_dir}
 mkdir -p ${install_dir}/inc && cp ${header_files} ${install_dir}/inc
+[ -f ${wsheader_files} ] && cp ${wsheader_files} ${install_dir}/inc
+
 mkdir -p ${install_dir}/cfg && cp ${cfg_dir}/${configFile} ${install_dir}/cfg/${configFile}
 mkdir -p ${install_dir}/bin && cp ${bin_files} ${install_dir}/bin && chmod a+x ${install_dir}/bin/*

@ -187,6 +193,7 @@ cp ${lib_files} ${install_dir}/driver
 # Copy connector
 connector_dir="${code_dir}/connector"
 mkdir -p ${install_dir}/connector
+[ -f ${wslib_files} ] && cp ${wslib_files} ${install_dir}/driver

 if [[ "$pagMode" != "lite" ]] && [[ "$cpuType" != "aarch32" ]]; then
  if [ "$osType" != "Darwin" ]; then
--- a/packaging/tools/makepkg.sh
+++ b/packaging/tools/makepkg.sh
@ -96,7 +96,7 @@ lib_files="${build_dir}/lib/libtaos.so.${version}"
 wslib_files="${build_dir}/lib/libtaosws.so."
 header_files="${code_dir}/include/client/taos.h ${code_dir}/include/common/taosdef.h ${code_dir}/include/util/taoserror.h ${code_dir}/include/libs/function/taosudf.h"

-wsheader_files="${code_dir}/inc/taosws.h"
+wsheader_files="${build_dir}/include/taosws.h"

 if [ "$dbName" != "taos" ]; then
  cfg_dir="${top_dir}/../enterprise/packaging/cfg"
@ -114,7 +114,7 @@ init_file_rpm=${script_dir}/../rpm/taosd
 mkdir -p ${install_dir}
 mkdir -p ${install_dir}/inc && cp ${header_files} ${install_dir}/inc

-${wsheader_files} ${install_dir}/inc || :
+[ -f ${wsheader_files} ] && cp ${wsheader_files} ${install_dir}/inc || :

 mkdir -p ${install_dir}/cfg && cp ${cfg_dir}/${configFile} ${install_dir}/cfg/${configFile}

--- a/packaging/tools/post.sh
+++ b/packaging/tools/post.sh
@ -82,22 +82,33 @@ function kill_taosd() {

 function install_include() {
    ${csudo}rm -f ${inc_link_dir}/taos.h ${inc_link_dir}/taosdef.h ${inc_link_dir}/taoserror.h ${inc_link_dir}/taosudf.h || :
+    ${csudo}rm -f ${inc_link_dir}/taosws.h
+
    ${csudo}ln -s ${inc_dir}/taos.h ${inc_link_dir}/taos.h
    ${csudo}ln -s ${inc_dir}/taosdef.h ${inc_link_dir}/taosdef.h
    ${csudo}ln -s ${inc_dir}/taoserror.h ${inc_link_dir}/taoserror.h
    ${csudo}ln -s ${inc_dir}/taosudf.h ${inc_link_dir}/taosudf.h    
+
+    [ -f ${inc_dir}/taosws.h ] && ${csudo}ln -s ${inc_dir}/taosudf.h ${inc_link_dir}/taosudf.h ||:
 }

 function install_lib() {
    ${csudo}rm -f ${lib_link_dir}/libtaos* || :
    ${csudo}rm -f ${lib64_link_dir}/libtaos* || :

+    ${csudo}rm -f ${lib_link_dir}/libtaosws* || :
+    ${csudo}rm -f ${lib64_link_dir}/libtaosws* || :
+
    ${csudo}ln -s ${lib_dir}/libtaos.* ${lib_link_dir}/libtaos.so.1
    ${csudo}ln -s ${lib_link_dir}/libtaos.so.1 ${lib_link_dir}/libtaos.so

+    [ -f ${lib_dir}/libtaosws.so ]${csudo}ln -s ${lib_dir}/libtaosws.so ${lib_link_dir}/libtaosws.so
+
    if [[ -d ${lib64_link_dir} && ! -e ${lib64_link_dir}/libtaos.so ]]; then
      ${csudo}ln -s ${lib_dir}/libtaos.* ${lib64_link_dir}/libtaos.so.1           || :
      ${csudo}ln -s ${lib64_link_dir}/libtaos.so.1 ${lib64_link_dir}/libtaos.so   || :
+
+      [ -ff ${lib_dir}/libtaosws.so ] && ${csudo}ln -s ${lib_dir}/libtaosws.so ${lib64_link_dir}/libtaosws.so           || :
    fi

    ${csudo}ldconfig
--- a/source/client/src/clientHb.c
+++ b/source/client/src/clientHb.c
@ -286,13 +286,10 @@ static int32_t hbAsyncCallBack(void *param, SDataBuf *pMsg, int32_t code) {
  if (pInst == NULL || NULL == *pInst) {
    taosThreadMutexUnlock(&appInfo.mutex);
    tscError("cluster not exist, key:%s", key);
-    taosMemoryFreeClear(param);
    tFreeClientHbBatchRsp(&pRsp);
    return -1;
  }

-  taosMemoryFreeClear(param);
-
  if (code != 0) {
    (*pInst)->onlineDnodes = ((*pInst)->totalDnodes ? 0 : -1);
    tscDebug("hb rsp error %s, update server status %d/%d", tstrerror(code), (*pInst)->onlineDnodes, (*pInst)->totalDnodes);
@ -716,6 +713,7 @@ static void *hbThreadFunc(void *param) {
      pInfo->msgInfo.len = tlen;
      pInfo->msgType = TDMT_MND_HEARTBEAT;
      pInfo->param = strdup(pAppHbMgr->key);
+      pInfo->paramFreeFp = taosMemoryFree;      
      pInfo->requestId = generateRequestId();
      pInfo->requestObjRefId = 0;

--- a/source/client/src/clientImpl.c
+++ b/source/client/src/clientImpl.c
@ -29,7 +29,6 @@

 static int32_t       initEpSetFromCfg(const char* firstEp, const char* secondEp, SCorEpSet* pEpSet);
 static SMsgSendInfo* buildConnectMsg(SRequestObj* pRequest);
-static void          destroySendMsgInfo(SMsgSendInfo* pMsgBody);

 static bool stringLengthCheck(const char* str, size_t maxsize) {
  if (str == NULL) {
@ -1215,13 +1214,6 @@ static SMsgSendInfo* buildConnectMsg(SRequestObj* pRequest) {
  return pMsgSendInfo;
 }

-static void destroySendMsgInfo(SMsgSendInfo* pMsgBody) {
-  assert(pMsgBody != NULL);
-  taosMemoryFreeClear(pMsgBody->target.dbFName);
-  taosMemoryFreeClear(pMsgBody->msgInfo.pData);
-  taosMemoryFreeClear(pMsgBody);
-}
-
 void updateTargetEpSet(SMsgSendInfo* pSendInfo, STscObj* pTscObj, SRpcMsg* pMsg, SEpSet* pEpSet) {
  if (NULL == pEpSet) {
    return;
@ -2101,8 +2093,9 @@ TAOS_RES* taosQueryImpl(TAOS* taos, const char* sql, bool validateOnly) {

  taosAsyncQueryImpl(*(int64_t*)taos, sql, syncQueryFn, param, validateOnly);
  tsem_wait(&param->sem);
-
-  param->pRequest->syncQuery = true;
+  if (param->pRequest != NULL) {
+    param->pRequest->syncQuery = true;
+  }
  return param->pRequest;
 #else
  size_t sqlLen = strlen(sql);
--- a/source/client/src/clientMsgHandler.c
+++ b/source/client/src/clientMsgHandler.c
@ -255,6 +255,8 @@ int32_t processDropDbRsp(void* param, SDataBuf* pMsg, int32_t code) {
    catalogRemoveDB(pCatalog, dropdbRsp.db, dropdbRsp.uid);
  }

+  taosMemoryFree(pMsg->pData);
+
  if (pRequest->body.queryFp != NULL) {
    pRequest->body.queryFp(pRequest->body.param, pRequest, code);
  } else {
@ -278,6 +280,8 @@ int32_t processAlterStbRsp(void* param, SDataBuf* pMsg, int32_t code) {
    pRequest->body.resInfo.execRes.res = alterRsp.pMeta;
  }

+  taosMemoryFree(pMsg->pData);
+
  if (pRequest->body.queryFp != NULL) {
    SExecResult* pRes = &pRequest->body.resInfo.execRes;

@ -387,6 +391,8 @@ int32_t processShowVariablesRsp(void* param, SDataBuf* pMsg, int32_t code) {
    tFreeSShowVariablesRsp(&rsp);
  }

+  taosMemoryFree(pMsg->pData);
+
  if (pRequest->body.queryFp != NULL) {
    pRequest->body.queryFp(pRequest->body.param, pRequest, code);
  } else {
--- a/source/client/src/clientSml.c
+++ b/source/client/src/clientSml.c
@ -268,7 +268,7 @@ static int32_t smlGenerateSchemaAction(SSchema *colField, SHashObj *colHash, SSm
    *actionNeeded = true;
  }
  if (*actionNeeded) {
-    uDebug("SML:0x%" PRIx64 " generate schema action. column name: %s, action: %d", info->id, colField->name,
+    uDebug("SML:0x%" PRIx64 " generate schema action. kv->name: %s, action: %d", info->id, kv->key,
           action->action);
  }
  return 0;
@ -436,6 +436,7 @@ static int32_t smlProcessSchemaAction(SSmlHandle *info, SSchema *schemaField, SH
                                      SSchemaAction *action, bool isTag) {
  int32_t code = TSDB_CODE_SUCCESS;
  for (int j = 0; j < taosArrayGetSize(cols); ++j) {
+    if(j == 0 && !isTag) continue;
    SSmlKv *kv = (SSmlKv *)taosArrayGetP(cols, j);
    bool    actionNeeded = false;
    code = smlGenerateSchemaAction(schemaField, schemaHash, kv, isTag, action, &actionNeeded, info);
@ -452,18 +453,25 @@ static int32_t smlProcessSchemaAction(SSmlHandle *info, SSchema *schemaField, SH
  return TSDB_CODE_SUCCESS;
 }

-static int32_t smlCheckMeta(SSchema *schema, int32_t length, SArray *cols) {
+static int32_t smlCheckMeta(SSchema *schema, int32_t length, SArray *cols, bool isTag) {
  SHashObj *hashTmp = taosHashInit(length, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BINARY), true, HASH_NO_LOCK);
-  for (uint16_t i = 0; i < length; i++) {
+  int32_t i = 0;
+  for ( ;i < length; i++) {
    taosHashPut(hashTmp, schema[i].name, strlen(schema[i].name), &i, SHORT_BYTES);
  }

-  for (int32_t i = 0; i < taosArrayGetSize(cols); i++) {
+  if (isTag){
+    i = 0;
+  } else {
+    i = 1;
+  }
+  for (; i < taosArrayGetSize(cols); i++) {
    SSmlKv *kv = (SSmlKv *)taosArrayGetP(cols, i);
    if (taosHashGet(hashTmp, kv->key, kv->keyLen) == NULL) {
      return -1;
    }
  }
+  taosHashCleanup(hashTmp);
  return 0;
 }

@ -523,7 +531,7 @@ static int32_t smlModifyDBSchemas(SSmlHandle *info) {
      }

      taosHashClear(hashTmp);
-      for (uint16_t i = 0; i < pTableMeta->tableInfo.numOfColumns; i++) {
+      for (uint16_t i = 1; i < pTableMeta->tableInfo.numOfColumns; i++) {
        taosHashPut(hashTmp, pTableMeta->schema[i].name, strlen(pTableMeta->schema[i].name), &i, SHORT_BYTES);
      }
      code = smlProcessSchemaAction(info, pTableMeta->schema, hashTmp, sTableData->cols, &schemaAction, false);
@ -551,12 +559,12 @@ static int32_t smlModifyDBSchemas(SSmlHandle *info) {

    if (needCheckMeta) {
      code = smlCheckMeta(&(pTableMeta->schema[pTableMeta->tableInfo.numOfColumns]), pTableMeta->tableInfo.numOfTags,
-                          sTableData->tags);
+                          sTableData->tags, true);
      if (code != TSDB_CODE_SUCCESS) {
        uError("SML:0x%" PRIx64 " check tag failed. super table name %s", info->id, (char *)superTable);
        goto end;
      }
-      code = smlCheckMeta(&(pTableMeta->schema[0]), pTableMeta->tableInfo.numOfColumns, sTableData->cols);
+      code = smlCheckMeta(&(pTableMeta->schema[0]), pTableMeta->tableInfo.numOfColumns, sTableData->cols, false);
      if (code != TSDB_CODE_SUCCESS) {
        uError("SML:0x%" PRIx64 " check cols failed. super table name %s", info->id, (char *)superTable);
        goto end;
@ -832,6 +840,7 @@ static int64_t smlParseOpenTsdbTime(SSmlHandle *info, const char *data, int32_t
 static int32_t smlParseTS(SSmlHandle *info, const char *data, int32_t len, SArray *cols) {
  int64_t ts = 0;
  if (info->protocol == TSDB_SML_LINE_PROTOCOL) {
+//    uError("SML:data:%s,len:%d", data, len);
    ts = smlParseInfluxTime(info, data, len);
  } else if (info->protocol == TSDB_SML_TELNET_PROTOCOL) {
    ts = smlParseOpenTsdbTime(info, data, len);
@ -1208,6 +1217,9 @@ static int32_t smlParseCols(const char *data, int32_t len, SArray *cols, char *c
    kv->value = value;
    kv->length = valueLen;
    if (isTag) {
+      if(valueLen > (TSDB_MAX_NCHAR_LEN - VARSTR_HEADER_SIZE) / TSDB_NCHAR_SIZE){
+        return TSDB_CODE_PAR_INVALID_VAR_COLUMN_LEN;
+      }
      kv->type = TSDB_DATA_TYPE_NCHAR;
    } else {
      int32_t ret = smlParseValue(kv, msg);
@ -2031,6 +2043,8 @@ static int32_t smlParseJSONString(SSmlHandle *info, cJSON *root, SSmlTableInfo *

 static int32_t smlParseInfluxLine(SSmlHandle *info, const char *sql) {
  SSmlLineInfo elements = {0};
+  uDebug("SML:0x%" PRIx64 " smlParseInfluxLine sql:%s, hello", info->id, sql);
+
  int          ret = smlParseInfluxString(sql, &elements, &info->msgBuf);
  if (ret != TSDB_CODE_SUCCESS) {
    uError("SML:0x%" PRIx64 " smlParseInfluxLine failed", info->id);
--- a/source/client/src/tmq.c
+++ b/source/client/src/tmq.c
@ -113,12 +113,6 @@ struct tmq_t {
  tsem_t rspSem;
 };

-struct tmq_raw_data {
-  void*   raw_meta;
-  int32_t raw_meta_len;
-  int16_t raw_meta_type;
-};
-
 enum {
  TMQ_VG_STATUS__IDLE = 0,
  TMQ_VG_STATUS__WAIT,
@ -510,6 +504,7 @@ static int32_t tmqSendCommitReq(tmq_t* tmq, SMqClientVg* pVg, SMqClientTopic* pT
  pMsgSendInfo->requestId = generateRequestId();
  pMsgSendInfo->requestObjRefId = 0;
  pMsgSendInfo->param = pParam;
+  pMsgSendInfo->paramFreeFp = taosMemoryFree;      
  pMsgSendInfo->fp = tmqCommitCb2;
  pMsgSendInfo->msgType = TDMT_VND_MQ_COMMIT_OFFSET;
  // send msg
@ -1918,16 +1913,15 @@ const char* tmq_get_table_name(TAOS_RES* res) {
  return NULL;
 }

-tmq_raw_data* tmq_get_raw_meta(TAOS_RES* res) {
-  if (TD_RES_TMQ_META(res)) {
-    tmq_raw_data*  raw = taosMemoryCalloc(1, sizeof(tmq_raw_data));
+int32_t tmq_get_raw_meta(TAOS_RES* res, tmq_raw_data *raw) {
+  if (TD_RES_TMQ_META(res) && raw) {
    SMqMetaRspObj* pMetaRspObj = (SMqMetaRspObj*)res;
    raw->raw_meta = pMetaRspObj->metaRsp.metaRsp;
    raw->raw_meta_len = pMetaRspObj->metaRsp.metaRspLen;
    raw->raw_meta_type = pMetaRspObj->metaRsp.resMsgType;
-    return raw;
+    return TSDB_CODE_SUCCESS;
  }
-  return NULL;
+  return TSDB_CODE_INVALID_PARA;
 }

 static char* buildCreateTableJson(SSchemaWrapper* schemaRow, SSchemaWrapper* schemaTag, char* name, int64_t id,
@ -2199,17 +2193,19 @@ static char* buildCreateCTableJson(STag* pTag, char* sname, char* name, SArray*
    cJSON* ttype = cJSON_CreateNumber(pTagVal->type);
    cJSON_AddItemToObject(tag, "type", ttype);

-    char* buf = NULL;
+    cJSON* tvalue = NULL;
    if (IS_VAR_DATA_TYPE(pTagVal->type)) {
-      buf = taosMemoryCalloc(pTagVal->nData + 1, 1);
+      char* buf = taosMemoryCalloc(pTagVal->nData + 3, 1);
+      if(!buf) goto end;
      dataConverToStr(buf, pTagVal->type, pTagVal->pData, pTagVal->nData, NULL);
+      tvalue = cJSON_CreateString(buf);
+      taosMemoryFree(buf);
    } else {
-      buf = taosMemoryCalloc(32, 1);
-      dataConverToStr(buf, pTagVal->type, &pTagVal->i64, tDataTypes[pTagVal->type].bytes, NULL);
+      double val = 0;
+      GET_TYPED_DATA(val, double, pTagVal->type, &pTagVal->i64);
+      tvalue = cJSON_CreateNumber(val);
    }

-    cJSON* tvalue = cJSON_CreateString(buf);
-    taosMemoryFree(buf);
    cJSON_AddItemToObject(tag, "value", tvalue);
    cJSON_AddItemToArray(tags, tag);
  }
@ -2935,23 +2931,23 @@ end:
  return code;
 }

-int32_t taos_write_raw_meta(TAOS* taos, tmq_raw_data* raw_meta) {
-  if (!taos || !raw_meta) {
+int32_t taos_write_raw_meta(TAOS *taos, tmq_raw_data raw_meta){
+  if (!taos) {
    return TSDB_CODE_INVALID_PARA;
  }

-  if (raw_meta->raw_meta_type == TDMT_VND_CREATE_STB) {
-    return taosCreateStb(taos, raw_meta->raw_meta, raw_meta->raw_meta_len);
-  } else if (raw_meta->raw_meta_type == TDMT_VND_ALTER_STB) {
-    return taosCreateStb(taos, raw_meta->raw_meta, raw_meta->raw_meta_len);
-  } else if (raw_meta->raw_meta_type == TDMT_VND_DROP_STB) {
-    return taosDropStb(taos, raw_meta->raw_meta, raw_meta->raw_meta_len);
-  } else if (raw_meta->raw_meta_type == TDMT_VND_CREATE_TABLE) {
-    return taosCreateTable(taos, raw_meta->raw_meta, raw_meta->raw_meta_len);
-  } else if (raw_meta->raw_meta_type == TDMT_VND_ALTER_TABLE) {
-    return taosAlterTable(taos, raw_meta->raw_meta, raw_meta->raw_meta_len);
-  } else if (raw_meta->raw_meta_type == TDMT_VND_DROP_TABLE) {
-    return taosDropTable(taos, raw_meta->raw_meta, raw_meta->raw_meta_len);
+  if(raw_meta.raw_meta_type == TDMT_VND_CREATE_STB) {
+    return taosCreateStb(taos, raw_meta.raw_meta, raw_meta.raw_meta_len);
+  }else if(raw_meta.raw_meta_type == TDMT_VND_ALTER_STB){
+    return taosCreateStb(taos, raw_meta.raw_meta, raw_meta.raw_meta_len);
+  }else if(raw_meta.raw_meta_type == TDMT_VND_DROP_STB){
+    return taosDropStb(taos, raw_meta.raw_meta, raw_meta.raw_meta_len);
+  }else if(raw_meta.raw_meta_type == TDMT_VND_CREATE_TABLE){
+    return taosCreateTable(taos, raw_meta.raw_meta, raw_meta.raw_meta_len);
+  }else if(raw_meta.raw_meta_type == TDMT_VND_ALTER_TABLE){
+    return taosAlterTable(taos, raw_meta.raw_meta, raw_meta.raw_meta_len);
+  }else if(raw_meta.raw_meta_type == TDMT_VND_DROP_TABLE){
+    return taosDropTable(taos, raw_meta.raw_meta, raw_meta.raw_meta_len);
  }
  return TSDB_CODE_INVALID_PARA;
 }
--- a/source/client/test/smlTest.cpp
+++ b/source/client/test/smlTest.cpp
--- a/source/common/src/tglobal.c
+++ b/source/common/src/tglobal.c
@ -556,7 +556,7 @@ static void taosSetSystemCfg(SConfig *pCfg) {
  osSetSystemLocale(locale, charset);

  bool enableCore = cfgGetItem(pCfg, "enableCoreFile")->bval;
-  taosSetConsoleEcho(enableCore);
+  taosSetCoreDump(enableCore);

  // todo
  tsVersion = 30000000;
@ -675,7 +675,7 @@ int32_t taosSetCfg(SConfig *pCfg, char *name) {
    case 'e': {
      if (strcasecmp("enableCoreFile", name) == 0) {
        bool enableCore = cfgGetItem(pCfg, "enableCoreFile")->bval;
-        taosSetConsoleEcho(enableCore);
+        taosSetCoreDump(enableCore);
      }
      break;
    }
--- a/source/dnode/mgmt/mgmt_dnode/src/dmHandle.c
+++ b/source/dnode/mgmt/mgmt_dnode/src/dmHandle.c
@ -49,9 +49,9 @@ static void dmProcessStatusRsp(SDnodeMgmt *pMgmt, SRpcMsg *pRsp) {
        dmUpdateEps(pMgmt->pData, statusRsp.pDnodeEps);
      }
    }
-    rpcFreeCont(pRsp->pCont);
    tFreeSStatusRsp(&statusRsp);
  }
+  rpcFreeCont(pRsp->pCont);
 }

 void dmSendStatusReq(SDnodeMgmt *pMgmt) {
--- a/source/dnode/mnode/impl/inc/mndDef.h
+++ b/source/dnode/mnode/impl/inc/mndDef.h
@ -35,6 +35,46 @@
 extern "C" {
 #endif

+typedef enum {
+  MND_OPER_CONNECT = 1,
+  MND_OPER_CREATE_ACCT,
+  MND_OPER_DROP_ACCT,
+  MND_OPER_ALTER_ACCT,
+  MND_OPER_CREATE_USER,
+  MND_OPER_DROP_USER,
+  MND_OPER_ALTER_USER,
+  MND_OPER_CREATE_BNODE,
+  MND_OPER_DROP_BNODE,
+  MND_OPER_CREATE_DNODE,
+  MND_OPER_DROP_DNODE,
+  MND_OPER_CONFIG_DNODE,
+  MND_OPER_CREATE_MNODE,
+  MND_OPER_DROP_MNODE,
+  MND_OPER_CREATE_QNODE,
+  MND_OPER_DROP_QNODE,
+  MND_OPER_CREATE_SNODE,
+  MND_OPER_DROP_SNODE,
+  MND_OPER_REDISTRIBUTE_VGROUP,
+  MND_OPER_MERGE_VGROUP,
+  MND_OPER_SPLIT_VGROUP,
+  MND_OPER_BALANCE_VGROUP,
+  MND_OPER_CREATE_FUNC,
+  MND_OPER_DROP_FUNC,
+  MND_OPER_KILL_TRANS,
+  MND_OPER_KILL_CONN,
+  MND_OPER_KILL_QUERY,
+  MND_OPER_CREATE_DB,
+  MND_OPER_ALTER_DB,
+  MND_OPER_DROP_DB,
+  MND_OPER_COMPACT_DB,
+  MND_OPER_TRIM_DB,
+  MND_OPER_USE_DB,
+  MND_OPER_WRITE_DB,
+  MND_OPER_READ_DB,
+  MND_OPER_READ_OR_WRITE_DB,
+  MND_OPER_SHOW_VARIBALES,
+} EOperType;
+
 typedef enum {
  MND_AUTH_ACCT_START = 0,
  MND_AUTH_ACCT_USER,
@ -109,9 +149,9 @@ typedef struct {
  ETrnPolicy     policy;
  ETrnConflct    conflict;
  ETrnExec       exec;
+  EOperType      oper;
  int32_t        code;
  int32_t        failedTimes;
-  SRpcHandleInfo rpcInfo;
  void*          rpcRsp;
  int32_t        rpcRspLen;
  int32_t        redoActionPos;
@ -130,6 +170,7 @@ typedef struct {
  int32_t        stopFunc;
  int32_t        paramLen;
  void*          param;
+  SArray*        pRpcArray;
 } STrans;

 typedef struct {
--- a/source/dnode/mnode/impl/inc/mndPrivilege.h
+++ b/source/dnode/mnode/impl/inc/mndPrivilege.h
@ -22,46 +22,6 @@
 extern "C" {
 #endif

-typedef enum {
-  MND_OPER_CONNECT = 1,
-  MND_OPER_CREATE_ACCT,
-  MND_OPER_DROP_ACCT,
-  MND_OPER_ALTER_ACCT,
-  MND_OPER_CREATE_USER,
-  MND_OPER_DROP_USER,
-  MND_OPER_ALTER_USER,
-  MND_OPER_CREATE_BNODE,
-  MND_OPER_DROP_BNODE,
-  MND_OPER_CREATE_DNODE,
-  MND_OPER_DROP_DNODE,
-  MND_OPER_CONFIG_DNODE,
-  MND_OPER_CREATE_MNODE,
-  MND_OPER_DROP_MNODE,
-  MND_OPER_CREATE_QNODE,
-  MND_OPER_DROP_QNODE,
-  MND_OPER_CREATE_SNODE,
-  MND_OPER_DROP_SNODE,
-  MND_OPER_REDISTRIBUTE_VGROUP,
-  MND_OPER_MERGE_VGROUP,
-  MND_OPER_SPLIT_VGROUP,
-  MND_OPER_BALANCE_VGROUP,
-  MND_OPER_CREATE_FUNC,
-  MND_OPER_DROP_FUNC,
-  MND_OPER_KILL_TRANS,
-  MND_OPER_KILL_CONN,
-  MND_OPER_KILL_QUERY,
-  MND_OPER_CREATE_DB,
-  MND_OPER_ALTER_DB,
-  MND_OPER_DROP_DB,
-  MND_OPER_COMPACT_DB,
-  MND_OPER_TRIM_DB,
-  MND_OPER_USE_DB,
-  MND_OPER_WRITE_DB,
-  MND_OPER_READ_DB,
-  MND_OPER_READ_OR_WRITE_DB,
-  MND_OPER_SHOW_VARIBALES,
-} EOperType;
-
 int32_t mndInitPrivilege(SMnode *pMnode);
 void    mndCleanupPrivilege(SMnode *pMnode);

--- a/source/dnode/mnode/impl/inc/mndTrans.h
+++ b/source/dnode/mnode/impl/inc/mndTrans.h
@ -73,12 +73,14 @@ void    mndTransSetRpcRsp(STrans *pTrans, void *pCont, int32_t contLen);
 void    mndTransSetCb(STrans *pTrans, ETrnFunc startFunc, ETrnFunc stopFunc, void *param, int32_t paramLen);
 void    mndTransSetDbName(STrans *pTrans, const char *dbname1, const char *dbname2);
 void    mndTransSetSerial(STrans *pTrans);
+void    mndTransSetOper(STrans *pTrans, EOperType oper);

 int32_t mndTransPrepare(SMnode *pMnode, STrans *pTrans);
 int32_t mndTransProcessRsp(SRpcMsg *pRsp);
 void    mndTransPullup(SMnode *pMnode);
 int32_t mndKillTrans(SMnode *pMnode, STrans *pTrans);
 void    mndTransExecute(SMnode *pMnode, STrans *pTrans);
+int32_t mndSetRpcInfoForDbTrans(SMnode *pMnode, SRpcMsg *pMsg, EOperType oper, const char *dbname);

 #ifdef __cplusplus
 }
--- a/source/dnode/mnode/impl/src/mndDb.c
+++ b/source/dnode/mnode/impl/src/mndDb.c
@ -487,6 +487,7 @@ static int32_t mndCreateDb(SMnode *pMnode, SRpcMsg *pReq, SCreateDbReq *pCreate,
  mDebug("trans:%d, used to create db:%s", pTrans->id, pCreate->db);

  mndTransSetDbName(pTrans, dbObj.name, NULL);
+  mndTransSetOper(pTrans, MND_OPER_CREATE_DB);
  if (mndSetCreateDbRedoLogs(pMnode, pTrans, &dbObj, pVgroups) != 0) goto _OVER;
  if (mndSetCreateDbUndoLogs(pMnode, pTrans, &dbObj, pVgroups) != 0) goto _OVER;
  if (mndSetCreateDbCommitLogs(pMnode, pTrans, &dbObj, pVgroups) != 0) goto _OVER;
@ -534,6 +535,14 @@ static int32_t mndProcessCreateDbReq(SRpcMsg *pReq) {
      terrno = TSDB_CODE_MND_DB_ALREADY_EXIST;
      goto _OVER;
    }
+  } else if (terrno == TSDB_CODE_SDB_OBJ_CREATING) {
+    if (mndSetRpcInfoForDbTrans(pMnode, pReq, MND_OPER_CREATE_DB, createReq.db) == 0) {
+      mDebug("db:%s, is creating and response after trans finished", createReq.db);
+      code = TSDB_CODE_ACTION_IN_PROGRESS;
+      goto _OVER;
+    } else {
+      goto _OVER;
+    }
  } else if (terrno != TSDB_CODE_MND_DB_NOT_EXIST) {
    goto _OVER;
  }
--- a/source/dnode/mnode/impl/src/mndScheduler.c
+++ b/source/dnode/mnode/impl/src/mndScheduler.c
@ -383,6 +383,11 @@ int32_t mndScheduleStream(SMnode* pMnode, SStreamObj* pStream) {
      // exec
      pInnerTask->execType = TASK_EXEC__PIPE;

+      SDbObj* pSourceDb = mndAcquireDb(pMnode, pStream->sourceDb);
+      ASSERT(pDbObj != NULL);
+      sdbRelease(pSdb, pSourceDb);
+      pInnerTask->numOfVgroups = pSourceDb->cfg.numOfVgroups;
+
      if (tsSchedStreamToSnode) {
        SSnodeObj* pSnode = mndSchedFetchOneSnode(pMnode);
        if (pSnode == NULL) {
--- a/source/dnode/mnode/impl/src/mndSync.c
+++ b/source/dnode/mnode/impl/src/mndSync.c
@ -144,9 +144,11 @@ int32_t mndSnapshotStartWrite(struct SSyncFSM *pFsm, void *pParam, void **ppWrit
 }

 int32_t mndSnapshotStopWrite(struct SSyncFSM *pFsm, void *pWriter, bool isApply, SSnapshot *pSnapshot) {
-  mInfo("stop to apply snapshot to sdb, apply:%d", isApply);
+  mInfo("stop to apply snapshot to sdb, apply:%d, index:%" PRId64 " term:%" PRIu64 " config:%" PRId64, isApply,
+        pSnapshot->lastApplyIndex, pSnapshot->lastApplyTerm, pSnapshot->lastApplyIndex);
  SMnode *pMnode = pFsm->data;
-  return sdbStopWrite(pMnode->pSdb, pWriter, isApply);
+  return sdbStopWrite(pMnode->pSdb, pWriter, isApply, pSnapshot->lastApplyIndex, pSnapshot->lastApplyTerm,
+                      pSnapshot->lastConfigIndex);
 }

 int32_t mndSnapshotDoWrite(struct SSyncFSM *pFsm, void *pWriter, void *pBuf, int32_t len) {
@ -157,7 +159,7 @@ int32_t mndSnapshotDoWrite(struct SSyncFSM *pFsm, void *pWriter, void *pBuf, int
 void mndLeaderTransfer(struct SSyncFSM *pFsm, const SRpcMsg *pMsg, SFsmCbMeta cbMeta) {
  SMnode *pMnode = pFsm->data;
  atomic_store_8(&(pMnode->syncMgmt.leaderTransferFinish), 1);
-  mDebug("vgId:1, mnd leader transfer finish");
+  mDebug("vgId:1, mnode leader transfer finish");
 }

 SSyncFSM *mndSyncMakeFsm(SMnode *pMnode) {
--- a/source/dnode/mnode/impl/src/mndTrans.c
+++ b/source/dnode/mnode/impl/src/mndTrans.c
@ -122,6 +122,10 @@ static SSdbRaw *mndTransActionEncode(STrans *pTrans) {
  SDB_SET_INT8(pRaw, dataPos, pTrans->policy, _OVER)
  SDB_SET_INT8(pRaw, dataPos, pTrans->conflict, _OVER)
  SDB_SET_INT8(pRaw, dataPos, pTrans->exec, _OVER)
+  SDB_SET_INT8(pRaw, dataPos, pTrans->oper, _OVER)
+  SDB_SET_INT8(pRaw, dataPos, 0, _OVER)
+  SDB_SET_INT8(pRaw, dataPos, 0, _OVER)
+  SDB_SET_INT8(pRaw, dataPos, 0, _OVER)
  SDB_SET_INT64(pRaw, dataPos, pTrans->createdTime, _OVER)
  SDB_SET_BINARY(pRaw, dataPos, pTrans->dbname1, TSDB_DB_FNAME_LEN, _OVER)
  SDB_SET_BINARY(pRaw, dataPos, pTrans->dbname2, TSDB_DB_FNAME_LEN, _OVER)
@ -269,15 +273,22 @@ static SSdbRow *mndTransActionDecode(SSdbRaw *pRaw) {
  int8_t policy = 0;
  int8_t conflict = 0;
  int8_t exec = 0;
+  int8_t oper = 0;
+  int8_t reserved = 0;
  int8_t actionType = 0;
  SDB_GET_INT8(pRaw, dataPos, &stage, _OVER)
  SDB_GET_INT8(pRaw, dataPos, &policy, _OVER)
  SDB_GET_INT8(pRaw, dataPos, &conflict, _OVER)
  SDB_GET_INT8(pRaw, dataPos, &exec, _OVER)
+  SDB_GET_INT8(pRaw, dataPos, &oper, _OVER)
+  SDB_GET_INT8(pRaw, dataPos, &reserved, _OVER)
+  SDB_GET_INT8(pRaw, dataPos, &reserved, _OVER)
+  SDB_GET_INT8(pRaw, dataPos, &reserved, _OVER)
  pTrans->stage = stage;
  pTrans->policy = policy;
  pTrans->conflict = conflict;
  pTrans->exec = exec;
+  pTrans->oper = oper;
  SDB_GET_INT64(pRaw, dataPos, &pTrans->createdTime, _OVER)
  SDB_GET_BINARY(pRaw, dataPos, pTrans->dbname1, TSDB_DB_FNAME_LEN, _OVER)
  SDB_GET_BINARY(pRaw, dataPos, pTrans->dbname2, TSDB_DB_FNAME_LEN, _OVER)
@ -495,6 +506,10 @@ static void mndTransDropData(STrans *pTrans) {
    mndTransDropActions(pTrans->commitActions);
    pTrans->commitActions = NULL;
  }
+  if (pTrans->pRpcArray != NULL) {
+    taosArrayDestroy(pTrans->pRpcArray);
+    pTrans->pRpcArray = NULL;
+  }
  if (pTrans->rpcRsp != NULL) {
    taosMemoryFree(pTrans->rpcRsp);
    pTrans->rpcRsp = NULL;
@ -585,14 +600,18 @@ STrans *mndTransCreate(SMnode *pMnode, ETrnPolicy policy, ETrnConflct conflict,
  pTrans->redoActions = taosArrayInit(TRANS_ARRAY_SIZE, sizeof(STransAction));
  pTrans->undoActions = taosArrayInit(TRANS_ARRAY_SIZE, sizeof(STransAction));
  pTrans->commitActions = taosArrayInit(TRANS_ARRAY_SIZE, sizeof(STransAction));
+  pTrans->pRpcArray = taosArrayInit(1, sizeof(SRpcHandleInfo));

-  if (pTrans->redoActions == NULL || pTrans->undoActions == NULL || pTrans->commitActions == NULL) {
+  if (pTrans->redoActions == NULL || pTrans->undoActions == NULL || pTrans->commitActions == NULL ||
+      pTrans->pRpcArray == NULL) {
    terrno = TSDB_CODE_OUT_OF_MEMORY;
    mError("failed to create transaction since %s", terrstr());
    return NULL;
  }

-  if (pReq != NULL) pTrans->rpcInfo = pReq->info;
+  if (pReq != NULL) {
+    taosArrayPush(pTrans->pRpcArray, &pReq->info);
+  }
  mTrace("trans:%d, local object is created, data:%p", pTrans->id, pTrans);
  return pTrans;
 }
@ -677,6 +696,31 @@ void mndTransSetCb(STrans *pTrans, ETrnFunc startFunc, ETrnFunc stopFunc, void *
  pTrans->paramLen = paramLen;
 }

+int32_t mndSetRpcInfoForDbTrans(SMnode *pMnode, SRpcMsg *pMsg, EOperType oper, const char *dbname) {
+  STrans *pTrans = NULL;
+  void   *pIter = NULL;
+  int32_t code = -1;
+
+  while (1) {
+    pIter = sdbFetch(pMnode->pSdb, SDB_TRANS, pIter, (void **)&pTrans);
+    if (pIter == NULL) break;
+
+    if (pTrans->oper == oper) {
+      if (strcasecmp(dbname, pTrans->dbname1) == 0) {
+        mDebug("trans:%d, db:%s oper:%d matched with input", pTrans->id, dbname, oper);
+        if (taosArrayPush(pTrans->pRpcArray, &pMsg->info) != NULL) {
+          code = 0;
+        }
+        sdbRelease(pMnode->pSdb, pTrans);
+        break;
+      }
+    }
+
+    sdbRelease(pMnode->pSdb, pTrans);
+  }
+  return code;
+}
+
 void mndTransSetDbName(STrans *pTrans, const char *dbname1, const char *dbname2) {
  if (dbname1 != NULL) {
    memcpy(pTrans->dbname1, dbname1, TSDB_DB_FNAME_LEN);
@ -688,6 +732,8 @@ void mndTransSetDbName(STrans *pTrans, const char *dbname1, const char *dbname2)

 void mndTransSetSerial(STrans *pTrans) { pTrans->exec = TRN_EXEC_SERIAL; }

+void mndTransSetOper(STrans *pTrans, EOperType oper) { pTrans->oper = oper; }
+
 static int32_t mndTransSync(SMnode *pMnode, STrans *pTrans) {
  SSdbRaw *pRaw = mndTransActionEncode(pTrans);
  if (pRaw == NULL) {
@ -711,7 +757,7 @@ static int32_t mndTransSync(SMnode *pMnode, STrans *pTrans) {

 static bool mndCheckDbConflict(const char *db, STrans *pTrans) {
  if (db[0] == 0) return false;
-  if (strcmp(db, pTrans->dbname1) == 0 || strcmp(db, pTrans->dbname2) == 0) return true;
+  if (strcasecmp(db, pTrans->dbname1) == 0 || strcasecmp(db, pTrans->dbname2) == 0) return true;
  return false;
 }

@ -784,9 +830,10 @@ int32_t mndTransPrepare(SMnode *pMnode, STrans *pTrans) {
    return -1;
  }

-  pNew->rpcInfo = pTrans->rpcInfo;
+  pNew->pRpcArray = pTrans->pRpcArray;
  pNew->rpcRsp = pTrans->rpcRsp;
  pNew->rpcRspLen = pTrans->rpcRspLen;
+  pTrans->pRpcArray = NULL;
  pTrans->rpcRsp = NULL;
  pTrans->rpcRspLen = 0;

@ -835,29 +882,34 @@ static void mndTransSendRpcRsp(SMnode *pMnode, STrans *pTrans) {
    }
  }

-  if (sendRsp && pTrans->rpcInfo.handle != NULL) {
-    mDebug("trans:%d, send rsp, code:0x%x stage:%s app:%p", pTrans->id, code, mndTransStr(pTrans->stage),
-           pTrans->rpcInfo.ahandle);
-    if (code == TSDB_CODE_RPC_NETWORK_UNAVAIL) {
-      code = TSDB_CODE_MND_TRANS_NETWORK_UNAVAILL;
-    }
-    SRpcMsg rspMsg = {.code = code, .info = pTrans->rpcInfo};
+  if (!sendRsp) return;

-    if (pTrans->rpcRspLen != 0) {
-      void *rpcCont = rpcMallocCont(pTrans->rpcRspLen);
-      if (rpcCont != NULL) {
-        memcpy(rpcCont, pTrans->rpcRsp, pTrans->rpcRspLen);
-        rspMsg.pCont = rpcCont;
-        rspMsg.contLen = pTrans->rpcRspLen;
+  int32_t size = taosArrayGetSize(pTrans->pRpcArray);
+  if (size <= 0) return;
+
+  for (int32_t i = 0; i < size; ++i) {
+    SRpcHandleInfo *pInfo = taosArrayGet(pTrans->pRpcArray, i);
+    if (pInfo->handle != NULL) {
+      mDebug("trans:%d, send rsp, code:0x%x stage:%s app:%p", pTrans->id, code, mndTransStr(pTrans->stage),
+             pInfo->ahandle);
+      if (code == TSDB_CODE_RPC_NETWORK_UNAVAIL) {
+        code = TSDB_CODE_MND_TRANS_NETWORK_UNAVAILL;
      }
-      taosMemoryFree(pTrans->rpcRsp);
-    }
+      SRpcMsg rspMsg = {.code = code, .info = *pInfo};

-    tmsgSendRsp(&rspMsg);
-    pTrans->rpcInfo.handle = NULL;
-    pTrans->rpcRsp = NULL;
-    pTrans->rpcRspLen = 0;
+      if (pTrans->rpcRspLen != 0) {
+        void *rpcCont = rpcMallocCont(pTrans->rpcRspLen);
+        if (rpcCont != NULL) {
+          memcpy(rpcCont, pTrans->rpcRsp, pTrans->rpcRspLen);
+          rspMsg.pCont = rpcCont;
+          rspMsg.contLen = pTrans->rpcRspLen;
+        }
+      }
+
+      tmsgSendRsp(&rspMsg);
+    }
  }
+  taosArrayClear(pTrans->pRpcArray);
 }

 int32_t mndTransProcessRsp(SRpcMsg *pRsp) {
--- a/source/dnode/mnode/impl/test/sdb/sdbTest.cpp
+++ b/source/dnode/mnode/impl/test/sdb/sdbTest.cpp
@ -925,7 +925,7 @@ TEST_F(MndTestSdb, 01_Read_Str) {
    }

    sdbStopRead(pSdb, pReader);
-    sdbStopWrite(pSdb, pWritter, true);
+    sdbStopWrite(pSdb, pWritter, true, -1, -1, -1);
  }

  ASSERT_EQ(sdbGetSize(pSdb, SDB_CONSUMER), 1);
--- a/source/dnode/mnode/sdb/inc/sdb.h
+++ b/source/dnode/mnode/sdb/inc/sdb.h
@ -394,7 +394,7 @@ int32_t sdbStopRead(SSdb *pSdb, SSdbIter *pIter);
 int32_t sdbDoRead(SSdb *pSdb, SSdbIter *pIter, void **ppBuf, int32_t *len);

 int32_t sdbStartWrite(SSdb *pSdb, SSdbIter **ppIter);
-int32_t sdbStopWrite(SSdb *pSdb, SSdbIter *pIter, bool isApply);
+int32_t sdbStopWrite(SSdb *pSdb, SSdbIter *pIter, bool isApply, int64_t index, int64_t term, int64_t config);
 int32_t sdbDoWrite(SSdb *pSdb, SSdbIter *pIter, void *pBuf, int32_t len);

 const char *sdbTableName(ESdbType type);
--- a/source/dnode/mnode/sdb/src/sdbFile.c
+++ b/source/dnode/mnode/sdb/src/sdbFile.c
@ -613,7 +613,7 @@ int32_t sdbStartWrite(SSdb *pSdb, SSdbIter **ppIter) {
  return 0;
 }

-int32_t sdbStopWrite(SSdb *pSdb, SSdbIter *pIter, bool isApply) {
+int32_t sdbStopWrite(SSdb *pSdb, SSdbIter *pIter, bool isApply, int64_t index, int64_t term, int64_t config) {
  int32_t code = 0;

  if (!isApply) {
@ -641,6 +641,19 @@ int32_t sdbStopWrite(SSdb *pSdb, SSdbIter *pIter, bool isApply) {
    return -1;
  }

+  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;
+  }
+
  mDebug("sdbiter:%p, successfully applyed to sdb", pIter);
  return 0;
 }
--- a/source/dnode/mnode/sdb/src/sdbHash.c
+++ b/source/dnode/mnode/sdb/src/sdbHash.c
@ -153,23 +153,22 @@ static int32_t sdbInsertRow(SSdb *pSdb, SHashObj *hash, SSdbRaw *pRaw, SSdbRow *
    return terrno;
  }

-  taosThreadRwlockUnlock(pLock);
-
  int32_t     code = 0;
  SdbInsertFp insertFp = pSdb->insertFps[pRow->type];
  if (insertFp != NULL) {
    code = (*insertFp)(pSdb, pRow->pObj);
    if (code != 0) {
      code = terrno;
-      taosThreadRwlockWrlock(pLock);
      taosHashRemove(hash, pRow->pObj, keySize);
-      taosThreadRwlockUnlock(pLock);
      sdbFreeRow(pSdb, pRow, false);
      terrno = code;
+      taosThreadRwlockUnlock(pLock);
      return terrno;
    }
  }

+  taosThreadRwlockUnlock(pLock);
+
  if (pSdb->keyTypes[pRow->type] == SDB_KEY_INT32) {
    pSdb->maxId[pRow->type] = TMAX(pSdb->maxId[pRow->type], *((int32_t *)pRow->pObj));
  }
@ -194,7 +193,6 @@ static int32_t sdbUpdateRow(SSdb *pSdb, SHashObj *hash, SSdbRaw *pRaw, SSdbRow *
  SSdbRow *pOldRow = *ppOldRow;
  pOldRow->status = pRaw->status;
  sdbPrintOper(pSdb, pOldRow, "update");
-  taosThreadRwlockUnlock(pLock);

  int32_t     code = 0;
  SdbUpdateFp updateFp = pSdb->updateFps[pNewRow->type];
@ -202,6 +200,7 @@ static int32_t sdbUpdateRow(SSdb *pSdb, SHashObj *hash, SSdbRaw *pRaw, SSdbRow *
    code = (*updateFp)(pSdb, pOldRow->pObj, pNewRow->pObj);
  }

+  taosThreadRwlockUnlock(pLock);
  sdbFreeRow(pSdb, pNewRow, false);

  pSdb->tableVer[pOldRow->type]++;
--- a/source/dnode/vnode/src/inc/tq.h
+++ b/source/dnode/vnode/src/inc/tq.h
@ -88,7 +88,7 @@ typedef struct {
    STqExecTb  execTb;
    STqExecDb  execDb;
  };
-
+  int32_t numOfCols;  // number of out pout column, temporarily used
 } STqExecHandle;

 typedef struct {
--- a/source/dnode/vnode/src/inc/tsdb.h
+++ b/source/dnode/vnode/src/inc/tsdb.h
@ -174,6 +174,10 @@ void    tsdbCalcColDataSMA(SColData *pColData, SColumnDataAgg *pColAgg);
 int32_t tsdbMemTableCreate(STsdb *pTsdb, SMemTable **ppMemTable);
 void    tsdbMemTableDestroy(SMemTable *pMemTable);
 void    tsdbGetTbDataFromMemTable(SMemTable *pMemTable, tb_uid_t suid, tb_uid_t uid, STbData **ppTbData);
+void    tsdbRefMemTable(SMemTable *pMemTable);
+void    tsdbUnrefMemTable(SMemTable *pMemTable);
+int32_t tsdbTakeMemSnapshot(STsdb *pTsdb, SMemTable **ppMem, SMemTable **ppIMem);
+void    tsdbUntakeMemSnapshot(STsdb *pTsdb, SMemTable *pMem, SMemTable *pIMem);
 // STbDataIter
 int32_t  tsdbTbDataIterCreate(STbData *pTbData, TSDBKEY *pFrom, int8_t backward, STbDataIter **ppIter);
 void    *tsdbTbDataIterDestroy(STbDataIter *pIter);
@ -186,6 +190,7 @@ int32_t tsdbGetNRowsInTbData(STbData *pTbData);
 typedef enum { TSDB_HEAD_FILE = 0, TSDB_DATA_FILE, TSDB_LAST_FILE, TSDB_SMA_FILE } EDataFileT;
 void    tsdbDataFileName(STsdb *pTsdb, SDFileSet *pDFileSet, EDataFileT ftype, char fname[]);
 bool    tsdbFileIsSame(SDFileSet *pDFileSet1, SDFileSet *pDFileSet2, EDataFileT ftype);
+bool    tsdbDelFileIsSame(SDelFile *pDelFile1, SDelFile *pDelFile2);
 int32_t tsdbUpdateDFileHdr(TdFilePtr pFD, SDFileSet *pSet, EDataFileT ftype);
 int32_t tsdbDFileRollback(STsdb *pTsdb, SDFileSet *pSet, EDataFileT ftype);
 int32_t tPutDataFileHdr(uint8_t *p, SDFileSet *pSet, EDataFileT ftype);
@ -272,23 +277,14 @@ typedef struct {
 } SRtn;

 struct STsdb {
-  char         *path;
-  SVnode       *pVnode;
-  TdThreadMutex mutex;
-  bool          repoLocked;
-  STsdbKeepCfg  keepCfg;
-  SMemTable    *mem;
-  SMemTable    *imem;
-  SRtn          rtn;
-  STsdbFS      *fs;
-  SLRUCache    *lruCache;
-};
-
-struct STable {
-  uint64_t  suid;
-  uint64_t  uid;
-  STSchema *pSchema;       // latest schema
-  STSchema *pCacheSchema;  // cached cache
+  char          *path;
+  SVnode        *pVnode;
+  STsdbKeepCfg   keepCfg;
+  TdThreadRwlock rwLock;
+  SMemTable     *mem;
+  SMemTable     *imem;
+  STsdbFS       *pFS;
+  SLRUCache     *lruCache;
 };

 struct TSDBKEY {
@ -329,21 +325,19 @@ struct STbData {
 };

 struct SMemTable {
-  SRWLatch latch;
-  STsdb   *pTsdb;
-  int32_t  nRef;
-  TSKEY    minKey;
-  TSKEY    maxKey;
-  int64_t  minVersion;
-  int64_t  maxVersion;
-  int64_t  nRow;
-  int64_t  nDel;
-  SArray  *aTbData;  // SArray<STbData*>
+  SRWLatch         latch;
+  STsdb           *pTsdb;
+  SVBufPool       *pPool;
+  volatile int32_t nRef;
+  TSKEY            minKey;
+  TSKEY            maxKey;
+  int64_t          minVersion;
+  int64_t          maxVersion;
+  int64_t          nRow;
+  int64_t          nDel;
+  SArray          *aTbData;  // SArray<STbData*>
 };

-int tsdbLockRepo(STsdb *pTsdb);
-int tsdbUnlockRepo(STsdb *pTsdb);
-
 struct TSDBROW {
  int8_t type;  // 0 for row from tsRow, 1 for row from block data
  union {
--- a/source/dnode/vnode/src/inc/vnd.h
+++ b/source/dnode/vnode/src/inc/vnd.h
@ -62,12 +62,13 @@ struct SVBufPoolNode {
 };

 struct SVBufPool {
-  SVBufPool*     next;
-  int64_t        nRef;
-  int64_t        size;
-  uint8_t*       ptr;
-  SVBufPoolNode* pTail;
-  SVBufPoolNode  node;
+  SVBufPool*       next;
+  SVnode*          pVnode;
+  volatile int32_t nRef;
+  int64_t          size;
+  uint8_t*         ptr;
+  SVBufPoolNode*   pTail;
+  SVBufPoolNode    node;
 };

 int32_t vnodeOpenBufPool(SVnode* pVnode, int64_t size);
@ -78,7 +79,7 @@ void    vnodeBufPoolReset(SVBufPool* pPool);
 int32_t vnodeQueryOpen(SVnode* pVnode);
 void    vnodeQueryClose(SVnode* pVnode);
 int32_t vnodeGetTableMeta(SVnode* pVnode, SRpcMsg* pMsg);
-int     vnodeGetTableCfg(SVnode *pVnode, SRpcMsg *pMsg);
+int     vnodeGetTableCfg(SVnode* pVnode, SRpcMsg* pMsg);

 // vnodeCommit.c
 int32_t vnodeBegin(SVnode* pVnode);
--- a/source/dnode/vnode/src/inc/vnodeInt.h
+++ b/source/dnode/vnode/src/inc/vnodeInt.h
@ -77,6 +77,8 @@ typedef struct SSnapDataHdr    SSnapDataHdr;
 // vnd.h
 void* vnodeBufPoolMalloc(SVBufPool* pPool, int size);
 void  vnodeBufPoolFree(SVBufPool* pPool, void* p);
+void  vnodeBufPoolRef(SVBufPool* pPool);
+void  vnodeBufPoolUnRef(SVBufPool* pPool);

 // meta
 typedef struct SMCtbCursor SMCtbCursor;
@ -247,34 +249,34 @@ struct STsdbKeepCfg {
 };

 struct SVnode {
-  char*      path;
-  SVnodeCfg  config;
-  SVState    state;
-  STfs*      pTfs;
-  SMsgCb     msgCb;
-  SVBufPool* pPool;
-  SVBufPool* inUse;
-  SVBufPool* onCommit;
-  SVBufPool* onRecycle;
-  SMeta*     pMeta;
-  SSma*      pSma;
-  STsdb*     pTsdb;
-  SWal*      pWal;
-  STQ*       pTq;
-  SSink*     pSink;
-  tsem_t     canCommit;
-  int64_t    sync;
-  int32_t    blockCount;
-  tsem_t     syncSem;
-  SQHandle*  pQuery;
+  char*         path;
+  SVnodeCfg     config;
+  SVState       state;
+  STfs*         pTfs;
+  SMsgCb        msgCb;
+  TdThreadMutex mutex;
+  TdThreadCond  poolNotEmpty;
+  SVBufPool*    pPool;
+  SVBufPool*    inUse;
+  SMeta*        pMeta;
+  SSma*         pSma;
+  STsdb*        pTsdb;
+  SWal*         pWal;
+  STQ*          pTq;
+  SSink*        pSink;
+  tsem_t        canCommit;
+  int64_t       sync;
+  int32_t       blockCount;
+  tsem_t        syncSem;
+  SQHandle*     pQuery;
 };

 #define TD_VID(PVNODE) ((PVNODE)->config.vgId)

 #define VND_TSDB(vnd)       ((vnd)->pTsdb)
 #define VND_RSMA0(vnd)      ((vnd)->pTsdb)
-#define VND_RSMA1(vnd)      ((vnd)->pSma->pRSmaTsdb1)
-#define VND_RSMA2(vnd)      ((vnd)->pSma->pRSmaTsdb2)
+#define VND_RSMA1(vnd)      ((vnd)->pSma->pRSmaTsdb[TSDB_RETENTION_L0])
+#define VND_RSMA2(vnd)      ((vnd)->pSma->pRSmaTsdb[TSDB_RETENTION_L1])
 #define VND_RETENTIONS(vnd) (&(vnd)->config.tsdbCfg.retentions)
 #define VND_IS_RSMA(v)      ((v)->config.isRsma == 1)
 #define VND_IS_TSMA(v)      ((v)->config.isTsma == 1)
@ -289,8 +291,7 @@ struct SSma {
  bool          locked;
  TdThreadMutex mutex;
  SVnode*       pVnode;
-  STsdb*        pRSmaTsdb1;
-  STsdb*        pRSmaTsdb2;
+  STsdb*        pRSmaTsdb[TSDB_RETENTION_L2];
  void*         pTSmaEnv;
  void*         pRSmaEnv;
 };
@ -305,8 +306,8 @@ struct SSma {
 #define SMA_TSMA_ENV(s)   ((s)->pTSmaEnv)
 #define SMA_RSMA_ENV(s)   ((s)->pRSmaEnv)
 #define SMA_RSMA_TSDB0(s) ((s)->pVnode->pTsdb)
-#define SMA_RSMA_TSDB1(s) ((s)->pRSmaTsdb1)
-#define SMA_RSMA_TSDB2(s) ((s)->pRSmaTsdb2)
+#define SMA_RSMA_TSDB1(s) ((s)->pRSmaTsdb[TSDB_RETENTION_L0])
+#define SMA_RSMA_TSDB2(s) ((s)->pRSmaTsdb[TSDB_RETENTION_L1])

 // sma
 void smaHandleRes(void* pVnode, int64_t smaId, const SArray* data);
--- a/source/dnode/vnode/src/meta/metaTable.c
+++ b/source/dnode/vnode/src/meta/metaTable.c
@ -371,6 +371,13 @@ int metaCreateTable(SMeta *pMeta, int64_t version, SVCreateTbReq *pReq) {
  }
  metaReaderClear(&mr);

+  if (pReq->type == TSDB_CHILD_TABLE) {
+    tb_uid_t suid = metaGetTableEntryUidByName(pMeta, pReq->ctb.name);
+    if (suid == 0) {
+      terrno = TSDB_CODE_PAR_TABLE_NOT_EXIST;
+      return -1;
+    }
+  }
  // build SMetaEntry
  me.version = version;
  me.type = pReq->type;
--- a/source/dnode/vnode/src/sma/smaOpen.c
+++ b/source/dnode/vnode/src/sma/smaOpen.c
@ -123,11 +123,9 @@ int32_t smaOpen(SVnode *pVnode) {
    }

    // restore the rsma
-#if 1
    if (rsmaRestore(pSma) < 0) {
      goto _err;
    }
-#endif
  }

  return 0;
--- a/source/dnode/vnode/src/sma/smaRollup.c
+++ b/source/dnode/vnode/src/sma/smaRollup.c
@ -605,7 +605,7 @@ static int32_t tdRSmaFetchAndSubmitResult(SRSmaInfoItem *pItem, STSchema *pTSche
    snprintf(flag, 10, "level %" PRIi8, pItem->level);
    blockDebugShowDataBlocks(pResult, flag);
 #endif
-    STsdb      *sinkTsdb = (pItem->level == TSDB_RETENTION_L1 ? pSma->pRSmaTsdb1 : pSma->pRSmaTsdb2);
+    STsdb      *sinkTsdb = (pItem->level == TSDB_RETENTION_L1 ? pSma->pRSmaTsdb[0] : pSma->pRSmaTsdb[1]);
    SSubmitReq *pReq = NULL;
    // TODO: the schema update should be handled
    if (buildSubmitReqFromDataBlock(&pReq, pResult, pTSchema, SMA_VID(pSma), suid) < 0) {
@ -949,13 +949,8 @@ _err:
 * @return int32_t
 */
 static int32_t tdRSmaRestoreTSDataReload(SSma *pSma, int64_t committed) {
-  // TODO
-  smaDebug("vgId:%d, rsma restore from %" PRIi64 ", ts data reload success", SMA_VID(pSma), committed);
+  // NOTHING TODO: the data would be restored from the unified WAL replay procedure
  return TSDB_CODE_SUCCESS;
-_err:
-  smaError("vgId:%d, rsma restore from %" PRIi64 ", ts data reload failed since %s", SMA_VID(pSma), committed,
-           terrstr());
-  return TSDB_CODE_FAILED;
 }

 int32_t tdProcessRSmaRestoreImpl(SSma *pSma) {
--- a/source/dnode/vnode/src/tq/tq.c
+++ b/source/dnode/vnode/src/tq/tq.c
@ -506,7 +506,7 @@ int32_t tqProcessVgChangeReq(STQ* pTq, char* msg, int32_t msgLen) {
            .initTqReader = true,
            .version = ver,
        };
-        pHandle->execHandle.execCol.task[i] = qCreateQueueExecTaskInfo(pHandle->execHandle.execCol.qmsg, &handle);
+        pHandle->execHandle.execCol.task[i] = qCreateQueueExecTaskInfo(pHandle->execHandle.execCol.qmsg, &handle, &pHandle->execHandle.numOfCols);
        ASSERT(pHandle->execHandle.execCol.task[i]);
        void* scanner = NULL;
        qExtractStreamScanner(pHandle->execHandle.execCol.task[i], &scanner);
@ -591,7 +591,11 @@ int32_t tqProcessTaskDeployReq(STQ* pTq, char* msg, int32_t msgLen) {
      };
      pTask->exec.executor = qCreateStreamExecTaskInfo(pTask->exec.qmsg, &handle);
    } else {
-      pTask->exec.executor = qCreateStreamExecTaskInfo(pTask->exec.qmsg, NULL);
+      SReadHandle mgHandle = {
+          .vnode = NULL,
+          .numOfVgroups = pTask->numOfVgroups,
+      };
+      pTask->exec.executor = qCreateStreamExecTaskInfo(pTask->exec.qmsg, &mgHandle);
    }
    ASSERT(pTask->exec.executor);
  }
--- a/source/dnode/vnode/src/tq/tqExec.c
+++ b/source/dnode/vnode/src/tq/tqExec.c
@ -15,7 +15,7 @@

 #include "tq.h"

-static int32_t tqAddBlockDataToRsp(const SSDataBlock* pBlock, SMqDataRsp* pRsp) {
+static int32_t tqAddBlockDataToRsp(const SSDataBlock* pBlock, SMqDataRsp* pRsp, int32_t numOfCols) {
  int32_t dataStrLen = sizeof(SRetrieveTableRsp) + blockGetEncodeSize(pBlock);
  void*   buf = taosMemoryCalloc(1, dataStrLen);
  if (buf == NULL) return -1;
@ -29,7 +29,7 @@ static int32_t tqAddBlockDataToRsp(const SSDataBlock* pBlock, SMqDataRsp* pRsp)

  // TODO enable compress
  int32_t actualLen = 0;
-  blockEncode(pBlock, pRetrieve->data, &actualLen, taosArrayGetSize(pBlock->pDataBlock), false);
+  blockEncode(pBlock, pRetrieve->data, &actualLen, numOfCols, false);
  actualLen += sizeof(SRetrieveTableRsp);
  ASSERT(actualLen <= dataStrLen);
  taosArrayPush(pRsp->blockDataLen, &actualLen);
@ -97,7 +97,7 @@ int64_t tqScan(STQ* pTq, const STqHandle* pHandle, SMqDataRsp* pRsp, STqOffsetVa
          pRsp->withTbName = 0;
        }
      }
-      tqAddBlockDataToRsp(pDataBlock, pRsp);
+      tqAddBlockDataToRsp(pDataBlock, pRsp, pExec->numOfCols);
      pRsp->blockNum++;
      if (pOffset->type == TMQ_OFFSET__LOG) {
        continue;
@ -202,7 +202,7 @@ int32_t tqLogScanExec(STQ* pTq, STqExecHandle* pExec, SSubmitReq* pReq, SMqDataR
          continue;
        }
      }
-      tqAddBlockDataToRsp(&block, pRsp);
+      tqAddBlockDataToRsp(&block, pRsp, taosArrayGetSize(block.pDataBlock));
      tqAddBlockSchemaToRsp(pExec, workerId, pRsp);
      pRsp->blockNum++;
    }
@ -221,7 +221,7 @@ int32_t tqLogScanExec(STQ* pTq, STqExecHandle* pExec, SSubmitReq* pReq, SMqDataR
          continue;
        }
      }
-      tqAddBlockDataToRsp(&block, pRsp);
+      tqAddBlockDataToRsp(&block, pRsp, taosArrayGetSize(block.pDataBlock));
      tqAddBlockSchemaToRsp(pExec, workerId, pRsp);
      pRsp->blockNum++;
    }
--- a/source/dnode/vnode/src/tq/tqMeta.c
+++ b/source/dnode/vnode/src/tq/tqMeta.c
@ -92,7 +92,8 @@ int32_t tqMetaOpen(STQ* pTq) {
            .initTqReader = true,
            .version = handle.snapshotVer,
        };
-        handle.execHandle.execCol.task[i] = qCreateQueueExecTaskInfo(handle.execHandle.execCol.qmsg, &reader);
+
+        handle.execHandle.execCol.task[i] = qCreateQueueExecTaskInfo(handle.execHandle.execCol.qmsg, &reader, &handle.execHandle.numOfCols);
        ASSERT(handle.execHandle.execCol.task[i]);
        void* scanner = NULL;
        qExtractStreamScanner(handle.execHandle.execCol.task[i], &scanner);
--- a/source/dnode/vnode/src/tq/tqSink.c
+++ b/source/dnode/vnode/src/tq/tqSink.c
@ -49,8 +49,11 @@ SSubmitReq* tdBlockToSubmit(const SArray* pBlocks, const STSchema* pTSchema, boo
      }

      SVCreateTbReq createTbReq = {0};
+      SName         name = {0};
+      tNameFromString(&name, stbFullName, T_NAME_ACCT | T_NAME_DB | T_NAME_TABLE);
+
      createTbReq.name = buildCtbNameByGroupId(stbFullName, pDataBlock->info.groupId);
-      createTbReq.ctb.name = strdup(stbFullName);
+      createTbReq.ctb.name = strdup((char*)tNameGetTableName(&name));  // strdup(stbFullName);
      createTbReq.flags = 0;
      createTbReq.type = TSDB_CHILD_TABLE;
      createTbReq.ctb.suid = suid;
--- a/source/dnode/vnode/src/tsdb/tsdbCache.c
+++ b/source/dnode/vnode/src/tsdb/tsdbCache.c
@ -464,7 +464,7 @@ static int32_t getNextRowFromFS(void *iter, TSDBROW **ppRow) {

  switch (state->state) {
    case SFSNEXTROW_FS:
-      state->aDFileSet = state->pTsdb->fs->cState->aDFileSet;
+      state->aDFileSet = state->pTsdb->pFS->cState->aDFileSet;
      state->nFileSet = taosArrayGetSize(state->aDFileSet);
      state->iFileSet = state->nFileSet;

@ -793,6 +793,9 @@ typedef struct {
  TSDBROW         memRow, imemRow, fsRow;

  TsdbNextRowState input[3];
+  SMemTable       *pMemTable;
+  SMemTable       *pIMemTable;
+  STsdb           *pTsdb;
 } CacheNextRowIter;

 static int32_t nextRowIterOpen(CacheNextRowIter *pIter, tb_uid_t uid, STsdb *pTsdb) {
@ -800,21 +803,25 @@ static int32_t nextRowIterOpen(CacheNextRowIter *pIter, tb_uid_t uid, STsdb *pTs

  tb_uid_t suid = getTableSuidByUid(uid, pTsdb);

+  tsdbTakeMemSnapshot(pTsdb, &pIter->pMemTable, &pIter->pIMemTable);
+
  STbData *pMem = NULL;
-  if (pTsdb->mem) {
-    tsdbGetTbDataFromMemTable(pTsdb->mem, suid, uid, &pMem);
+  if (pIter->pMemTable) {
+    tsdbGetTbDataFromMemTable(pIter->pMemTable, suid, uid, &pMem);
  }

  STbData *pIMem = NULL;
-  if (pTsdb->imem) {
-    tsdbGetTbDataFromMemTable(pTsdb->imem, suid, uid, &pIMem);
+  if (pIter->pIMemTable) {
+    tsdbGetTbDataFromMemTable(pIter->pIMemTable, suid, uid, &pIMem);
  }

+  pIter->pTsdb = pTsdb;
+
  pIter->pSkyline = taosArrayInit(32, sizeof(TSDBKEY));

  SDelIdx delIdx;

-  SDelFile *pDelFile = tsdbFSStateGetDelFile(pTsdb->fs->cState);
+  SDelFile *pDelFile = tsdbFSStateGetDelFile(pTsdb->pFS->cState);
  if (pDelFile) {
    SDelFReader *pDelFReader;

@ -878,6 +885,8 @@ static int32_t nextRowIterClose(CacheNextRowIter *pIter) {
    taosArrayDestroy(pIter->pSkyline);
  }

+  tsdbUntakeMemSnapshot(pIter->pTsdb, pIter->pMemTable, pIter->pIMemTable);
+
  return code;
 _err:
  return code;
@ -1189,7 +1198,7 @@ static int32_t mergeLastRow(tb_uid_t uid, STsdb *pTsdb, bool *dup, STSRow **ppRo

  SDelIdx delIdx;

-  SDelFile *pDelFile = tsdbFSStateGetDelFile(pTsdb->fs->cState);
+  SDelFile *pDelFile = tsdbFSStateGetDelFile(pTsdb->pFS->cState);
  if (pDelFile) {
    SDelFReader *pDelFReader;

@ -1377,7 +1386,7 @@ static int32_t mergeLast(tb_uid_t uid, STsdb *pTsdb, SArray **ppLastArray) {

  SDelIdx delIdx;

-  SDelFile *pDelFile = tsdbFSStateGetDelFile(pTsdb->fs->cState);
+  SDelFile *pDelFile = tsdbFSStateGetDelFile(pTsdb->pFS->cState);
  if (pDelFile) {
    SDelFReader *pDelFReader;

--- a/source/dnode/vnode/src/tsdb/tsdbCommit.c
+++ b/source/dnode/vnode/src/tsdb/tsdbCommit.c
@ -64,9 +64,26 @@ int32_t tsdbBegin(STsdb *pTsdb) {

  if (!pTsdb) return code;

-  code = tsdbMemTableCreate(pTsdb, &pTsdb->mem);
+  SMemTable *pMemTable;
+  code = tsdbMemTableCreate(pTsdb, &pMemTable);
  if (code) goto _err;

+  // lock
+  code = taosThreadRwlockWrlock(&pTsdb->rwLock);
+  if (code) {
+    code = TAOS_SYSTEM_ERROR(code);
+    goto _err;
+  }
+
+  pTsdb->mem = pMemTable;
+
+  // unlock
+  code = taosThreadRwlockUnlock(&pTsdb->rwLock);
+  if (code) {
+    code = TAOS_SYSTEM_ERROR(code);
+    goto _err;
+  }
+
  return code;

 _err:
@ -83,9 +100,11 @@ int32_t tsdbCommit(STsdb *pTsdb) {

  // check
  if (pMemTable->nRow == 0 && pMemTable->nDel == 0) {
-    // TODO: lock?
+    taosThreadRwlockWrlock(&pTsdb->rwLock);
    pTsdb->mem = NULL;
-    tsdbMemTableDestroy(pMemTable);
+    taosThreadRwlockUnlock(&pTsdb->rwLock);
+
+    tsdbUnrefMemTable(pMemTable);
    goto _exit;
  }

@ -139,7 +158,7 @@ static int32_t tsdbCommitDelStart(SCommitter *pCommitter) {
    goto _err;
  }

-  SDelFile *pDelFileR = pTsdb->fs->nState->pDelFile;
+  SDelFile *pDelFileR = pTsdb->pFS->nState->pDelFile;
  if (pDelFileR) {
    code = tsdbDelFReaderOpen(&pCommitter->pDelFReader, pDelFileR, pTsdb, NULL);
    if (code) goto _err;
@ -228,7 +247,7 @@ static int32_t tsdbCommitDelEnd(SCommitter *pCommitter) {
  code = tsdbUpdateDelFileHdr(pCommitter->pDelFWriter);
  if (code) goto _err;

-  code = tsdbFSStateUpsertDelFile(pTsdb->fs->nState, &pCommitter->pDelFWriter->fDel);
+  code = tsdbFSStateUpsertDelFile(pTsdb->pFS->nState, &pCommitter->pDelFWriter->fDel);
  if (code) goto _err;

  code = tsdbDelFWriterClose(&pCommitter->pDelFWriter, 1);
@ -263,7 +282,7 @@ static int32_t tsdbCommitFileDataStart(SCommitter *pCommitter) {
  taosArrayClear(pCommitter->aBlockIdx);
  tMapDataReset(&pCommitter->oBlockMap);
  tBlockDataReset(&pCommitter->oBlockData);
-  pRSet = tsdbFSStateGetDFileSet(pTsdb->fs->nState, pCommitter->commitFid, TD_EQ);
+  pRSet = tsdbFSStateGetDFileSet(pTsdb->pFS->nState, pCommitter->commitFid, TD_EQ);
  if (pRSet) {
    code = tsdbDataFReaderOpen(&pCommitter->pReader, pTsdb, pRSet);
    if (code) goto _err;
@ -836,7 +855,7 @@ static int32_t tsdbCommitFileDataEnd(SCommitter *pCommitter) {
  if (code) goto _err;

  // upsert SDFileSet
-  code = tsdbFSStateUpsertDFileSet(pCommitter->pTsdb->fs->nState, tsdbDataFWriterGetWSet(pCommitter->pWriter));
+  code = tsdbFSStateUpsertDFileSet(pCommitter->pTsdb->pFS->nState, tsdbDataFWriterGetWSet(pCommitter->pWriter));
  if (code) goto _err;

  // close and sync
@ -941,10 +960,10 @@ static int32_t tsdbStartCommit(STsdb *pTsdb, SCommitter *pCommitter) {
  memset(pCommitter, 0, sizeof(*pCommitter));
  ASSERT(pTsdb->mem && pTsdb->imem == NULL);

-  // lock();
+  taosThreadRwlockWrlock(&pTsdb->rwLock);
  pTsdb->imem = pTsdb->mem;
  pTsdb->mem = NULL;
-  // unlock();
+  taosThreadRwlockUnlock(&pTsdb->rwLock);

  pCommitter->pTsdb = pTsdb;
  pCommitter->commitID = pTsdb->pVnode->state.commitID;
@ -954,7 +973,7 @@ static int32_t tsdbStartCommit(STsdb *pTsdb, SCommitter *pCommitter) {
  pCommitter->maxRow = pTsdb->pVnode->config.tsdbCfg.maxRows;
  pCommitter->cmprAlg = pTsdb->pVnode->config.tsdbCfg.compression;

-  code = tsdbFSBegin(pTsdb->fs);
+  code = tsdbFSBegin(pTsdb->pFS);
  if (code) goto _err;

  return code;
@ -1135,13 +1154,16 @@ static int32_t tsdbEndCommit(SCommitter *pCommitter, int32_t eno) {
  SMemTable *pMemTable = pTsdb->imem;

  if (eno == 0) {
-    code = tsdbFSCommit(pTsdb->fs);
+    code = tsdbFSCommit(pTsdb->pFS);
  } else {
-    code = tsdbFSRollback(pTsdb->fs);
+    code = tsdbFSRollback(pTsdb->pFS);
  }

-  tsdbMemTableDestroy(pMemTable);
+  taosThreadRwlockWrlock(&pTsdb->rwLock);
  pTsdb->imem = NULL;
+  taosThreadRwlockUnlock(&pTsdb->rwLock);
+
+  tsdbUnrefMemTable(pMemTable);

  tsdbInfo("vgId:%d tsdb end commit", TD_VID(pTsdb->pVnode));
  return code;
--- a/source/dnode/vnode/src/tsdb/tsdbFS.c
+++ b/source/dnode/vnode/src/tsdb/tsdbFS.c
@ -245,7 +245,7 @@ static int32_t tsdbApplyDelFileChange(STsdbFS *pFS, SDelFile *pFrom, SDelFile *p
  char    fname[TSDB_FILENAME_LEN];

  if (pFrom && pTo) {
-    if (pFrom != pTo) {
+    if (!tsdbDelFileIsSame(pFrom, pTo)) {
      tsdbDelFileName(pFS->pTsdb, pFrom, fname);
      if (taosRemoveFile(fname) < 0) {
        code = TAOS_SYSTEM_ERROR(errno);
--- a/source/dnode/vnode/src/tsdb/tsdbFile.c
+++ b/source/dnode/vnode/src/tsdb/tsdbFile.c
@ -140,6 +140,8 @@ bool tsdbFileIsSame(SDFileSet *pDFileSet1, SDFileSet *pDFileSet2, EDataFileT fty
  }
 }

+bool tsdbDelFileIsSame(SDelFile *pDelFile1, SDelFile *pDelFile2) { return pDelFile1->commitID == pDelFile2->commitID; }
+
 int32_t tsdbUpdateDFileHdr(TdFilePtr pFD, SDFileSet *pSet, EDataFileT ftype) {
  int32_t code = 0;
  int64_t n;
--- a/source/dnode/vnode/src/tsdb/tsdbMemTable.c
+++ b/source/dnode/vnode/src/tsdb/tsdbMemTable.c
@ -41,6 +41,7 @@ int32_t tsdbMemTableCreate(STsdb *pTsdb, SMemTable **ppMemTable) {
  }
  taosInitRWLatch(&pMemTable->latch);
  pMemTable->pTsdb = pTsdb;
+  pMemTable->pPool = pTsdb->pVnode->inUse;
  pMemTable->nRef = 1;
  pMemTable->minKey = TSKEY_MAX;
  pMemTable->maxKey = TSKEY_MIN;
@ -54,6 +55,7 @@ int32_t tsdbMemTableCreate(STsdb *pTsdb, SMemTable **ppMemTable) {
    taosMemoryFree(pMemTable);
    goto _err;
  }
+  vnodeBufPoolRef(pMemTable->pPool);

  *ppMemTable = pMemTable;
  return code;
@ -65,6 +67,7 @@ _err:

 void tsdbMemTableDestroy(SMemTable *pMemTable) {
  if (pMemTable) {
+    vnodeBufPoolUnRef(pMemTable->pPool);
    taosArrayDestroy(pMemTable->aTbData);
    taosMemoryFree(pMemTable);
  }
@ -590,3 +593,58 @@ _err:
 }

 int32_t tsdbGetNRowsInTbData(STbData *pTbData) { return pTbData->sl.size; }
+
+void tsdbRefMemTable(SMemTable *pMemTable) {
+  int32_t nRef = atomic_fetch_add_32(&pMemTable->nRef, 1);
+  ASSERT(nRef > 0);
+}
+
+void tsdbUnrefMemTable(SMemTable *pMemTable) {
+  int32_t nRef = atomic_sub_fetch_32(&pMemTable->nRef, 1);
+  if (nRef == 0) {
+    tsdbMemTableDestroy(pMemTable);
+  }
+}
+
+int32_t tsdbTakeMemSnapshot(STsdb *pTsdb, SMemTable **ppMem, SMemTable **ppIMem) {
+  int32_t code = 0;
+
+  // lock
+  code = taosThreadRwlockRdlock(&pTsdb->rwLock);
+  if (code) {
+    code = TAOS_SYSTEM_ERROR(code);
+    goto _exit;
+  }
+
+  // take snapshot
+  *ppMem = pTsdb->mem;
+  *ppIMem = pTsdb->imem;
+
+  if (*ppMem) {
+    tsdbRefMemTable(*ppMem);
+  }
+
+  if (*ppIMem) {
+    tsdbRefMemTable(*ppIMem);
+  }
+
+  // unlock
+  code = taosThreadRwlockUnlock(&pTsdb->rwLock);
+  if (code) {
+    code = TAOS_SYSTEM_ERROR(code);
+    goto _exit;
+  }
+
+_exit:
+  return code;
+}
+
+void tsdbUntakeMemSnapshot(STsdb *pTsdb, SMemTable *pMem, SMemTable *pIMem) {
+  if (pMem) {
+    tsdbUnrefMemTable(pMem);
+  }
+
+  if (pIMem) {
+    tsdbUnrefMemTable(pIMem);
+  }
+}
--- a/source/dnode/vnode/src/tsdb/tsdbOpen.c
+++ b/source/dnode/vnode/src/tsdb/tsdbOpen.c
@ -54,8 +54,7 @@ int tsdbOpen(SVnode *pVnode, STsdb **ppTsdb, const char *dir, STsdbKeepCfg *pKee
  sprintf(pTsdb->path, "%s%s%s", pVnode->path, TD_DIRSEP, dir);
  taosRealPath(pTsdb->path, NULL, slen);
  pTsdb->pVnode = pVnode;
-  pTsdb->repoLocked = false;
-  taosThreadMutexInit(&pTsdb->mutex, NULL);
+  taosThreadRwlockInit(&pTsdb->rwLock, NULL);
  if (!pKeepCfg) {
    tsdbSetKeepCfg(&pTsdb->keepCfg, &pVnode->config.tsdbCfg);
  } else {
@ -67,7 +66,7 @@ int tsdbOpen(SVnode *pVnode, STsdb **ppTsdb, const char *dir, STsdbKeepCfg *pKee
  tfsMkdir(pVnode->pTfs, pTsdb->path);

  // open tsdb
-  if (tsdbFSOpen(pTsdb, &pTsdb->fs) < 0) {
+  if (tsdbFSOpen(pTsdb, &pTsdb->pFS) < 0) {
    goto _err;
  }

@ -88,33 +87,10 @@ _err:

 int tsdbClose(STsdb **pTsdb) {
  if (*pTsdb) {
-    taosThreadMutexDestroy(&(*pTsdb)->mutex);
-    tsdbFSClose((*pTsdb)->fs);
+    taosThreadRwlockDestroy(&(*pTsdb)->rwLock);
+    tsdbFSClose((*pTsdb)->pFS);
    tsdbCloseCache((*pTsdb)->lruCache);
    taosMemoryFreeClear(*pTsdb);
  }
  return 0;
 }
-
-int tsdbLockRepo(STsdb *pTsdb) {
-  int code = taosThreadMutexLock(&pTsdb->mutex);
-  if (code != 0) {
-    tsdbError("vgId:%d, failed to lock tsdb since %s", TD_VID(pTsdb->pVnode), strerror(errno));
-    terrno = TAOS_SYSTEM_ERROR(code);
-    return -1;
-  }
-  pTsdb->repoLocked = true;
-  return 0;
-}
-
-int tsdbUnlockRepo(STsdb *pTsdb) {
-  // ASSERT(IS_REPO_LOCKED(pTsdb));
-  pTsdb->repoLocked = false;
-  int code = taosThreadMutexUnlock(&pTsdb->mutex);
-  if (code != 0) {
-    tsdbError("vgId:%d, failed to unlock tsdb since %s", TD_VID(pTsdb->pVnode), strerror(errno));
-    terrno = TAOS_SYSTEM_ERROR(code);
-    return -1;
-  }
-  return 0;
-}
--- a/source/dnode/vnode/src/tsdb/tsdbRead.c
+++ b/source/dnode/vnode/src/tsdb/tsdbRead.c
@ -118,6 +118,8 @@ struct STsdbReader {
  char*              idStr;  // query info handle, for debug purpose
  int32_t            type;   // query type: 1. retrieve all data blocks, 2. retrieve direct prev|next rows
  SBlockLoadSuppInfo suppInfo;
+  SMemTable*         pMem;
+  SMemTable*         pIMem;

  SIOCostSummary cost;
  STSchema*      pSchema;
@ -1453,38 +1455,71 @@ static bool keyOverlapFileBlock(TSDBKEY key, SBlock* pBlock, SVersionRange* pVer
         (pBlock->minVersion <= pVerRange->maxVer);
 }

-static bool overlapWithDelSkyline(STableBlockScanInfo* pBlockScanInfo, const SBlock* pBlock, int32_t order) {
-  if (pBlockScanInfo->delSkyline == NULL) {
-    return false;
-  }

-  TSDBKEY* pFirst = taosArrayGet(pBlockScanInfo->delSkyline, 0);
-  TSDBKEY* pLast = taosArrayGetLast(pBlockScanInfo->delSkyline);
-
-  // ts is not overlap
-  if (pBlock->minKey.ts > pLast->ts || pBlock->maxKey.ts < pFirst->ts) {
-    return false;
-  }
-
-  int32_t step = ASCENDING_TRAVERSE(order) ? 1 : -1;
-
-  // version is not overlap
+static bool doCheckforDatablockOverlap(STableBlockScanInfo* pBlockScanInfo, const SBlock* pBlock) {
  size_t num = taosArrayGetSize(pBlockScanInfo->delSkyline);
-  for (int32_t i = pBlockScanInfo->fileDelIndex; i < num; i += step) {
+
+  for (int32_t i = pBlockScanInfo->fileDelIndex; i < num; i += 1) {
    TSDBKEY* p = taosArrayGet(pBlockScanInfo->delSkyline, i);
    if (p->ts >= pBlock->minKey.ts && p->ts <= pBlock->maxKey.ts) {
      if (p->version >= pBlock->minVersion) {
        return true;
      }
-    } else if (p->ts > pBlock->maxKey.ts) {
+    } else if (p->ts < pBlock->minKey.ts) {  // p->ts < pBlock->minKey.ts
+      if (p->version >= pBlock->minVersion) {
+        if (i < num - 1) {
+          TSDBKEY* pnext = taosArrayGet(pBlockScanInfo->delSkyline, i + 1);
+          if (i + 1 == num - 1) {  // pnext is the last point
+            if (pnext->ts >= pBlock->minKey.ts) {
+              return true;
+            }
+          } else {
+            if (pnext->ts >= pBlock->minKey.ts && pnext->version >= pBlock->minVersion) {
+              return true;
+            }
+          }
+        } else {  // it must be the last point
+          ASSERT(p->version == 0);
+        }
+      }
+    } else {  // (p->ts > pBlock->maxKey.ts) {
      return false;
    }
  }

-  ASSERT(0);
  return false;
 }

+static bool overlapWithDelSkyline(STableBlockScanInfo* pBlockScanInfo, const SBlock* pBlock, int32_t order) {
+  if (pBlockScanInfo->delSkyline == NULL) {
+    return false;
+  }
+
+  // ts is not overlap
+  TSDBKEY* pFirst = taosArrayGet(pBlockScanInfo->delSkyline, 0);
+  TSDBKEY* pLast = taosArrayGetLast(pBlockScanInfo->delSkyline);
+  if (pBlock->minKey.ts > pLast->ts || pBlock->maxKey.ts < pFirst->ts) {
+    return false;
+  }
+
+  // version is not overlap
+  if (ASCENDING_TRAVERSE(order)) {
+    return doCheckforDatablockOverlap(pBlockScanInfo, pBlock);
+  } else {
+    int32_t index = pBlockScanInfo->fileDelIndex;
+    while(1) {
+      TSDBKEY* p = taosArrayGet(pBlockScanInfo->delSkyline, index);
+      if (p->ts > pBlock->minKey.ts && index > 0) {
+        index -= 1;
+      } else {  // find the first point that is smaller than the minKey.ts of dataBlock.
+        break;
+      }
+    }
+
+    return doCheckforDatablockOverlap(pBlockScanInfo, pBlock);
+  }
+}
+
 // 1. the version of all rows should be less than the endVersion
 // 2. current block should not overlap with next neighbor block
 // 3. current timestamp should not be overlap with each other
@ -1847,8 +1882,8 @@ static int32_t initMemDataIterator(STableBlockScanInfo* pBlockScanInfo, STsdbRea
  int32_t backward = (!ASCENDING_TRAVERSE(pReader->order));

  STbData* d = NULL;
-  if (pReader->pTsdb->mem != NULL) {
-    tsdbGetTbDataFromMemTable(pReader->pTsdb->mem, pReader->suid, pBlockScanInfo->uid, &d);
+  if (pReader->pMem != NULL) {
+    tsdbGetTbDataFromMemTable(pReader->pMem, pReader->suid, pBlockScanInfo->uid, &d);
    if (d != NULL) {
      code = tsdbTbDataIterCreate(d, &startKey, backward, &pBlockScanInfo->iter.iter);
      if (code == TSDB_CODE_SUCCESS) {
@ -1868,8 +1903,8 @@ static int32_t initMemDataIterator(STableBlockScanInfo* pBlockScanInfo, STsdbRea
  }

  STbData* di = NULL;
-  if (pReader->pTsdb->imem != NULL) {
-    tsdbGetTbDataFromMemTable(pReader->pTsdb->imem, pReader->suid, pBlockScanInfo->uid, &di);
+  if (pReader->pIMem != NULL) {
+    tsdbGetTbDataFromMemTable(pReader->pIMem, pReader->suid, pBlockScanInfo->uid, &di);
    if (di != NULL) {
      code = tsdbTbDataIterCreate(di, &startKey, backward, &pBlockScanInfo->iiter.iter);
      if (code == TSDB_CODE_SUCCESS) {
@ -1905,7 +1940,7 @@ int32_t initDelSkylineIterator(STableBlockScanInfo* pBlockScanInfo, STsdbReader*

  SArray* pDelData = taosArrayInit(4, sizeof(SDelData));

-  SDelFile* pDelFile = tsdbFSStateGetDelFile(pTsdb->fs->cState);
+  SDelFile* pDelFile = tsdbFSStateGetDelFile(pTsdb->pFS->cState);
  if (pDelFile) {
    SDelFReader* pDelFReader = NULL;
    code = tsdbDelFReaderOpen(&pDelFReader, pDelFile, pTsdb, NULL);
@ -1926,9 +1961,11 @@ int32_t initDelSkylineIterator(STableBlockScanInfo* pBlockScanInfo, STsdbReader*
    SDelIdx  idx = {.suid = pReader->suid, .uid = pBlockScanInfo->uid};
    SDelIdx* pIdx = taosArraySearch(aDelIdx, &idx, tCmprDelIdx, TD_EQ);

-    code = tsdbReadDelData(pDelFReader, pIdx, pDelData, NULL);
-    if (code != TSDB_CODE_SUCCESS) {
-      goto _err;
+    if (pIdx != NULL) {
+      code = tsdbReadDelData(pDelFReader, pIdx, pDelData, NULL);
+      if (code != TSDB_CODE_SUCCESS) {
+        goto _err;
+      }
    }
  }

@ -2795,7 +2832,7 @@ int32_t tsdbReaderOpen(SVnode* pVnode, SQueryTableDataCond* pCond, SArray* pTabl

  SDataBlockIter* pBlockIter = &pReader->status.blockIter;

-  STsdbFSState* pFState = pReader->pTsdb->fs->cState;
+  STsdbFSState* pFState = pReader->pTsdb->pFS->cState;
  initFilesetIterator(&pReader->status.fileIter, pFState, pReader->order, pReader->idStr);
  resetDataBlockIterator(&pReader->status.blockIter, pReader->order);

@ -2809,6 +2846,8 @@ int32_t tsdbReaderOpen(SVnode* pVnode, SQueryTableDataCond* pCond, SArray* pTabl
    }
  }

+  tsdbTakeMemSnapshot(pReader->pTsdb, &pReader->pMem, &pReader->pIMem);
+
  tsdbDebug("%p total numOfTable:%d in this query %s", pReader, numOfTables, pReader->idStr);
  return code;

@ -2824,6 +2863,8 @@ void tsdbReaderClose(STsdbReader* pReader) {

  SBlockLoadSuppInfo* pSupInfo = &pReader->suppInfo;

+  tsdbUntakeMemSnapshot(pReader->pTsdb, pReader->pMem, pReader->pIMem);
+
  taosMemoryFreeClear(pSupInfo->plist);
  taosMemoryFree(pSupInfo->colIds);

@ -3042,7 +3083,7 @@ int32_t tsdbReaderReset(STsdbReader* pReader, SQueryTableDataCond* pCond) {

  tsdbDataFReaderClose(&pReader->pFileReader);

-  STsdbFSState* pFState = pReader->pTsdb->fs->cState;
+  STsdbFSState* pFState = pReader->pTsdb->pFS->cState;
  initFilesetIterator(&pReader->status.fileIter, pFState, pReader->order, pReader->idStr);
  resetDataBlockIterator(&pReader->status.blockIter, pReader->order);
  resetDataBlockScanInfo(pReader->status.pTableMap);
--- a/source/dnode/vnode/src/tsdb/tsdbReaderWriter.c
+++ b/source/dnode/vnode/src/tsdb/tsdbReaderWriter.c
@ -246,7 +246,7 @@ int32_t tsdbDelFReaderOpen(SDelFReader **ppReader, SDelFile *pFile, STsdb *pTsdb

  tsdbDelFileName(pTsdb, pFile, fname);
  pDelFReader->pReadH = taosOpenFile(fname, TD_FILE_READ);
-  if (pDelFReader == NULL) {
+  if (pDelFReader->pReadH == NULL) {
    code = TAOS_SYSTEM_ERROR(errno);
    taosMemoryFree(pDelFReader);
    goto _err;
--- a/source/dnode/vnode/src/tsdb/tsdbRetention.c
+++ b/source/dnode/vnode/src/tsdb/tsdbRetention.c
@ -20,10 +20,10 @@ static int32_t tsdbDoRetentionImpl(STsdb *pTsdb, int64_t now, int8_t try, int8_t
  STsdbFSState *pState;

  if (try) {
-    pState = pTsdb->fs->cState;
+    pState = pTsdb->pFS->cState;
    *canDo = 0;
  } else {
-    pState = pTsdb->fs->nState;
+    pState = pTsdb->pFS->nState;
  }

  for (int32_t iSet = 0; iSet < taosArrayGetSize(pState->aDFileSet); iSet++) {
@ -83,7 +83,7 @@ int32_t tsdbDoRetention(STsdb *pTsdb, int64_t now) {
  if (!canDo) goto _exit;

  // begin
-  code = tsdbFSBegin(pTsdb->fs);
+  code = tsdbFSBegin(pTsdb->pFS);
  if (code) goto _err;

  // do retention
@ -91,7 +91,7 @@ int32_t tsdbDoRetention(STsdb *pTsdb, int64_t now) {
  if (code) goto _err;

  // commit
-  code = tsdbFSCommit(pTsdb->fs);
+  code = tsdbFSCommit(pTsdb->pFS);
  if (code) goto _err;

 _exit:
@ -99,6 +99,6 @@ _exit:

 _err:
  tsdbError("vgId:%d tsdb do retention failed since %s", TD_VID(pTsdb->pVnode), tstrerror(code));
-  tsdbFSRollback(pTsdb->fs);
+  tsdbFSRollback(pTsdb->pFS);
  return code;
 }
--- a/source/dnode/vnode/src/tsdb/tsdbSnapshot.c
+++ b/source/dnode/vnode/src/tsdb/tsdbSnapshot.c
@ -45,7 +45,7 @@ static int32_t tsdbSnapReadData(STsdbSnapReader* pReader, uint8_t** ppData) {

  while (true) {
    if (pReader->pDataFReader == NULL) {
-      SDFileSet* pSet = tsdbFSStateGetDFileSet(pTsdb->fs->cState, pReader->fid, TD_GT);
+      SDFileSet* pSet = tsdbFSStateGetDFileSet(pTsdb->pFS->cState, pReader->fid, TD_GT);

      if (pSet == NULL) goto _exit;

@ -159,7 +159,7 @@ _err:
 static int32_t tsdbSnapReadDel(STsdbSnapReader* pReader, uint8_t** ppData) {
  int32_t   code = 0;
  STsdb*    pTsdb = pReader->pTsdb;
-  SDelFile* pDelFile = pTsdb->fs->cState->pDelFile;
+  SDelFile* pDelFile = pTsdb->pFS->cState->pDelFile;

  if (pReader->pDelFReader == NULL) {
    if (pDelFile == NULL) {
@ -798,7 +798,7 @@ static int32_t tsdbSnapWriteDataEnd(STsdbSnapWriter* pWriter) {
  code = tsdbWriteBlockIdx(pWriter->pDataFWriter, pWriter->aBlockIdxW, NULL);
  if (code) goto _err;

-  code = tsdbFSStateUpsertDFileSet(pTsdb->fs->nState, tsdbDataFWriterGetWSet(pWriter->pDataFWriter));
+  code = tsdbFSStateUpsertDFileSet(pTsdb->pFS->nState, tsdbDataFWriterGetWSet(pWriter->pDataFWriter));
  if (code) goto _err;

  code = tsdbDataFWriterClose(&pWriter->pDataFWriter, 1);
@ -843,7 +843,7 @@ static int32_t tsdbSnapWriteData(STsdbSnapWriter* pWriter, uint8_t* pData, uint3
    pWriter->fid = fid;

    // read
-    SDFileSet* pSet = tsdbFSStateGetDFileSet(pTsdb->fs->nState, fid, TD_EQ);
+    SDFileSet* pSet = tsdbFSStateGetDFileSet(pTsdb->pFS->nState, fid, TD_EQ);
    if (pSet) {
      code = tsdbDataFReaderOpen(&pWriter->pDataFReader, pTsdb, pSet);
      if (code) goto _err;
@ -907,7 +907,7 @@ static int32_t tsdbSnapWriteDel(STsdbSnapWriter* pWriter, uint8_t* pData, uint32
  STsdb*  pTsdb = pWriter->pTsdb;

  if (pWriter->pDelFWriter == NULL) {
-    SDelFile* pDelFile = tsdbFSStateGetDelFile(pTsdb->fs->nState);
+    SDelFile* pDelFile = tsdbFSStateGetDelFile(pTsdb->pFS->nState);

    // reader
    if (pDelFile) {
@ -1017,7 +1017,7 @@ static int32_t tsdbSnapWriteDelEnd(STsdbSnapWriter* pWriter) {
  code = tsdbUpdateDelFileHdr(pWriter->pDelFWriter);
  if (code) goto _err;

-  code = tsdbFSStateUpsertDelFile(pTsdb->fs->nState, &pWriter->pDelFWriter->fDel);
+  code = tsdbFSStateUpsertDelFile(pTsdb->pFS->nState, &pWriter->pDelFWriter->fDel);
  if (code) goto _err;

  code = tsdbDelFWriterClose(&pWriter->pDelFWriter, 1);
@ -1096,7 +1096,7 @@ int32_t tsdbSnapWriterOpen(STsdb* pTsdb, int64_t sver, int64_t ever, STsdbSnapWr
    goto _err;
  }

-  code = tsdbFSBegin(pTsdb->fs);
+  code = tsdbFSBegin(pTsdb->pFS);
  if (code) goto _err;

  *ppWriter = pWriter;
@ -1113,7 +1113,7 @@ int32_t tsdbSnapWriterClose(STsdbSnapWriter** ppWriter, int8_t rollback) {
  STsdbSnapWriter* pWriter = *ppWriter;

  if (rollback) {
-    code = tsdbFSRollback(pWriter->pTsdb->fs);
+    code = tsdbFSRollback(pWriter->pTsdb->pFS);
    if (code) goto _err;
  } else {
    code = tsdbSnapWriteDataEnd(pWriter);
@ -1122,7 +1122,7 @@ int32_t tsdbSnapWriterClose(STsdbSnapWriter** ppWriter, int8_t rollback) {
    code = tsdbSnapWriteDelEnd(pWriter);
    if (code) goto _err;

-    code = tsdbFSCommit(pWriter->pTsdb->fs);
+    code = tsdbFSCommit(pWriter->pTsdb->pFS);
    if (code) goto _err;
  }

--- a/source/dnode/vnode/src/vnd/vnodeBufPool.c
+++ b/source/dnode/vnode/src/vnd/vnodeBufPool.c
@ -17,7 +17,7 @@

 /* ------------------------ STRUCTURES ------------------------ */

-static int vnodeBufPoolCreate(int64_t size, SVBufPool **ppPool);
+static int vnodeBufPoolCreate(SVnode *pVnode, int64_t size, SVBufPool **ppPool);
 static int vnodeBufPoolDestroy(SVBufPool *pPool);

 int vnodeOpenBufPool(SVnode *pVnode, int64_t size) {
@ -28,7 +28,7 @@ int vnodeOpenBufPool(SVnode *pVnode, int64_t size) {

  for (int i = 0; i < 3; i++) {
    // create pool
-    ret = vnodeBufPoolCreate(size, &pPool);
+    ret = vnodeBufPoolCreate(pVnode, size, &pPool);
    if (ret < 0) {
      vError("vgId:%d, failed to open vnode buffer pool since %s", TD_VID(pVnode), tstrerror(terrno));
      vnodeCloseBufPool(pVnode);
@ -120,7 +120,7 @@ void vnodeBufPoolFree(SVBufPool *pPool, void *p) {
 }

 // STATIC METHODS -------------------
-static int vnodeBufPoolCreate(int64_t size, SVBufPool **ppPool) {
+static int vnodeBufPoolCreate(SVnode *pVnode, int64_t size, SVBufPool **ppPool) {
  SVBufPool *pPool;

  pPool = taosMemoryMalloc(sizeof(SVBufPool) + size);
@ -130,6 +130,7 @@ static int vnodeBufPoolCreate(int64_t size, SVBufPool **ppPool) {
  }

  pPool->next = NULL;
+  pPool->pVnode = pVnode;
  pPool->nRef = 0;
  pPool->size = 0;
  pPool->ptr = pPool->node.data;
@ -146,4 +147,26 @@ static int vnodeBufPoolDestroy(SVBufPool *pPool) {
  vnodeBufPoolReset(pPool);
  taosMemoryFree(pPool);
  return 0;
+}
+
+void vnodeBufPoolRef(SVBufPool *pPool) {
+  int32_t nRef = atomic_fetch_add_32(&pPool->nRef, 1);
+  ASSERT(nRef > 0);
+}
+
+void vnodeBufPoolUnRef(SVBufPool *pPool) {
+  int32_t nRef = atomic_sub_fetch_32(&pPool->nRef, 1);
+  if (nRef == 0) {
+    SVnode *pVnode = pPool->pVnode;
+
+    vnodeBufPoolReset(pPool);
+
+    taosThreadMutexLock(&pVnode->mutex);
+
+    pPool->next = pVnode->pPool;
+    pVnode->pPool = pPool;
+    taosThreadCondSignal(&pVnode->poolNotEmpty);
+
+    taosThreadMutexUnlock(&pVnode->mutex);
+  }
 }
--- a/source/dnode/vnode/src/vnd/vnodeCommit.c
+++ b/source/dnode/vnode/src/vnd/vnodeCommit.c
@ -15,7 +15,7 @@

 #include "vnd.h"

-#define VND_INFO_FNAME "vnode.json"
+#define VND_INFO_FNAME     "vnode.json"
 #define VND_INFO_FNAME_TMP "vnode_tmp.json"

 static int  vnodeEncodeInfo(const SVnodeInfo *pInfo, char **ppData);
@ -27,18 +27,18 @@ static void vnodeWaitCommit(SVnode *pVnode);

 int vnodeBegin(SVnode *pVnode) {
  // alloc buffer pool
-  /* pthread_mutex_lock(); */
+  taosThreadMutexLock(&pVnode->mutex);

  while (pVnode->pPool == NULL) {
-    /* pthread_cond_wait(); */
+    taosThreadCondWait(&pVnode->poolNotEmpty, &pVnode->mutex);
  }

  pVnode->inUse = pVnode->pPool;
+  pVnode->inUse->nRef = 1;
  pVnode->pPool = pVnode->inUse->next;
  pVnode->inUse->next = NULL;
-  /* ref pVnode->inUse buffer pool */

-  /* pthread_mutex_unlock(); */
+  taosThreadMutexUnlock(&pVnode->mutex);

  pVnode->state.commitID++;
  // begin meta
@ -217,7 +217,7 @@ int vnodeCommit(SVnode *pVnode) {
  vInfo("vgId:%d, start to commit, commit ID:%" PRId64 " version:%" PRId64, TD_VID(pVnode), pVnode->state.commitID,
        pVnode->state.applied);

-  pVnode->onCommit = pVnode->inUse;
+  vnodeBufPoolUnRef(pVnode->inUse);
  pVnode->inUse = NULL;

  // save info
@ -284,10 +284,6 @@ int vnodeCommit(SVnode *pVnode) {

  // apply the commit (TODO)
  walEndSnapshot(pVnode->pWal);
-  vnodeBufPoolReset(pVnode->onCommit);
-  pVnode->onCommit->next = pVnode->pPool;
-  pVnode->pPool = pVnode->onCommit;
-  pVnode->onCommit = NULL;

  vInfo("vgId:%d, commit over", TD_VID(pVnode));

--- a/source/dnode/vnode/src/vnd/vnodeOpen.c
+++ b/source/dnode/vnode/src/vnd/vnodeOpen.c
@ -89,6 +89,8 @@ SVnode *vnodeOpen(const char *path, STfs *pTfs, SMsgCb msgCb) {

  tsem_init(&pVnode->syncSem, 0, 0);
  tsem_init(&(pVnode->canCommit), 0, 1);
+  taosThreadMutexInit(&pVnode->mutex, NULL);
+  taosThreadCondInit(&pVnode->poolNotEmpty, NULL);

  // open buffer pool
  if (vnodeOpenBufPool(pVnode, pVnode->config.isHeap ? 0 : pVnode->config.szBuf / 3) < 0) {
@ -195,6 +197,8 @@ void vnodeClose(SVnode *pVnode) {
    // destroy handle
    tsem_destroy(&(pVnode->canCommit));
    tsem_destroy(&pVnode->syncSem);
+    taosThreadCondDestroy(&pVnode->poolNotEmpty);
+    taosThreadMutexDestroy(&pVnode->mutex);
    taosMemoryFree(pVnode);
  }
 }
--- a/source/dnode/vnode/src/vnd/vnodeSnapshot.c
+++ b/source/dnode/vnode/src/vnd/vnodeSnapshot.c
@ -27,6 +27,8 @@ struct SVSnapReader {
  // tsdb
  int8_t           tsdbDone;
  STsdbSnapReader *pTsdbReader;
+  // rsma
+  int8_t rsmaDone[TSDB_RETENTION_L2];
 };

 int32_t vnodeSnapReaderOpen(SVnode *pVnode, int64_t sver, int64_t ever, SVSnapReader **ppReader) {
@ -115,6 +117,42 @@ int32_t vnodeSnapRead(SVSnapReader *pReader, uint8_t **ppData, uint32_t *nData)
    }
  }

+  // RSMA ==============
+#if 0
+  if (VND_IS_RSMA(pReader->pVnode)) {
+    // RSMA1/RSMA2
+    for (int32_t i = 0; i < TSDB_RETENTION_L2; ++i) {
+      if (!pReader->rsmaDone[i]) {
+        if (!pReader->pVnode->pSma->pRSmaTsdb[i]) {
+          // no valid tsdb 
+          pReader->rsmaDone[i] = 1;
+          continue;
+        }
+        if (pReader->pTsdbReader == NULL) {
+          code = tsdbSnapReaderOpen(pReader->pVnode->pSma->pRSmaTsdb[i], pReader->sver, pReader->ever,
+                                    &pReader->pTsdbReader);
+          if (code) goto _err;
+        }
+
+        code = tsdbSnapRead(pReader->pTsdbReader, ppData);
+        if (code) {
+          goto _err;
+        } else {
+          if (*ppData) {
+            goto _exit;
+          } else {
+            pReader->tsdbDone = 1;
+            code = tsdbSnapReaderClose(&pReader->pTsdbReader);
+            if (code) goto _err;
+          }
+        }
+      }
+    }
+    // QTaskInfoFile
+    // TODO ...
+  }
+#endif
+
  *ppData = NULL;
  *nData = 0;

--- a/source/libs/catalog/inc/catalogInt.h
+++ b/source/libs/catalog/inc/catalogInt.h
@ -532,6 +532,14 @@ typedef struct SCtgOperation {
  }                                                       \
 } while (0)

+#define CTG_API_JENTER() do {                             \
+  CTG_API_DEBUG("CTG API enter %s", __FUNCTION__);        \
+  CTG_LOCK(CTG_READ, &gCtgMgmt.lock);                     \
+  if (atomic_load_8((int8_t*)&gCtgMgmt.exit)) {           \
+    CTG_ERR_JRET(TSDB_CODE_CTG_OUT_OF_SERVICE);           \
+  }                                                       \
+} while (0)
+

 #define CTG_API_LEAVE_NOLOCK(c) do {                        \
    int32_t __code = c;                                     \
--- a/source/libs/catalog/src/ctgRemote.c
+++ b/source/libs/catalog/src/ctgRemote.c
@ -244,10 +244,11 @@ int32_t ctgProcessRspMsg(void* out, int32_t reqType, char* msg, int32_t msgSize,
 int32_t ctgHandleMsgCallback(void *param, SDataBuf *pMsg, int32_t rspCode) {
  SCtgTaskCallbackParam* cbParam = (SCtgTaskCallbackParam*)param;
  int32_t code = 0;
+  SCtgJob* pJob = NULL;
  
-  CTG_API_ENTER();
+  CTG_API_JENTER();

-  SCtgJob* pJob = taosAcquireRef(gCtgMgmt.jobPool, cbParam->refId);
+  pJob = taosAcquireRef(gCtgMgmt.jobPool, cbParam->refId);
  if (NULL == pJob) {
    qDebug("ctg job refId 0x%" PRIx64 " already dropped", cbParam->refId);
    goto _return;
@ -266,8 +267,6 @@ _return:
  if (pJob) {
    taosReleaseRef(gCtgMgmt.jobPool, cbParam->refId);
  }
-  
-  taosMemoryFree(param);

  CTG_API_LEAVE(code);
 }
@ -293,6 +292,7 @@ int32_t ctgMakeMsgSendInfo(SCtgTask* pTask, int32_t msgType, SMsgSendInfo **pMsg
  param->taskId = pTask->taskId;

  msgSendInfo->param = param;
+  msgSendInfo->paramFreeFp = taosMemoryFree;
  msgSendInfo->fp = ctgHandleMsgCallback;

  *pMsgSendInfo = msgSendInfo;
--- a/source/libs/executor/inc/executorimpl.h
+++ b/source/libs/executor/inc/executorimpl.h
@ -159,27 +159,28 @@ typedef struct {
  int64_t recoverEndVer;
 } SStreamTaskInfo;

+typedef struct {
+  char*           tablename;
+  char*           dbname;
+  int32_t         tversion;
+  SSchemaWrapper* sw;
+} SSchemaInfo;
+
 typedef struct SExecTaskInfo {
-  STaskIdInfo     id;
-  uint32_t        status;
-  STimeWindow     window;
-  STaskCostInfo   cost;
-  int64_t         owner;  // if it is in execution
-  int32_t         code;
+  STaskIdInfo      id;
+  uint32_t         status;
+  STimeWindow      window;
+  STaskCostInfo    cost;
+  int64_t          owner;  // if it is in execution
+  int32_t          code;

-  SStreamTaskInfo streamInfo;
-
-  struct {
-    char*           tablename;
-    char*           dbname;
-    int32_t         tversion;
-    SSchemaWrapper* sw;
-  } schemaVer;
-
-  STableListInfo        tableqinfoList;  // this is a table list
-  const char*           sql;             // query sql string
-  jmp_buf               env;             // jump to this position when error happens.
-  EOPTR_EXEC_MODEL      execModel;       // operator execution model [batch model|stream model]
+  SStreamTaskInfo  streamInfo;
+  SSchemaInfo      schemaInfo;
+  STableListInfo   tableqinfoList;  // this is a table list
+  const char*      sql;             // query sql string
+  jmp_buf          env;             // jump to this position when error happens.
+  EOPTR_EXEC_MODEL execModel;       // operator execution model [batch model|stream model]
+  SSubplan*        pSubplan;
  struct SOperatorInfo* pRoot;
 } SExecTaskInfo;

@ -248,13 +249,13 @@ typedef struct SLoadRemoteDataInfo {
 } SLoadRemoteDataInfo;

 typedef struct SLimitInfo {
-    SLimit             limit;
-    SLimit             slimit;
-    uint64_t           currentGroupId;
-    int64_t            remainGroupOffset;
-    int64_t            numOfOutputGroups;
-    int64_t            remainOffset;
-    int64_t            numOfOutputRows;
+  SLimit              limit;
+  SLimit              slimit;
+  uint64_t            currentGroupId;
+  int64_t             remainGroupOffset;
+  int64_t             numOfOutputGroups;
+  int64_t             remainOffset;
+  int64_t             numOfOutputRows;
 } SLimitInfo;

 typedef struct SExchangeInfo {
--- a/source/libs/executor/src/cachescanoperator.c
+++ b/source/libs/executor/src/cachescanoperator.c
@ -227,14 +227,14 @@ int32_t extractTargetSlotId(const SArray* pColMatchInfo, SExecTaskInfo* pTaskInf

  for (int32_t i = 0; i < numOfCols; ++i) {
    SColMatchInfo* pColMatch = taosArrayGet(pColMatchInfo, i);
-    for (int32_t j = 0; j < pTaskInfo->schemaVer.sw->nCols; ++j) {
-      if (pColMatch->colId == pTaskInfo->schemaVer.sw->pSchema[j].colId &&
+    for (int32_t j = 0; j < pTaskInfo->schemaInfo.sw->nCols; ++j) {
+      if (pColMatch->colId == pTaskInfo->schemaInfo.sw->pSchema[j].colId &&
          pColMatch->colId == PRIMARYKEY_TIMESTAMP_COL_ID) {
        (*pSlotIds)[pColMatch->targetSlotId] = -1;
        break;
      }

-      if (pColMatch->colId == pTaskInfo->schemaVer.sw->pSchema[j].colId) {
+      if (pColMatch->colId == pTaskInfo->schemaInfo.sw->pSchema[j].colId) {
        (*pSlotIds)[pColMatch->targetSlotId] = j;
        break;
      }
--- a/source/libs/executor/src/dataDispatcher.c
+++ b/source/libs/executor/src/dataDispatcher.c
@ -67,15 +67,15 @@ static bool needCompress(const SSDataBlock* pData, int32_t numOfCols) {
  return false;
 }

+// clang-format off
 // data format:
-// +----------------+--------------+----------+--------------------------------------------+--------------------------------------+-------------+-----------+-------------+-----------+
-// |SDataCacheEntry | total length | group id | col1_schema | col2_schema | col3_schema ...| column#1 length, column#2
-// length ... | col1 bitmap | col1 data | col2 bitmap | col2 data | .... |                |  (4 bytes)   |(8 bytes)
-// |(sizeof(int16_t)+sizeof(int32_t))*numOfCols | sizeof(int32_t) * numOfCols          | actual size |           |
-// actual size |           |
-// +----------------+--------------+----------+--------------------------------------------+--------------------------------------+-------------+-----------+-------------+-----------+
+// +----------------+--------------+-----------------+--------------------------------------------+------------------------------------+-------------+-----------+-------------+-----------+
+// |SDataCacheEntry | total length |    group id     | col1_schema | col2_schema | col3_schema... | column#1 length, column#2 length...| col1 bitmap | col1 data | col2 bitmap | col2 data | .... |                |  (4 bytes)   |(8 bytes)
+// |                |sizeof(int32) |sizeof(uint64_t) |(sizeof(int16_t)+sizeof(int32_t))*numOfCols | sizeof(int32_t) * numOfCols        | actual size |           |
+// +----------------+--------------+-----------------+--------------------------------------------+------------------------------------+-------------+-----------+-------------+-----------+
 // The length of bitmap is decided by number of rows of this data block, and the length of each column data is
 // recorded in the first segment, next to the struct header
+// clang-format on
 static void toDataCacheEntry(SDataDispatchHandle* pHandle, const SInputData* pInput, SDataDispatchBuf* pBuf) {
  int32_t numOfCols = 0;
  SNode*  pNode;
--- a/source/libs/executor/src/dataInserter.c
+++ b/source/libs/executor/src/dataInserter.c
@ -91,7 +91,6 @@ _return:
  tsem_post(&pInserter->ready);

  taosMemoryFree(pMsg->pData);
-  taosMemoryFree(param);
  
  return TSDB_CODE_SUCCESS;
 }
@ -110,6 +109,7 @@ static int32_t sendSubmitRequest(SDataInserterHandle* pInserter, SSubmitReq* pMs
  pParam->pInserter = pInserter;

  pMsgSendInfo->param = pParam;
+  pMsgSendInfo->paramFreeFp = taosMemoryFree;        
  pMsgSendInfo->msgInfo.pData = pMsg;
  pMsgSendInfo->msgInfo.len = ntohl(pMsg->length);
  pMsgSendInfo->msgType = TDMT_VND_SUBMIT;
--- a/source/libs/executor/src/executor.c
+++ b/source/libs/executor/src/executor.c
@ -115,27 +115,40 @@ int32_t qSetMultiStreamInput(qTaskInfo_t tinfo, const void* pBlocks, size_t numO
  return code;
 }

-qTaskInfo_t qCreateQueueExecTaskInfo(void* msg, SReadHandle* readers) {
+qTaskInfo_t qCreateQueueExecTaskInfo(void* msg, SReadHandle* readers, int32_t* numOfCols) {
  if (msg == NULL) {
    // TODO create raw scan
    return NULL;
  }

-  struct SSubplan* plan = NULL;
-  int32_t          code = qStringToSubplan(msg, &plan);
+  struct SSubplan* pPlan = NULL;
+  int32_t          code = qStringToSubplan(msg, &pPlan);
  if (code != TSDB_CODE_SUCCESS) {
    terrno = code;
    return NULL;
  }

  qTaskInfo_t pTaskInfo = NULL;
-  code = qCreateExecTask(readers, 0, 0, plan, &pTaskInfo, NULL, NULL, OPTR_EXEC_MODEL_QUEUE);
+  code = qCreateExecTask(readers, 0, 0, pPlan, &pTaskInfo, NULL, NULL, OPTR_EXEC_MODEL_QUEUE);
  if (code != TSDB_CODE_SUCCESS) {
-    // TODO: destroy SSubplan & pTaskInfo
+    nodesDestroyNode((SNode*)pPlan);
+    qDestroyTask(pTaskInfo);
    terrno = code;
    return NULL;
  }

+  // extract the number of output columns
+  SDataBlockDescNode* pDescNode = pPlan->pNode->pOutputDataBlockDesc;
+  *numOfCols = 0;
+
+  SNode*  pNode;
+  FOREACH(pNode, pDescNode->pSlots) {
+    SSlotDescNode* pSlotDesc = (SSlotDescNode*)pNode;
+    if (pSlotDesc->output) {
+      ++(*numOfCols);
+    }
+  }
+
  return pTaskInfo;
 }

@ -146,17 +159,18 @@ qTaskInfo_t qCreateStreamExecTaskInfo(void* msg, SReadHandle* readers) {

  /*qDebugL("stream task string %s", (const char*)msg);*/

-  struct SSubplan* plan = NULL;
-  int32_t          code = qStringToSubplan(msg, &plan);
+  struct SSubplan* pPlan = NULL;
+  int32_t          code = qStringToSubplan(msg, &pPlan);
  if (code != TSDB_CODE_SUCCESS) {
    terrno = code;
    return NULL;
  }

  qTaskInfo_t pTaskInfo = NULL;
-  code = qCreateExecTask(readers, 0, 0, plan, &pTaskInfo, NULL, NULL, OPTR_EXEC_MODEL_STREAM);
+  code = qCreateExecTask(readers, 0, 0, pPlan, &pTaskInfo, NULL, NULL, OPTR_EXEC_MODEL_STREAM);
  if (code != TSDB_CODE_SUCCESS) {
-    // TODO: destroy SSubplan & pTaskInfo
+    nodesDestroyNode((SNode*)pPlan);
+    qDestroyTask(pTaskInfo);
    terrno = code;
    return NULL;
  }
@ -239,19 +253,19 @@ int32_t qGetQueryTableSchemaVersion(qTaskInfo_t tinfo, char* dbName, char* table
  ASSERT(tinfo != NULL && dbName != NULL && tableName != NULL);
  SExecTaskInfo* pTaskInfo = (SExecTaskInfo*)tinfo;

-  if (pTaskInfo->schemaVer.sw == NULL) {
+  if (pTaskInfo->schemaInfo.sw == NULL) {
    return TSDB_CODE_SUCCESS;
  }

-  *sversion = pTaskInfo->schemaVer.sw->version;
-  *tversion = pTaskInfo->schemaVer.tversion;
-  if (pTaskInfo->schemaVer.dbname) {
-    strcpy(dbName, pTaskInfo->schemaVer.dbname);
+  *sversion = pTaskInfo->schemaInfo.sw->version;
+  *tversion = pTaskInfo->schemaInfo.tversion;
+  if (pTaskInfo->schemaInfo.dbname) {
+    strcpy(dbName, pTaskInfo->schemaInfo.dbname);
  } else {
    dbName[0] = 0;
  }
-  if (pTaskInfo->schemaVer.tablename) {
-    strcpy(tableName, pTaskInfo->schemaVer.tablename);
+  if (pTaskInfo->schemaInfo.tablename) {
+    strcpy(tableName, pTaskInfo->schemaInfo.tablename);
  } else {
    tableName[0] = 0;
  }
--- a/source/libs/executor/src/executorMain.c
+++ b/source/libs/executor/src/executorMain.c
@ -39,6 +39,7 @@ int32_t qCreateExecTask(SReadHandle* readHandle, int32_t vgId, uint64_t taskId,

  taosThreadOnce(&initPoolOnce, initRefPool);
  atexit(cleanupRefPool);
+
  int32_t code = createExecTaskInfoImpl(pSubplan, pTask, readHandle, taskId, sql, model);
  if (code != TSDB_CODE_SUCCESS) {
    goto _error;
--- a/source/libs/executor/src/executorimpl.c
+++ b/source/libs/executor/src/executorimpl.c
@ -1337,7 +1337,9 @@ void doFilter(const SNode* pFilterNode, SSDataBlock* pBlock) {
  if (pFilterNode == NULL) {
    return;
  }
-
+  if (pBlock->info.rows == 0) {
+    return;
+  }
  SFilterInfo* filter = NULL;

  // todo move to the initialization function
@ -1994,16 +1996,9 @@ int32_t loadRemoteDataCallback(void* param, SDataBuf* pMsg, int32_t code) {
  tsem_post(&pExchangeInfo->ready);
  taosReleaseRef(exchangeObjRefPool, pWrapper->exchangeId);

-  taosMemoryFree(pWrapper);
  return TSDB_CODE_SUCCESS;
 }

-static void destroySendMsgInfo(SMsgSendInfo* pMsgBody) {
-  assert(pMsgBody != NULL);
-  taosMemoryFreeClear(pMsgBody->msgInfo.pData);
-  taosMemoryFreeClear(pMsgBody);
-}
-
 void qProcessRspMsg(void* parent, SRpcMsg* pMsg, SEpSet* pEpSet) {
  SMsgSendInfo* pSendInfo = (SMsgSendInfo*)pMsg->info.ahandle;
  assert(pMsg->info.ahandle != NULL);
@ -2063,6 +2058,7 @@ static int32_t doSendFetchDataRequest(SExchangeInfo* pExchangeInfo, SExecTaskInf
  pWrapper->sourceIndex = sourceIndex;

  pMsgSendInfo->param = pWrapper;
+  pMsgSendInfo->paramFreeFp = taosMemoryFree;
  pMsgSendInfo->msgInfo.pData = pMsg;
  pMsgSendInfo->msgInfo.len = sizeof(SResFetchReq);
  pMsgSendInfo->msgType = pSource->fetchMsgType;
@ -4127,7 +4123,7 @@ static SExecTaskInfo* createExecTaskInfo(uint64_t queryId, uint64_t taskId, EOPT
  SExecTaskInfo* pTaskInfo = taosMemoryCalloc(1, sizeof(SExecTaskInfo));
  setTaskStatus(pTaskInfo, TASK_NOT_COMPLETED);

-  pTaskInfo->schemaVer.dbname = strdup(dbFName);
+  pTaskInfo->schemaInfo.dbname = strdup(dbFName);
  pTaskInfo->cost.created = taosGetTimestampMs();
  pTaskInfo->id.queryId = queryId;
  pTaskInfo->execModel = model;
@ -4153,35 +4149,35 @@ int32_t extractTableSchemaInfo(SReadHandle* pHandle, uint64_t uid, SExecTaskInfo
    return terrno;
  }

-  pTaskInfo->schemaVer.tablename = strdup(mr.me.name);
+  pTaskInfo->schemaInfo.tablename = strdup(mr.me.name);

  if (mr.me.type == TSDB_SUPER_TABLE) {
-    pTaskInfo->schemaVer.sw = tCloneSSchemaWrapper(&mr.me.stbEntry.schemaRow);
-    pTaskInfo->schemaVer.tversion = mr.me.stbEntry.schemaTag.version;
+    pTaskInfo->schemaInfo.sw = tCloneSSchemaWrapper(&mr.me.stbEntry.schemaRow);
+    pTaskInfo->schemaInfo.tversion = mr.me.stbEntry.schemaTag.version;
  } else if (mr.me.type == TSDB_CHILD_TABLE) {
    tDecoderClear(&mr.coder);

    tb_uid_t suid = mr.me.ctbEntry.suid;
    metaGetTableEntryByUid(&mr, suid);
-    pTaskInfo->schemaVer.sw = tCloneSSchemaWrapper(&mr.me.stbEntry.schemaRow);
-    pTaskInfo->schemaVer.tversion = mr.me.stbEntry.schemaTag.version;
+    pTaskInfo->schemaInfo.sw = tCloneSSchemaWrapper(&mr.me.stbEntry.schemaRow);
+    pTaskInfo->schemaInfo.tversion = mr.me.stbEntry.schemaTag.version;
  } else {
-    pTaskInfo->schemaVer.sw = tCloneSSchemaWrapper(&mr.me.ntbEntry.schemaRow);
+    pTaskInfo->schemaInfo.sw = tCloneSSchemaWrapper(&mr.me.ntbEntry.schemaRow);
  }

  metaReaderClear(&mr);
  return TSDB_CODE_SUCCESS;
 }

-static void cleanupTableSchemaInfo(SExecTaskInfo* pTaskInfo) {
-  taosMemoryFreeClear(pTaskInfo->schemaVer.dbname);
-  if (pTaskInfo->schemaVer.sw == NULL) {
+static void cleanupTableSchemaInfo(SSchemaInfo* pSchemaInfo) {
+  taosMemoryFreeClear(pSchemaInfo->dbname);
+  if (pSchemaInfo->sw == NULL) {
    return;
  }

-  taosMemoryFree(pTaskInfo->schemaVer.sw->pSchema);
-  taosMemoryFree(pTaskInfo->schemaVer.sw);
-  taosMemoryFree(pTaskInfo->schemaVer.tablename);
+  taosMemoryFree(pSchemaInfo->tablename);
+  taosMemoryFree(pSchemaInfo->sw->pSchema);
+  taosMemoryFree(pSchemaInfo->sw);
 }

 static int32_t sortTableGroup(STableListInfo* pTableListInfo, int32_t groupNum) {
@ -4424,7 +4420,7 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
            .calTrigger = pTableScanNode->triggerType,
            .maxTs = INT64_MIN,
      };
-      if (pHandle) {
+      if (pHandle->vnode) {
        int32_t code = createScanTableListInfo(&pTableScanNode->scan, pTableScanNode->pGroupTags,
                                               pTableScanNode->groupSort, pHandle, pTableListInfo, pTagCond, pTagIndexCond, GET_TASKID(pTaskInfo));
        if (code) {
@ -4590,7 +4586,7 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
    int32_t children = 0;
    pOptr = createStreamFinalIntervalOperatorInfo(ops[0], pPhyNode, pTaskInfo, children);
  } else if (QUERY_NODE_PHYSICAL_PLAN_STREAM_FINAL_INTERVAL == type) {
-    int32_t children = 1;
+    int32_t children = pHandle->numOfVgroups;
    pOptr = createStreamFinalIntervalOperatorInfo(ops[0], pPhyNode, pTaskInfo, children);
  } else if (QUERY_NODE_PHYSICAL_PLAN_SORT == type) {
    pOptr = createSortOperatorInfo(ops[0], (SSortPhysiNode*)pPhyNode, pTaskInfo);
@ -4939,6 +4935,7 @@ int32_t createExecTaskInfoImpl(SSubplan* pPlan, SExecTaskInfo** pTaskInfo, SRead
  }

  (*pTaskInfo)->sql = sql;
+  (*pTaskInfo)->pSubplan = pPlan;
  (*pTaskInfo)->pRoot = createOperatorTree(pPlan->pNode, *pTaskInfo, pHandle, &(*pTaskInfo)->tableqinfoList, pPlan->pTagCond, pPlan->pTagIndexCond, pPlan->user);

  if (NULL == (*pTaskInfo)->pRoot) {
@ -4977,7 +4974,9 @@ void doDestroyTask(SExecTaskInfo* pTaskInfo) {

  doDestroyTableList(&pTaskInfo->tableqinfoList);
  destroyOperatorInfo(pTaskInfo->pRoot);
-  cleanupTableSchemaInfo(pTaskInfo);
+  cleanupTableSchemaInfo(&pTaskInfo->schemaInfo);
+
+  nodesDestroyNode((SNode*)pTaskInfo->pSubplan);

  taosMemoryFreeClear(pTaskInfo->sql);
  taosMemoryFreeClear(pTaskInfo->id.str);
--- a/source/libs/executor/src/scanoperator.c
+++ b/source/libs/executor/src/scanoperator.c
@ -1196,7 +1196,6 @@ static int32_t setBlockIntoRes(SStreamScanInfo* pInfo, const SSDataBlock* pBlock
      if (pResCol->info.colId == pColMatchInfo->colId) {
        SColumnInfoData* pDst = taosArrayGet(pInfo->pRes->pDataBlock, pColMatchInfo->targetSlotId);
        colDataAssign(pDst, pResCol, pBlock->info.rows, &pInfo->pRes->info);
-        //        taosArraySet(pInfo->pRes->pDataBlock, pColMatchInfo->targetSlotId, pResCol);
        colExists = true;
        break;
      }
@ -1460,6 +1459,7 @@ static SSDataBlock* doStreamScan(SOperatorInfo* pOperator) {
        }
      }
    }
+
    qDebug("scan rows: %d", pBlockInfo->rows);
    return (pBlockInfo->rows == 0) ? NULL : pInfo->pRes;

@ -1532,8 +1532,7 @@ SOperatorInfo* createStreamScanOperatorInfo(SReadHandle* pHandle, STableScanPhys
    goto _error;
  }

-  SScanPhysiNode* pScanPhyNode = &pTableScanNode->scan;
-
+  SScanPhysiNode*     pScanPhyNode = &pTableScanNode->scan;
  SDataBlockDescNode* pDescNode = pScanPhyNode->node.pOutputDataBlockDesc;

  pInfo->pTagCond = pTagCond;
@ -1564,7 +1563,7 @@ SOperatorInfo* createStreamScanOperatorInfo(SReadHandle* pHandle, STableScanPhys
    goto _error;
  }

-  if (pHandle) {
+  if (pHandle->vnode) {
    SOperatorInfo*  pTableScanOp = createTableScanOperatorInfo(pTableScanNode, pHandle, pTaskInfo);
    STableScanInfo* pTSInfo = (STableScanInfo*)pTableScanOp->info;
    if (pHandle->version > 0) {
--- a/source/libs/executor/src/timewindowoperator.c
+++ b/source/libs/executor/src/timewindowoperator.c
@ -1543,6 +1543,7 @@ void destroyStreamFinalIntervalOperatorInfo(void* param, int32_t numOfOutput) {
  taosArrayDestroy(pInfo->pPullWins);
  blockDataDestroy(pInfo->pPullDataRes);
  taosArrayDestroy(pInfo->pRecycledPages);
+  blockDataDestroy(pInfo->pUpdateRes);

  if (pInfo->pChildren) {
    int32_t size = taosArrayGetSize(pInfo->pChildren);
--- a/source/libs/function/src/builtins.c
+++ b/source/libs/function/src/builtins.c
@ -1646,9 +1646,9 @@ static int32_t translateCast(SFunctionNode* pFunc, char* pErrBuf, int32_t len) {
  if (IS_VAR_DATA_TYPE(para2Type)) {
    para2Bytes -= VARSTR_HEADER_SIZE;
  }
-  if (para2Bytes <= 0 || para2Bytes > 1000) {  // cast dst var type length limits to 1000
+  if (para2Bytes <= 0 || para2Bytes > 4096) {  // cast dst var type length limits to 4096 bytes
    return buildFuncErrMsg(pErrBuf, len, TSDB_CODE_FUNC_FUNTION_ERROR,
-                           "CAST function converted length should be in range [0, 1000]");
+                           "CAST function converted length should be in range [0, 4096] bytes");
  }

  // add database precision as param
--- a/source/libs/function/src/builtinsimpl.c
+++ b/source/libs/function/src/builtinsimpl.c
@ -5098,12 +5098,7 @@ bool modeFunctionSetup(SqlFunctionCtx* pCtx, SResultRowEntryInfo* pResInfo) {
  return true;
 }

-static void doModeAdd(SModeInfo* pInfo, char* data, bool isNull) {
-  // ignore null elements
-  if (isNull) {
-    return;
-  }
-
+static void doModeAdd(SModeInfo* pInfo, char* data) {
  int32_t     hashKeyBytes = IS_VAR_DATA_TYPE(pInfo->colType) ? varDataTLen(data) : pInfo->colBytes;
  SModeItem** pHashItem = taosHashGet(pInfo->pHash, data, hashKeyBytes);
  if (pHashItem == NULL) {
@ -5128,10 +5123,16 @@ int32_t modeFunction(SqlFunctionCtx* pCtx) {
  SColumnInfoData* pInputCol = pInput->pData[0];
  SColumnInfoData* pOutput = (SColumnInfoData*)pCtx->pOutput;

+  int32_t numOfElems = 0;
  int32_t startOffset = pCtx->offset;
  for (int32_t i = pInput->startRowIndex; i < pInput->numOfRows + pInput->startRowIndex; ++i) {
    char* data = colDataGetData(pInputCol, i);
-    doModeAdd(pInfo, data, colDataIsNull_s(pInputCol, i));
+    if (colDataIsNull_s(pInputCol, i)) {
+      continue;
+    }
+
+    numOfElems++;
+    doModeAdd(pInfo, data);

    if (sizeof(SModeInfo) + pInfo->numOfPoints * (sizeof(SModeItem) + pInfo->colBytes) >= MODE_MAX_RESULT_SIZE) {
      taosHashCleanup(pInfo->pHash);
@ -5139,7 +5140,7 @@ int32_t modeFunction(SqlFunctionCtx* pCtx) {
    }
  }

-  SET_VAL(pResInfo, 1, 1);
+  SET_VAL(pResInfo, numOfElems, 1);

  return TSDB_CODE_SUCCESS;
 }
@ -5661,7 +5662,12 @@ int32_t derivativeFunction(SqlFunctionCtx* pCtx) {
        double r = ((v - pDerivInfo->prevValue) * pDerivInfo->tsWindow) / (tsList[i] - pDerivInfo->prevTs);
        if (pDerivInfo->ignoreNegative && r < 0) {
        } else {
-          colDataAppend(pOutput, pos, (const char*)&r, false);
+          if (isinf(r) || isnan(r)) {
+            colDataAppendNULL(pOutput, pos);
+          } else {
+            colDataAppend(pOutput, pos, (const char*)&r, false);
+          }
+
          if (pTsOutput != NULL) {
            colDataAppendInt64(pTsOutput, pos, &tsList[i]);
          }
@ -5688,7 +5694,12 @@ int32_t derivativeFunction(SqlFunctionCtx* pCtx) {
        double r = ((pDerivInfo->prevValue - v) * pDerivInfo->tsWindow) / (pDerivInfo->prevTs - tsList[i]);
        if (pDerivInfo->ignoreNegative && r < 0) {
        } else {
-          colDataAppend(pOutput, pos, (const char*)&r, false);
+          if (isinf(r) || isnan(r)) {
+            colDataAppendNULL(pOutput, pos);
+          } else {
+            colDataAppend(pOutput, pos, (const char*)&r, false);
+          }
+
          if (pTsOutput != NULL) {
            colDataAppendInt64(pTsOutput, pos, &pDerivInfo->prevTs);
          }
--- a/source/libs/nodes/src/nodesUtilFuncs.c
+++ b/source/libs/nodes/src/nodesUtilFuncs.c
@ -1923,15 +1923,18 @@ int32_t nodesPartitionCond(SNode** pCondition, SNode** pPrimaryKeyCond, SNode**
    return partitionLogicCond(pCondition, pPrimaryKeyCond, pTagIndexCond, pTagCond, pOtherCond);
  }

+  bool needOutput = false;
  switch (classifyCondition(*pCondition)) {
    case COND_TYPE_PRIMARY_KEY:
      if (NULL != pPrimaryKeyCond) {
        *pPrimaryKeyCond = *pCondition;
+        needOutput = true;
      }
      break;
    case COND_TYPE_TAG_INDEX:
      if (NULL != pTagIndexCond) {
        *pTagIndexCond = *pCondition;
+        needOutput = true;
      }
      if (NULL != pTagCond) {
        SNode* pTempCond = *pCondition;
@ -1942,21 +1945,26 @@ int32_t nodesPartitionCond(SNode** pCondition, SNode** pPrimaryKeyCond, SNode**
          }
        }
        *pTagCond = pTempCond;
+        needOutput = true;
      }
      break;
    case COND_TYPE_TAG:
      if (NULL != pTagCond) {
        *pTagCond = *pCondition;
+        needOutput = true;
      }
      break;
    case COND_TYPE_NORMAL:
    default:
      if (NULL != pOtherCond) {
        *pOtherCond = *pCondition;
+        needOutput = true;
      }
      break;
  }
-  *pCondition = NULL;
+  if (needOutput) {
+    *pCondition = NULL;
+  }

  return TSDB_CODE_SUCCESS;
 }
--- a/source/libs/parser/src/parInsert.c
+++ b/source/libs/parser/src/parInsert.c
@ -2161,7 +2161,7 @@ static void smlDestroyTableHandle(void* pHandle) {
  tdDestroySVCreateTbReq(&handle->createTblReq);
 }

-static int32_t smlBoundColumnData(SArray* cols, SParsedDataColInfo* pColList, SSchema* pSchema) {
+static int32_t smlBoundColumnData(SArray* cols, SParsedDataColInfo* pColList, SSchema* pSchema, bool isTag) {
  col_id_t nCols = pColList->numOfCols;

  pColList->numOfBound = 0;
@ -2177,7 +2177,8 @@ static int32_t smlBoundColumnData(SArray* cols, SParsedDataColInfo* pColList, SS
    SSmlKv*  kv = taosArrayGetP(cols, i);
    SToken   sToken = {.n = kv->keyLen, .z = (char*)kv->key};
    col_id_t t = lastColIdx + 1;
-    col_id_t index = findCol(&sToken, t, nCols, pSchema);
+    col_id_t index = ((t == 0 && !isTag) ? 0 : findCol(&sToken, t, nCols, pSchema));
+    uDebug("SML, index:%d, t:%d, ncols:%d, kv->name:%s", index, t, nCols, kv->key);
    if (index < 0 && t > 0) {
      index = findCol(&sToken, 0, t, pSchema);
      isOrdered = false;
@ -2312,7 +2313,7 @@ int32_t smlBindData(void* handle, SArray* tags, SArray* colsSchema, SArray* cols
  smlDestroyTableHandle(&smlHandle->tableExecHandle);  // free for each table
  SSchema* pTagsSchema = getTableTagSchema(pTableMeta);
  setBoundColumnInfo(&smlHandle->tableExecHandle.tags, pTagsSchema, getNumOfTags(pTableMeta));
-  int ret = smlBoundColumnData(tags, &smlHandle->tableExecHandle.tags, pTagsSchema);
+  int ret = smlBoundColumnData(tags, &smlHandle->tableExecHandle.tags, pTagsSchema, true);
  if (ret != TSDB_CODE_SUCCESS) {
    buildInvalidOperationMsg(&pBuf, "bound tags error");
    return ret;
@ -2343,7 +2344,7 @@ int32_t smlBindData(void* handle, SArray* tags, SArray* colsSchema, SArray* cols

  SSchema* pSchema = getTableColumnSchema(pTableMeta);

-  ret = smlBoundColumnData(colsSchema, &pDataBlock->boundColumnInfo, pSchema);
+  ret = smlBoundColumnData(colsSchema, &pDataBlock->boundColumnInfo, pSchema, false);
  if (ret != TSDB_CODE_SUCCESS) {
    buildInvalidOperationMsg(&pBuf, "bound cols error");
    return ret;
@ -2384,7 +2385,7 @@ int32_t smlBindData(void* handle, SArray* tags, SArray* colsSchema, SArray* cols
      if (format) {
        if (j < rowDataSize) {
          kv = taosArrayGetP(rowData, j);
-          if (rowDataSize != spd->numOfBound &&
+          if (rowDataSize != spd->numOfBound && j != 0 &&
              (kv->keyLen != strlen(pColSchema->name) || strncmp(kv->key, pColSchema->name, kv->keyLen) != 0)) {
            kv = NULL;
          } else {
@ -2401,7 +2402,9 @@ int32_t smlBindData(void* handle, SArray* tags, SArray* colsSchema, SArray* cols
      } else {
        int32_t colLen = kv->length;
        if (pColSchema->type == TSDB_DATA_TYPE_TIMESTAMP) {
+//          uError("SML:data before:%ld, precision:%d", kv->i, pTableMeta->tableInfo.precision);
          kv->i = convertTimePrecision(kv->i, TSDB_TIME_PRECISION_NANO, pTableMeta->tableInfo.precision);
+//          uError("SML:data after:%ld, precision:%d", kv->i, pTableMeta->tableInfo.precision);
        }

        if (IS_VAR_DATA_TYPE(kv->type)) {
--- a/source/libs/qcom/src/queryUtil.c
+++ b/source/libs/qcom/src/queryUtil.c
@ -138,6 +138,16 @@ int32_t taosAsyncExec(__async_exec_fn_t execFn, void* execParam, int32_t* code)
  return 0;
 }

+void destroySendMsgInfo(SMsgSendInfo* pMsgBody) {
+  assert(pMsgBody != NULL);
+  taosMemoryFreeClear(pMsgBody->target.dbFName);
+  taosMemoryFreeClear(pMsgBody->msgInfo.pData);
+  if (pMsgBody->paramFreeFp) {
+    (*pMsgBody->paramFreeFp)(pMsgBody->param);
+  }
+  taosMemoryFreeClear(pMsgBody);
+}
+
 int32_t asyncSendMsgToServerExt(void* pTransporter, SEpSet* epSet, int64_t* pTransporterId, const SMsgSendInfo* pInfo,
                                bool persistHandle, void* rpcCtx) {
  char* pMsg = rpcMallocCont(pInfo->msgInfo.len);
--- a/source/libs/qworker/inc/qwInt.h
+++ b/source/libs/qworker/inc/qwInt.h
@ -150,7 +150,6 @@ typedef struct SQWTaskCtx {

  void     *taskHandle;
  void     *sinkHandle;
-  SSubplan *plan;
  STbVerInfo tbInfo;
 } SQWTaskCtx;

--- a/source/libs/qworker/src/qwUtil.c
+++ b/source/libs/qworker/src/qwUtil.c
@ -306,11 +306,6 @@ void qwFreeTaskCtx(SQWTaskCtx *ctx) {
    dsDestroyDataSinker(ctx->sinkHandle);
    ctx->sinkHandle = NULL;
  }
-
-  if (ctx->plan) {
-    nodesDestroyNode((SNode*)ctx->plan);
-    ctx->plan = NULL;
-  }
 }

 int32_t qwDropTaskCtx(QW_FPARAMS_DEF) {
@ -327,7 +322,6 @@ int32_t qwDropTaskCtx(QW_FPARAMS_DEF) {

  atomic_store_ptr(&ctx->taskHandle, NULL);
  atomic_store_ptr(&ctx->sinkHandle, NULL);
-  atomic_store_ptr(&ctx->plan, NULL);

  QW_SET_EVENT_PROCESSED(ctx, QW_EVENT_DROP);

--- a/source/libs/qworker/src/qworker.c
+++ b/source/libs/qworker/src/qworker.c
@ -522,8 +522,6 @@ int32_t qwProcessQuery(QW_FPARAMS_DEF, SQWMsg *qwMsg, const char* sql) {
    QW_ERR_JRET(code);
  }

-  ctx->plan = plan;
-
  code = qCreateExecTask(qwMsg->node, mgmt->nodeId, tId, plan, &pTaskInfo, &sinkHandle, sql, OPTR_EXEC_MODEL_BATCH);
  if (code) {
    QW_TASK_ELOG("qCreateExecTask failed, code:%x - %s", code, tstrerror(code));
@ -928,8 +926,6 @@ int32_t qwProcessDelete(QW_FPARAMS_DEF, SQWMsg *qwMsg, SDeleteRes *pRes) {
    QW_ERR_JRET(code);
  }

-  ctx.plan = plan;
-  
  code = qCreateExecTask(qwMsg->node, mgmt->nodeId, tId, plan, &pTaskInfo, &sinkHandle, NULL, OPTR_EXEC_MODEL_BATCH);
  if (code) {
    QW_TASK_ELOG("qCreateExecTask failed, code:%x - %s", code, tstrerror(code));
--- a/source/libs/scheduler/inc/schInt.h
+++ b/source/libs/scheduler/inc/schInt.h
@ -277,7 +277,7 @@ typedef struct SSchJob {
  bool               fetched;
  int32_t            resNumOfRows;
  SSchResInfo        userRes;
-  const char        *sql;
+  char              *sql;
  SQueryProfileSummary summary;
 } SSchJob;

@ -461,7 +461,6 @@ int32_t schJobFetchRows(SSchJob *pJob);
 int32_t schJobFetchRowsA(SSchJob *pJob);
 int32_t schUpdateTaskHandle(SSchJob *pJob, SSchTask *pTask, bool dropExecNode, void *handle, int32_t execId);
 int32_t schProcessOnTaskStatusRsp(SQueryNodeEpId* pEpId, SArray* pStatusList);
-void    schFreeSMsgSendInfo(SMsgSendInfo *msgSendInfo);
 char*   schGetOpStr(SCH_OP_TYPE type);
 int32_t schBeginOperation(SSchJob *pJob, SCH_OP_TYPE type, bool sync);
 int32_t schInitJob(int64_t *pJobId, SSchedulerReq *pReq);
--- a/source/libs/scheduler/src/schJob.c
+++ b/source/libs/scheduler/src/schJob.c
@ -675,6 +675,7 @@ void schFreeJobImpl(void *job) {

  taosMemoryFreeClear(pJob->userRes.execRes);
  taosMemoryFreeClear(pJob->fetchRes);
+  taosMemoryFreeClear(pJob->sql);
  taosMemoryFree(pJob);

  int32_t jobNum = atomic_sub_fetch_32(&schMgmt.jobNum, 1);
@ -718,7 +719,9 @@ int32_t schInitJob(int64_t *pJobId, SSchedulerReq *pReq) {

  pJob->attr.explainMode = pReq->pDag->explainInfo.mode;
  pJob->conn = *pReq->pConn;
-  pJob->sql = pReq->sql;
+  if (pReq->sql) {
+    pJob->sql = strdup(pReq->sql);
+  }
  pJob->pDag = pReq->pDag;
  pJob->chkKillFp = pReq->chkKillFp;
  pJob->chkKillParam = pReq->chkKillParam;
--- a/source/libs/scheduler/src/schRemote.c
+++ b/source/libs/scheduler/src/schRemote.c
@ -386,7 +386,6 @@ int32_t schHandleCallback(void *param, SDataBuf *pMsg, int32_t rspCode) {
  schProcessOnCbEnd(pJob, pTask, code);

  taosMemoryFreeClear(pMsg->pData);
-  taosMemoryFreeClear(param);

  qDebug("end to handle rsp msg, type:%s, handle:%p, code:%s", TMSG_INFO(pMsg->msgType), pMsg->handle,
         tstrerror(rspCode));
@ -398,7 +397,6 @@ int32_t schHandleDropCallback(void *param, SDataBuf *pMsg, int32_t code) {
  SSchTaskCallbackParam *pParam = (SSchTaskCallbackParam *)param;
  qDebug("QID:0x%" PRIx64 ",TID:0x%" PRIx64 " drop task rsp received, code:0x%x", pParam->queryId, pParam->taskId,
         code);
-  taosMemoryFreeClear(param);
  return TSDB_CODE_SUCCESS;
 }

@ -447,8 +445,8 @@ int32_t schHandleHbCallback(void *param, SDataBuf *pMsg, int32_t code) {
  SCH_ERR_JRET(schProcessOnTaskStatusRsp(&rsp.epId, rsp.taskStatus));

 _return:
+
  tFreeSSchedulerHbRsp(&rsp);
-  taosMemoryFree(param);
  taosMemoryFree(pMsg->pData);
  SCH_RET(code);
 }
@ -514,7 +512,9 @@ int32_t schGenerateCallBackInfo(SSchJob *pJob, SSchTask *pTask, void *msg, uint3
    SCH_ERR_RET(TSDB_CODE_QRY_OUT_OF_MEMORY);
  }

+  msgSendInfo->paramFreeFp = taosMemoryFree;      
  SCH_ERR_JRET(schMakeCallbackParam(pJob, pTask, msgType, isHb, trans, &msgSendInfo->param));
+
  SCH_ERR_JRET(schGetCallbackFp(msgType, &msgSendInfo->fp));

  if (pJob) {
@ -535,7 +535,7 @@ int32_t schGenerateCallBackInfo(SSchJob *pJob, SSchTask *pTask, void *msg, uint3

 _return:

-  schFreeSMsgSendInfo(msgSendInfo);
+  destroySendMsgInfo(msgSendInfo);

  SCH_RET(code);
 }
@ -676,6 +676,7 @@ int32_t schMakeHbRpcCtx(SSchJob *pJob, SSchTask *pTask, SRpcCtx *pCtx) {
  param->pTrans = pJob->conn.pTrans;

  pMsgSendInfo->param = param;
+  pMsgSendInfo->paramFreeFp = taosMemoryFree;      
  pMsgSendInfo->fp = fp;

  SRpcCtxVal ctxVal = {.val = pMsgSendInfo, .clone = schCloneSMsgSendInfo};
@ -795,6 +796,7 @@ int32_t schCloneSMsgSendInfo(void *src, void **dst) {
  pDst->param = NULL;

  SCH_ERR_JRET(schCloneCallbackParam(pSrc->param, (SSchCallbackParamHeader **)&pDst->param));
+  pDst->paramFreeFp = taosMemoryFree;      

  *dst = pDst;

@ -861,8 +863,7 @@ _return:
  }

  if (pMsgSendInfo) {
-    taosMemoryFreeClear(pMsgSendInfo->param);
-    taosMemoryFreeClear(pMsgSendInfo);
+    destroySendMsgInfo(pMsgSendInfo);
  }

  SCH_RET(code);
--- a/source/libs/scheduler/src/schUtil.c
+++ b/source/libs/scheduler/src/schUtil.c
@ -50,6 +50,12 @@ char* schGetOpStr(SCH_OP_TYPE type) {
  }
 }

+void schFreeHbTrans(SSchHbTrans *pTrans) {
+  rpcReleaseHandle(pTrans->trans.pHandle, TAOS_CONN_CLIENT);
+
+  schFreeRpcCtx(&pTrans->rpcCtx);  
+}
+
 void schCleanClusterHb(void* pTrans) {
  SCH_LOCK(SCH_WRITE, &schMgmt.hbLock);

@ -57,7 +63,7 @@ void schCleanClusterHb(void* pTrans) {
  while (hb) {
    if (hb->trans.pTrans == pTrans) {
      SQueryNodeEpId* pEpId = taosHashGetKey(hb, NULL);
-      rpcReleaseHandle(hb->trans.pHandle, TAOS_CONN_CLIENT);
+      schFreeHbTrans(hb);
      taosHashRemove(schMgmt.hbConnections, pEpId, sizeof(SQueryNodeEpId));
    }
    
@ -68,8 +74,6 @@ void schCleanClusterHb(void* pTrans) {
 }

 int32_t schRemoveHbConnection(SSchJob *pJob, SSchTask *pTask, SQueryNodeEpId *epId) {
-  return TSDB_CODE_SUCCESS; // TODO ENABLE IT WHEN RPC IS READY
-  
  int32_t     code = 0;

  SCH_LOCK(SCH_WRITE, &schMgmt.hbLock);
@ -82,7 +86,7 @@ int32_t schRemoveHbConnection(SSchJob *pJob, SSchTask *pTask, SQueryNodeEpId *ep

  int64_t taskNum = atomic_load_64(&hb->taskNum);
  if (taskNum <= 0) {
-    rpcReleaseHandle(hb->trans.pHandle, TAOS_CONN_CLIENT);
+    schFreeHbTrans(hb);
    taosHashRemove(schMgmt.hbConnections, epId, sizeof(SQueryNodeEpId));
  }
  SCH_UNLOCK(SCH_WRITE, &schMgmt.hbLock);
@ -265,9 +269,7 @@ void schFreeRpcCtxVal(const void *arg) {
  }

  SMsgSendInfo *pMsgSendInfo = (SMsgSendInfo *)arg;
-  taosMemoryFreeClear(pMsgSendInfo->param);
-  taosMemoryFreeClear(pMsgSendInfo->msgInfo.pData);  
-  taosMemoryFreeClear(pMsgSendInfo);
+  destroySendMsgInfo(pMsgSendInfo);
 }

 void schFreeRpcCtx(SRpcCtx *pCtx) {
@ -290,15 +292,6 @@ void schFreeRpcCtx(SRpcCtx *pCtx) {
  }
 }

-void schFreeSMsgSendInfo(SMsgSendInfo *msgSendInfo) {
-  if (NULL == msgSendInfo) {
-    return;
-  }
-
-  taosMemoryFree(msgSendInfo->param);
-  taosMemoryFree(msgSendInfo);
-}
-
 int32_t schGetTaskFromList(SHashObj *pTaskList, uint64_t taskId, SSchTask **pTask) {
  int32_t s = taosHashGetSize(pTaskList);
  if (s <= 0) {
--- a/source/libs/stream/src/streamTask.c
+++ b/source/libs/stream/src/streamTask.c
@ -64,6 +64,7 @@ int32_t tEncodeSStreamTask(SEncoder* pEncoder, const SStreamTask* pTask) {
  if (tEncodeI32(pEncoder, pTask->selfChildId) < 0) return -1;
  if (tEncodeI32(pEncoder, pTask->nodeId) < 0) return -1;
  if (tEncodeSEpSet(pEncoder, &pTask->epSet) < 0) return -1;
+  if (tEncodeI32(pEncoder, pTask->numOfVgroups) < 0) return -1;

  int32_t epSz = taosArrayGetSize(pTask->childEpInfo);
  if (tEncodeI32(pEncoder, epSz) < 0) return -1;
@ -118,6 +119,7 @@ int32_t tDecodeSStreamTask(SDecoder* pDecoder, SStreamTask* pTask) {
  if (tDecodeI32(pDecoder, &pTask->selfChildId) < 0) return -1;
  if (tDecodeI32(pDecoder, &pTask->nodeId) < 0) return -1;
  if (tDecodeSEpSet(pDecoder, &pTask->epSet) < 0) return -1;
+  if (tDecodeI32(pDecoder, &pTask->numOfVgroups) < 0) return -1;

  int32_t epSz;
  if (tDecodeI32(pDecoder, &epSz) < 0) return -1;
--- a/source/libs/sync/src/syncElection.c
+++ b/source/libs/sync/src/syncElection.c
@ -125,7 +125,7 @@ int32_t syncNodeRequestVote(SSyncNode* pSyncNode, const SRaftId* destRaftId, con
    syncUtilU642Addr(destRaftId->addr, host, sizeof(host), &port);
    sDebug("vgId:%d, send sync-request-vote to %s:%d, {term:%" PRIu64 ", last-index:%" PRId64 ", last-term:%" PRIu64
           "}",
-           pSyncNode->vgId, host, port, pMsg->term, pMsg->lastLogTerm, pMsg->lastLogIndex);
+           pSyncNode->vgId, host, port, pMsg->term, pMsg->lastLogIndex, pMsg->lastLogTerm);
  } while (0);

  SRpcMsg rpcMsg;
--- a/source/libs/sync/src/syncMain.c
+++ b/source/libs/sync/src/syncMain.c
@ -1539,20 +1539,21 @@ void syncNodeEventLog(const SSyncNode* pSyncNode, char* str) {
    if (pSyncNode != NULL && pSyncNode->pRaftCfg != NULL && pSyncNode->pRaftStore != NULL) {
      snprintf(logBuf, sizeof(logBuf),
               "vgId:%d, sync %s %s, term:%" PRIu64 ", commit:%" PRId64 ", first:%" PRId64 ", last:%" PRId64
-               ", snapshot:%" PRId64
+               ", snapshot:%" PRId64 ", snapshot-term:%" PRIu64
               ", standby:%d, "
               "strategy:%d, batch:%d, "
               "replica-num:%d, "
               "lconfig:%" PRId64 ", changing:%d, restore:%d, %s",
               pSyncNode->vgId, syncUtilState2String(pSyncNode->state), str, pSyncNode->pRaftStore->currentTerm,
-               pSyncNode->commitIndex, logBeginIndex, logLastIndex, snapshot.lastApplyIndex,
+               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);
    } else {
      snprintf(logBuf, sizeof(logBuf), "%s", str);
    }
-    sDebug("%s", logBuf);
+    // sDebug("%s", logBuf);
+    sInfo("%s", logBuf);

  } else {
    int   len = 256 + userStrLen;
@ -1560,20 +1561,21 @@ void syncNodeEventLog(const SSyncNode* pSyncNode, char* str) {
    if (pSyncNode != NULL && pSyncNode->pRaftCfg != NULL && pSyncNode->pRaftStore != NULL) {
      snprintf(s, len,
               "vgId:%d, sync %s %s, term:%" PRIu64 ", commit:%" PRId64 ", first:%" PRId64 ", last:%" PRId64
-               ", snapshot:%" PRId64
+               ", snapshot:%" PRId64 ", snapshot-term:%" PRIu64
               ", standby:%d, "
               "strategy:%d, batch:%d, "
               "replica-num:%d, "
               "lconfig:%" PRId64 ", changing:%d, restore:%d, %s",
               pSyncNode->vgId, syncUtilState2String(pSyncNode->state), str, pSyncNode->pRaftStore->currentTerm,
-               pSyncNode->commitIndex, logBeginIndex, logLastIndex, snapshot.lastApplyIndex,
+               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);
    } else {
      snprintf(s, len, "%s", str);
    }
-    sDebug("%s", s);
+    // sDebug("%s", s);
+    sInfo("%s", s);
    taosMemoryFree(s);
  }

@ -1606,14 +1608,16 @@ void syncNodeErrorLog(const SSyncNode* pSyncNode, char* str) {
    if (pSyncNode != NULL && pSyncNode->pRaftCfg != NULL && pSyncNode->pRaftStore != NULL) {
      snprintf(logBuf, sizeof(logBuf),
               "vgId:%d, sync %s %s, term:%" PRIu64 ", commit:%" PRId64 ", first:%" PRId64 ", last:%" PRId64
-               ", snapshot:%" PRId64
+               ", snapshot:%" PRId64 ", snapshot-term:%" PRIu64
               ", standby:%d, "
+               "strategy:%d, batch:%d, "
               "replica-num:%d, "
               "lconfig:%" PRId64 ", changing:%d, restore:%d, %s",
               pSyncNode->vgId, syncUtilState2String(pSyncNode->state), str, pSyncNode->pRaftStore->currentTerm,
-               pSyncNode->commitIndex, logBeginIndex, logLastIndex, snapshot.lastApplyIndex,
-               pSyncNode->pRaftCfg->isStandBy, pSyncNode->replicaNum, pSyncNode->pRaftCfg->lastConfigIndex,
-               pSyncNode->changing, pSyncNode->restoreFinish, printStr);
+               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);
    } else {
      snprintf(logBuf, sizeof(logBuf), "%s", str);
    }
@ -1625,14 +1629,16 @@ void syncNodeErrorLog(const SSyncNode* pSyncNode, char* str) {
    if (pSyncNode != NULL && pSyncNode->pRaftCfg != NULL && pSyncNode->pRaftStore != NULL) {
      snprintf(s, len,
               "vgId:%d, sync %s %s, term:%" PRIu64 ", commit:%" PRId64 ", first:%" PRId64 ", last:%" PRId64
-               ", snapshot:%" PRId64
+               ", snapshot:%" PRId64 ", snapshot-term:%" PRIu64
               ", standby:%d, "
+               "strategy:%d, batch:%d, "
               "replica-num:%d, "
               "lconfig:%" PRId64 ", changing:%d, restore:%d, %s",
               pSyncNode->vgId, syncUtilState2String(pSyncNode->state), str, pSyncNode->pRaftStore->currentTerm,
-               pSyncNode->commitIndex, logBeginIndex, logLastIndex, snapshot.lastApplyIndex,
-               pSyncNode->pRaftCfg->isStandBy, pSyncNode->replicaNum, pSyncNode->pRaftCfg->lastConfigIndex,
-               pSyncNode->changing, pSyncNode->restoreFinish, printStr);
+               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);
    } else {
      snprintf(s, len, "%s", str);
    }
--- a/source/libs/sync/src/syncRaftLog.c
+++ b/source/libs/sync/src/syncRaftLog.c
@ -313,7 +313,7 @@ static int32_t raftLogGetEntry(struct SSyncLogStore* pLogStore, SyncIndex index,
      snprintf(logBuf, sizeof(logBuf), "wal read error, index:%" PRId64 ", err:%d %X, msg:%s, syserr:%d, sysmsg:%s",
               index, err, err, errStr, sysErr, sysErrStr);
      if (terrno == TSDB_CODE_WAL_LOG_NOT_EXIST) {
-        syncNodeEventLog(pData->pSyncNode, logBuf);
+        // syncNodeEventLog(pData->pSyncNode, logBuf);
      } else {
        syncNodeErrorLog(pData->pSyncNode, logBuf);
      }
@ -499,7 +499,7 @@ SSyncRaftEntry* logStoreGetEntry(SSyncLogStore* pLogStore, SyncIndex index) {
        snprintf(logBuf, sizeof(logBuf), "wal read error, index:%" PRId64 ", err:%d %X, msg:%s, syserr:%d, sysmsg:%s",
                 index, err, err, errStr, sysErr, sysErrStr);
        if (terrno == TSDB_CODE_WAL_LOG_NOT_EXIST) {
-          syncNodeEventLog(pData->pSyncNode, logBuf);
+          // syncNodeEventLog(pData->pSyncNode, logBuf);
        } else {
          syncNodeErrorLog(pData->pSyncNode, logBuf);
        }
--- a/source/libs/sync/src/syncRequestVote.c
+++ b/source/libs/sync/src/syncRequestVote.c
@ -159,16 +159,53 @@ static bool syncNodeOnRequestVoteLogOK(SSyncNode* pSyncNode, SyncRequestVote* pM
  SyncIndex myLastIndex = syncNodeGetLastIndex(pSyncNode);

  if (myLastTerm == SYNC_TERM_INVALID) {
+    do {
+      char logBuf[128];
+      snprintf(logBuf, sizeof(logBuf),
+               "logok:0, {my-lterm:%" PRIu64 ", my-lindex:%" PRId64 ", recv-lterm:%" PRIu64 ", recv-lindex:%" PRId64
+               ", recv-term:%" PRIu64 "}",
+               myLastTerm, myLastIndex, pMsg->lastLogTerm, pMsg->lastLogIndex, pMsg->term);
+      syncNodeEventLog(pSyncNode, logBuf);
+    } while (0);
+
    return false;
  }

  if (pMsg->lastLogTerm > myLastTerm) {
+    do {
+      char logBuf[128];
+      snprintf(logBuf, sizeof(logBuf),
+               "logok:1, {my-lterm:%" PRIu64 ", my-lindex:%" PRId64 ", recv-lterm:%" PRIu64 ", recv-lindex:%" PRId64
+               ", recv-term:%" PRIu64 "}",
+               myLastTerm, myLastIndex, pMsg->lastLogTerm, pMsg->lastLogIndex, pMsg->term);
+      syncNodeEventLog(pSyncNode, logBuf);
+    } while (0);
+
    return true;
  }
+
  if (pMsg->lastLogTerm == myLastTerm && pMsg->lastLogIndex >= myLastIndex) {
+    do {
+      char logBuf[128];
+      snprintf(logBuf, sizeof(logBuf),
+               "logok:1, {my-lterm:%" PRIu64 ", my-lindex:%" PRId64 ", recv-lterm:%" PRIu64 ", recv-lindex:%" PRId64
+               ", recv-term:%" PRIu64 "}",
+               myLastTerm, myLastIndex, pMsg->lastLogTerm, pMsg->lastLogIndex, pMsg->term);
+      syncNodeEventLog(pSyncNode, logBuf);
+    } while (0);
+
    return true;
  }

+  do {
+    char logBuf[128];
+    snprintf(logBuf, sizeof(logBuf),
+             "logok:0, {my-lterm:%" PRIu64 ", my-lindex:%" PRId64 ", recv-lterm:%" PRIu64 ", recv-lindex:%" PRId64
+             ", recv-term:%" PRIu64 "}",
+             myLastTerm, myLastIndex, pMsg->lastLogTerm, pMsg->lastLogIndex, pMsg->term);
+    syncNodeEventLog(pSyncNode, logBuf);
+  } while (0);
+
  return false;
 }

@ -224,8 +261,8 @@ int32_t syncNodeOnRequestVoteSnapshotCb(SSyncNode* ths, SyncRequestVote* pMsg) {
    uint16_t port;
    syncUtilU642Addr(pMsg->srcId.addr, host, sizeof(host), &port);
    snprintf(logBuf, sizeof(logBuf),
-             "recv sync-request-vote from %s:%d, term:%" PRIu64 ", lindex:%" PRId64 ", lterm:%" PRIu64
-             ", reply-grant:%d",
+             "recv sync-request-vote from %s:%d, {term:%" PRIu64 ", lindex:%" PRId64 ", lterm:%" PRIu64
+             ", reply-grant:%d}",
             host, port, pMsg->term, pMsg->lastLogIndex, pMsg->lastLogTerm, pReply->voteGranted);
    syncNodeEventLog(ths, logBuf);
  } while (0);
--- a/source/libs/sync/src/syncRespMgr.c
+++ b/source/libs/sync/src/syncRespMgr.c
@ -120,15 +120,18 @@ void syncRespClean(SSyncRespMgr *pObj) {
 void syncRespCleanByTTL(SSyncRespMgr *pObj, int64_t ttl) {
  SRespStub *pStub = (SRespStub *)taosHashIterate(pObj->pRespHash, NULL);
  int        cnt = 0;
+  int        sum = 0;
  SSyncNode *pSyncNode = pObj->data;

  SArray *delIndexArray = taosArrayInit(0, sizeof(uint64_t));
  ASSERT(delIndexArray != NULL);
+  sDebug("vgId:%d, resp mgr begin clean by ttl", pSyncNode->vgId);

  while (pStub) {
    size_t    len;
-    void *    key = taosHashGetKey(pStub, &len);
+    void     *key = taosHashGetKey(pStub, &len);
    uint64_t *pSeqNum = (uint64_t *)key;
+    sum++;

    int64_t nowMS = taosGetTimestampMs();
    if (nowMS - pStub->createTime > ttl) {
@ -155,7 +158,7 @@ void syncRespCleanByTTL(SSyncRespMgr *pObj, int64_t ttl) {
  }

  int32_t arraySize = taosArrayGetSize(delIndexArray);
-  sDebug("vgId:%d, resp mgr clean by ttl, cnt:%d, array-size:%d", pSyncNode->vgId, cnt, arraySize);
+  sDebug("vgId:%d, resp mgr end clean by ttl, sum:%d, cnt:%d, array-size:%d", pSyncNode->vgId, sum, cnt, arraySize);

  for (int32_t i = 0; i < arraySize; ++i) {
    uint64_t *pSeqNum = taosArrayGet(delIndexArray, i);
--- a/tests/pytest/crash_gen/crash_gen_main.py
+++ b/tests/pytest/crash_gen/crash_gen_main.py
@ -30,6 +30,7 @@ import argparse
 import sys
 import os
 import io
+import datetime
 import signal
 import traceback
 import requests
@ -1107,14 +1108,20 @@ class Database:
    # TODO: fix the error as result of above: "tsdb timestamp is out of range"
    @classmethod
    def setupLastTick(cls):
-        t1 = datetime.datetime(2020, 6, 1)
+        # start time will be auto generated , start at 10 years ago  local time 
+        local_time = datetime.datetime.utcnow().strftime('%Y-%m-%d %H:%M:%S.%f')[:-16]
+        local_epoch_time = [int(i) for i in local_time.split("-")]
+        #local_epoch_time will be such as : [2022, 7, 18]
+
+        t1 = datetime.datetime(local_epoch_time[0]-5, local_epoch_time[1], local_epoch_time[2])
        t2 = datetime.datetime.now()
        # maybe a very large number, takes 69 years to exceed Python int range
        elSec = int(t2.timestamp() - t1.timestamp())
        elSec2 = (elSec % (8 * 12 * 30 * 24 * 60 * 60 / 500)) * \
            500  # a number representing seconds within 10 years
        # print("elSec = {}".format(elSec))
-        t3 = datetime.datetime(2012, 1, 1)  # default "keep" is 10 years
+
+        t3 = datetime.datetime(local_epoch_time[0]-10, local_epoch_time[1], local_epoch_time[2])  # default "keep" is 10 years
        t4 = datetime.datetime.fromtimestamp(
            t3.timestamp() + elSec2)  # see explanation above
        Logging.debug("Setting up TICKS to start from: {}".format(t4))
--- a/tests/pytest/util/dnodes.py
+++ b/tests/pytest/util/dnodes.py
@ -488,10 +488,13 @@ class TDDnode:
            psCmd = "ps -ef|grep -w %s| grep -v grep | awk '{print $2}'" % toBeKilled
            processID = subprocess.check_output(
                psCmd, shell=True).decode("utf-8")
-
+            
+            onlyKillOnceWindows = 0
            while(processID):
-                killCmd = "kill -INT %s > /dev/null 2>&1" % processID
-                os.system(killCmd)
+                if not platform.system().lower() == 'windows' or (onlyKillOnceWindows == 0 and platform.system().lower() == 'windows'):
+                    killCmd = "kill -INT %s > /dev/null 2>&1" % processID
+                    os.system(killCmd)
+                    onlyKillOnceWindows = 1
                time.sleep(1)
                processID = subprocess.check_output(
                    psCmd, shell=True).decode("utf-8")
@ -524,9 +527,12 @@ class TDDnode:
            processID = subprocess.check_output(
                psCmd, shell=True).decode("utf-8")

+            onlyKillOnceWindows = 0
            while(processID):
-                killCmd = "kill -INT %s > /dev/null 2>&1" % processID
-                os.system(killCmd)
+                if not platform.system().lower() == 'windows' or (onlyKillOnceWindows == 0 and platform.system().lower() == 'windows'):
+                    killCmd = "kill -INT %s > /dev/null 2>&1" % processID
+                    os.system(killCmd)
+                    onlyKillOnceWindows = 1
                time.sleep(1)
                processID = subprocess.check_output(
                    psCmd, shell=True).decode("utf-8")
@ -550,9 +556,12 @@ class TDDnode:
            processID = subprocess.check_output(
                psCmd, shell=True).decode("utf-8")

+            onlyKillOnceWindows = 0
            while(processID):
-                killCmd = "kill -KILL %s > /dev/null 2>&1" % processID
-                os.system(killCmd)
+                if not platform.system().lower() == 'windows' or (onlyKillOnceWindows == 0 and platform.system().lower() == 'windows'):
+                    killCmd = "kill -KILL %s > /dev/null 2>&1" % processID
+                    os.system(killCmd)
+                    onlyKillOnceWindows = 1
                time.sleep(1)
                processID = subprocess.check_output(
                    psCmd, shell=True).decode("utf-8")
--- a/tests/script/jenkins/basic.txt
+++ b/tests/script/jenkins/basic.txt
@ -56,6 +56,7 @@
 # unsupport ./test.sh -f tsim/dnode/redistribute_vgroup_replica3_v2.sim
 # unsupport ./test.sh -f tsim/dnode/redistribute_vgroup_replica3_v3.sim
 # unsupport ./test.sh -f tsim/dnode/vnode_clean.sim
+./test.sh -f tsim/dnode/use_dropped_dnode.sim 

 # ---- import
 ./test.sh -f tsim/import/basic.sim
@ -248,8 +249,8 @@
 ./test.sh -f tsim/stream/sliding.sim

 # ---- transaction
- ./test.sh -f tsim/trans/lossdata1.sim
- ./test.sh -f tsim/trans/create_db.sim
+./test.sh -f tsim/trans/lossdata1.sim
+./test.sh -f tsim/trans/create_db.sim

 # ---- tmq
 ./test.sh -f tsim/tmq/basic1.sim
--- a/tests/script/tmp/data.sim
+++ b/tests/script/tmp/data.sim
@ -1,18 +1,125 @@
 system sh/stop_dnodes.sh
 system sh/deploy.sh -n dnode1 -i 1
-system sh/cfg.sh -n dnode1 -c telemetryReporting -v 1
-system sh/cfg.sh -n dnode1 -c telemetryInterval -v 1
-system sh/cfg.sh -n dnode1 -c telemetryServer -v localhost
-system sh/cfg.sh -n dnode1 -c telemetryPort -v 80
-
-return
+system sh/deploy.sh -n dnode2 -i 2
+system sh/deploy.sh -n dnode3 -i 3
+system sh/deploy.sh -n dnode4 -i 4
 system sh/exec.sh -n dnode1 -s start
+system sh/exec.sh -n dnode2 -s start
+system sh/exec.sh -n dnode3 -s start
+system sh/exec.sh -n dnode4 -s start
 sql connect

-sql create database db 
-sql create table db.tb (ts timestamp, i int)
-sql insert into db.tb values(now, 1)
+print =============== step1: create dnodes
+sql create dnode $hostname port 7200
+sql create dnode $hostname port 7300
+sql create dnode $hostname port 7400

+$x = 0
+step1:
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+	  print ====> dnode not ready!
+		return -1
+	endi
+sql show dnodes
+print ===> rows: $rows 
+print ===> $data00 $data01 $data02 $data03 $data04 $data05
+print ===> $data10 $data11 $data12 $data13 $data14 $data15
+print ===> $data20 $data21 $data22 $data23 $data24 $data25
+print ===> $data30 $data31 $data32 $data33 $data24 $data35
+if $rows != 4 then
+  return -1
+endi
+if $data(1)[4] != ready then
+  goto step1
+endi
+if $data(2)[4] != ready then
+  goto step1
+endi
+if $data(3)[4] != ready then
+  goto step1
+endi
+if $data(4)[4] != ready then
+  goto step1
+endi
+
+print =============== step2: create database
+sql create database db vgroups 1 replica 3
+sql show databases
+if $rows != 3 then 
+  return -1
+endi
+if $data(db)[4] != 3 then 
+  return -1
+endi
+
+sql show dnodes
+if $data(2)[2] != 1 then
+  return -1
+endi
+if $data(3)[2] != 1 then
+  return -1
+endi
+if $data(4)[2] != 1 then
+  return -1
+endi
+
+# vnodes
+sql show dnodes
+if $data(2)[2] != 1 then
+  return -1
+endi
+if $data(3)[2] != 1 then
+  return -1
+endi
+if $data(4)[2] != 1 then
+  return -1
+endi
+
+# v1_dnode
+$hasleader = 0
+$x = 0
+step2:
+	$x = $x + 1
+	sleep 1000
+	if $x == 20 then
+	  print ====> dnode not ready!
+		return -1
+	endi
+sql show db.vgroups
+print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08
+if $data(2)[3] != 2 then
+  return -1
+endi
+if $data(2)[5] != 3 then
+  return -1
+endi
+if $data(2)[7] != 4 then
+  return -1
+endi
+if $data(2)[4] == leader then
+  $hasleader = 1
+endi
+if $data(2)[6] == leader then
+  $hasleader = 1
+endi
+if $data(2)[8] == leader then
+  $hasleader = 1
+endi
+if $hasleader != 1 then
+  goto step2
+endi
+
+sql use db;
+sql create table stb (ts timestamp, c int) tags (t int);
+sql create table t0 using stb tags (0);
+sql insert into t0 values(now, 1);
+sql show db.stables;
+sql show db.tables;
+sql show db.vgroups;
+
+return
 print ======== start back
 run_back tmp/back.sim

--- a/tests/script/tmp/monitor.sim
+++ b/tests/script/tmp/monitor.sim
@ -6,32 +6,21 @@ system sh/cfg.sh -n dnode1 -c monitorInterval -v 1
 system sh/cfg.sh -n dnode1 -c monitorComp -v 1
 #system sh/cfg.sh -n dnode1 -c supportVnodes -v 128

+#system sh/cfg.sh -n dnode1 -c telemetryReporting -v 1
+#system sh/cfg.sh -n dnode1 -c telemetryInterval -v 1
+#system sh/cfg.sh -n dnode1 -c telemetryServer -v localhost
+#system sh/cfg.sh -n dnode1 -c telemetryPort -v 80
+
 system sh/exec.sh -n dnode1 -s start
 sql connect

 print =============== show dnodes
-sleep 2000
 sql create database db vgroups 2;
 sql use db;
 sql create table db.stb (ts timestamp, c1 int, c2 binary(4)) tags(t1 int, t2 binary(16)) comment "abd";
-sleep 2000

 print =============== create drop qnode 1
 sql create qnode on dnode 1
 sql create snode on dnode 1
 sql create bnode on dnode 1

-return
-print =============== restart
-system sh/exec.sh -n dnode1 -s stop -x SIGINT
-system sh/exec.sh -n dnode1 -s start
-
-
-return 
-system sh/deploy.sh -n dnode2 -i 2
-system sh/exec.sh -n dnode2 -s start
-system sh/exec.sh -n dnode2 -s stop -x SIGINT
-system sh/exec.sh -n dnode2 -s start
-
-system sh/exec.sh -n dnode1 -s stop -x SIGINT
-system sh/exec.sh -n dnode2 -s stop -x SIGINT
--- a/tests/script/tsim/db/alter_replica_31.sim
+++ b/tests/script/tsim/db/alter_replica_31.sim
@ -111,15 +111,6 @@ if $hasleader != 1 then
  goto step2
 endi

-# sql use db;
-# sql create table stb (ts timestamp, c int) tags (t int);
-# sql create table t0 using stb tags (0);
-# sql insert into t0 values(now, 1);
-# sql show db.stables;
-# sql show db.tables;
-# sql show db.vgroups;
-return
-
 sql create table db.stb (ts timestamp, c1 int, c2 binary(4)) tags(t1 int, t2 binary(16)) comment "abd"
 sql create table db.ctb using db.stb tags(101, "102")
 sql insert into db.ctb values(now, 1, "2")
--- a/tests/script/tsim/dnode/use_dropped_dnode.sim
+++ b/tests/script/tsim/dnode/use_dropped_dnode.sim
@ -0,0 +1,133 @@
+system sh/stop_dnodes.sh
+system sh/deploy.sh -n dnode1 -i 1
+system sh/deploy.sh -n dnode2 -i 2
+system sh/deploy.sh -n dnode3 -i 3
+system sh/exec.sh -n dnode1 -s start
+system sh/exec.sh -n dnode2 -s start
+system sh/exec.sh -n dnode3 -s start
+sql connect
+
+print =============== step1 create dnode2
+sql create dnode $hostname port 7200
+sql create dnode $hostname port 7300
+
+$x = 0
+step1:
+	$ = $x + 1
+	sleep 1000
+	if $x == 10 then
+	  print ====> dnode not online!
+		return -1
+	endi
+sql show dnodes
+print ===> $data00 $data01 $data02 $data03 $data04 $data05
+print ===> $data10 $data11 $data12 $data13 $data14 $data15
+print ===> $data20 $data21 $data22 $data23 $data24 $data25
+if $rows != 3 then
+  return -1
+endi
+if $data(1)[4] != ready then
+  goto step1
+endi
+if $data(2)[4] != ready then
+  goto step1
+endi
+if $data(3)[4] != ready then
+  goto step1
+endi
+
+print =============== step2 drop dnode 3
+sql drop dnode 3
+sql create dnode $hostname port 7300
+
+$x = 0
+step2:
+	$ = $x + 1
+	sleep 1000
+	if $x == 10 then
+	  print ====> dnode not online!
+		return -1
+	endi
+sql show dnodes
+print ===> $data00 $data01 $data02 $data03 $data04 $data05
+print ===> $data10 $data11 $data12 $data13 $data14 $data15
+print ===> $data20 $data21 $data22 $data23 $data24 $data25
+print ===> $data30 $data31 $data32 $data33 $data34 $data35
+if $rows != 3 then
+  return -1
+endi
+if $data(1)[4] != ready then
+  goto step2
+endi
+if $data(2)[4] != ready then
+  goto step2
+endi
+if $data(3)[4] != null then
+  goto step2
+endi
+if $data(4)[4] != offline then
+  goto step2
+endi
+
+print =============== step3: create dnode 4
+sleep 1000
+system sh/exec.sh -n dnode3 -s stop -x SIGINT
+system sh/deploy.sh -n dnode3 -i 3
+system sh/exec.sh -n dnode3 -s start
+
+$x = 0
+step3:
+	$ = $x + 1
+	sleep 1000
+	if $x == 10 then
+	  print ====> dnode not online!
+		return -1
+	endi
+sql show dnodes
+print ===> $data00 $data01 $data02 $data03 $data04 $data05
+print ===> $data10 $data11 $data12 $data13 $data14 $data15
+print ===> $data20 $data21 $data22 $data23 $data24 $data25
+print ===> $data30 $data31 $data32 $data33 $data34 $data35
+if $rows != 3 then
+  return -1
+endi
+if $data(1)[4] != ready then
+  goto step3
+endi
+if $data(2)[4] != ready then
+  goto step3
+endi
+if $data(3)[4] != null then
+  goto step3
+endi
+if $data(4)[4] != ready then
+  goto step3
+endi
+
+print =============== step4: create mnode 4
+sql create mnode on dnode 4
+
+$x = 0
+step4:
+	$ = $x + 1
+	sleep 1000
+	if $x == 10 then
+	  print ====> dnode not online!
+		return -1
+	endi
+sql show mnodes
+print ===> $data00 $data01 $data02 $data03 $data04 $data05
+print ===> $data10 $data11 $data12 $data13 $data14 $data15
+if $rows != 2 then
+  return -1
+endi
+if $data(1)[2] != leader then
+  goto step4
+endi
+if $data(4)[2] != follower then
+ goto step4
+endi
+
+system sh/exec.sh -n dnode1 -s stop -x SIGINT
+system sh/exec.sh -n dnode2 -s stop -x SIGINT
+system sh/exec.sh -n dnode3 -s stop -x SIGINT
--- a/tests/script/tsim/valgrind/checkError1.sim
+++ b/tests/script/tsim/valgrind/checkError1.sim
@ -152,7 +152,7 @@ endi

 system_content sh/checkValgrind.sh -n dnode2
 print cmd return result ----> [ $system_content ]
-if $system_content > 4 then
+if $system_content > 0 then
  return -1
 endi 

--- a/tests/system-test/0-others/cachemodel.py
+++ b/tests/system-test/0-others/cachemodel.py
@ -56,6 +56,16 @@ class TDTestCase:
        }
        return numbers.get(value, 'other')

+    def getCacheModelNum(self,str):
+        numbers = {
+            "none" : 0,
+            "last_row" : 1,
+            "last_value" : 2,
+            "both" : 3
+
+        }
+        return numbers.get(str, 'other')
+
    def prepare_datas(self):
        for i in range(4):
            str = self.getCacheModelStr(i)
@ -69,7 +79,7 @@ class TDTestCase:
                tdSql.execute(" insert into tb1 values(now , %d, %f)" %(k,k*10) )
                tdSql.execute(" insert into tb2 values(now , %d, %f)" %(k,k*10) )

-    def check_cache_last_sets(self):
+    def check_cachemodel_sets(self):
        
        
        # check cache_last value for database 
@ -84,52 +94,54 @@ class TDTestCase:
            # print(cache_last_value)
            if dbname in ["information_schema" , "performance_schema"]:
                continue
-            cache_lasts[dbname]=cache_last_value
+            cache_lasts[dbname]=self.getCacheModelNum(cache_last_value)

        
        # cache_last_set value 
        for k , v in cache_lasts.items():
            
-            if k=="testdb_"+str(v):
-                tdLog.info(" database %s cache_last value check pass, value is %s "%(k,v) )
+            if k=="testdb_"+str(self.getCacheModelStr(v)):
+                tdLog.info(" database %s cache_last value check pass, value is %s "%(k,self.getCacheModelStr(v)) )
            else:
-                tdLog.exit(" database %s cache_last value check fail, value is %s "%(k,v) )
+                tdLog.exit(" database %s cache_last value check fail, value is %s "%(k,self.getCacheModelStr(v)) )

        # # check storage layer implementation


-        # buildPath = self.getBuildPath()
-        # if (buildPath == ""):
-        #     tdLog.exit("taosd not found!")
-        # else:
-        #     tdLog.info("taosd found in %s" % buildPath)
-        # dataPath = buildPath + "/../sim/dnode1/data"
-        # abs_vnodePath = os.path.abspath(dataPath)+"/vnode/"
-        # tdLog.info("abs_vnodePath: %s" % abs_vnodePath)
+        buildPath = self.getBuildPath()
+        if (buildPath == ""):
+            tdLog.exit("taosd not found!")
+        else:
+            tdLog.info("taosd found in %s" % buildPath)
+        dataPath = buildPath + "/../sim/dnode1/data"
+        abs_vnodePath = os.path.abspath(dataPath)+"/vnode/"
+        tdLog.info("abs_vnodePath: %s" % abs_vnodePath)
        
-        # tdSql.query(" show dnodes ")
-        # dnode_id  = tdSql.queryResult[0][0]
+        tdSql.query(" show dnodes ")
+        dnode_id  = tdSql.queryResult[0][0]

-        # for dbname in cache_lasts.keys():
-        #     print(dbname)
-        #     tdSql.execute(" use %s" % dbname)
-        #     tdSql.query(" show vgroups ")
-        #     vgroups_infos = tdSql.queryResult
-        #     for vgroup_info in vgroups_infos:
-        #         vnode_json = abs_vnodePath + "/vnode" +f"{vgroup_info[0]}/" + "vnode.json"
-        #         vnode_info_of_db = f"cat {vnode_json}" 
-        #         vnode_info = subprocess.check_output(vnode_info_of_db, shell=True).decode("utf-8")
-        #         infoDict = json.loads(vnode_info)
-        #         vnode_json_of_dbname = f"{dnode_id}."+ dbname
-        #         config = infoDict["config"]
-        #         if infoDict["config"]["dbname"] == vnode_json_of_dbname:
-        #             if "cachelast" in infoDict["config"]:
-        #                 if int(infoDict["config"]["cachelast"]) != cache_lasts[dbname]:
-        #                     tdLog.exit("cachelast value is error in vnode.json of vnode%d "%(vgroup_info[0]))
-        #             else:
-        #                 tdLog.exit("cachelast not found in vnode.json of vnode%d "%(vgroup_info[0]))
+        for dbname in cache_lasts.keys():
+            # print(dbname)
+            tdSql.execute(" use %s" % dbname)
+            tdSql.query(" show vgroups ")
+            vgroups_infos = tdSql.queryResult
+            for vgroup_info in vgroups_infos:
+                vnode_json = abs_vnodePath + "/vnode" +f"{vgroup_info[0]}/" + "vnode.json"
+                vnode_info_of_db = f"cat {vnode_json}" 
+                vnode_info = subprocess.check_output(vnode_info_of_db, shell=True).decode("utf-8")
+                infoDict = json.loads(vnode_info)
+                vnode_json_of_dbname = f"{dnode_id}."+ dbname
+                config = infoDict["config"]
+                if infoDict["config"]["dbname"] == vnode_json_of_dbname:
+                    if "cacheLast" in infoDict["config"]:
+                        if int(infoDict["config"]["cacheLast"]) != cache_lasts[dbname]:
+                            tdLog.exit("cachemodel value is error in vnode.json of vnode%d "%(vgroup_info[0]))
+                        else:
+                            tdLog.info("cachemodel value is success in vnode.json of vnode%d "%(vgroup_info[0]))
+                    else:
+                        tdLog.exit("cacheLast not found in vnode.json of vnode%d "%(vgroup_info[0]))

-    def restart_check_cache_last_sets(self):
+    def restart_check_cachemodel_sets(self):
        
        for i in range(3):
            tdSql.query("show dnodes")
@ -137,14 +149,14 @@ class TDTestCase:
            tdDnodes.stop(index)
            tdDnodes.start(index)
            time.sleep(3)
-            self.check_cache_last_sets()
+            self.check_cachemodel_sets()


    def run(self):  # sourcery skip: extract-duplicate-method, remove-redundant-fstring
        self.illegal_params()
        self.prepare_datas()
-        self.check_cache_last_sets()
-        self.restart_check_cache_last_sets()
+        self.check_cachemodel_sets()
+        self.restart_check_cachemodel_sets()
       
    def stop(self):
        tdSql.close()
--- a/tests/system-test/0-others/sysinfo.py
+++ b/tests/system-test/0-others/sysinfo.py
@ -24,7 +24,7 @@ class TDTestCase:
        tdLog.debug("start to execute %s" % __file__)
        tdSql.init(conn.cursor())
        self.dbname = 'db'
-        self.delaytime = 10
+        self.delaytime = 3
    def get_database_info(self):
        tdSql.query('select database()')
        tdSql.checkData(0,0,None)
@ -42,12 +42,13 @@ class TDTestCase:
            tdSql.checkData(0,0,version_info)
        
    def get_server_status(self):
+        sleep(self.delaytime)
        tdSql.query('select server_status()')
        tdSql.checkData(0,0,1)
        #!for bug
-        # tdDnodes.stoptaosd(1)
-        # sleep(self.delaytime)
-        # tdSql.error('select server_status()')
+        tdDnodes.stoptaosd(1)
+        sleep(self.delaytime)
+        tdSql.error('select server_status()')
        
    def run(self):
        self.get_database_info()
--- a/tests/system-test/1-insert/delete_data.py
+++ b/tests/system-test/1-insert/delete_data.py
@ -0,0 +1,190 @@
+###################################################################
+#           Copyright (c) 2016 by TAOS Technologies, Inc.
+#                     All rights reserved.
+#
+#  This file is proprietary and confidential to TAOS Technologies.
+#  No part of this file may be reproduced, stored, transmitted,
+#  disclosed or used in any form or by any means other than as
+#  expressly provided by the written permission from Jianhui Tao
+#
+###################################################################
+
+# -*- coding: utf-8 -*-
+
+import random
+import string
+
+from numpy import logspace
+from util import constant
+from util.log import *
+from util.cases import *
+from util.sql import *
+from util.common import *
+from util.sqlset import TDSetSql
+
+class TDTestCase:
+    def init(self, conn, logSql):
+        tdLog.debug("start to execute %s" % __file__)
+        tdSql.init(conn.cursor(),logSql)
+        self.dbname = 'db_test'
+        self.setsql = TDSetSql()
+        self.ntbname = 'ntb'
+        self.rowNum = 10
+        self.tbnum = 20
+        self.ts = 1537146000000
+        self.binary_str = 'taosdata'
+        self.nchar_str = '涛思数据'
+        self.str_length = 20
+        self.column_dict = {
+            'col1': 'tinyint',
+            'col2': 'smallint',
+            'col3': 'int',
+            'col4': 'bigint',
+            'col5': 'tinyint unsigned',
+            'col6': 'smallint unsigned',
+            'col7': 'int unsigned',
+            'col8': 'bigint unsigned',
+            'col9': 'float',
+            'col10': 'double',
+            'col11': 'bool',
+            'col12': f'binary({self.str_length})',
+            'col13': f'nchar({self.str_length})',
+            
+        }
+        self.tinyint_val = random.randint(constant.TINYINT_MIN,constant.TINYINT_MAX)
+        self.smallint_val = random.randint(constant.SMALLINT_MIN,constant.SMALLINT_MAX)
+        self.int_val = random.randint(constant.INT_MIN,constant.INT_MAX)
+        self.bigint_val = random.randint(constant.BIGINT_MIN,constant.BIGINT_MAX)
+        self.untingint_val = random.randint(constant.TINYINT_UN_MIN,constant.TINYINT_UN_MAX)
+        self.unsmallint_val = random.randint(constant.SMALLINT_UN_MIN,constant.SMALLINT_UN_MAX)
+        self.unint_val = random.randint(constant.INT_UN_MIN,constant.INT_MAX)
+        self.unbigint_val = random.randint(constant.BIGINT_UN_MIN,constant.BIGINT_UN_MAX)
+        self.float_val = random.uniform(constant.FLOAT_MIN,constant.FLOAT_MAX)
+        self.double_val = random.uniform(constant.DOUBLE_MIN*(1E-300),constant.DOUBLE_MAX*(1E-300))
+        self.bool_val = random.randint(0,100)%2
+        self.binary_val = tdCom.getLongName(random.randint(0,self.str_length))
+        self.nchar_val = tdCom.getLongName(random.randint(0,self.str_length))
+        self.base_data = {
+            'tinyint':self.tinyint_val,
+            'smallint':self.smallint_val,
+            'int':self.int_val,
+            'bigint':self.bigint_val,
+            'tinyint unsigned':self.untingint_val,
+            'smallint unsigned':self.unsmallint_val,
+            'int unsigned':self.unint_val,
+            'bigint unsigned':self.unbigint_val,
+            'bool':self.bool_val,
+            'float':self.float_val,
+            'double':self.double_val,
+            'binary':self.binary_val,
+            'nchar':self.nchar_val
+                    }
+    def insert_base_data(self,col_type,tbname,rows,base_data):
+        for i in range(rows):
+            if col_type.lower() == 'tinyint':
+                tdSql.execute(f'insert into {tbname} values({self.ts+i},{base_data["tinyint"]})')
+            elif col_type.lower() == 'smallint':
+                tdSql.execute(f'insert into {tbname} values({self.ts+i},{base_data["smallint"]})')
+            elif col_type.lower() == 'int':
+                tdSql.execute(f'insert into {tbname} values({self.ts+i},{base_data["int"]})')
+            elif col_type.lower() == 'bigint':
+                tdSql.execute(f'insert into {tbname} values({self.ts+i},{base_data["bigint"]})')
+            elif col_type.lower() == 'tinyint unsigned':
+                tdSql.execute(f'insert into {tbname} values({self.ts+i},{base_data["tinyint unsigned"]})')
+            elif col_type.lower() == 'smallint unsigned':
+                tdSql.execute(f'insert into {tbname} values({self.ts+i},{base_data["smallint unsigned"]})')
+            elif col_type.lower() == 'int unsigned':
+                tdSql.execute(f'insert into {tbname} values({self.ts+i},{base_data["int unsigned"]})')
+            elif col_type.lower() == 'bigint unsigned':
+                tdSql.execute(f'insert into {tbname} values({self.ts+i},{base_data["bigint unsigned"]})')
+            elif col_type.lower() == 'bool':
+                tdSql.execute(f'insert into {tbname} values({self.ts+i},{base_data["bool"]})')    
+            elif col_type.lower() == 'float':
+                tdSql.execute(f'insert into {tbname} values({self.ts+i},{base_data["float"]})')      
+            elif col_type.lower() == 'double':
+                tdSql.execute(f'insert into {tbname} values({self.ts+i},{base_data["double"]})')
+            elif 'binary' in col_type.lower():
+                tdSql.execute(f'''insert into {tbname} values({self.ts+i},"{base_data['binary']}")''')
+            elif 'nchar' in col_type.lower():
+                tdSql.execute(f'''insert into {tbname} values({self.ts+i},"{base_data['nchar']}")''')        
+
+    def delete_all_data(self,tbname,col_type,row_num,base_data,dbname):
+        tdSql.execute(f'delete from {tbname}')
+        tdSql.execute(f'flush database {dbname}')
+        tdSql.execute('reset query cache')
+        tdSql.query(f'select * from {tbname}')
+        tdSql.checkRows(0)
+        self.insert_base_data(col_type,tbname,row_num,base_data)
+        tdSql.execute(f'flush database {dbname}')
+        tdSql.execute('reset query cache')
+        tdSql.query(f'select * from {tbname}')
+        tdSql.checkRows(row_num)
+    def delete_one_row(self,tbname,column_type,column_name,base_data,dbname):
+        tdSql.execute(f'delete from {tbname} where ts={self.ts}')
+        tdSql.execute(f'flush database {dbname}')
+        tdSql.execute('reset query cache')
+        tdSql.query(f'select {column_name} from {tbname}')
+        tdSql.checkRows(self.rowNum-1)
+        tdSql.query(f'select {column_name} from {tbname} where ts={self.ts}')
+        tdSql.checkRows(0)
+        if 'binary' in column_type.lower():
+            tdSql.execute(f'''insert into {tbname} values({self.ts},"{base_data['binary']}")''')
+        elif 'nchar' in column_type.lower():
+            tdSql.execute(f'''insert into {tbname} values({self.ts},"{base_data['nchar']}")''')
+        else:
+            tdSql.execute(f'insert into {tbname} values({self.ts},{base_data[column_type]})')
+        tdSql.query(f'select {column_name} from {tbname} where ts={self.ts}')
+        if column_type.lower() == 'float' or column_type.lower() == 'double':
+            if abs(tdSql.queryResult[0][0] - base_data[column_type]) / base_data[column_type] <= 0.0001:
+                tdSql.checkEqual(tdSql.queryResult[0][0],tdSql.queryResult[0][0])
+            else:
+                tdLog.exit(f'{column_type} data check failure')
+        elif 'binary' in column_type.lower():
+            tdSql.checkEqual(tdSql.queryResult[0][0],base_data['binary'])
+        elif 'nchar' in column_type.lower():
+            tdSql.checkEqual(tdSql.queryResult[0][0],base_data['nchar'])
+        else:
+            tdSql.checkEqual(tdSql.queryResult[0][0],base_data[column_type])
+    
+    def delete_rows(self):
+
+
+        pass
+    def delete_error(self,tbname,column_name,column_type,base_data):
+        for error_list in ['',f'ts = {self.ts} and',f'ts = {self.ts} or']:
+            if 'binary' in column_type.lower():
+                tdSql.error(f'''delete from {tbname} where {error_list} {column_name} ="{base_data['binary']}"''')
+            elif 'nchar' in column_type.lower():
+                tdSql.error(f'''delete from {tbname} where {error_list} {column_name} ="{base_data['nchar']}"''')
+            else:
+                tdSql.error('delete from {tbname} where {error_list} {column_name} = {base_data[column_type]}')
+                
+    def delete_data_ntb(self):
+        tdSql.execute(f'create database if not exists {self.dbname}')
+        tdSql.execute(f'use {self.dbname}')
+        for col_name,col_type in self.column_dict.items():
+            tdSql.execute(f'create table {self.ntbname} (ts timestamp,{col_name} {col_type})')
+            self.insert_base_data(col_type,self.ntbname,self.rowNum,self.base_data)
+            self.delete_one_row(self.ntbname,col_type,col_name,self.base_data,self.dbname)
+            self.delete_all_data(self.ntbname,col_type,self.rowNum,self.base_data,self.dbname)
+            self.delete_error(self.ntbname,col_name,col_type,self.base_data)
+            for i in range(self.rowNum):
+                tdSql.execute(f'delete from {self.ntbname} where ts>{self.ts+i}')
+                tdSql.execute(f'flush database {self.dbname}')
+                tdSql.execute('reset query cache')
+                tdSql.query(f'select {col_name} from {self.ntbname}')
+                tdSql.checkRows(i+1)
+                self.insert_base_data(col_type,self.ntbname,self.rowNum,self.base_data)
+            
+            tdSql.execute(f'drop table {self.ntbname}')
+        
+    def run(self):
+        self.delete_data_ntb()
+        
+
+    def stop(self):
+        tdSql.close()
+        tdLog.success("%s successfully executed" % __file__)
+
+tdCases.addWindows(__file__, TDTestCase())
+tdCases.addLinux(__file__, TDTestCase())
--- a/tests/system-test/1-insert/insert_drop.py
+++ b/tests/system-test/1-insert/insert_drop.py
@ -0,0 +1,49 @@
+# -*- coding: utf-8 -*-
+
+import sys
+from util.log import *
+from util.cases import *
+from util.sql import *
+import threading
+
+
+class TDTestCase:
+    def init(self, conn, logSql):
+        tdLog.debug("start to execute %s" % __file__)
+        tdSql.init(conn.cursor(), logSql)
+
+    def genMultiThreadSeq(self, sql_list):
+        tlist = list()
+        for insert_sql in sql_list:
+            t = threading.Thread(target=tdSql.execute, args=(insert_sql,))
+            tlist.append(t)
+        return tlist
+
+    def multiThreadRun(self, tlist):
+        for t in tlist:
+            t.start()
+        for t in tlist:
+            t.join()
+
+    def run(self):
+        tdSql.prepare()
+        tdSql.execute('create database if not exists test;')
+        tdSql.execute('create table test.stb (ts timestamp, c11 int, c12 float ) TAGS(t11 int, t12 int );')
+        tdSql.execute('create table test.tb using test.stb TAGS (1, 1);')
+        sql_list = list()
+        for i in range(5):
+            sql = f'insert into test.tb values (now-{i}m, {i}, {i});'
+            sql_list.append(sql)
+        sql_list.append(f'drop database test;')
+        tlist = self.genMultiThreadSeq(sql_list)
+        self.multiThreadRun(tlist)
+        tdSql.query(f'show databases')
+        
+
+    def stop(self):
+        tdSql.close()
+        tdLog.success("%s successfully executed" % __file__)
+
+
+tdCases.addWindows(__file__, TDTestCase())
+tdCases.addLinux(__file__, TDTestCase())
--- a/Show More
+++ b/Show More