Merge branch 'feat/tsdb_refact' of https://github.com/taosdata/TDengine into feat/tsdb_refact

include/common/ttokendef.h

@@ -205,65 +205,65 @@
 #define TK_SPLIT          187
 #define TK_SYNCDB         188
 #define TK_DELETE         189
-#define TK_NULL           190
-#define TK_NK_QUESTION    191
-#define TK_NK_ARROW       192
-#define TK_ROWTS          193
-#define TK_TBNAME         194
-#define TK_QSTARTTS       195
-#define TK_QENDTS         196
-#define TK_WSTARTTS       197
-#define TK_WENDTS         198
-#define TK_WDURATION      199
-#define TK_CAST           200
-#define TK_NOW            201
-#define TK_TODAY          202
-#define TK_TIMEZONE       203
-#define TK_CLIENT_VERSION 204
-#define TK_SERVER_VERSION 205
-#define TK_SERVER_STATUS  206
-#define TK_CURRENT_USER   207
-#define TK_COUNT          208
-#define TK_LAST_ROW       209
-#define TK_BETWEEN        210
-#define TK_IS             211
-#define TK_NK_LT          212
-#define TK_NK_GT          213
-#define TK_NK_LE          214
-#define TK_NK_GE          215
-#define TK_NK_NE          216
-#define TK_MATCH          217
-#define TK_NMATCH         218
-#define TK_CONTAINS       219
-#define TK_JOIN           220
-#define TK_INNER          221
-#define TK_SELECT         222
-#define TK_DISTINCT       223
-#define TK_WHERE          224
-#define TK_PARTITION      225
-#define TK_BY             226
-#define TK_SESSION        227
-#define TK_STATE_WINDOW   228
-#define TK_SLIDING        229
-#define TK_FILL           230
-#define TK_VALUE          231
-#define TK_NONE           232
-#define TK_PREV           233
-#define TK_LINEAR         234
-#define TK_NEXT           235
-#define TK_HAVING         236
-#define TK_RANGE          237
-#define TK_EVERY          238
-#define TK_ORDER          239
-#define TK_SLIMIT         240
-#define TK_SOFFSET        241
-#define TK_LIMIT          242
-#define TK_OFFSET         243
-#define TK_ASC            244
-#define TK_NULLS          245
-#define TK_ID             246
-#define TK_NK_BITNOT      247
-#define TK_INSERT         248
+#define TK_INSERT         190
+#define TK_NULL           191
+#define TK_NK_QUESTION    192
+#define TK_NK_ARROW       193
+#define TK_ROWTS          194
+#define TK_TBNAME         195
+#define TK_QSTARTTS       196
+#define TK_QENDTS         197
+#define TK_WSTARTTS       198
+#define TK_WENDTS         199
+#define TK_WDURATION      200
+#define TK_CAST           201
+#define TK_NOW            202
+#define TK_TODAY          203
+#define TK_TIMEZONE       204
+#define TK_CLIENT_VERSION 205
+#define TK_SERVER_VERSION 206
+#define TK_SERVER_STATUS  207
+#define TK_CURRENT_USER   208
+#define TK_COUNT          209
+#define TK_LAST_ROW       210
+#define TK_BETWEEN        211
+#define TK_IS             212
+#define TK_NK_LT          213
+#define TK_NK_GT          214
+#define TK_NK_LE          215
+#define TK_NK_GE          216
+#define TK_NK_NE          217
+#define TK_MATCH          218
+#define TK_NMATCH         219
+#define TK_CONTAINS       220
+#define TK_JOIN           221
+#define TK_INNER          222
+#define TK_SELECT         223
+#define TK_DISTINCT       224
+#define TK_WHERE          225
+#define TK_PARTITION      226
+#define TK_BY             227
+#define TK_SESSION        228
+#define TK_STATE_WINDOW   229
+#define TK_SLIDING        230
+#define TK_FILL           231
+#define TK_VALUE          232
+#define TK_NONE           233
+#define TK_PREV           234
+#define TK_LINEAR         235
+#define TK_NEXT           236
+#define TK_HAVING         237
+#define TK_RANGE          238
+#define TK_EVERY          239
+#define TK_ORDER          240
+#define TK_SLIMIT         241
+#define TK_SOFFSET        242
+#define TK_LIMIT          243
+#define TK_OFFSET         244
+#define TK_ASC            245
+#define TK_NULLS          246
+#define TK_ID             247
+#define TK_NK_BITNOT      248
 #define TK_VALUES         249
 #define TK_IMPORT         250
 #define TK_NK_SEMI        251

include/libs/index/index.h

@@ -208,7 +208,7 @@ int32_t doFilterTag(const SNode* pFilterNode, SIndexMetaArg* metaArg, SArray* re
  * destory index env
  *
  */
-void indexCleanUp();
+void indexCleanup();
 
 #ifdef __cplusplus
 }

include/libs/nodes/nodes.h

@@ -195,6 +195,7 @@ typedef enum ENodeType {
   QUERY_NODE_KILL_QUERY_STMT,
   QUERY_NODE_KILL_TRANSACTION_STMT,
   QUERY_NODE_DELETE_STMT,
+  QUERY_NODE_INSERT_STMT,
   QUERY_NODE_QUERY,
 
   // logic plan node
@@ -248,6 +249,7 @@ typedef enum ENodeType {
   QUERY_NODE_PHYSICAL_PLAN_INTERP_FUNC,
   QUERY_NODE_PHYSICAL_PLAN_DISPATCH,
   QUERY_NODE_PHYSICAL_PLAN_INSERT,
+  QUERY_NODE_PHYSICAL_PLAN_QUERY_INSERT,
   QUERY_NODE_PHYSICAL_PLAN_DELETE,
   QUERY_NODE_PHYSICAL_SUBPLAN,
   QUERY_NODE_PHYSICAL_PLAN

include/libs/nodes/plannodes.h

@@ -131,6 +131,7 @@ typedef struct SVnodeModifyLogicNode {
   int8_t           tableType;  // table type
   char             tableFName[TSDB_TABLE_FNAME_LEN];
   STimeWindow      deleteTimeRange;
+  SVgroupsInfo*    pVgroupList;
 } SVnodeModifyLogicNode;
 
 typedef struct SExchangeLogicNode {
@@ -456,6 +457,15 @@ typedef struct SDataInserterNode {
   char*         pData;
 } SDataInserterNode;
 
+typedef struct SQueryInserterNode {
+  SDataSinkNode sink;
+  uint64_t      tableId;
+  int8_t        tableType;  // table type
+  char          tableFName[TSDB_TABLE_FNAME_LEN];
+  int32_t       vgId;
+  SEpSet        epSet;
+} SQueryInserterNode;
+
 typedef struct SDataDeleterNode {
   SDataSinkNode sink;
   uint64_t      tableId;

include/libs/nodes/querynodes.h

@@ -302,6 +302,14 @@ typedef struct SDeleteStmt {
   bool        deleteZeroRows;
 } SDeleteStmt;
 
+typedef struct SInsertStmt {
+  ENodeType  type;  // QUERY_NODE_INSERT_STMT
+  SNode*     pTable;
+  SNodeList* pCols;
+  SNode*     pQuery;
+  uint8_t    precision;
+} SInsertStmt;
+
 typedef enum {
   PAYLOAD_TYPE_KV = 0,
   PAYLOAD_TYPE_RAW = 1,

include/libs/parser/parser.h

@@ -56,7 +56,7 @@ typedef struct SParseContext {
 } SParseContext;
 
 int32_t qParseSql(SParseContext* pCxt, SQuery** pQuery);
-bool    qIsInsertSql(const char* pStr, size_t length);
+bool    qIsInsertValuesSql(const char* pStr, size_t length);
 
 // for async mode
 int32_t qParseSqlSyntax(SParseContext* pCxt, SQuery** pQuery, struct SCatalogReq* pCatalogReq);

include/libs/qcom/query.h

@@ -29,12 +29,13 @@ extern "C" {
 
 typedef enum {
   JOB_TASK_STATUS_NULL = 0,
-  JOB_TASK_STATUS_NOT_START = 1,
-  JOB_TASK_STATUS_EXECUTING,
-  JOB_TASK_STATUS_PARTIAL_SUCCEED,
-  JOB_TASK_STATUS_SUCCEED,
-  JOB_TASK_STATUS_FAILED,
-  JOB_TASK_STATUS_DROPPING,
+  JOB_TASK_STATUS_INIT,
+  JOB_TASK_STATUS_EXEC,
+  JOB_TASK_STATUS_PART_SUCC,
+  JOB_TASK_STATUS_SUCC,
+  JOB_TASK_STATUS_FAIL,
+  JOB_TASK_STATUS_DROP,
+  JOB_TASK_STATUS_MAX,
 } EJobTaskType;
 
 typedef enum {
@@ -59,10 +60,6 @@ typedef struct STableComInfo {
   int32_t  rowSize;       // row size of the schema
 } STableComInfo;
 
-typedef struct SQueryExecRes {
-  int32_t msgType;
-  void*   res;
-} SQueryExecRes;
 
 typedef struct SIndexMeta {
 #if defined(WINDOWS) || defined(_TD_DARWIN_64)
@@ -71,6 +68,13 @@ typedef struct SIndexMeta {
 
 } SIndexMeta;
 
+typedef struct SExecResult {
+  int32_t         code;
+  uint64_t        numOfRows;
+  int32_t         msgType;
+  void*           res;
+} SExecResult;
+
 typedef struct STbVerInfo {
   char    tbFName[TSDB_TABLE_FNAME_LEN];
   int32_t sversion;
@@ -210,7 +214,7 @@ char*   jobTaskStatusStr(int32_t status);
 
 SSchema createSchema(int8_t type, int32_t bytes, col_id_t colId, const char* name);
 
-void    destroyQueryExecRes(SQueryExecRes* pRes);
+void    destroyQueryExecRes(SExecResult* pRes);
 int32_t dataConverToStr(char* str, int type, void* buf, int32_t bufSize, int32_t* len);
 char*   parseTagDatatoJson(void* p);
 int32_t cloneTableMeta(STableMeta* pSrc, STableMeta** pDst);

include/libs/scheduler/scheduler.h

@@ -53,12 +53,6 @@ typedef struct SQueryProfileSummary {
   uint64_t resultSize;   // generated result size in Kb.
 } SQueryProfileSummary;
 
-typedef struct SQueryResult {
-  int32_t         code;
-  uint64_t        numOfRows;
-  SQueryExecRes   res;
-} SQueryResult;
-
 typedef struct STaskInfo {
   SQueryNodeAddr addr;
   SSubQueryMsg  *msg;
@@ -69,50 +63,34 @@ typedef struct SSchdFetchParam {
   int32_t* code;
 } SSchdFetchParam;
 
-typedef void (*schedulerExecFp)(SQueryResult* pResult, void* param, int32_t code);
+typedef void (*schedulerExecFp)(SExecResult* pResult, void* param, int32_t code);
 typedef void (*schedulerFetchFp)(void* pResult, void* param, int32_t code);
 typedef bool (*schedulerChkKillFp)(void* param);
 
 typedef struct SSchedulerReq {
+  bool                  syncReq;
   SRequestConnInfo     *pConn;
   SArray               *pNodeList;
   SQueryPlan           *pDag;
   const char           *sql;
   int64_t               startTs;
   schedulerExecFp       execFp;
-  void*                 execParam;
+  schedulerFetchFp      fetchFp;
+  void*                 cbParam;
   schedulerChkKillFp    chkKillFp;
   void*                 chkKillParam;
+  SExecResult*          pExecRes;
+  void**                pFetchRes;
 } SSchedulerReq;
 
 
 int32_t schedulerInit(SSchedulerCfg *cfg);
 
-/**
- * Process the query job, generated according to the query physical plan.
- * This is a synchronized API, and is also thread-safety.
- * @param nodeList  Qnode/Vnode address list, element is SQueryNodeAddr
- * @return
- */
-int32_t schedulerExecJob(SSchedulerReq *pReq, int64_t *pJob, SQueryResult *pRes);
-
-/**
- * Process the query job, generated according to the query physical plan.
- * This is a asynchronized API, and is also thread-safety.
- * @param pNodeList  Qnode/Vnode address list, element is SQueryNodeAddr
- * @return
- */
-  int32_t schedulerAsyncExecJob(SSchedulerReq *pReq, int64_t *pJob);
+int32_t schedulerExecJob(SSchedulerReq *pReq, int64_t *pJob);
 
-/**
- * Fetch query result from the remote query executor
- * @param pJob
- * @param data
- * @return
- */
-int32_t schedulerFetchRows(int64_t job, void **data);
+int32_t schedulerFetchRows(int64_t jobId, SSchedulerReq *pReq);
 
-void schedulerAsyncFetchRows(int64_t job, schedulerFetchFp fp, void* param);
+void schedulerFetchRowsA(int64_t job, schedulerFetchFp fp, void* param);
 
 int32_t schedulerGetTasksStatus(int64_t job, SArray *pSub);
 
@@ -134,7 +112,7 @@ void schedulerFreeJob(int64_t* job, int32_t errCode);
 
 void schedulerDestroy(void);
 
-void schdExecCallback(SQueryResult* pResult, void* param, int32_t code);
+void schdExecCallback(SExecResult* pResult, void* param, int32_t code);
 
 #ifdef __cplusplus
 }

include/libs/sync/sync.h

@@ -26,7 +26,7 @@ extern "C" {
 
 extern bool gRaftDetailLog;
 
-#define SYNC_MAX_BATCH_SIZE 100
+#define SYNC_MAX_BATCH_SIZE 500
 #define SYNC_INDEX_BEGIN 0
 #define SYNC_INDEX_INVALID -1
 #define SYNC_TERM_INVALID 0xFFFFFFFFFFFFFFFF

include/os/osMemory.h

@@ -32,7 +32,7 @@ extern "C" {
 void *taosMemoryMalloc(int32_t size);
 void *taosMemoryCalloc(int32_t num, int32_t size);
 void *taosMemoryRealloc(void *ptr, int32_t size);
-void *taosMemoryStrDup(void *ptr);
+void *taosMemoryStrDup(const char *ptr);
 void taosMemoryFree(void *ptr);
 int32_t taosMemorySize(void *ptr);
 void taosPrintBackTrace();

include/util/taoserror.h

@@ -390,10 +390,10 @@ int32_t* taosGetErrno();
 #define TSDB_CODE_QRY_TASK_MSG_ERROR            TAOS_DEF_ERROR_CODE(0, 0x0719)
 #define TSDB_CODE_QRY_JOB_FREED                 TAOS_DEF_ERROR_CODE(0, 0x071A)
 #define TSDB_CODE_QRY_TASK_STATUS_ERROR         TAOS_DEF_ERROR_CODE(0, 0x071B)
-//json
 #define TSDB_CODE_QRY_JSON_IN_ERROR             TAOS_DEF_ERROR_CODE(0, 0x071C)
 #define TSDB_CODE_QRY_JSON_NOT_SUPPORT_ERROR    TAOS_DEF_ERROR_CODE(0, 0x071D)
 #define TSDB_CODE_QRY_JSON_IN_GROUP_ERROR       TAOS_DEF_ERROR_CODE(0, 0x071E)
+#define TSDB_CODE_QRY_JOB_NOT_EXIST             TAOS_DEF_ERROR_CODE(0, 0x071F)
 
 // grant
 #define TSDB_CODE_GRANT_EXPIRED                 TAOS_DEF_ERROR_CODE(0, 0x0800)

source/client/inc/clientInt.h

@@ -156,7 +156,7 @@ typedef struct SResultColumn {
 } SResultColumn;
 
 typedef struct SReqResultInfo {
-  SQueryExecRes  execRes;
+  SExecResult    execRes;
   const char*    pRspMsg;
   const char*    pData;
   TAOS_FIELD*    fields;      // todo, column names are not needed.

source/client/src/clientImpl.c

@@ -627,22 +627,26 @@ _return:
 int32_t scheduleQuery(SRequestObj* pRequest, SQueryPlan* pDag, SArray* pNodeList) {
   void* pTransporter = pRequest->pTscObj->pAppInfo->pTransporter;
 
-  SQueryResult     res = {0};
+  SExecResult     res = {0};
   SRequestConnInfo conn = {.pTrans = pRequest->pTscObj->pAppInfo->pTransporter,
                            .requestId = pRequest->requestId,
                            .requestObjRefId = pRequest->self};
-  SSchedulerReq    req = {.pConn = &conn,
-                       .pNodeList = pNodeList,
-                       .pDag = pDag,
-                       .sql = pRequest->sqlstr,
-                       .startTs = pRequest->metric.start,
-                       .execFp = NULL,
-                       .execParam = NULL,
-                       .chkKillFp = chkRequestKilled,
-                       .chkKillParam = (void*)pRequest->self};
-
-  int32_t code = schedulerExecJob(&req, &pRequest->body.queryJob, &res);
-  pRequest->body.resInfo.execRes = res.res;
+  SSchedulerReq    req = {
+    .syncReq = true,
+    .pConn = &conn,
+    .pNodeList = pNodeList,
+    .pDag = pDag,
+    .sql = pRequest->sqlstr,
+    .startTs = pRequest->metric.start,
+    .execFp = NULL,
+    .cbParam = NULL,
+    .chkKillFp = chkRequestKilled,
+    .chkKillParam = (void*)pRequest->self,
+    .pExecRes = &res,
+  };
+
+  int32_t code = schedulerExecJob(&req, &pRequest->body.queryJob);
+  memcpy(&pRequest->body.resInfo.execRes, &res, sizeof(res));
 
   if (code != TSDB_CODE_SUCCESS) {
     schedulerFreeJob(&pRequest->body.queryJob, 0);
@@ -753,7 +757,7 @@ int32_t handleQueryExecRsp(SRequestObj* pRequest) {
   }
 
   SEpSet         epset = getEpSet_s(&pAppInfo->mgmtEp);
-  SQueryExecRes* pRes = &pRequest->body.resInfo.execRes;
+  SExecResult* pRes = &pRequest->body.resInfo.execRes;
 
   switch (pRes->msgType) {
     case TDMT_VND_ALTER_TABLE:
@@ -779,10 +783,10 @@ int32_t handleQueryExecRsp(SRequestObj* pRequest) {
   return code;
 }
 
-void schedulerExecCb(SQueryResult* pResult, void* param, int32_t code) {
+void schedulerExecCb(SExecResult* pResult, void* param, int32_t code) {
   SRequestObj* pRequest = (SRequestObj*)param;
   pRequest->code = code;
-  pRequest->body.resInfo.execRes = pResult->res;
+  memcpy(&pRequest->body.resInfo.execRes, pResult, sizeof(*pResult));
 
   if (TDMT_VND_SUBMIT == pRequest->type || TDMT_VND_DELETE == pRequest->type ||
       TDMT_VND_CREATE_TABLE == pRequest->type) {
@@ -939,16 +943,20 @@ void launchAsyncQuery(SRequestObj* pRequest, SQuery* pQuery, SMetaData* pResultM
 
         SRequestConnInfo conn = {
             .pTrans = pAppInfo->pTransporter, .requestId = pRequest->requestId, .requestObjRefId = pRequest->self};
-        SSchedulerReq req = {.pConn = &conn,
-                             .pNodeList = pNodeList,
-                             .pDag = pDag,
-                             .sql = pRequest->sqlstr,
-                             .startTs = pRequest->metric.start,
-                             .execFp = schedulerExecCb,
-                             .execParam = pRequest,
-                             .chkKillFp = chkRequestKilled,
-                             .chkKillParam = (void*)pRequest->self};
-        code = schedulerAsyncExecJob(&req, &pRequest->body.queryJob);
+        SSchedulerReq req = {
+          .syncReq = false,
+          .pConn = &conn,
+          .pNodeList = pNodeList,
+          .pDag = pDag,
+          .sql = pRequest->sqlstr,
+          .startTs = pRequest->metric.start,
+          .execFp = schedulerExecCb,
+          .cbParam = pRequest,
+          .chkKillFp = chkRequestKilled,
+          .chkKillParam = (void*)pRequest->self,
+          .pExecRes = NULL,
+        };
+        code = schedulerExecJob(&req, &pRequest->body.queryJob);
         taosArrayDestroy(pNodeList);
       } else {
         tscDebug("0x%" PRIx64 " plan not executed, code:%s 0x%" PRIx64, pRequest->self, tstrerror(code),
@@ -1387,7 +1395,11 @@ void* doFetchRows(SRequestObj* pRequest, bool setupOneRowPtr, bool convertUcs4)
     }
 
     SReqResultInfo* pResInfo = &pRequest->body.resInfo;
-    pRequest->code = schedulerFetchRows(pRequest->body.queryJob, (void**)&pResInfo->pData);
+    SSchedulerReq req = {
+      .syncReq = true,
+      .pFetchRes = (void**)&pResInfo->pData,
+    };
+    pRequest->code = schedulerFetchRows(pRequest->body.queryJob, &req);
     if (pRequest->code != TSDB_CODE_SUCCESS) {
       pResultInfo->numOfRows = 0;
       return NULL;

source/client/src/clientMain.c

@@ -858,7 +858,12 @@ void taos_fetch_rows_a(TAOS_RES *res, __taos_async_fn_t fp, void *param) {
     }
   }
 
-  schedulerAsyncFetchRows(pRequest->body.queryJob, fetchCallback, pRequest);
+  SSchedulerReq req = {
+    .syncReq = false,
+    .fetchFp = fetchCallback,
+    .cbParam = pRequest,
+  };
+  schedulerFetchRows(pRequest->body.queryJob, &req);
 }
 
 void taos_fetch_raw_block_a(TAOS_RES *res, __taos_async_fn_t fp, void *param) {

source/client/src/clientMsgHandler.c

@@ -266,7 +266,7 @@ int32_t processAlterStbRsp(void* param, SDataBuf* pMsg, int32_t code) {
   }
 
   if (pRequest->body.queryFp != NULL) {
-    SQueryExecRes* pRes = &pRequest->body.resInfo.execRes;
+    SExecResult* pRes = &pRequest->body.resInfo.execRes;
 
     if (code == TSDB_CODE_SUCCESS) {
       SCatalog* pCatalog = NULL;

source/client/src/clientStmt.c

@@ -324,9 +324,9 @@ int32_t stmtCleanSQLInfo(STscStmt* pStmt) {
 }
 
 int32_t stmtRebuildDataBlock(STscStmt* pStmt, STableDataBlocks* pDataBlock, STableDataBlocks** newBlock, uint64_t uid) {
-  SEpSet      ep = getEpSet_s(&pStmt->taos->pAppInfo->mgmtEp);
-  SVgroupInfo vgInfo = {0};
-  SRequestConnInfo conn = {.pTrans = pStmt->taos->pAppInfo->pTransporter, 
+  SEpSet           ep = getEpSet_s(&pStmt->taos->pAppInfo->mgmtEp);
+  SVgroupInfo      vgInfo = {0};
+  SRequestConnInfo conn = {.pTrans = pStmt->taos->pAppInfo->pTransporter,
                            .requestId = pStmt->exec.pRequest->requestId,
                            .requestObjRefId = pStmt->exec.pRequest->self,
                            .mgmtEps = getEpSet_s(&pStmt->taos->pAppInfo->mgmtEp)};
@@ -391,13 +391,12 @@ int32_t stmtGetFromCache(STscStmt* pStmt) {
     STMT_RET(stmtCleanBindInfo(pStmt));
   }
 
-  STableMeta* pTableMeta = NULL;
-  SRequestConnInfo conn = {.pTrans = pStmt->taos->pAppInfo->pTransporter, 
+  STableMeta*      pTableMeta = NULL;
+  SRequestConnInfo conn = {.pTrans = pStmt->taos->pAppInfo->pTransporter,
                            .requestId = pStmt->exec.pRequest->requestId,
                            .requestObjRefId = pStmt->exec.pRequest->self,
                            .mgmtEps = getEpSet_s(&pStmt->taos->pAppInfo->mgmtEp)};
-  int32_t     code =
-      catalogGetTableMeta(pStmt->pCatalog, &conn, &pStmt->bInfo.sname, &pTableMeta);
+  int32_t          code = catalogGetTableMeta(pStmt->pCatalog, &conn, &pStmt->bInfo.sname, &pTableMeta);
   if (TSDB_CODE_PAR_TABLE_NOT_EXIST == code) {
     STMT_ERR_RET(stmtCleanBindInfo(pStmt));
 
@@ -849,7 +848,7 @@ int stmtIsInsert(TAOS_STMT* stmt, int* insert) {
   if (pStmt->sql.type) {
     *insert = (STMT_TYPE_INSERT == pStmt->sql.type || STMT_TYPE_MULTI_INSERT == pStmt->sql.type);
   } else {
-    *insert = qIsInsertSql(pStmt->sql.sqlStr, 0);
+    *insert = qIsInsertValuesSql(pStmt->sql.sqlStr, pStmt->sql.sqlLen);
   }
 
   return TSDB_CODE_SUCCESS;
@@ -861,7 +860,7 @@ int stmtGetTagFields(TAOS_STMT* stmt, int* nums, TAOS_FIELD_E** fields) {
   if (STMT_TYPE_QUERY == pStmt->sql.type) {
     STMT_RET(TSDB_CODE_TSC_STMT_API_ERROR);
   }
-  
+
   STMT_ERR_RET(stmtSwitchStatus(pStmt, STMT_FETCH_FIELDS));
 
   if (pStmt->bInfo.needParse && pStmt->sql.runTimes && pStmt->sql.type > 0 &&
@@ -893,7 +892,7 @@ int stmtGetColFields(TAOS_STMT* stmt, int* nums, TAOS_FIELD_E** fields) {
   if (STMT_TYPE_QUERY == pStmt->sql.type) {
     STMT_RET(TSDB_CODE_TSC_STMT_API_ERROR);
   }
-  
+
   STMT_ERR_RET(stmtSwitchStatus(pStmt, STMT_FETCH_FIELDS));
 
   if (pStmt->bInfo.needParse && pStmt->sql.runTimes && pStmt->sql.type > 0 &&
@@ -919,7 +918,6 @@ int stmtGetColFields(TAOS_STMT* stmt, int* nums, TAOS_FIELD_E** fields) {
   return TSDB_CODE_SUCCESS;
 }
 
-
 int stmtGetParamNum(TAOS_STMT* stmt, int* nums) {
   STscStmt* pStmt = (STscStmt*)stmt;
 
@@ -952,13 +950,13 @@ int stmtGetParamNum(TAOS_STMT* stmt, int* nums) {
   return TSDB_CODE_SUCCESS;
 }
 
-int stmtGetParam(TAOS_STMT *stmt, int idx, int *type, int *bytes) {
+int stmtGetParam(TAOS_STMT* stmt, int idx, int* type, int* bytes) {
   STscStmt* pStmt = (STscStmt*)stmt;
 
   if (STMT_TYPE_QUERY == pStmt->sql.type) {
     STMT_RET(TSDB_CODE_TSC_STMT_API_ERROR);
   }
-  
+
   STMT_ERR_RET(stmtSwitchStatus(pStmt, STMT_FETCH_FIELDS));
 
   if (pStmt->bInfo.needParse && pStmt->sql.runTimes && pStmt->sql.type > 0 &&
@@ -979,8 +977,8 @@ int stmtGetParam(TAOS_STMT *stmt, int idx, int *type, int *bytes) {
     STMT_ERR_RET(stmtParseSql(pStmt));
   }
 
-  int32_t nums = 0;
-  TAOS_FIELD_E *pField = NULL;
+  int32_t       nums = 0;
+  TAOS_FIELD_E* pField = NULL;
   STMT_ERR_RET(stmtFetchColFields(stmt, &nums, &pField));
   if (idx >= nums) {
     tscError("idx %d is too big", idx);

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

@@ -16,6 +16,7 @@
 #define _DEFAULT_SOURCE
 #include "dmMgmt.h"
 #include "dmNodes.h"
+#include "index.h"
 #include "qworker.h"
 
 static bool dmRequireNode(SDnode *pDnode, SMgmtWrapper *pWrapper) {
@@ -213,6 +214,7 @@ void dmCleanupDnode(SDnode *pDnode) {
   dmCleanupServer(pDnode);
   dmClearVars(pDnode);
   rpcCleanup();
+  indexCleanup();
   dDebug("dnode is closed, ptr:%p", pDnode);
 }
 

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

@@ -1373,9 +1373,9 @@ char *buildRetension(SArray *pRetension) {
 static void dumpDbInfoData(SSDataBlock *pBlock, SDbObj *pDb, SShowObj *pShow, int32_t rows, int64_t numOfTables,
                            bool sysDb, ESdbStatus objStatus, bool sysinfo) {
   int32_t cols = 0;
+  int32_t bytes = pShow->pMeta->pSchemas[cols].bytes;
+  char   *buf = taosMemoryMalloc(bytes);
 
-  int32_t     bytes = pShow->pMeta->pSchemas[cols].bytes;
-  char       *buf = taosMemoryMalloc(bytes);
   const char *name = mndGetDbStr(pDb->name);
   if (name != NULL) {
     STR_WITH_MAXSIZE_TO_VARSTR(buf, name, bytes);
@@ -1383,11 +1383,11 @@ static void dumpDbInfoData(SSDataBlock *pBlock, SDbObj *pDb, SShowObj *pShow, in
     STR_WITH_MAXSIZE_TO_VARSTR(buf, "NULL", bytes);
   }
 
-  char *status = "ready";
-  if (objStatus == SDB_STATUS_CREATING) status = "creating";
-  if (objStatus == SDB_STATUS_DROPPING) status = "dropping";
-  char statusB[24] = {0};
-  STR_WITH_SIZE_TO_VARSTR(statusB, status, strlen(status));
+  char *statusStr = "ready";
+  if (objStatus == SDB_STATUS_CREATING) statusStr = "creating";
+  if (objStatus == SDB_STATUS_DROPPING) statusStr = "dropping";
+  char statusVstr[24] = {0};
+  STR_WITH_SIZE_TO_VARSTR(statusVstr, statusStr, strlen(statusStr));
 
   if (sysDb || !sysinfo) {
     for (int32_t i = 0; i < pShow->numOfColumns; ++i) {
@@ -1397,7 +1397,7 @@ static void dumpDbInfoData(SSDataBlock *pBlock, SDbObj *pDb, SShowObj *pShow, in
       } else if (i == 3) {
         colDataAppend(pColInfo, rows, (const char *)&numOfTables, false);
       } else if (i == 20) {
-        colDataAppend(pColInfo, rows, statusB, false);
+        colDataAppend(pColInfo, rows, statusVstr, false);
       } else {
         colDataAppendNULL(pColInfo, rows);
       }
@@ -1405,7 +1405,6 @@ static void dumpDbInfoData(SSDataBlock *pBlock, SDbObj *pDb, SShowObj *pShow, in
   } else {
     SColumnInfoData *pColInfo = taosArrayGet(pBlock->pDataBlock, cols++);
     colDataAppend(pColInfo, rows, buf, false);
-    taosMemoryFree(buf);
 
     pColInfo = taosArrayGet(pBlock->pDataBlock, cols++);
     colDataAppend(pColInfo, rows, (const char *)&pDb->createdTime, false);
@@ -1419,30 +1418,29 @@ static void dumpDbInfoData(SSDataBlock *pBlock, SDbObj *pDb, SShowObj *pShow, in
     pColInfo = taosArrayGet(pBlock->pDataBlock, cols++);
     colDataAppend(pColInfo, rows, (const char *)&pDb->cfg.replications, false);
 
-    const char *src = pDb->cfg.strict ? "strict" : "no_strict";
-    char        strict[24] = {0};
-    STR_WITH_SIZE_TO_VARSTR(strict, src, strlen(src));
+    const char *strictStr = pDb->cfg.strict ? "strict" : "no_strict";
+    char        strictVstr[24] = {0};
+    STR_WITH_SIZE_TO_VARSTR(strictVstr, strictStr, strlen(strictStr));
     pColInfo = taosArrayGet(pBlock->pDataBlock, cols++);
-    colDataAppend(pColInfo, rows, (const char *)strict, false);
+    colDataAppend(pColInfo, rows, (const char *)strictVstr, false);
 
-    char    tmp[128] = {0};
-    int32_t len = 0;
-    len = sprintf(&tmp[VARSTR_HEADER_SIZE], "%dm", pDb->cfg.daysPerFile);
-    varDataSetLen(tmp, len);
+    char    durationVstr[128] = {0};
+    int32_t len = sprintf(&durationVstr[VARSTR_HEADER_SIZE], "%dm", pDb->cfg.daysPerFile);
+    varDataSetLen(durationVstr, len);
     pColInfo = taosArrayGet(pBlock->pDataBlock, cols++);
-    colDataAppend(pColInfo, rows, (const char *)tmp, false);
+    colDataAppend(pColInfo, rows, (const char *)durationVstr, false);
 
+    char keepVstr[128] = {0};
     if (pDb->cfg.daysToKeep0 > pDb->cfg.daysToKeep1 || pDb->cfg.daysToKeep0 > pDb->cfg.daysToKeep2) {
-      len = sprintf(&tmp[VARSTR_HEADER_SIZE], "%dm,%dm,%dm", pDb->cfg.daysToKeep1, pDb->cfg.daysToKeep2,
+      len = sprintf(&keepVstr[VARSTR_HEADER_SIZE], "%dm,%dm,%dm", pDb->cfg.daysToKeep1, pDb->cfg.daysToKeep2,
                     pDb->cfg.daysToKeep0);
     } else {
-      len = sprintf(&tmp[VARSTR_HEADER_SIZE], "%dm,%dm,%dm", pDb->cfg.daysToKeep0, pDb->cfg.daysToKeep1,
+      len = sprintf(&keepVstr[VARSTR_HEADER_SIZE], "%dm,%dm,%dm", pDb->cfg.daysToKeep0, pDb->cfg.daysToKeep1,
                     pDb->cfg.daysToKeep2);
     }
-
-    varDataSetLen(tmp, len);
+    varDataSetLen(keepVstr, len);
     pColInfo = taosArrayGet(pBlock->pDataBlock, cols++);
-    colDataAppend(pColInfo, rows, (const char *)tmp, false);
+    colDataAppend(pColInfo, rows, (const char *)keepVstr, false);
 
     pColInfo = taosArrayGet(pBlock->pDataBlock, cols++);
     colDataAppend(pColInfo, rows, (const char *)&pDb->cfg.buffer, false);
@@ -1469,68 +1467,49 @@ static void dumpDbInfoData(SSDataBlock *pBlock, SDbObj *pDb, SShowObj *pShow, in
     colDataAppend(pColInfo, rows, (const char *)&pDb->cfg.compression, false);
 
     pColInfo = taosArrayGet(pBlock->pDataBlock, cols++);
-
-    STR_WITH_SIZE_TO_VARSTR(strict, src, strlen(src));
-#if 0
-    char cacheModel[24] = {0};
-    bool null = false;
-    if (pDb->cfg.cacheLastRow == 0) {
-      STR_TO_VARSTR(cacheModel, "no_cache");
-    } else if (pDb->cfg.cacheLastRow == 1) {
-      STR_TO_VARSTR(cacheModel, "last_row_cache")
-    } else {
-      null = true;
-    }
-    colDataAppend(pColInfo, rows, cacheModel, null);
-#endif
     colDataAppend(pColInfo, rows, (const char *)&pDb->cfg.cacheLastRow, false);
 
-    char *prec = NULL;
+    const char *precStr = NULL;
     switch (pDb->cfg.precision) {
       case TSDB_TIME_PRECISION_MILLI:
-        prec = TSDB_TIME_PRECISION_MILLI_STR;
+        precStr = TSDB_TIME_PRECISION_MILLI_STR;
         break;
       case TSDB_TIME_PRECISION_MICRO:
-        prec = TSDB_TIME_PRECISION_MICRO_STR;
+        precStr = TSDB_TIME_PRECISION_MICRO_STR;
         break;
       case TSDB_TIME_PRECISION_NANO:
-        prec = TSDB_TIME_PRECISION_NANO_STR;
+        precStr = TSDB_TIME_PRECISION_NANO_STR;
         break;
       default:
-        prec = "none";
+        precStr = "none";
         break;
     }
-
-    char t[10] = {0};
-    STR_WITH_SIZE_TO_VARSTR(t, prec, 2);
+    char precVstr[10] = {0};
+    STR_WITH_SIZE_TO_VARSTR(precVstr, precStr, 2);
     pColInfo = taosArrayGet(pBlock->pDataBlock, cols++);
-    colDataAppend(pColInfo, rows, (const char *)t, false);
+    colDataAppend(pColInfo, rows, (const char *)precVstr, false);
 
     pColInfo = taosArrayGet(pBlock->pDataBlock, cols++);
     colDataAppend(pColInfo, rows, (const char *)&pDb->cfg.numOfStables, false);
 
     pColInfo = taosArrayGet(pBlock->pDataBlock, cols++);
-    colDataAppend(pColInfo, rows, (const char *)statusB, false);
-
-    //    pColInfo = taosArrayGet(pBlock->pDataBlock, cols++);
-    //    colDataAppend(pColInfo, rows, (const char *)&pDb->cfg.schemaless, false);
-
-    char *p = buildRetension(pDb->cfg.pRetensions);
+    colDataAppend(pColInfo, rows, (const char *)statusVstr, false);
 
+    char *rentensionVstr = buildRetension(pDb->cfg.pRetensions);
     pColInfo = taosArrayGet(pBlock->pDataBlock, cols);
-    if (p == NULL) {
+    if (rentensionVstr == NULL) {
       colDataAppendNULL(pColInfo, rows);
     } else {
-      colDataAppend(pColInfo, rows, (const char *)p, false);
-      taosMemoryFree(p);
+      colDataAppend(pColInfo, rows, (const char *)rentensionVstr, false);
+      taosMemoryFree(rentensionVstr);
     }
   }
+
+  taosMemoryFree(buf);
 }
 
 static void setInformationSchemaDbCfg(SDbObj *pDbObj) {
-  ASSERT(pDbObj != NULL);
-  strncpy(pDbObj->name, TSDB_INFORMATION_SCHEMA_DB, tListLen(pDbObj->name));
-
+  tstrncpy(pDbObj->name, TSDB_INFORMATION_SCHEMA_DB, tListLen(pDbObj->name));
   pDbObj->createdTime = 0;
   pDbObj->cfg.numOfVgroups = 0;
   pDbObj->cfg.strict = 1;
@@ -1539,9 +1518,7 @@ static void setInformationSchemaDbCfg(SDbObj *pDbObj) {
 }
 
 static void setPerfSchemaDbCfg(SDbObj *pDbObj) {
-  ASSERT(pDbObj != NULL);
-  strncpy(pDbObj->name, TSDB_PERFORMANCE_SCHEMA_DB, tListLen(pDbObj->name));
-
+  tstrncpy(pDbObj->name, TSDB_PERFORMANCE_SCHEMA_DB, tListLen(pDbObj->name));
   pDbObj->createdTime = 0;
   pDbObj->cfg.numOfVgroups = 0;
   pDbObj->cfg.strict = 1;
@@ -1585,14 +1562,11 @@ static int32_t mndRetrieveDbs(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pBloc
 
   while (numOfRows < rowsCapacity) {
     pShow->pIter = sdbFetchAll(pSdb, SDB_DB, pShow->pIter, (void **)&pDb, &objStatus);
-    if (pShow->pIter == NULL) {
-      break;
-    }
+    if (pShow->pIter == NULL) break;
 
     if (mndCheckDbPrivilege(pMnode, pReq->info.conn.user, MND_OPER_READ_OR_WRITE_DB, pDb) == 0) {
       int32_t numOfTables = 0;
       sdbTraverse(pSdb, SDB_VGROUP, mndGetTablesOfDbFp, &numOfTables, NULL, NULL);
-
       dumpDbInfoData(pBlock, pDb, pShow, numOfRows, numOfTables, false, objStatus, sysinfo);
       numOfRows++;
     }

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

@@ -75,25 +75,13 @@ int32_t tsdbLastRowReaderOpen(void* pVnode, int32_t type, SArray* pTableIdList,
   STableKeyInfo* pKeyInfo = taosArrayGet(pTableIdList, 0);
   p->pSchema = metaGetTbTSchema(p->pVnode->pMeta, pKeyInfo->uid, -1);
   p->pTableList = pTableIdList;
-#if 0
-  for(int32_t i = 0; i < p->numOfCols; ++i) {
-    for(int32_t j = 0; j < p->pSchema->numOfCols; ++j) {
-      if (colId[i] == PRIMARYKEY_TIMESTAMP_COL_ID) {
-        p->pSlotIds[i] = -1;
-        break;
-      }
-
-      if (colId[i] == p->pSchema->columns[j].colId) {
-        p->pSlotIds[i] = j;
-        break;
-      }
-    }
 
+  for(int32_t i = 0; i < p->numOfCols; ++i) {
     if (IS_VAR_DATA_TYPE(colId[i])) {
-      p->transferBuf[i] = taosMemoryMalloc(p->pSchema->columns[p->pSlotIds[i]].bytes);
+      p->transferBuf[i] = taosMemoryMalloc(p->pSchema->columns[i].bytes);
     }
   }
-#endif
+
   *pReader = p;
   return TSDB_CODE_SUCCESS;
 }

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

@@ -527,7 +527,7 @@ static int32_t tsdbCommitTableMemData(SCommitter *pCommitter, STbDataIter *pIter
     // crash on CI, use the block following
     if (pRow) {
       TSDBKEY key = TSDBROW_KEY(pRow);
-      if (tsdbKeyCmprFn(&TSDBROW_KEY(pRow), &toKey) >= 0) {
+      if (tsdbKeyCmprFn(&key, &toKey) >= 0) {
         pRow = NULL;
       }
     }

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

@@ -212,6 +212,20 @@ static SHashObj* createDataBlockScanInfo(STsdbReader* pTsdbReader, const STableK
   return pTableMap;
 }
 
+static void resetDataBlockScanInfo(SHashObj* pTableMap) {
+  STableBlockScanInfo* p = NULL;
+
+  while((p = taosHashIterate(pTableMap, p)) != NULL) {
+    p->iterInit = false;
+    p->iiter.hasVal = false;
+    if (p->iter.iter != NULL)  {
+      tsdbTbDataIterDestroy(p->iter.iter);
+    }
+
+    taosArrayDestroy(p->delSkyline);
+  }
+}
+
 static bool isEmptyQueryTimeWindow(STimeWindow* pWindow) {
   ASSERT(pWindow != NULL);
   return pWindow->skey > pWindow->ekey;
@@ -3157,37 +3171,6 @@ bool tsdbNextDataBlock(STsdbReader* pReader) {
   } else {
     ASSERT(0);
   }
-  // if (pReader->loadType == BLOCK_LOAD_TABLE_SEQ_ORDER) {
-  //   return loadDataBlockFromTableSeq(pReader);
-  // } else {  // loadType == RR and Offset Order
-  //   if (pReader->checkFiles) {
-  //     // check if the query range overlaps with the file data block
-  //     bool exists = true;
-  //     int32_t code = buildBlockFromFiles(pReader, &exists);
-  //     if (code != TSDB_CODE_SUCCESS) {
-  //       pReader->activeIndex = 0;
-  //       pReader->checkFiles = false;
-
-  //       return false;
-  //     }
-
-  //     if (exists) {
-  //       pReader->cost.checkForNextTime += (taosGetTimestampUs() - stime);
-  //       return exists;
-  //     }
-
-  //     pReader->activeIndex = 0;
-  //     pReader->checkFiles = false;
-  //   }
-
-  //   // TODO: opt by consider the scan order
-  //   bool ret = doHasDataInBuffer(pReader);
-  //   terrno = TSDB_CODE_SUCCESS;
-
-  //   elapsedTime = taosGetTimestampUs() - stime;
-  //   pReader->cost.checkForNextTime += elapsedTime;
-  //   return ret;
-  // }
   return false;
 }
 
@@ -3305,10 +3288,10 @@ int32_t tsdbReaderReset(STsdbReader* pReader, SQueryTableDataCond* pCond, int32_
 
   setQueryTimewindow(pReader, pCond, tWinIdx);
 
-  pReader->order = pCond->order;
-  pReader->type = BLOCK_LOAD_OFFSET_ORDER;
+  pReader->order               = pCond->order;
+  pReader->type                = BLOCK_LOAD_OFFSET_ORDER;
   pReader->status.loadFromFile = true;
-  pReader->status.pTableIter = NULL;
+  pReader->status.pTableIter   = NULL;
 
   pReader->window = updateQueryTimeWindow(pReader->pTsdb, &pCond->twindows[tWinIdx]);
 
@@ -3323,6 +3306,7 @@ int32_t tsdbReaderReset(STsdbReader* pReader, SQueryTableDataCond* pCond, int32_
   STsdbFSState* pFState = pReader->pTsdb->fs->cState;
   initFilesetIterator(&pReader->status.fileIter, pFState, pReader->order, pReader->idStr);
   resetDataBlockIterator(&pReader->status.blockIter, pReader->order);
+  resetDataBlockScanInfo(pReader->status.pTableMap);
 
   int32_t code = 0;
   // no data in files, let's try buffer in memory

source/libs/executor/inc/executorimpl.h

@@ -240,7 +240,7 @@ typedef struct SColMatchInfo {
   int32_t srcSlotId;     // source slot id
   int32_t colId;
   int32_t targetSlotId;
-  bool    output;
+  bool    output;        // todo remove this?
   bool    reserved;
   int32_t matchType;     // determinate the source according to col id or slot id
 } SColMatchInfo;

source/libs/executor/src/dataInserter.c

+/*
+ * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
+ *
+ * This program is free software: you can use, redistribute, and/or modify
+ * it under the terms of the GNU Affero General Public License, version 3
+ * or later ("AGPL"), as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "dataSinkInt.h"
+#include "dataSinkMgt.h"
+#include "executorimpl.h"
+#include "planner.h"
+#include "tcompression.h"
+#include "tdatablock.h"
+#include "tglobal.h"
+#include "tqueue.h"
+
+extern SDataSinkStat gDataSinkStat;
+
+typedef struct SDataInserterBuf {
+  int32_t useSize;
+  int32_t allocSize;
+  char*   pData;
+} SDataInserterBuf;
+
+typedef struct SDataCacheEntry {
+  int32_t dataLen;
+  int32_t numOfRows;
+  int32_t numOfCols;
+  int8_t  compressed;
+  char    data[];
+} SDataCacheEntry;
+
+typedef struct SDataInserterHandle {
+  SDataSinkHandle     sink;
+  SDataSinkManager*   pManager;
+  SDataBlockDescNode* pSchema;
+  SDataDeleterNode*   pDeleter;
+  SDeleterParam*      pParam;
+  STaosQueue*         pDataBlocks;
+  SDataInserterBuf    nextOutput;
+  int32_t             status;
+  bool                queryEnd;
+  uint64_t            useconds;
+  uint64_t            cachedSize;
+  TdThreadMutex       mutex;
+} SDataInserterHandle;
+
+static bool needCompress(const SSDataBlock* pData, int32_t numOfCols) {
+  if (tsCompressColData < 0 || 0 == pData->info.rows) {
+    return false;
+  }
+
+  for (int32_t col = 0; col < numOfCols; ++col) {
+    SColumnInfoData* pColRes = taosArrayGet(pData->pDataBlock, col);
+    int32_t          colSize = pColRes->info.bytes * pData->info.rows;
+    if (NEEDTO_COMPRESS_QUERY(colSize)) {
+      return true;
+    }
+  }
+
+  return false;
+}
+
+static void toDataCacheEntry(SDataInserterHandle* pHandle, const SInputData* pInput, SDataInserterBuf* pBuf) {
+  int32_t numOfCols = LIST_LENGTH(pHandle->pSchema->pSlots);
+
+  SDataCacheEntry* pEntry = (SDataCacheEntry*)pBuf->pData;
+  pEntry->compressed = 0;
+  pEntry->numOfRows = pInput->pData->info.rows;
+  pEntry->numOfCols = taosArrayGetSize(pInput->pData->pDataBlock);
+  pEntry->dataLen = sizeof(SDeleterRes);
+
+  ASSERT(1 == pEntry->numOfRows);
+  ASSERT(1 == pEntry->numOfCols);
+
+  pBuf->useSize = sizeof(SDataCacheEntry);
+
+  SColumnInfoData* pColRes = (SColumnInfoData*)taosArrayGet(pInput->pData->pDataBlock, 0);
+
+  SDeleterRes* pRes = (SDeleterRes*)pEntry->data;
+  pRes->suid = pHandle->pParam->suid;
+  pRes->uidList = pHandle->pParam->pUidList;
+  pRes->skey = pHandle->pDeleter->deleteTimeRange.skey;
+  pRes->ekey = pHandle->pDeleter->deleteTimeRange.ekey;
+  pRes->affectedRows = *(int64_t*)pColRes->pData;
+
+  pBuf->useSize += pEntry->dataLen;
+  
+  atomic_add_fetch_64(&pHandle->cachedSize, pEntry->dataLen); 
+  atomic_add_fetch_64(&gDataSinkStat.cachedSize, pEntry->dataLen); 
+}
+
+static bool allocBuf(SDataInserterHandle* pDeleter, const SInputData* pInput, SDataInserterBuf* pBuf) {
+  uint32_t capacity = pDeleter->pManager->cfg.maxDataBlockNumPerQuery;
+  if (taosQueueItemSize(pDeleter->pDataBlocks) > capacity) {
+    qError("SinkNode queue is full, no capacity, max:%d, current:%d, no capacity", capacity,
+           taosQueueItemSize(pDeleter->pDataBlocks));
+    return false;
+  }
+
+  pBuf->allocSize = sizeof(SDataCacheEntry) + sizeof(SDeleterRes);
+
+  pBuf->pData = taosMemoryMalloc(pBuf->allocSize);
+  if (pBuf->pData == NULL) {
+    qError("SinkNode failed to malloc memory, size:%d, code:%d", pBuf->allocSize, TAOS_SYSTEM_ERROR(errno));
+  }
+
+  return NULL != pBuf->pData;
+}
+
+static int32_t updateStatus(SDataInserterHandle* pDeleter) {
+  taosThreadMutexLock(&pDeleter->mutex);
+  int32_t blockNums = taosQueueItemSize(pDeleter->pDataBlocks);
+  int32_t status =
+      (0 == blockNums ? DS_BUF_EMPTY
+                      : (blockNums < pDeleter->pManager->cfg.maxDataBlockNumPerQuery ? DS_BUF_LOW : DS_BUF_FULL));
+  pDeleter->status = status;
+  taosThreadMutexUnlock(&pDeleter->mutex);
+  return status;
+}
+
+static int32_t getStatus(SDataInserterHandle* pDeleter) {
+  taosThreadMutexLock(&pDeleter->mutex);
+  int32_t status = pDeleter->status;
+  taosThreadMutexUnlock(&pDeleter->mutex);
+  return status;
+}
+
+static int32_t putDataBlock(SDataSinkHandle* pHandle, const SInputData* pInput, bool* pContinue) {
+  SDataInserterHandle* pDeleter = (SDataInserterHandle*)pHandle;
+  SDataInserterBuf*    pBuf = taosAllocateQitem(sizeof(SDataInserterBuf), DEF_QITEM);
+  if (NULL == pBuf || !allocBuf(pDeleter, pInput, pBuf)) {
+    return TSDB_CODE_QRY_OUT_OF_MEMORY;
+  }
+  toDataCacheEntry(pDeleter, pInput, pBuf);
+  taosWriteQitem(pDeleter->pDataBlocks, pBuf);
+  *pContinue = (DS_BUF_LOW == updateStatus(pDeleter) ? true : false);
+  return TSDB_CODE_SUCCESS;
+}
+
+static void endPut(struct SDataSinkHandle* pHandle, uint64_t useconds) {
+  SDataInserterHandle* pDeleter = (SDataInserterHandle*)pHandle;
+  taosThreadMutexLock(&pDeleter->mutex);
+  pDeleter->queryEnd = true;
+  pDeleter->useconds = useconds;
+  taosThreadMutexUnlock(&pDeleter->mutex);
+}
+
+static void getDataLength(SDataSinkHandle* pHandle, int32_t* pLen, bool* pQueryEnd) {
+  SDataInserterHandle* pDeleter = (SDataInserterHandle*)pHandle;
+  if (taosQueueEmpty(pDeleter->pDataBlocks)) {
+    *pQueryEnd = pDeleter->queryEnd;
+    *pLen = 0;
+    return;
+  }
+
+  SDataInserterBuf* pBuf = NULL;
+  taosReadQitem(pDeleter->pDataBlocks, (void**)&pBuf);
+  memcpy(&pDeleter->nextOutput, pBuf, sizeof(SDataInserterBuf));
+  taosFreeQitem(pBuf);
+  *pLen = ((SDataCacheEntry*)(pDeleter->nextOutput.pData))->dataLen;
+  *pQueryEnd = pDeleter->queryEnd;
+  qDebug("got data len %d, row num %d in sink", *pLen, ((SDataCacheEntry*)(pDeleter->nextOutput.pData))->numOfRows);
+}
+
+static int32_t getDataBlock(SDataSinkHandle* pHandle, SOutputData* pOutput) {
+  SDataInserterHandle* pDeleter = (SDataInserterHandle*)pHandle;
+  if (NULL == pDeleter->nextOutput.pData) {
+    assert(pDeleter->queryEnd);
+    pOutput->useconds = pDeleter->useconds;
+    pOutput->precision = pDeleter->pSchema->precision;
+    pOutput->bufStatus = DS_BUF_EMPTY;
+    pOutput->queryEnd = pDeleter->queryEnd;
+    return TSDB_CODE_SUCCESS;
+  }
+  SDataCacheEntry* pEntry = (SDataCacheEntry*)(pDeleter->nextOutput.pData);
+  memcpy(pOutput->pData, pEntry->data, pEntry->dataLen);
+  pOutput->numOfRows = pEntry->numOfRows;
+  pOutput->numOfCols = pEntry->numOfCols;
+  pOutput->compressed = pEntry->compressed;
+
+  atomic_sub_fetch_64(&pDeleter->cachedSize, pEntry->dataLen);  
+  atomic_sub_fetch_64(&gDataSinkStat.cachedSize, pEntry->dataLen); 
+
+  taosMemoryFreeClear(pDeleter->nextOutput.pData);  // todo persistent
+  pOutput->bufStatus = updateStatus(pDeleter);
+  taosThreadMutexLock(&pDeleter->mutex);
+  pOutput->queryEnd = pDeleter->queryEnd;
+  pOutput->useconds = pDeleter->useconds;
+  pOutput->precision = pDeleter->pSchema->precision;
+  taosThreadMutexUnlock(&pDeleter->mutex);
+  
+  return TSDB_CODE_SUCCESS;
+}
+
+static int32_t destroyDataSinker(SDataSinkHandle* pHandle) {
+  SDataInserterHandle* pDeleter = (SDataInserterHandle*)pHandle;
+  atomic_sub_fetch_64(&gDataSinkStat.cachedSize, pDeleter->cachedSize);
+  taosMemoryFreeClear(pDeleter->nextOutput.pData);
+  while (!taosQueueEmpty(pDeleter->pDataBlocks)) {
+    SDataInserterBuf* pBuf = NULL;
+    taosReadQitem(pDeleter->pDataBlocks, (void**)&pBuf);
+    taosMemoryFreeClear(pBuf->pData);
+    taosFreeQitem(pBuf);
+  }
+  taosCloseQueue(pDeleter->pDataBlocks);
+  taosThreadMutexDestroy(&pDeleter->mutex);
+  return TSDB_CODE_SUCCESS;
+}
+
+static int32_t getCacheSize(struct SDataSinkHandle* pHandle, uint64_t* size) {
+  SDataInserterHandle* pDispatcher = (SDataInserterHandle*)pHandle;
+
+  *size = atomic_load_64(&pDispatcher->cachedSize);
+  return TSDB_CODE_SUCCESS;
+}
+
+int32_t createDataInserter(SDataSinkManager* pManager, const SDataSinkNode* pDataSink, DataSinkHandle* pHandle, void *pParam) {
+  SDataInserterHandle* inserter = taosMemoryCalloc(1, sizeof(SDataInserterHandle));
+  if (NULL == inserter) {
+    terrno = TSDB_CODE_QRY_OUT_OF_MEMORY;
+    return TSDB_CODE_QRY_OUT_OF_MEMORY;
+  }
+
+  SDataDeleterNode* pDeleterNode = (SDataDeleterNode *)pDataSink;
+  inserter->sink.fPut = putDataBlock;
+  inserter->sink.fEndPut = endPut;
+  inserter->sink.fGetLen = getDataLength;
+  inserter->sink.fGetData = getDataBlock;
+  inserter->sink.fDestroy = destroyDataSinker;
+  inserter->sink.fGetCacheSize = getCacheSize;
+  inserter->pManager = pManager;
+  inserter->pDeleter = pDeleterNode;
+  inserter->pSchema = pDataSink->pInputDataBlockDesc;
+  inserter->pParam = pParam;
+  inserter->status = DS_BUF_EMPTY;
+  inserter->queryEnd = false;
+  inserter->pDataBlocks = taosOpenQueue();
+  taosThreadMutexInit(&inserter->mutex, NULL);
+  if (NULL == inserter->pDataBlocks) {
+    terrno = TSDB_CODE_QRY_OUT_OF_MEMORY;
+    return TSDB_CODE_QRY_OUT_OF_MEMORY;
+  }
+  *pHandle = inserter;
+  return TSDB_CODE_SUCCESS;
+}

source/libs/executor/src/executil.c

@@ -704,7 +704,7 @@ void relocateColumnData(SSDataBlock* pBlock, const SArray* pColMatchInfo, SArray
   while (i < numOfSrcCols && j < taosArrayGetSize(pColMatchInfo)) {
     SColumnInfoData* p = taosArrayGet(pCols, i);
     SColMatchInfo*   pmInfo = taosArrayGet(pColMatchInfo, j);
-    if (!outputEveryColumn && !pmInfo->output) {
+    if (!outputEveryColumn && pmInfo->reserved) {
       j++;
       continue;
     }

source/libs/executor/src/executorimpl.c

@@ -3992,9 +3992,9 @@ int32_t extractTableSchemaVersion(SReadHandle* pHandle, uint64_t uid, SExecTaskI
   SMetaReader mr = {0};
   metaReaderInit(&mr, pHandle->meta, 0);
   int32_t code = metaGetTableEntryByUid(&mr, uid);
-  if (code) {
+  if (code != TSDB_CODE_SUCCESS) {
     metaReaderClear(&mr);
-    return code;
+    return terrno;
   }
 
   pTaskInfo->schemaVer.tablename = strdup(mr.me.name);
@@ -4307,6 +4307,10 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
 //      }
 
       int32_t code = extractTableSchemaVersion(pHandle, pScanNode->uid, pTaskInfo);
+      if (code != TSDB_CODE_SUCCESS) {
+        pTaskInfo->code = code;
+        return NULL;
+      }
 
       pTableListInfo->pTableList = taosArrayInit(4, sizeof(STableKeyInfo));
       if (pScanNode->tableType == TSDB_SUPER_TABLE) {

source/libs/executor/src/scanoperator.c

@@ -2438,8 +2438,9 @@ int32_t stopGroupTableMergeScan(SOperatorInfo* pOperator) {
     STsdbReader* reader = taosArrayGetP(pInfo->dataReaders, i);
     tsdbReaderClose(reader);
   }
-  taosArrayDestroy(pInfo->dataReaders);
 
+  taosArrayDestroy(pInfo->dataReaders);
+  pInfo->dataReaders = NULL;
   return TSDB_CODE_SUCCESS;
 }
 

source/libs/function/src/builtins.c

@@ -605,7 +605,7 @@ static int32_t translateTopBot(SFunctionNode* pFunc, char* pErrBuf, int32_t len)
   }
 
   SValueNode* pValue = (SValueNode*)pParamNode1;
-  if (pValue->node.resType.type != TSDB_DATA_TYPE_BIGINT) {
+  if (!IS_INTEGER_TYPE(pValue->node.resType.type)) {
     return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
   }
 
@@ -2210,7 +2210,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
   {
     .name = "last_row",
     .type = FUNCTION_TYPE_LAST_ROW,
-    .classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_MULTI_RES_FUNC | FUNC_MGT_TIMELINE_FUNC,
+    .classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_MULTI_RES_FUNC | FUNC_MGT_SELECT_FUNC | FUNC_MGT_TIMELINE_FUNC,
     .translateFunc = translateFirstLast,
     .getEnvFunc   = getFirstLastFuncEnv,
     .initFunc     = functionSetup,

source/libs/function/src/builtinsimpl.c

@@ -90,12 +90,14 @@ typedef struct SStddevRes {
   double  result;
   int64_t count;
   union {
-    double  quadraticDSum;
-    int64_t quadraticISum;
+    double   quadraticDSum;
+    int64_t  quadraticISum;
+    uint64_t quadraticUSum;
   };
   union {
-    double  dsum;
-    int64_t isum;
+    double   dsum;
+    int64_t  isum;
+    uint64_t usum;
   };
   int16_t type;
 } SStddevRes;
@@ -1729,6 +1731,68 @@ int32_t stddevFunction(SqlFunctionCtx* pCtx) {
       break;
     }
 
+    case TSDB_DATA_TYPE_UTINYINT: {
+      uint8_t* plist = (uint8_t*)pCol->pData;
+      for (int32_t i = start; i < numOfRows + start; ++i) {
+        if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) {
+          continue;
+        }
+
+        numOfElem += 1;
+        pStddevRes->count += 1;
+        pStddevRes->usum += plist[i];
+        pStddevRes->quadraticISum += plist[i] * plist[i];
+      }
+
+      break;
+    }
+
+    case TSDB_DATA_TYPE_USMALLINT: {
+      uint16_t* plist = (uint16_t*)pCol->pData;
+      for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) {
+        if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) {
+          continue;
+        }
+
+        numOfElem += 1;
+        pStddevRes->count += 1;
+        pStddevRes->usum += plist[i];
+        pStddevRes->quadraticISum += plist[i] * plist[i];
+      }
+      break;
+    }
+
+    case TSDB_DATA_TYPE_UINT: {
+      uint32_t* plist = (uint32_t*)pCol->pData;
+      for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) {
+        if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) {
+          continue;
+        }
+
+        numOfElem += 1;
+        pStddevRes->count += 1;
+        pStddevRes->usum += plist[i];
+        pStddevRes->quadraticISum += plist[i] * plist[i];
+      }
+
+      break;
+    }
+
+    case TSDB_DATA_TYPE_UBIGINT: {
+      uint64_t* plist = (uint64_t*)pCol->pData;
+      for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) {
+        if (pCol->hasNull && colDataIsNull_f(pCol->nullbitmap, i)) {
+          continue;
+        }
+
+        numOfElem += 1;
+        pStddevRes->count += 1;
+        pStddevRes->usum += plist[i];
+        pStddevRes->quadraticISum += plist[i] * plist[i];
+      }
+      break;
+    }
+
     case TSDB_DATA_TYPE_FLOAT: {
       float* plist = (float*)pCol->pData;
       for (int32_t i = start; i < numOfRows + pInput->startRowIndex; ++i) {
@@ -1771,9 +1835,12 @@ _stddev_over:
 
 static void stddevTransferInfo(SStddevRes* pInput, SStddevRes* pOutput) {
   pOutput->type = pInput->type;
-  if (IS_INTEGER_TYPE(pOutput->type)) {
+  if (IS_SIGNED_NUMERIC_TYPE(pOutput->type)) {
     pOutput->quadraticISum += pInput->quadraticISum;
     pOutput->isum += pInput->isum;
+  } else if (IS_UNSIGNED_NUMERIC_TYPE(pOutput->type)) {
+    pOutput->quadraticUSum += pInput->quadraticUSum;
+    pOutput->usum += pInput->usum;
   } else {
     pOutput->quadraticDSum += pInput->quadraticDSum;
     pOutput->dsum += pInput->dsum;
@@ -1848,6 +1915,22 @@ int32_t stddevInvertFunction(SqlFunctionCtx* pCtx) {
       LIST_STDDEV_SUB_N(pStddevRes->isum, int64_t);
       break;
     }
+    case TSDB_DATA_TYPE_UTINYINT: {
+      LIST_STDDEV_SUB_N(pStddevRes->isum, uint8_t);
+      break;
+    }
+    case TSDB_DATA_TYPE_USMALLINT: {
+      LIST_STDDEV_SUB_N(pStddevRes->isum, uint16_t);
+      break;
+    }
+    case TSDB_DATA_TYPE_UINT: {
+      LIST_STDDEV_SUB_N(pStddevRes->isum, uint32_t);
+      break;
+    }
+    case TSDB_DATA_TYPE_UBIGINT: {
+      LIST_STDDEV_SUB_N(pStddevRes->isum, uint64_t);
+      break;
+    }
     case TSDB_DATA_TYPE_FLOAT: {
       LIST_STDDEV_SUB_N(pStddevRes->dsum, float);
       break;
@@ -1871,9 +1954,12 @@ int32_t stddevFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) {
   int32_t               type = pStddevRes->type;
   double                avg;
 
-  if (IS_INTEGER_TYPE(type)) {
+  if (IS_SIGNED_NUMERIC_TYPE(type)) {
     avg = pStddevRes->isum / ((double)pStddevRes->count);
     pStddevRes->result = sqrt(fabs(pStddevRes->quadraticISum / ((double)pStddevRes->count) - avg * avg));
+  } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) {
+    avg = pStddevRes->usum / ((double)pStddevRes->count);
+    pStddevRes->result = sqrt(fabs(pStddevRes->quadraticUSum / ((double)pStddevRes->count) - avg * avg));
   } else {
     avg = pStddevRes->dsum / ((double)pStddevRes->count);
     pStddevRes->result = sqrt(fabs(pStddevRes->quadraticDSum / ((double)pStddevRes->count) - avg * avg));
@@ -1913,9 +1999,12 @@ int32_t stddevCombine(SqlFunctionCtx* pDestCtx, SqlFunctionCtx* pSourceCtx) {
   SResultRowEntryInfo* pSResInfo = GET_RES_INFO(pSourceCtx);
   SStddevRes*          pSBuf = GET_ROWCELL_INTERBUF(pSResInfo);
 
-  if (IS_INTEGER_TYPE(type)) {
+  if (IS_SIGNED_NUMERIC_TYPE(type)) {
     pDBuf->isum += pSBuf->isum;
     pDBuf->quadraticISum += pSBuf->quadraticISum;
+  } else if (IS_UNSIGNED_NUMERIC_TYPE(type)) {
+    pDBuf->usum += pSBuf->usum;
+    pDBuf->quadraticUSum += pSBuf->quadraticUSum;
   } else {
     pDBuf->dsum += pSBuf->dsum;
     pDBuf->quadraticDSum += pSBuf->quadraticDSum;

source/libs/index/src/index.c

@@ -65,9 +65,10 @@ void indexInit() {
   indexQhandle = taosInitScheduler(INDEX_QUEUE_SIZE, INDEX_NUM_OF_THREADS, "index");
   indexRefMgt = taosOpenRef(10, indexDestroy);
 }
-void indexCleanUp() {
+void indexCleanup() {
   // refacto later
   taosCleanUpScheduler(indexQhandle);
+  taosCloseRef(indexRefMgt);
 }
 
 typedef struct SIdxColInfo {

source/libs/nodes/src/nodesCloneFuncs.c

@@ -165,7 +165,7 @@ static int32_t valueNodeCopy(const SValueNode* pSrc, SValueNode* pDst) {
       memcpy(pDst->datum.p, pSrc->datum.p, len);
       break;
     }
-    case TSDB_DATA_TYPE_JSON:{
+    case TSDB_DATA_TYPE_JSON: {
       int32_t len = getJsonValueLen(pSrc->datum.p);
       pDst->datum.p = taosMemoryCalloc(1, len);
       if (NULL == pDst->datum.p) {
@@ -397,6 +397,7 @@ static int32_t logicVnodeModifCopy(const SVnodeModifyLogicNode* pSrc, SVnodeModi
   COPY_SCALAR_FIELD(tableType);
   COPY_CHAR_ARRAY_FIELD(tableFName);
   COPY_OBJECT_FIELD(deleteTimeRange, sizeof(STimeWindow));
+  CLONE_OBJECT_FIELD(pVgroupList, vgroupsInfoClone);
   return TSDB_CODE_SUCCESS;
 }
 

source/libs/nodes/src/nodesCodeFuncs.c

@@ -19,8 +19,8 @@
 #include "query.h"
 #include "querynodes.h"
 #include "taoserror.h"
-#include "tjson.h"
 #include "tdatablock.h"
+#include "tjson.h"
 
 static int32_t nodeToJson(const void* pObj, SJson* pJson);
 static int32_t jsonToNode(const SJson* pJson, void* pObj);
@@ -179,6 +179,8 @@ const char* nodesNodeName(ENodeType type) {
       return "ShowVnodeStmt";
     case QUERY_NODE_DELETE_STMT:
       return "DeleteStmt";
+    case QUERY_NODE_INSERT_STMT:
+      return "InsertStmt";
     case QUERY_NODE_LOGIC_PLAN_SCAN:
       return "LogicScan";
     case QUERY_NODE_LOGIC_PLAN_JOIN:
@@ -271,6 +273,8 @@ const char* nodesNodeName(ENodeType type) {
       return "PhysiDispatch";
     case QUERY_NODE_PHYSICAL_PLAN_INSERT:
       return "PhysiInsert";
+    case QUERY_NODE_PHYSICAL_PLAN_QUERY_INSERT:
+      return "PhysiQueryInsert";
     case QUERY_NODE_PHYSICAL_PLAN_DELETE:
       return "PhysiDelete";
     case QUERY_NODE_PHYSICAL_SUBPLAN:
@@ -2210,6 +2214,58 @@ static int32_t physiDispatchNodeToJson(const void* pObj, SJson* pJson) { return
 
 static int32_t jsonToPhysiDispatchNode(const SJson* pJson, void* pObj) { return jsonToPhysicDataSinkNode(pJson, pObj); }
 
+static const char* jkQueryInsertPhysiPlanTableId = "TableId";
+static const char* jkQueryInsertPhysiPlanTableType = "TableType";
+static const char* jkQueryInsertPhysiPlanTableFName = "TableFName";
+static const char* jkQueryInsertPhysiPlanVgId = "VgId";
+static const char* jkQueryInsertPhysiPlanEpSet = "EpSet";
+
+static int32_t physiQueryInsertNodeToJson(const void* pObj, SJson* pJson) {
+  const SQueryInserterNode* pNode = (const SQueryInserterNode*)pObj;
+
+  int32_t code = physicDataSinkNodeToJson(pObj, pJson);
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonAddIntegerToObject(pJson, jkQueryInsertPhysiPlanTableId, pNode->tableId);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonAddIntegerToObject(pJson, jkQueryInsertPhysiPlanTableType, pNode->tableType);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonAddStringToObject(pJson, jkQueryInsertPhysiPlanTableFName, pNode->tableFName);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonAddIntegerToObject(pJson, jkQueryInsertPhysiPlanVgId, pNode->vgId);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonAddObject(pJson, jkQueryInsertPhysiPlanEpSet, epSetToJson, &pNode->epSet);
+  }
+
+  return code;
+}
+
+static int32_t jsonToPhysiQueryInsertNode(const SJson* pJson, void* pObj) {
+  SQueryInserterNode* pNode = (SQueryInserterNode*)pObj;
+
+  int32_t code = jsonToPhysicDataSinkNode(pJson, pObj);
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonGetUBigIntValue(pJson, jkQueryInsertPhysiPlanTableId, &pNode->tableId);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonGetTinyIntValue(pJson, jkQueryInsertPhysiPlanTableType, &pNode->tableType);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonGetStringValue(pJson, jkQueryInsertPhysiPlanTableFName, pNode->tableFName);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonGetIntValue(pJson, jkQueryInsertPhysiPlanVgId, &pNode->vgId);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonToObject(pJson, jkQueryInsertPhysiPlanEpSet, jsonToEpSet, &pNode->epSet);
+  }
+
+  return code;
+}
+
 static const char* jkDeletePhysiPlanTableId = "TableId";
 static const char* jkDeletePhysiPlanTableType = "TableType";
 static const char* jkDeletePhysiPlanTableFName = "TableFName";
@@ -2641,9 +2697,9 @@ static int32_t datumToJson(const void* pObj, SJson* pJson) {
     case TSDB_DATA_TYPE_VARBINARY:
       code = tjsonAddStringToObject(pJson, jkValueDatum, varDataVal(pNode->datum.p));
       break;
-    case TSDB_DATA_TYPE_JSON:{
+    case TSDB_DATA_TYPE_JSON: {
       int32_t len = getJsonValueLen(pNode->datum.p);
-      char* buf = taosMemoryCalloc( len * 2 + 1, sizeof(char));
+      char*   buf = taosMemoryCalloc(len * 2 + 1, sizeof(char));
       code = taosHexEncode(pNode->datum.p, buf, len);
       if (code != TSDB_CODE_SUCCESS) {
         taosMemoryFree(buf);
@@ -2775,7 +2831,7 @@ static int32_t jsonToDatum(const SJson* pJson, void* pObj) {
       }
       break;
     }
-    case TSDB_DATA_TYPE_JSON:{
+    case TSDB_DATA_TYPE_JSON: {
       pNode->datum.p = taosMemoryCalloc(1, pNode->node.resType.bytes);
       if (NULL == pNode->datum.p) {
         code = TSDB_CODE_OUT_OF_MEMORY;
@@ -4232,6 +4288,8 @@ static int32_t specificNodeToJson(const void* pObj, SJson* pJson) {
       return physiDispatchNodeToJson(pObj, pJson);
     case QUERY_NODE_PHYSICAL_PLAN_INSERT:
       break;
+    case QUERY_NODE_PHYSICAL_PLAN_QUERY_INSERT:
+      return physiQueryInsertNodeToJson(pObj, pJson);
     case QUERY_NODE_PHYSICAL_PLAN_DELETE:
       return physiDeleteNodeToJson(pObj, pJson);
     case QUERY_NODE_PHYSICAL_SUBPLAN:
@@ -4374,6 +4432,8 @@ static int32_t jsonToSpecificNode(const SJson* pJson, void* pObj) {
       return jsonToPhysiInterpFuncNode(pJson, pObj);
     case QUERY_NODE_PHYSICAL_PLAN_DISPATCH:
       return jsonToPhysiDispatchNode(pJson, pObj);
+    case QUERY_NODE_PHYSICAL_PLAN_QUERY_INSERT:
+      return jsonToPhysiQueryInsertNode(pJson, pObj);
     case QUERY_NODE_PHYSICAL_PLAN_DELETE:
       return jsonToPhysiDeleteNode(pJson, pObj);
     case QUERY_NODE_PHYSICAL_SUBPLAN:

source/libs/nodes/src/nodesUtilFuncs.c

@@ -231,6 +231,8 @@ SNode* nodesMakeNode(ENodeType type) {
       return makeNode(type, sizeof(SKillStmt));
     case QUERY_NODE_DELETE_STMT:
       return makeNode(type, sizeof(SDeleteStmt));
+    case QUERY_NODE_INSERT_STMT:
+      return makeNode(type, sizeof(SInsertStmt));
     case QUERY_NODE_QUERY:
       return makeNode(type, sizeof(SQuery));
     case QUERY_NODE_LOGIC_PLAN_SCAN:
@@ -327,6 +329,8 @@ SNode* nodesMakeNode(ENodeType type) {
       return makeNode(type, sizeof(SDataDispatcherNode));
     case QUERY_NODE_PHYSICAL_PLAN_INSERT:
       return makeNode(type, sizeof(SDataInserterNode));
+    case QUERY_NODE_PHYSICAL_PLAN_QUERY_INSERT:
+      return makeNode(type, sizeof(SQueryInserterNode));
     case QUERY_NODE_PHYSICAL_PLAN_DELETE:
       return makeNode(type, sizeof(SDataDeleterNode));
     case QUERY_NODE_PHYSICAL_SUBPLAN:
@@ -694,6 +698,13 @@ void nodesDestroyNode(SNode* pNode) {
       nodesDestroyNode(pStmt->pTagCond);
       break;
     }
+    case QUERY_NODE_INSERT_STMT: {
+      SInsertStmt* pStmt = (SInsertStmt*)pNode;
+      nodesDestroyNode(pStmt->pTable);
+      nodesDestroyList(pStmt->pCols);
+      nodesDestroyNode(pStmt->pQuery);
+      break;
+    }
     case QUERY_NODE_QUERY: {
       SQuery* pQuery = (SQuery*)pNode;
       nodesDestroyNode(pQuery->pRoot);
@@ -929,6 +940,11 @@ void nodesDestroyNode(SNode* pNode) {
       taosMemoryFreeClear(pSink->pData);
       break;
     }
+    case QUERY_NODE_PHYSICAL_PLAN_QUERY_INSERT: {
+      SQueryInserterNode* pSink = (SQueryInserterNode*)pNode;
+      destroyDataSinkNode((SDataSinkNode*)pSink);
+      break;
+    }
     case QUERY_NODE_PHYSICAL_PLAN_DELETE: {
       SDataDeleterNode* pSink = (SDataDeleterNode*)pNode;
       destroyDataSinkNode((SDataSinkNode*)pSink);
@@ -1528,7 +1544,6 @@ int32_t nodesCollectColumnsFromNode(SNode* node, const char* pTableAlias, EColle
   }
 
   return TSDB_CODE_SUCCESS;
-
 }
 
 typedef struct SCollectFuncsCxt {

source/libs/parser/inc/parAst.h

@@ -211,6 +211,7 @@ SNode* createSyncdbStmt(SAstCreateContext* pCxt, const SToken* pDbName);
 SNode* createGrantStmt(SAstCreateContext* pCxt, int64_t privileges, SToken* pDbName, SToken* pUserName);
 SNode* createRevokeStmt(SAstCreateContext* pCxt, int64_t privileges, SToken* pDbName, SToken* pUserName);
 SNode* createDeleteStmt(SAstCreateContext* pCxt, SNode* pTable, SNode* pWhere);
+SNode* createInsertStmt(SAstCreateContext* pCxt, SNode* pTable, SNodeList* pCols, SNode* pQuery);
 
 #ifdef __cplusplus
 }

source/libs/parser/inc/sql.y

@@ -260,7 +260,7 @@ multi_create_clause(A) ::= multi_create_clause(B) create_subtable_clause(C).
 
 create_subtable_clause(A) ::=
   not_exists_opt(B) full_table_name(C) USING full_table_name(D)
-  specific_tags_opt(E) TAGS NK_LP expression_list(F) NK_RP table_options(G).      { A = createCreateSubTableClause(pCxt, B, C, D, E, F, G); }
+  specific_cols_opt(E) TAGS NK_LP expression_list(F) NK_RP table_options(G).      { A = createCreateSubTableClause(pCxt, B, C, D, E, F, G); }
 
 %type multi_drop_clause                                                           { SNodeList* }
 %destructor multi_drop_clause                                                     { nodesDestroyList($$); }
@@ -269,10 +269,10 @@ multi_drop_clause(A) ::= multi_drop_clause(B) drop_table_clause(C).
 
 drop_table_clause(A) ::= exists_opt(B) full_table_name(C).                        { A = createDropTableClause(pCxt, B, C); }
 
-%type specific_tags_opt                                                           { SNodeList* }
-%destructor specific_tags_opt                                                     { nodesDestroyList($$); }
-specific_tags_opt(A) ::= .                                                        { A = NULL; }
-specific_tags_opt(A) ::= NK_LP col_name_list(B) NK_RP.                            { A = B; }
+%type specific_cols_opt                                                           { SNodeList* }
+%destructor specific_cols_opt                                                     { nodesDestroyList($$); }
+specific_cols_opt(A) ::= .                                                        { A = NULL; }
+specific_cols_opt(A) ::= NK_LP col_name_list(B) NK_RP.                            { A = B; }
 
 full_table_name(A) ::= table_name(B).                                             { A = createRealTableNode(pCxt, NULL, &B, NULL); }
 full_table_name(A) ::= db_name(B) NK_DOT table_name(C).                           { A = createRealTableNode(pCxt, &B, &C, NULL); }
@@ -516,6 +516,9 @@ cmd ::= DELETE FROM full_table_name(A) where_clause_opt(B).
 /************************************************ select **************************************************************/
 cmd ::= query_expression(A).                                                      { pCxt->pRootNode = A; }
 
+/************************************************ insert **************************************************************/
+cmd ::= INSERT INTO full_table_name(A) specific_cols_opt(B) query_expression(C).  { pCxt->pRootNode = createInsertStmt(pCxt, A, B, C); }
+
 /************************************************ literal *************************************************************/
 literal(A) ::= NK_INTEGER(B).                                                     { A = createRawExprNode(pCxt, &B, createValueNode(pCxt, TSDB_DATA_TYPE_BIGINT, &B)); }
 literal(A) ::= NK_FLOAT(B).                                                       { A = createRawExprNode(pCxt, &B, createValueNode(pCxt, TSDB_DATA_TYPE_DOUBLE, &B)); }
@@ -974,4 +977,4 @@ null_ordering_opt(A) ::= .
 null_ordering_opt(A) ::= NULLS FIRST.                                             { A = NULL_ORDER_FIRST; }
 null_ordering_opt(A) ::= NULLS LAST.                                              { A = NULL_ORDER_LAST; }
 
-%fallback ID NK_BITNOT INSERT VALUES IMPORT NK_SEMI FILE.
+%fallback ID NK_BITNOT VALUES IMPORT NK_SEMI FILE.

source/libs/parser/src/parAstCreater.c

@@ -1673,3 +1673,13 @@ SNode* createDeleteStmt(SAstCreateContext* pCxt, SNode* pTable, SNode* pWhere) {
   }
   return (SNode*)pStmt;
 }
+
+SNode* createInsertStmt(SAstCreateContext* pCxt, SNode* pTable, SNodeList* pCols, SNode* pQuery) {
+  CHECK_PARSER_STATUS(pCxt);
+  SInsertStmt* pStmt = (SInsertStmt*)nodesMakeNode(QUERY_NODE_INSERT_STMT);
+  CHECK_OUT_OF_MEM(pStmt);
+  pStmt->pTable = pTable;
+  pStmt->pCols = pCols;
+  pStmt->pQuery = pQuery;
+  return (SNode*)pStmt;
+}

source/libs/parser/src/parAstParser.c

@@ -451,6 +451,14 @@ static int32_t collectMetaKeyFromDelete(SCollectMetaKeyCxt* pCxt, SDeleteStmt* p
   return collectMetaKeyFromRealTableImpl(pCxt, (SRealTableNode*)pStmt->pFromTable, AUTH_TYPE_WRITE);
 }
 
+static int32_t collectMetaKeyFromInsert(SCollectMetaKeyCxt* pCxt, SInsertStmt* pStmt) {
+  int32_t code = collectMetaKeyFromRealTableImpl(pCxt, (SRealTableNode*)pStmt->pTable, AUTH_TYPE_WRITE);
+  if (TSDB_CODE_SUCCESS == code) {
+    code = collectMetaKeyFromQuery(pCxt, pStmt->pQuery);
+  }
+  return code;
+}
+
 static int32_t collectMetaKeyFromShowBlockDist(SCollectMetaKeyCxt* pCxt, SShowTableDistributedStmt* pStmt) {
   SName name = {.type = TSDB_TABLE_NAME_T, .acctId = pCxt->pParseCxt->acctId};
   strcpy(name.dbname, pStmt->dbName);
@@ -560,6 +568,8 @@ static int32_t collectMetaKeyFromQuery(SCollectMetaKeyCxt* pCxt, SNode* pStmt) {
       return collectMetaKeyFromShowTransactions(pCxt, (SShowStmt*)pStmt);
     case QUERY_NODE_DELETE_STMT:
       return collectMetaKeyFromDelete(pCxt, (SDeleteStmt*)pStmt);
+    case QUERY_NODE_INSERT_STMT:
+      return collectMetaKeyFromInsert(pCxt, (SInsertStmt*)pStmt);
     case QUERY_NODE_SHOW_TABLE_DISTRIBUTED_STMT:
       return collectMetaKeyFromShowBlockDist(pCxt, (SShowTableDistributedStmt*)pStmt);
     case QUERY_NODE_SHOW_SUBSCRIPTIONS_STMT:

source/libs/parser/src/parAuthenticator.c

@@ -39,7 +39,7 @@ static int32_t checkAuth(SAuthCxt* pCxt, const char* pDbName, AUTH_TYPE type) {
   if (NULL != pCxt->pMetaCache) {
     code = getUserAuthFromCache(pCxt->pMetaCache, pParseCxt->pUser, dbFname, type, &pass);
   } else {
-    SRequestConnInfo conn = {.pTrans = pParseCxt->pTransporter, 
+    SRequestConnInfo conn = {.pTrans = pParseCxt->pTransporter,
                              .requestId = pParseCxt->requestId,
                              .requestObjRefId = pParseCxt->requestRid,
                              .mgmtEps = pParseCxt->mgmtEpSet};
@@ -88,6 +88,14 @@ static int32_t authDelete(SAuthCxt* pCxt, SDeleteStmt* pDelete) {
   return checkAuth(pCxt, ((SRealTableNode*)pDelete->pFromTable)->table.dbName, AUTH_TYPE_WRITE);
 }
 
+static int32_t authInsert(SAuthCxt* pCxt, SInsertStmt* pInsert) {
+  int32_t code = checkAuth(pCxt, ((SRealTableNode*)pInsert->pTable)->table.dbName, AUTH_TYPE_WRITE);
+  if (TSDB_CODE_SUCCESS == code) {
+    code = authQuery(pCxt, pInsert->pQuery);
+  }
+  return code;
+}
+
 static int32_t authQuery(SAuthCxt* pCxt, SNode* pStmt) {
   switch (nodeType(pStmt)) {
     case QUERY_NODE_SET_OPERATOR:
@@ -98,6 +106,8 @@ static int32_t authQuery(SAuthCxt* pCxt, SNode* pStmt) {
       return authDropUser(pCxt, (SDropUserStmt*)pStmt);
     case QUERY_NODE_DELETE_STMT:
       return authDelete(pCxt, (SDeleteStmt*)pStmt);
+    case QUERY_NODE_INSERT_STMT:
+      return authInsert(pCxt, (SInsertStmt*)pStmt);
     default:
       break;
   }

source/libs/parser/src/parCalcConst.c

@@ -300,6 +300,14 @@ static int32_t calcConstDelete(SCalcConstContext* pCxt, SDeleteStmt* pDelete) {
   return code;
 }
 
+static int32_t calcConstInsert(SCalcConstContext* pCxt, SInsertStmt* pInsert) {
+  int32_t code = calcConstFromTable(pCxt, pInsert->pTable);
+  if (TSDB_CODE_SUCCESS == code) {
+    code = calcConstQuery(pCxt, pInsert->pQuery, false);
+  }
+  return code;
+}
+
 static int32_t calcConstQuery(SCalcConstContext* pCxt, SNode* pStmt, bool subquery) {
   int32_t code = TSDB_CODE_SUCCESS;
   switch (nodeType(pStmt)) {
@@ -320,6 +328,9 @@ static int32_t calcConstQuery(SCalcConstContext* pCxt, SNode* pStmt, bool subque
     case QUERY_NODE_DELETE_STMT:
       code = calcConstDelete(pCxt, (SDeleteStmt*)pStmt);
       break;
+    case QUERY_NODE_INSERT_STMT:
+      code = calcConstInsert(pCxt, (SInsertStmt*)pStmt);
+      break;
     default:
       break;
   }

source/libs/parser/src/parInsert.c

@@ -110,7 +110,7 @@ typedef struct SMemParam {
 static int32_t skipInsertInto(char** pSql, SMsgBuf* pMsg) {
   SToken sToken;
   NEXT_TOKEN(*pSql, sToken);
-  if (TK_INSERT != sToken.type) {
+  if (TK_INSERT != sToken.type && TK_IMPORT != sToken.type) {
     return buildSyntaxErrMsg(pMsg, "keyword INSERT is expected", sToken.z);
   }
   NEXT_TOKEN(*pSql, sToken);

source/libs/parser/src/parTranslater.c

@@ -2839,6 +2839,21 @@ static int32_t translateDelete(STranslateContext* pCxt, SDeleteStmt* pDelete) {
   return code;
 }
 
+static int32_t translateInsert(STranslateContext* pCxt, SInsertStmt* pInsert) {
+  pCxt->pCurrStmt = (SNode*)pInsert;
+  int32_t code = translateFrom(pCxt, pInsert->pTable);
+  if (TSDB_CODE_SUCCESS == code) {
+    code = translateExprList(pCxt, pInsert->pCols);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = resetTranslateNamespace(pCxt);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = translateQuery(pCxt, pInsert->pQuery);
+  }
+  return code;
+}
+
 static int64_t getUnitPerMinute(uint8_t precision) {
   switch (precision) {
     case TSDB_TIME_PRECISION_MILLI:
@@ -4608,6 +4623,9 @@ static int32_t translateQuery(STranslateContext* pCxt, SNode* pNode) {
     case QUERY_NODE_DELETE_STMT:
       code = translateDelete(pCxt, (SDeleteStmt*)pNode);
       break;
+    case QUERY_NODE_INSERT_STMT:
+      code = translateInsert(pCxt, (SInsertStmt*)pNode);
+      break;
     case QUERY_NODE_CREATE_DATABASE_STMT:
       code = translateCreateDatabase(pCxt, (SCreateDatabaseStmt*)pNode);
       break;
@@ -6288,6 +6306,10 @@ static int32_t setQuery(STranslateContext* pCxt, SQuery* pQuery) {
       pQuery->execMode = QUERY_EXEC_MODE_SCHEDULE;
       pQuery->msgType = TDMT_VND_DELETE;
       break;
+    case QUERY_NODE_INSERT_STMT:
+      pQuery->execMode = QUERY_EXEC_MODE_SCHEDULE;
+      pQuery->msgType = TDMT_VND_SUBMIT;
+      break;
     case QUERY_NODE_VNODE_MODIF_STMT:
       pQuery->execMode = QUERY_EXEC_MODE_SCHEDULE;
       pQuery->msgType = toMsgType(((SVnodeModifOpStmt*)pQuery->pRoot)->sqlNodeType);

source/libs/parser/src/parser.c

@@ -19,19 +19,28 @@
 #include "parInt.h"
 #include "parToken.h"
 
-bool qIsInsertSql(const char* pStr, size_t length) {
+bool qIsInsertValuesSql(const char* pStr, size_t length) {
   if (NULL == pStr) {
     return false;
   }
 
+  const char* pSql = pStr;
+
   int32_t index = 0;
+  SToken  t = tStrGetToken((char*)pStr, &index, false);
+  if (TK_INSERT != t.type && TK_IMPORT != t.type) {
+    return false;
+  }
 
   do {
-    SToken t0 = tStrGetToken((char*)pStr, &index, false);
-    if (t0.type != TK_NK_LP) {
-      return t0.type == TK_INSERT || t0.type == TK_IMPORT;
+    pStr += index;
+    index = 0;
+    t = tStrGetToken((char*)pStr, &index, false);
+    if (TK_USING == t.type || TK_VALUES == t.type) {
+      return true;
     }
-  } while (1);
+  } while (pStr - pSql < length);
+  return false;
 }
 
 static int32_t analyseSemantic(SParseContext* pCxt, SQuery* pQuery, SParseMetaCache* pMetaCache) {
@@ -148,7 +157,7 @@ static void rewriteExprAlias(SNode* pRoot) {
 
 int32_t qParseSql(SParseContext* pCxt, SQuery** pQuery) {
   int32_t code = TSDB_CODE_SUCCESS;
-  if (qIsInsertSql(pCxt->pSql, pCxt->sqlLen)) {
+  if (qIsInsertValuesSql(pCxt->pSql, pCxt->sqlLen)) {
     code = parseInsertSql(pCxt, pQuery, NULL);
   } else {
     code = parseSqlIntoAst(pCxt, pQuery);
@@ -160,7 +169,7 @@ int32_t qParseSql(SParseContext* pCxt, SQuery** pQuery) {
 int32_t qParseSqlSyntax(SParseContext* pCxt, SQuery** pQuery, struct SCatalogReq* pCatalogReq) {
   SParseMetaCache metaCache = {0};
   int32_t         code = TSDB_CODE_SUCCESS;
-  if (qIsInsertSql(pCxt->pSql, pCxt->sqlLen)) {
+  if (qIsInsertValuesSql(pCxt->pSql, pCxt->sqlLen)) {
     code = parseInsertSyntax(pCxt, pQuery, &metaCache);
   } else {
     code = parseSqlSyntax(pCxt, pQuery, &metaCache);

source/libs/parser/test/parExplainToSyncdbTest.cpp

@@ -40,6 +40,12 @@ TEST_F(ParserExplainToSyncdbTest, grant) {
   run("GRANT READ, WRITE ON test.* TO wxy");
 }
 
+TEST_F(ParserExplainToSyncdbTest, insert) {
+  useDb("root", "test");
+
+  run("INSERT INTO t1 SELECT * FROM t1");
+}
+
 // todo kill connection
 // todo kill query
 // todo kill stream

source/libs/planner/src/planLogicCreater.c

@@ -26,6 +26,7 @@ typedef int32_t (*FCreateLogicNode)(SLogicPlanContext*, void*, SLogicNode**);
 typedef int32_t (*FCreateSelectLogicNode)(SLogicPlanContext*, SSelectStmt*, SLogicNode**);
 typedef int32_t (*FCreateSetOpLogicNode)(SLogicPlanContext*, SSetOperator*, SLogicNode**);
 typedef int32_t (*FCreateDeleteLogicNode)(SLogicPlanContext*, SDeleteStmt*, SLogicNode**);
+typedef int32_t (*FCreateInsertLogicNode)(SLogicPlanContext*, SInsertStmt*, SLogicNode**);
 
 static int32_t doCreateLogicNodeByTable(SLogicPlanContext* pCxt, SSelectStmt* pSelect, SNode* pTable,
                                         SLogicNode** pLogicNode);
@@ -1264,6 +1265,47 @@ static int32_t createDeleteLogicNode(SLogicPlanContext* pCxt, SDeleteStmt* pDele
   return code;
 }
 
+static int32_t creatInsertRootLogicNode(SLogicPlanContext* pCxt, SInsertStmt* pInsert, FCreateInsertLogicNode func,
+                                        SLogicNode** pRoot) {
+  return createRootLogicNode(pCxt, pInsert, pInsert->precision, (FCreateLogicNode)func, pRoot);
+}
+
+static int32_t createVnodeModifLogicNodeByInsert(SLogicPlanContext* pCxt, SInsertStmt* pInsert,
+                                                 SLogicNode** pLogicNode) {
+  SVnodeModifyLogicNode* pModify = (SVnodeModifyLogicNode*)nodesMakeNode(QUERY_NODE_LOGIC_PLAN_VNODE_MODIFY);
+  if (NULL == pModify) {
+    return TSDB_CODE_OUT_OF_MEMORY;
+  }
+
+  SRealTableNode* pRealTable = (SRealTableNode*)pInsert->pTable;
+
+  pModify->modifyType = MODIFY_TABLE_TYPE_INSERT;
+  pModify->tableId = pRealTable->pMeta->uid;
+  pModify->tableType = pRealTable->pMeta->tableType;
+  snprintf(pModify->tableFName, sizeof(pModify->tableFName), "%d.%s.%s", pCxt->pPlanCxt->acctId,
+           pRealTable->table.dbName, pRealTable->table.tableName);
+  TSWAP(pModify->pVgroupList, pRealTable->pVgroupList);
+
+  *pLogicNode = (SLogicNode*)pModify;
+  return TSDB_CODE_SUCCESS;
+}
+
+static int32_t createInsertLogicNode(SLogicPlanContext* pCxt, SInsertStmt* pInsert, SLogicNode** pLogicNode) {
+  SLogicNode* pRoot = NULL;
+  int32_t     code = createQueryLogicNode(pCxt, pInsert->pQuery, &pRoot);
+  if (TSDB_CODE_SUCCESS == code) {
+    code = creatInsertRootLogicNode(pCxt, pInsert, createVnodeModifLogicNodeByInsert, &pRoot);
+  }
+
+  if (TSDB_CODE_SUCCESS == code) {
+    *pLogicNode = pRoot;
+  } else {
+    nodesDestroyNode((SNode*)pRoot);
+  }
+
+  return code;
+}
+
 static int32_t createQueryLogicNode(SLogicPlanContext* pCxt, SNode* pStmt, SLogicNode** pLogicNode) {
   switch (nodeType(pStmt)) {
     case QUERY_NODE_SELECT_STMT:
@@ -1276,6 +1318,8 @@ static int32_t createQueryLogicNode(SLogicPlanContext* pCxt, SNode* pStmt, SLogi
       return createSetOperatorLogicNode(pCxt, (SSetOperator*)pStmt, pLogicNode);
     case QUERY_NODE_DELETE_STMT:
       return createDeleteLogicNode(pCxt, (SDeleteStmt*)pStmt, pLogicNode);
+    case QUERY_NODE_INSERT_STMT:
+      return createInsertLogicNode(pCxt, (SInsertStmt*)pStmt, pLogicNode);
     default:
       break;
   }

source/libs/planner/src/planOptimizer.c

@@ -485,7 +485,7 @@ static int32_t pushDownCondOptPushCondToProject(SOptimizeContext* pCxt, SProject
   return pushDownCondOptAppendCond(&pProject->node.pConditions, pCond);
 }
 
-static int32_t pushDownCondOptPushCondToJoin(SOptimizeContext* pCxt, SJoinLogicNode * pJoin, SNode** pCond) {
+static int32_t pushDownCondOptPushCondToJoin(SOptimizeContext* pCxt, SJoinLogicNode* pJoin, SNode** pCond) {
   return pushDownCondOptAppendCond(&pJoin->node.pConditions, pCond);
 }
 
@@ -557,9 +557,9 @@ static int32_t pushDownCondOptCheckJoinOnCond(SOptimizeContext* pCxt, SJoinLogic
 static int32_t pushDownCondOptPartJoinOnCondLogicCond(SJoinLogicNode* pJoin, SNode** ppMergeCond, SNode** ppOnCond) {
   SLogicConditionNode* pLogicCond = (SLogicConditionNode*)(pJoin->pOnConditions);
 
-  int32_t code = TSDB_CODE_SUCCESS;
+  int32_t    code = TSDB_CODE_SUCCESS;
   SNodeList* pOnConds = NULL;
-  SNode* pCond = NULL;
+  SNode*     pCond = NULL;
   FOREACH(pCond, pLogicCond->pParameterList) {
     if (pushDownCondOptIsPriKeyEqualCond(pJoin, pCond)) {
       *ppMergeCond = nodesCloneNode(pCond);
@@ -604,8 +604,8 @@ static int32_t pushDownCondOptPartJoinOnCond(SJoinLogicNode* pJoin, SNode** ppMe
 
 static int32_t pushDownCondOptJoinExtractMergeCond(SOptimizeContext* pCxt, SJoinLogicNode* pJoin) {
   int32_t code = pushDownCondOptCheckJoinOnCond(pCxt, pJoin);
-  SNode* pJoinMergeCond = NULL;
-  SNode* pJoinOnCond = NULL;
+  SNode*  pJoinMergeCond = NULL;
+  SNode*  pJoinOnCond = NULL;
   if (TSDB_CODE_SUCCESS == code) {
     code = pushDownCondOptPartJoinOnCond(pJoin, &pJoinMergeCond, &pJoinOnCond);
   }
@@ -820,12 +820,12 @@ static int32_t pushDownCondOptDealAgg(SOptimizeContext* pCxt, SAggLogicNode* pAg
 
 typedef struct SRewriteProjCondContext {
   SProjectLogicNode* pProj;
-  int32_t errCode;
-}SRewriteProjCondContext;
+  int32_t            errCode;
+} SRewriteProjCondContext;
 
 static EDealRes rewriteProjectCondForPushDownImpl(SNode** ppNode, void* pContext) {
   SRewriteProjCondContext* pCxt = pContext;
-  SProjectLogicNode* pProj = pCxt->pProj;
+  SProjectLogicNode*       pProj = pCxt->pProj;
   if (QUERY_NODE_COLUMN == nodeType(*ppNode)) {
     SNode* pTarget = NULL;
     FOREACH(pTarget, pProj->node.pTargets) {
@@ -840,18 +840,19 @@ static EDealRes rewriteProjectCondForPushDownImpl(SNode** ppNode, void* pContext
             }
             nodesDestroyNode(*ppNode);
             *ppNode = pExpr;
-          } // end if expr alias name equal column name
-        } // end for each project
-      } // end if target node equals cond column node
-    } // end for each targets
+          }  // end if expr alias name equal column name
+        }    // end for each project
+      }      // end if target node equals cond column node
+    }        // end for each targets
     return DEAL_RES_IGNORE_CHILD;
   }
   return DEAL_RES_CONTINUE;
 }
 
-static int32_t rewriteProjectCondForPushDown(SOptimizeContext* pCxt, SProjectLogicNode* pProject, SNode** ppProjectCond) {
+static int32_t rewriteProjectCondForPushDown(SOptimizeContext* pCxt, SProjectLogicNode* pProject,
+                                             SNode** ppProjectCond) {
   SRewriteProjCondContext cxt = {.pProj = pProject, .errCode = TSDB_CODE_SUCCESS};
-  SNode* pProjectCond = pProject->node.pConditions;
+  SNode*                  pProjectCond = pProject->node.pConditions;
   nodesRewriteExpr(&pProjectCond, rewriteProjectCondForPushDownImpl, &cxt);
   *ppProjectCond = pProjectCond;
   pProject->node.pConditions = NULL;
@@ -873,7 +874,7 @@ static int32_t pushDownCondOptDealProject(SOptimizeContext* pCxt, SProjectLogicN
   }
 
   int32_t code = TSDB_CODE_SUCCESS;
-  SNode* pProjCond = NULL;
+  SNode*  pProjCond = NULL;
   code = rewriteProjectCondForPushDown(pCxt, pProject, &pProjCond);
   if (TSDB_CODE_SUCCESS == code) {
     SLogicNode* pChild = (SLogicNode*)nodesListGetNode(pProject->node.pChildren, 0);
@@ -2082,13 +2083,18 @@ static const int32_t optimizeRuleNum = (sizeof(optimizeRuleSet) / sizeof(SOptimi
 static void dumpLogicSubplan(const char* pRuleName, SLogicSubplan* pSubplan) {
   char* pStr = NULL;
   nodesNodeToString((SNode*)pSubplan, false, &pStr, NULL);
-  qDebugL("apply optimize %s rule: %s", pRuleName, pStr);
+  if (NULL == pRuleName) {
+    qDebugL("before optimize: %s", pStr);
+  } else {
+    qDebugL("apply optimize %s rule: %s", pRuleName, pStr);
+  }
   taosMemoryFree(pStr);
 }
 
 static int32_t applyOptimizeRule(SPlanContext* pCxt, SLogicSubplan* pLogicSubplan) {
   SOptimizeContext cxt = {.pPlanCxt = pCxt, .optimized = false};
   bool             optimized = false;
+  dumpLogicSubplan(NULL, pLogicSubplan);
   do {
     optimized = false;
     for (int32_t i = 0; i < optimizeRuleNum; ++i) {

source/libs/planner/src/planPhysiCreater.c

@@ -632,8 +632,8 @@ static int32_t createJoinPhysiNode(SPhysiPlanContext* pCxt, SNodeList* pChildren
   }
 
   if (TSDB_CODE_SUCCESS == code && NULL != pJoinLogicNode->pOnConditions) {
-    code = setNodeSlotId(pCxt, ((SPhysiNode*)pJoin)->pOutputDataBlockDesc->dataBlockId, -1, pJoinLogicNode->pOnConditions,
-                         &pJoin->pOnConditions);
+    code = setNodeSlotId(pCxt, ((SPhysiNode*)pJoin)->pOutputDataBlockDesc->dataBlockId, -1,
+                         pJoinLogicNode->pOnConditions, &pJoin->pOnConditions);
   }
 
   if (TSDB_CODE_SUCCESS == code) {
@@ -1496,12 +1496,60 @@ static SSubplan* makeSubplan(SPhysiPlanContext* pCxt, SLogicSubplan* pLogicSubpl
   return pSubplan;
 }
 
-static int32_t buildInsertSubplan(SPhysiPlanContext* pCxt, SVnodeModifyLogicNode* pModify, SSubplan* pSubplan) {
+static int32_t buildInsertValuesSubplan(SPhysiPlanContext* pCxt, SVnodeModifyLogicNode* pModify, SSubplan* pSubplan) {
   pSubplan->msgType = pModify->msgType;
   pSubplan->execNode.epSet = pModify->pVgDataBlocks->vg.epSet;
   return createDataInserter(pCxt, pModify->pVgDataBlocks, &pSubplan->pDataSink);
 }
 
+static int32_t createQueryInserter(SPhysiPlanContext* pCxt, SVnodeModifyLogicNode* pModify, SSubplan* pSubplan,
+                                   SDataSinkNode** pSink) {
+  SQueryInserterNode* pInserter = (SQueryInserterNode*)nodesMakeNode(QUERY_NODE_PHYSICAL_PLAN_QUERY_INSERT);
+  if (NULL == pInserter) {
+    return TSDB_CODE_OUT_OF_MEMORY;
+  }
+
+  pInserter->tableId = pModify->tableId;
+  pInserter->tableType = pModify->tableType;
+  strcpy(pInserter->tableFName, pModify->tableFName);
+  pInserter->vgId = pModify->pVgroupList->vgroups[0].vgId;
+  pInserter->epSet = pModify->pVgroupList->vgroups[0].epSet;
+  vgroupInfoToNodeAddr(pModify->pVgroupList->vgroups, &pSubplan->execNode);
+
+  int32_t code = TSDB_CODE_SUCCESS;
+
+  pInserter->sink.pInputDataBlockDesc =
+      (SDataBlockDescNode*)nodesCloneNode((SNode*)pSubplan->pNode->pOutputDataBlockDesc);
+  if (NULL == pInserter->sink.pInputDataBlockDesc) {
+    code = TSDB_CODE_OUT_OF_MEMORY;
+  }
+
+  if (TSDB_CODE_SUCCESS == code) {
+    *pSink = (SDataSinkNode*)pInserter;
+  } else {
+    nodesDestroyNode((SNode*)pInserter);
+  }
+
+  return TSDB_CODE_SUCCESS;
+}
+
+static int32_t buildInsertSelectSubplan(SPhysiPlanContext* pCxt, SVnodeModifyLogicNode* pModify, SSubplan* pSubplan) {
+  int32_t code =
+      createPhysiNode(pCxt, (SLogicNode*)nodesListGetNode(pModify->node.pChildren, 0), pSubplan, &pSubplan->pNode);
+  if (TSDB_CODE_SUCCESS == code) {
+    code = createQueryInserter(pCxt, pModify, pSubplan, &pSubplan->pDataSink);
+  }
+  pSubplan->msgType = TDMT_VND_SUBMIT;
+  return code;
+}
+
+static int32_t buildInsertSubplan(SPhysiPlanContext* pCxt, SVnodeModifyLogicNode* pModify, SSubplan* pSubplan) {
+  if (NULL == pModify->node.pChildren) {
+    return buildInsertValuesSubplan(pCxt, pModify, pSubplan);
+  }
+  return buildInsertSelectSubplan(pCxt, pModify, pSubplan);
+}
+
 static int32_t createDataDeleter(SPhysiPlanContext* pCxt, SVnodeModifyLogicNode* pModify, const SPhysiNode* pRoot,
                                  SDataSinkNode** pSink) {
   SDataDeleterNode* pDeleter = (SDataDeleterNode*)nodesMakeNode(QUERY_NODE_PHYSICAL_PLAN_DELETE);

source/libs/planner/src/planScaleOut.c

@@ -82,29 +82,41 @@ static int32_t scaleOutByVgroups(SScaleOutContext* pCxt, SLogicSubplan* pSubplan
   return code;
 }
 
-static int32_t scaleOutForModify(SScaleOutContext* pCxt, SLogicSubplan* pSubplan, int32_t level, SNodeList* pGroup) {
+static int32_t scaleOutForMerge(SScaleOutContext* pCxt, SLogicSubplan* pSubplan, int32_t level, SNodeList* pGroup) {
+  return nodesListStrictAppend(pGroup, (SNode*)singleCloneSubLogicPlan(pCxt, pSubplan, level));
+}
+
+static int32_t scaleOutForInsertValues(SScaleOutContext* pCxt, SLogicSubplan* pSubplan, int32_t level,
+                                       SNodeList* pGroup) {
   SVnodeModifyLogicNode* pNode = (SVnodeModifyLogicNode*)pSubplan->pNode;
-  if (MODIFY_TABLE_TYPE_DELETE == pNode->modifyType) {
-    return scaleOutByVgroups(pCxt, pSubplan, level, pGroup);
-  } else {
-    size_t numOfVgroups = taosArrayGetSize(pNode->pDataBlocks);
-    for (int32_t i = 0; i < numOfVgroups; ++i) {
-      SLogicSubplan* pNewSubplan = singleCloneSubLogicPlan(pCxt, pSubplan, level);
-      if (NULL == pNewSubplan) {
-        return TSDB_CODE_OUT_OF_MEMORY;
-      }
-      ((SVnodeModifyLogicNode*)pNewSubplan->pNode)->pVgDataBlocks =
-          (SVgDataBlocks*)taosArrayGetP(pNode->pDataBlocks, i);
-      if (TSDB_CODE_SUCCESS != nodesListStrictAppend(pGroup, (SNode*)pNewSubplan)) {
-        return TSDB_CODE_OUT_OF_MEMORY;
-      }
+  size_t                 numOfVgroups = taosArrayGetSize(pNode->pDataBlocks);
+  for (int32_t i = 0; i < numOfVgroups; ++i) {
+    SLogicSubplan* pNewSubplan = singleCloneSubLogicPlan(pCxt, pSubplan, level);
+    if (NULL == pNewSubplan) {
+      return TSDB_CODE_OUT_OF_MEMORY;
+    }
+    ((SVnodeModifyLogicNode*)pNewSubplan->pNode)->pVgDataBlocks = (SVgDataBlocks*)taosArrayGetP(pNode->pDataBlocks, i);
+    if (TSDB_CODE_SUCCESS != nodesListStrictAppend(pGroup, (SNode*)pNewSubplan)) {
+      return TSDB_CODE_OUT_OF_MEMORY;
     }
-    return TSDB_CODE_SUCCESS;
   }
+  return TSDB_CODE_SUCCESS;
 }
 
-static int32_t scaleOutForMerge(SScaleOutContext* pCxt, SLogicSubplan* pSubplan, int32_t level, SNodeList* pGroup) {
-  return nodesListStrictAppend(pGroup, (SNode*)singleCloneSubLogicPlan(pCxt, pSubplan, level));
+static int32_t scaleOutForInsert(SScaleOutContext* pCxt, SLogicSubplan* pSubplan, int32_t level, SNodeList* pGroup) {
+  SVnodeModifyLogicNode* pNode = (SVnodeModifyLogicNode*)pSubplan->pNode;
+  if (NULL == pNode->node.pChildren) {
+    return scaleOutForInsertValues(pCxt, pSubplan, level, pGroup);
+  }
+  return scaleOutForMerge(pCxt, pSubplan, level, pGroup);
+}
+
+static int32_t scaleOutForModify(SScaleOutContext* pCxt, SLogicSubplan* pSubplan, int32_t level, SNodeList* pGroup) {
+  SVnodeModifyLogicNode* pNode = (SVnodeModifyLogicNode*)pSubplan->pNode;
+  if (MODIFY_TABLE_TYPE_DELETE == pNode->modifyType) {
+    return scaleOutByVgroups(pCxt, pSubplan, level, pGroup);
+  }
+  return scaleOutForInsert(pCxt, pSubplan, level, pGroup);
 }
 
 static int32_t scaleOutForScan(SScaleOutContext* pCxt, SLogicSubplan* pSubplan, int32_t level, SNodeList* pGroup) {

source/libs/planner/src/planSpliter.c

@@ -20,6 +20,7 @@
 #define SPLIT_FLAG_MASK(n) (1 << n)
 
 #define SPLIT_FLAG_STABLE_SPLIT SPLIT_FLAG_MASK(0)
+#define SPLIT_FLAG_INSERT_SPLIT SPLIT_FLAG_MASK(1)
 
 #define SPLIT_FLAG_SET_MASK(val, mask)  (val) |= (mask)
 #define SPLIT_FLAG_TEST_MASK(val, mask) (((val) & (mask)) != 0)
@@ -1196,6 +1197,41 @@ static int32_t smaIndexSplit(SSplitContext* pCxt, SLogicSubplan* pSubplan) {
   return code;
 }
 
+typedef struct SInsertSelectSplitInfo {
+  SLogicNode*    pQueryRoot;
+  SLogicSubplan* pSubplan;
+} SInsertSelectSplitInfo;
+
+static bool insSelSplFindSplitNode(SSplitContext* pCxt, SLogicSubplan* pSubplan, SLogicNode* pNode,
+                                   SInsertSelectSplitInfo* pInfo) {
+  if (QUERY_NODE_LOGIC_PLAN_VNODE_MODIFY == nodeType(pNode) && 1 == LIST_LENGTH(pNode->pChildren) &&
+      MODIFY_TABLE_TYPE_INSERT == ((SVnodeModifyLogicNode*)pNode)->modifyType) {
+    pInfo->pQueryRoot = (SLogicNode*)nodesListGetNode(pNode->pChildren, 0);
+    pInfo->pSubplan = pSubplan;
+    return true;
+  }
+  return false;
+}
+
+static int32_t insertSelectSplit(SSplitContext* pCxt, SLogicSubplan* pSubplan) {
+  SInsertSelectSplitInfo info = {0};
+  if (!splMatch(pCxt, pSubplan, SPLIT_FLAG_INSERT_SPLIT, (FSplFindSplitNode)insSelSplFindSplitNode, &info)) {
+    return TSDB_CODE_SUCCESS;
+  }
+
+  int32_t code = splCreateExchangeNodeForSubplan(pCxt, info.pSubplan, info.pQueryRoot, info.pSubplan->subplanType);
+  if (TSDB_CODE_SUCCESS == code) {
+    code = nodesListMakeStrictAppend(&info.pSubplan->pChildren, (SNode*)splCreateScanSubplan(pCxt, info.pQueryRoot, 0));
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    info.pSubplan->subplanType = SUBPLAN_TYPE_MODIFY;
+    SPLIT_FLAG_SET_MASK(info.pSubplan->splitFlag, SPLIT_FLAG_INSERT_SPLIT);
+  }
+  ++(pCxt->groupId);
+  pCxt->split = true;
+  return code;
+}
+
 typedef struct SQnodeSplitInfo {
   SLogicNode*    pSplitNode;
   SLogicSubplan* pSubplan;
@@ -1249,7 +1285,8 @@ static const SSplitRule splitRuleSet[] = {
   {.pName = "SingleTableJoinSplit", .splitFunc = singleTableJoinSplit},
   {.pName = "UnionAllSplit",        .splitFunc = unionAllSplit},
   {.pName = "UnionDistinctSplit",   .splitFunc = unionDistinctSplit},
-  {.pName = "SmaIndexSplit",        .splitFunc = smaIndexSplit}
+  {.pName = "SmaIndexSplit",        .splitFunc = smaIndexSplit},
+  {.pName = "InsertSelectSplit",    .splitFunc = insertSelectSplit}
 };
 // clang-format on
 
@@ -1258,7 +1295,11 @@ static const int32_t splitRuleNum = (sizeof(splitRuleSet) / sizeof(SSplitRule));
 static void dumpLogicSubplan(const char* pRuleName, SLogicSubplan* pSubplan) {
   char* pStr = NULL;
   nodesNodeToString((SNode*)pSubplan, false, &pStr, NULL);
-  qDebugL("apply split %s rule: %s", pRuleName, pStr);
+  if (NULL == pRuleName) {
+    qDebugL("before split: %s", pStr);
+  } else {
+    qDebugL("apply split %s rule: %s", pRuleName, pStr);
+  }
   taosMemoryFree(pStr);
 }
 
@@ -1266,6 +1307,7 @@ static int32_t applySplitRule(SPlanContext* pCxt, SLogicSubplan* pSubplan) {
   SSplitContext cxt = {
       .pPlanCxt = pCxt, .queryId = pSubplan->id.queryId, .groupId = pSubplan->id.groupId + 1, .split = false};
   bool split = false;
+  dumpLogicSubplan(NULL, pSubplan);
   do {
     split = false;
     for (int32_t i = 0; i < splitRuleNum; ++i) {
@@ -1293,8 +1335,16 @@ static void setVgroupsInfo(SLogicNode* pNode, SLogicSubplan* pSubplan) {
   FOREACH(pChild, pNode->pChildren) { setVgroupsInfo((SLogicNode*)pChild, pSubplan); }
 }
 
+static bool needSplitSubplan(SLogicSubplan* pLogicSubplan) {
+  if (QUERY_NODE_LOGIC_PLAN_VNODE_MODIFY != nodeType(pLogicSubplan->pNode)) {
+    return true;
+  }
+  SVnodeModifyLogicNode* pModify = (SVnodeModifyLogicNode*)pLogicSubplan->pNode;
+  return (MODIFY_TABLE_TYPE_INSERT == pModify->modifyType && NULL != pModify->node.pChildren);
+}
+
 int32_t splitLogicPlan(SPlanContext* pCxt, SLogicSubplan* pLogicSubplan) {
-  if (QUERY_NODE_LOGIC_PLAN_VNODE_MODIFY == nodeType(pLogicSubplan->pNode)) {
+  if (!needSplitSubplan(pLogicSubplan)) {
     setVgroupsInfo(pLogicSubplan->pNode, pLogicSubplan);
     return TSDB_CODE_SUCCESS;
   }

source/libs/planner/test/planOtherTest.cpp

@@ -91,3 +91,9 @@ TEST_F(PlanOtherTest, delete) {
 
   run("DELETE FROM st1 WHERE ts > now - 2d and ts < now - 1d AND tag1 = 10");
 }
+
+TEST_F(PlanOtherTest, insert) {
+  useDb("root", "test");
+
+  run("INSERT INTO t1 SELECT * FROM t1");
+}

source/libs/qcom/src/queryUtil.c

@@ -171,17 +171,17 @@ char* jobTaskStatusStr(int32_t status) {
   switch (status) {
     case JOB_TASK_STATUS_NULL:
       return "NULL";
-    case JOB_TASK_STATUS_NOT_START:
-      return "NOT_START";
-    case JOB_TASK_STATUS_EXECUTING:
+    case JOB_TASK_STATUS_INIT:
+      return "INIT";
+    case JOB_TASK_STATUS_EXEC:
       return "EXECUTING";
-    case JOB_TASK_STATUS_PARTIAL_SUCCEED:
+    case JOB_TASK_STATUS_PART_SUCC:
       return "PARTIAL_SUCCEED";
-    case JOB_TASK_STATUS_SUCCEED:
+    case JOB_TASK_STATUS_SUCC:
       return "SUCCEED";
-    case JOB_TASK_STATUS_FAILED:
+    case JOB_TASK_STATUS_FAIL:
       return "FAILED";
-    case JOB_TASK_STATUS_DROPPING:
+    case JOB_TASK_STATUS_DROP:
       return "DROPPING";
     default:
       break;
@@ -200,7 +200,7 @@ SSchema createSchema(int8_t type, int32_t bytes, col_id_t colId, const char* nam
   return s;
 }
 
-void destroyQueryExecRes(SQueryExecRes* pRes) {
+void destroyQueryExecRes(SExecResult* pRes) {
   if (NULL == pRes || NULL == pRes->res) {
     return;
   }

source/libs/qworker/inc/qwInt.h

@@ -226,8 +226,8 @@ typedef struct SQWorkerMgmt {
 #define QW_TASK_NOT_EXIST(code)     (TSDB_CODE_QRY_SCH_NOT_EXIST == (code) || TSDB_CODE_QRY_TASK_NOT_EXIST == (code))
 #define QW_TASK_ALREADY_EXIST(code) (TSDB_CODE_QRY_TASK_ALREADY_EXIST == (code))
 #define QW_TASK_READY(status)                                                                                      \
-  (status == JOB_TASK_STATUS_SUCCEED || status == JOB_TASK_STATUS_FAILED || status == JOB_TASK_STATUS_CANCELLED || \
-   status == JOB_TASK_STATUS_PARTIAL_SUCCEED)
+  (status == JOB_TASK_STATUS_SUCC || status == JOB_TASK_STATUS_FAIL || status == JOB_TASK_STATUS_CANCELLED || \
+   status == JOB_TASK_STATUS_PART_SUCC)
 #define QW_SET_QTID(id, qId, tId, eId)                              \
   do {                                                              \
     *(uint64_t *)(id) = (qId);                                      \

source/libs/qworker/src/qwDbg.c

@@ -19,7 +19,7 @@ int32_t qwDbgValidateStatus(QW_FPARAMS_DEF, int8_t oriStatus, int8_t newStatus,
   int32_t code = 0;
 
   if (oriStatus == newStatus) {
-    if (newStatus == JOB_TASK_STATUS_EXECUTING || newStatus == JOB_TASK_STATUS_FAILED) {
+    if (newStatus == JOB_TASK_STATUS_EXEC || newStatus == JOB_TASK_STATUS_FAIL) {
       *ignore = true;
       return TSDB_CODE_SUCCESS;
     }
@@ -29,47 +29,47 @@ int32_t qwDbgValidateStatus(QW_FPARAMS_DEF, int8_t oriStatus, int8_t newStatus,
 
   switch (oriStatus) {
     case JOB_TASK_STATUS_NULL:
-      if (newStatus != JOB_TASK_STATUS_EXECUTING && newStatus != JOB_TASK_STATUS_FAILED &&
-          newStatus != JOB_TASK_STATUS_NOT_START) {
+      if (newStatus != JOB_TASK_STATUS_EXEC && newStatus != JOB_TASK_STATUS_FAIL &&
+          newStatus != JOB_TASK_STATUS_INIT) {
         QW_ERR_JRET(TSDB_CODE_QRY_APP_ERROR);
       }
 
       break;
-    case JOB_TASK_STATUS_NOT_START:
-      if (newStatus != JOB_TASK_STATUS_DROPPING && newStatus != JOB_TASK_STATUS_EXECUTING
-        && newStatus != JOB_TASK_STATUS_FAILED) {
+    case JOB_TASK_STATUS_INIT:
+      if (newStatus != JOB_TASK_STATUS_DROP && newStatus != JOB_TASK_STATUS_EXEC
+        && newStatus != JOB_TASK_STATUS_FAIL) {
         QW_ERR_JRET(TSDB_CODE_QRY_APP_ERROR);
       }
 
       break;
-    case JOB_TASK_STATUS_EXECUTING:
-      if (newStatus != JOB_TASK_STATUS_PARTIAL_SUCCEED && newStatus != JOB_TASK_STATUS_SUCCEED &&
-          newStatus != JOB_TASK_STATUS_FAILED && newStatus != JOB_TASK_STATUS_DROPPING) {
+    case JOB_TASK_STATUS_EXEC:
+      if (newStatus != JOB_TASK_STATUS_PART_SUCC && newStatus != JOB_TASK_STATUS_SUCC &&
+          newStatus != JOB_TASK_STATUS_FAIL && newStatus != JOB_TASK_STATUS_DROP) {
         QW_ERR_JRET(TSDB_CODE_QRY_APP_ERROR);
       }
 
       break;
-    case JOB_TASK_STATUS_PARTIAL_SUCCEED:
-      if (newStatus != JOB_TASK_STATUS_EXECUTING && newStatus != JOB_TASK_STATUS_SUCCEED &&
-          newStatus != JOB_TASK_STATUS_FAILED && newStatus != JOB_TASK_STATUS_DROPPING) {
+    case JOB_TASK_STATUS_PART_SUCC:
+      if (newStatus != JOB_TASK_STATUS_EXEC && newStatus != JOB_TASK_STATUS_SUCC &&
+          newStatus != JOB_TASK_STATUS_FAIL && newStatus != JOB_TASK_STATUS_DROP) {
         QW_ERR_JRET(TSDB_CODE_QRY_APP_ERROR);
       }
 
       break;
-    case JOB_TASK_STATUS_SUCCEED:
-      if (newStatus != JOB_TASK_STATUS_DROPPING && newStatus != JOB_TASK_STATUS_FAILED) {
+    case JOB_TASK_STATUS_SUCC:
+      if (newStatus != JOB_TASK_STATUS_DROP && newStatus != JOB_TASK_STATUS_FAIL) {
         QW_ERR_JRET(TSDB_CODE_QRY_APP_ERROR);
       }
 
       break;
-    case JOB_TASK_STATUS_FAILED:
-      if (newStatus != JOB_TASK_STATUS_DROPPING) {
+    case JOB_TASK_STATUS_FAIL:
+      if (newStatus != JOB_TASK_STATUS_DROP) {
         QW_ERR_JRET(TSDB_CODE_QRY_APP_ERROR);
       }
       break;
 
-    case JOB_TASK_STATUS_DROPPING:
-      if (newStatus != JOB_TASK_STATUS_FAILED && newStatus != JOB_TASK_STATUS_PARTIAL_SUCCEED) {
+    case JOB_TASK_STATUS_DROP:
+      if (newStatus != JOB_TASK_STATUS_FAIL && newStatus != JOB_TASK_STATUS_PART_SUCC) {
         QW_ERR_JRET(TSDB_CODE_QRY_APP_ERROR);
       }
       break;

source/libs/qworker/src/qwUtil.c

@@ -433,8 +433,8 @@ void qwSetHbParam(int64_t refId, SQWHbParam **pParam) {
 }
 
 void qwSaveTbVersionInfo(qTaskInfo_t pTaskInfo, SQWTaskCtx *ctx) {
-  char dbFName[TSDB_DB_FNAME_LEN];
-  char tbName[TSDB_TABLE_NAME_LEN];
+  char dbFName[TSDB_DB_FNAME_LEN] = {0};
+  char tbName[TSDB_TABLE_NAME_LEN] = {0};
 
   qGetQueriedTableSchemaVersion(pTaskInfo, dbFName, tbName, &ctx->tbInfo.sversion, &ctx->tbInfo.tversion);
 

source/libs/qworker/src/qworker.c

@@ -206,7 +206,7 @@ int32_t qwGetQueryResFromSink(QW_FPARAMS_DEF, SQWTaskCtx *ctx, int32_t *dataLen,
 
       QW_TASK_DLOG_E("no data in sink and query end");
 
-      qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_SUCCEED);
+      qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_SUCC);
       QW_ERR_RET(qwMallocFetchRsp(len, &rsp));
 
       *rspMsg = rsp;
@@ -236,7 +236,7 @@ int32_t qwGetQueryResFromSink(QW_FPARAMS_DEF, SQWTaskCtx *ctx, int32_t *dataLen,
 
   if (DS_BUF_EMPTY == pOutput->bufStatus && pOutput->queryEnd) {
     QW_TASK_DLOG_E("task all data fetched, done");
-    qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_SUCCEED);
+    qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_SUCC);
   }
 
   return TSDB_CODE_SUCCESS;
@@ -319,7 +319,7 @@ int32_t qwHandlePrePhaseEvents(QW_FPARAMS_DEF, int8_t phase, SQWPhaseInput *inpu
         break;
       }
 
-      QW_ERR_JRET(qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_EXECUTING));
+      QW_ERR_JRET(qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_EXEC));
       break;
     }
     case QW_PHASE_PRE_FETCH: {
@@ -436,7 +436,7 @@ int32_t qwHandlePostPhaseEvents(QW_FPARAMS_DEF, int8_t phase, SQWPhaseInput *inp
 _return:
 
   if (TSDB_CODE_SUCCESS == code && QW_PHASE_POST_QUERY == phase) {
-    qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_PARTIAL_SUCCEED);
+    qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_PART_SUCC);
   }
 
   if (rspConnection) {
@@ -456,7 +456,7 @@ _return:
   }
 
   if (code) {
-    qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_FAILED);
+    qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_FAIL);
   }
 
   QW_TASK_DLOG("end to handle event at phase %s, code:%x - %s", qwPhaseStr(phase), code, tstrerror(code));
@@ -488,7 +488,7 @@ int32_t qwPrerocessQuery(QW_FPARAMS_DEF, SQWMsg *qwMsg) {
 
   ctx->ctrlConnInfo = qwMsg->connInfo;
 
-  QW_ERR_JRET(qwAddTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_NOT_START));
+  QW_ERR_JRET(qwAddTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_INIT));
 
 _return:
 
@@ -687,7 +687,7 @@ int32_t qwProcessFetch(QW_FPARAMS_DEF, SQWMsg *qwMsg) {
     if (QW_IS_QUERY_RUNNING(ctx)) {
       atomic_store_8((int8_t *)&ctx->queryContinue, 1);
     } else if (0 == atomic_load_8((int8_t *)&ctx->queryInQueue)) {
-      qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_EXECUTING);
+      qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_EXEC);
 
       atomic_store_8((int8_t *)&ctx->queryInQueue, 1);
 
@@ -738,7 +738,7 @@ int32_t qwProcessDrop(QW_FPARAMS_DEF, SQWMsg *qwMsg) {
 
   if (QW_IS_QUERY_RUNNING(ctx)) {
     QW_ERR_JRET(qwKillTaskHandle(QW_FPARAMS(), ctx));
-    qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_DROPPING);
+    qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_DROP);
   } else if (ctx->phase > 0) {
     QW_ERR_JRET(qwDropTask(QW_FPARAMS()));
     rsped = true;
@@ -759,7 +759,7 @@ _return:
       QW_UPDATE_RSP_CODE(ctx, code);
     }
 
-    qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_FAILED);
+    qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_FAIL);
   }
 
   if (locked) {

source/libs/scheduler/inc/schedulerInt.h → source/libs/scheduler/inc/schInt.h

@@ -52,6 +52,7 @@ typedef enum {
   SCH_OP_NULL = 0,
   SCH_OP_EXEC,
   SCH_OP_FETCH,
+  SCH_OP_GET_STATUS,
 } SCH_OP_TYPE;
 
 typedef struct SSchTrans {
@@ -97,13 +98,30 @@ typedef struct SSchStat {
 } SSchStat;
 
 typedef struct SSchResInfo {
-  SQueryResult*          queryRes;
+  SExecResult*           execRes;
   void**                 fetchRes;
   schedulerExecFp        execFp; 
   schedulerFetchFp       fetchFp; 
-  void*                  userParam;
+  void*                  cbParam;
 } SSchResInfo;
 
+typedef struct SSchOpEvent {
+  SCH_OP_TYPE    type;
+  bool           begin;
+  SSchedulerReq *pReq;
+} SSchOpEvent;
+
+typedef int32_t (*schStatusEnterFp)(void* pHandle, void* pParam);
+typedef int32_t (*schStatusLeaveFp)(void* pHandle, void* pParam);
+typedef int32_t (*schStatusEventFp)(void* pHandle, void* pParam, void* pEvent);
+
+typedef struct SSchStatusFps {
+  EJobTaskType     status;
+  schStatusEnterFp enterFp;
+  schStatusLeaveFp leaveFp;
+  schStatusEventFp eventFp;
+} SSchStatusFps;
+
 typedef struct SSchedulerMgmt {
   uint64_t        taskId; // sequential taksId
   uint64_t        sId;    // schedulerId
@@ -157,7 +175,7 @@ typedef struct SSchLevel {
   int32_t         taskNum;
   int32_t         taskLaunchedNum;
   int32_t         taskDoneNum;
-  SArray         *subTasks;      // Element is SQueryTask
+  SArray         *subTasks;      // Element is SSchTask
 } SSchLevel;
 
 typedef struct SSchTaskProfile {
@@ -195,12 +213,13 @@ typedef struct SSchTask {
 typedef struct SSchJobAttr {
   EExplainMode explainMode;
   bool         queryJob;
+  bool         needFetch;
   bool         needFlowCtrl;
 } SSchJobAttr;
 
 typedef struct {
   int32_t     op;
-  bool        sync;
+  bool        syncReq;
 } SSchOpStatus;
 
 typedef struct SSchJob {
@@ -231,7 +250,7 @@ typedef struct SSchJob {
   SSchTask          *fetchTask;
   int32_t            errCode;
   SRWLatch           resLock;
-  SQueryExecRes      execRes;
+  SExecResult        execRes;
   void              *resData;         //TODO free it or not
   int32_t            resNumOfRows;
   SSchResInfo        userRes;
@@ -292,19 +311,19 @@ extern SSchedulerMgmt schMgmt;
 #define SCH_GET_JOB_STATUS(job) atomic_load_8(&(job)->status)
 #define SCH_GET_JOB_STATUS_STR(job) jobTaskStatusStr(SCH_GET_JOB_STATUS(job))
 
-#define SCH_JOB_IN_SYNC_OP(job) ((job)->opStatus.op && (job)->opStatus.sync)
-#define SCH_JOB_IN_ASYNC_EXEC_OP(job) (((job)->opStatus.op == SCH_OP_EXEC) && (!(job)->opStatus.sync))
-#define SCH_JOB_IN_ASYNC_FETCH_OP(job) (((job)->opStatus.op == SCH_OP_FETCH) && (!(job)->opStatus.sync))
+#define SCH_JOB_IN_SYNC_OP(job) ((job)->opStatus.op && (job)->opStatus.syncReq)
+#define SCH_JOB_IN_ASYNC_EXEC_OP(job) ((SCH_OP_EXEC == atomic_val_compare_exchange_32(&(job)->opStatus.op, SCH_OP_EXEC, SCH_OP_NULL)) && (!(job)->opStatus.syncReq))
+#define SCH_JOB_IN_ASYNC_FETCH_OP(job) ((SCH_OP_FETCH == atomic_val_compare_exchange_32(&(job)->opStatus.op, SCH_OP_FETCH, SCH_OP_NULL)) && (!(job)->opStatus.syncReq))
 
 #define SCH_SET_JOB_NEED_FLOW_CTRL(_job) (_job)->attr.needFlowCtrl = true
 #define SCH_JOB_NEED_FLOW_CTRL(_job) ((_job)->attr.needFlowCtrl)
 #define SCH_TASK_NEED_FLOW_CTRL(_job, _task) (SCH_IS_DATA_SRC_QRY_TASK(_task) && SCH_JOB_NEED_FLOW_CTRL(_job) && SCH_IS_LEVEL_UNFINISHED((_task)->level))
 
-#define SCH_SET_JOB_TYPE(_job, type) (_job)->attr.queryJob = ((type) != SUBPLAN_TYPE_MODIFY)
+#define SCH_SET_JOB_TYPE(_job, type) do { if ((type) != SUBPLAN_TYPE_MODIFY) { (_job)->attr.queryJob = true; } } while (0)
 #define SCH_IS_QUERY_JOB(_job) ((_job)->attr.queryJob) 
-#define SCH_JOB_NEED_FETCH(_job) SCH_IS_QUERY_JOB(_job)
-#define SCH_IS_WAIT_ALL_JOB(_job) (!SCH_IS_QUERY_JOB(_job))
-#define SCH_IS_NEED_DROP_JOB(_job) (SCH_IS_QUERY_JOB(_job))
+#define SCH_JOB_NEED_FETCH(_job) ((_job)->attr.needFetch)
+#define SCH_JOB_NEED_WAIT(_job) (!SCH_IS_QUERY_JOB(_job))
+#define SCH_JOB_NEED_DROP(_job) (SCH_IS_QUERY_JOB(_job))
 #define SCH_IS_EXPLAIN_JOB(_job) (EXPLAIN_MODE_ANALYZE == (_job)->attr.explainMode)
 #define SCH_NETWORK_ERR(_code) ((_code) == TSDB_CODE_RPC_BROKEN_LINK || (_code) == TSDB_CODE_RPC_NETWORK_UNAVAIL)
 #define SCH_SUB_TASK_NETWORK_ERR(_code, _len) (SCH_NETWORK_ERR(_code) && ((_len) > 0))
@@ -329,9 +348,10 @@ extern SSchedulerMgmt schMgmt;
 #define SCH_TASK_WLOG(param, ...) \
   qWarn("QID:0x%" PRIx64 ",TID:0x%" PRIx64 ",EID:%d " param, pJob->queryId, SCH_TASK_ID(pTask), SCH_TASK_EID(pTask),__VA_ARGS__)
 
-#define SCH_ERR_RET(c) do { int32_t _code = c; if (_code != TSDB_CODE_SUCCESS) { terrno = _code; return _code; } } while (0)
-#define SCH_RET(c) do { int32_t _code = c; if (_code != TSDB_CODE_SUCCESS) { terrno = _code; } return _code; } while (0)
-#define SCH_ERR_JRET(c) do { code = c; if (code != TSDB_CODE_SUCCESS) { terrno = code; goto _return; } } while (0)
+#define SCH_SET_ERRNO(_err) do { if (TSDB_CODE_SCH_IGNORE_ERROR != (_err)) { terrno = (_err); } } while (0)
+#define SCH_ERR_RET(c) do { int32_t _code = c; if (_code != TSDB_CODE_SUCCESS) { SCH_SET_ERRNO(_code); return _code; } } while (0)
+#define SCH_RET(c) do { int32_t _code = c; if (_code != TSDB_CODE_SUCCESS) { SCH_SET_ERRNO(_code); } return _code; } while (0)
+#define SCH_ERR_JRET(c) do { code = c; if (code != TSDB_CODE_SUCCESS) { SCH_SET_ERRNO(code); goto _return; } } while (0)
 
 #define SCH_LOCK(type, _lock) (SCH_READ == (type) ? taosRLockLatch(_lock) : taosWLockLatch(_lock))
 #define SCH_UNLOCK(type, _lock) (SCH_READ == (type) ? taosRUnLockLatch(_lock) : taosWUnLockLatch(_lock))
@@ -349,7 +369,7 @@ int32_t schDecTaskFlowQuota(SSchJob *pJob, SSchTask *pTask);
 int32_t schCheckIncTaskFlowQuota(SSchJob *pJob, SSchTask *pTask, bool *enough);
 int32_t schLaunchTasksInFlowCtrlList(SSchJob *pJob, SSchTask *pTask);
 int32_t schLaunchTaskImpl(SSchJob *pJob, SSchTask *pTask);
-int32_t schFetchFromRemote(SSchJob *pJob);
+int32_t schLaunchFetchTask(SSchJob *pJob);
 int32_t schProcessOnTaskFailure(SSchJob *pJob, SSchTask *pTask, int32_t errCode);
 int32_t schBuildAndSendHbMsg(SQueryNodeEpId *nodeEpId, SArray* taskAction);
 int32_t schCloneSMsgSendInfo(void *src, void **dst);
@@ -371,25 +391,45 @@ void    schFreeRpcCtxVal(const void *arg);
 int32_t schMakeBrokenLinkVal(SSchJob *pJob, SSchTask *pTask, SRpcBrokenlinkVal *brokenVal, bool isHb);
 int32_t schAppendTaskExecNode(SSchJob *pJob, SSchTask *pTask, SQueryNodeAddr *addr, int32_t execId);
 int32_t schExecStaticExplainJob(SSchedulerReq *pReq, int64_t *job, bool sync);
-int32_t schExecJobImpl(SSchedulerReq *pReq, SSchJob *pJob, bool sync);
 int32_t schUpdateJobStatus(SSchJob *pJob, int8_t newStatus);
 int32_t schCancelJob(SSchJob *pJob);
 int32_t schProcessOnJobDropped(SSchJob *pJob, int32_t errCode);
 uint64_t schGenTaskId(void);
 void    schCloseJobRef(void);
-int32_t schExecJob(SSchedulerReq *pReq, int64_t *pJob, SQueryResult *pRes);
 int32_t schAsyncExecJob(SSchedulerReq *pReq, int64_t *pJob);
-int32_t schFetchRows(SSchJob *pJob);
-int32_t schAsyncFetchRows(SSchJob *pJob);
+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(SSchedulerReq *pReq, SSchJob **pSchJob);
-int32_t schSetJobQueryRes(SSchJob* pJob, SQueryResult* pRes);
+int32_t schInitJob(int64_t *pJobId, SSchedulerReq *pReq);
+int32_t schExecJob(SSchJob *pJob, SSchedulerReq *pReq);
+int32_t schDumpJobExecRes(SSchJob* pJob, SExecResult* pRes);
 int32_t schUpdateTaskCandidateAddr(SSchJob *pJob, SSchTask *pTask, SEpSet* pEpSet);
 int32_t schHandleRedirect(SSchJob *pJob, SSchTask *pTask, SDataBuf* pData, int32_t rspCode);
+void    schProcessOnOpEnd(SSchJob *pJob, SCH_OP_TYPE type, SSchedulerReq* pReq, int32_t errCode);
+int32_t schProcessOnOpBegin(SSchJob* pJob, SCH_OP_TYPE type, SSchedulerReq* pReq);
+void    schProcessOnCbEnd(SSchJob *pJob, SSchTask *pTask, int32_t errCode);
+int32_t schProcessOnCbBegin(SSchJob** job, SSchTask** task, uint64_t qId, int64_t rId, uint64_t tId);
+void    schDropTaskOnExecNode(SSchJob *pJob, SSchTask *pTask);
+bool    schJobDone(SSchJob *pJob);
+int32_t schRemoveTaskFromExecList(SSchJob *pJob, SSchTask *pTask);
+int32_t schLaunchJobLowerLevel(SSchJob *pJob, SSchTask *pTask);
+int32_t schSwitchJobStatus(SSchJob* pJob, int32_t status, void* param);
+int32_t schHandleOpBeginEvent(int64_t jobId, SSchJob** job, SCH_OP_TYPE type, SSchedulerReq* pReq);
+int32_t schHandleOpEndEvent(SSchJob* pJob, SCH_OP_TYPE type, SSchedulerReq* pReq, int32_t errCode);
+int32_t schHandleTaskRetry(SSchJob *pJob, SSchTask *pTask);
+void    schUpdateJobErrCode(SSchJob *pJob, int32_t errCode);
+int32_t schTaskCheckSetRetry(SSchJob *pJob, SSchTask *pTask, int32_t errCode, bool *needRetry);
+int32_t schProcessOnJobFailure(SSchJob *pJob, int32_t errCode);
+int32_t schProcessOnJobPartialSuccess(SSchJob *pJob);
+void    schFreeTask(SSchJob *pJob, SSchTask *pTask);
+void    schDropTaskInHashList(SSchJob *pJob, SHashObj *list);
+int32_t schLaunchLevelTasks(SSchJob *pJob, SSchLevel *level);
+int32_t schGetTaskFromList(SHashObj *pTaskList, uint64_t taskId, SSchTask **pTask);
+int32_t schInitTask(SSchJob *pJob, SSchTask *pTask, SSubplan *pPlan, SSchLevel *pLevel);
 
 
 #ifdef __cplusplus

source/libs/scheduler/src/schDbg.c

@@ -14,16 +14,16 @@
  */
 
 #include "query.h"
-#include "schedulerInt.h"
+#include "schInt.h"
 
 tsem_t schdRspSem;
 
-void schdExecCallback(SQueryResult* pResult, void* param, int32_t code) {
+void schdExecCallback(SExecResult* pResult, void* param, int32_t code) {
   if (code) {
     pResult->code = code;
   }
   
-  *(SQueryResult*)param = *pResult;
+  *(SExecResult*)param = *pResult;
 
   taosMemoryFree(pResult);
 

source/libs/scheduler/src/schFlowCtrl.c

@@ -13,7 +13,7 @@
  * along with this program. If not, see <http://www.gnu.org/licenses/>.
  */
 
-#include "schedulerInt.h"
+#include "schInt.h"
 #include "tmsg.h"
 #include "query.h"
 #include "catalog.h"

source/libs/scheduler/src/schRemote.c

@@ -16,7 +16,7 @@
 #include "catalog.h"
 #include "command.h"
 #include "query.h"
-#include "schedulerInt.h"
+#include "schInt.h"
 #include "tmsg.h"
 #include "tref.h"
 #include "trpc.h"
@@ -36,7 +36,7 @@ int32_t schValidateReceivedMsgType(SSchJob *pJob, SSchTask *pTask, int32_t msgTy
                       TMSG_INFO(msgType));
       }
 
-      if (taskStatus != JOB_TASK_STATUS_EXECUTING && taskStatus != JOB_TASK_STATUS_PARTIAL_SUCCEED) {
+      if (taskStatus != JOB_TASK_STATUS_EXEC && taskStatus != JOB_TASK_STATUS_PART_SUCC) {
         SCH_TASK_DLOG("rsp msg conflicted with task status, status:%s, rspType:%s", jobTaskStatusStr(taskStatus),
                       TMSG_INFO(msgType));
       }
@@ -50,7 +50,7 @@ int32_t schValidateReceivedMsgType(SSchJob *pJob, SSchTask *pTask, int32_t msgTy
         SCH_ERR_RET(TSDB_CODE_SCH_STATUS_ERROR);
       }
 
-      if (taskStatus != JOB_TASK_STATUS_EXECUTING && taskStatus != JOB_TASK_STATUS_PARTIAL_SUCCEED) {
+      if (taskStatus != JOB_TASK_STATUS_EXEC && taskStatus != JOB_TASK_STATUS_PART_SUCC) {
         SCH_TASK_ELOG("rsp msg conflicted with task status, status:%s, rspType:%s", jobTaskStatusStr(taskStatus),
                       TMSG_INFO(msgType));
         SCH_ERR_RET(TSDB_CODE_SCH_STATUS_ERROR);
@@ -76,7 +76,7 @@ int32_t schValidateReceivedMsgType(SSchJob *pJob, SSchTask *pTask, int32_t msgTy
     SCH_ERR_RET(TSDB_CODE_SCH_STATUS_ERROR);
   }
 
-  if (taskStatus != JOB_TASK_STATUS_EXECUTING && taskStatus != JOB_TASK_STATUS_PARTIAL_SUCCEED) {
+  if (taskStatus != JOB_TASK_STATUS_EXEC && taskStatus != JOB_TASK_STATUS_PART_SUCC) {
     SCH_TASK_ELOG("rsp msg conflicted with task status, status:%s, rspType:%s", jobTaskStatusStr(taskStatus),
                   TMSG_INFO(msgType));
     SCH_ERR_RET(TSDB_CODE_SCH_STATUS_ERROR);
@@ -88,9 +88,21 @@ int32_t schValidateReceivedMsgType(SSchJob *pJob, SSchTask *pTask, int32_t msgTy
 }
 
 // Note: no more task error processing, handled in function internal
-int32_t schHandleResponseMsg(SSchJob *pJob, SSchTask *pTask, int32_t msgType, char *msg, int32_t msgSize,
-                             int32_t rspCode) {
+int32_t schHandleResponseMsg(SSchJob *pJob, SSchTask *pTask, int32_t execId, SDataBuf *pMsg, int32_t rspCode) {
   int32_t code = 0;
+  char *msg = pMsg->pData;
+  int32_t msgSize = pMsg->len;
+  int32_t msgType = pMsg->msgType;
+
+  bool dropExecNode = (msgType == TDMT_SCH_LINK_BROKEN || SCH_NETWORK_ERR(rspCode));
+  SCH_ERR_JRET(schUpdateTaskHandle(pJob, pTask, dropExecNode, pMsg->handle, execId));
+
+  SCH_ERR_JRET(schValidateReceivedMsgType(pJob, pTask, msgType));
+
+  int32_t reqType = IsReq(pMsg) ? pMsg->msgType : (pMsg->msgType - 1);
+  if (SCH_NEED_REDIRECT(reqType, rspCode, pMsg->len)) {
+    SCH_RET(schHandleRedirect(pJob, pTask, (SDataBuf *)pMsg, rspCode));
+  }
 
   switch (msgType) {
     case TDMT_VND_COMMIT_RSP: {
@@ -313,7 +325,7 @@ int32_t schHandleResponseMsg(SSchJob *pJob, SSchTask *pTask, int32_t msgType, ch
           return TSDB_CODE_SUCCESS;
         }
 
-        SCH_ERR_JRET(schFetchFromRemote(pJob));
+        SCH_ERR_JRET(schLaunchFetchTask(pJob));
 
         taosMemoryFreeClear(msg);
 
@@ -330,7 +342,7 @@ int32_t schHandleResponseMsg(SSchJob *pJob, SSchTask *pTask, int32_t msgType, ch
       atomic_add_fetch_32(&pJob->resNumOfRows, htonl(rsp->numOfRows));
 
       if (rsp->completed) {
-        SCH_SET_TASK_STATUS(pTask, JOB_TASK_STATUS_SUCCEED);
+        SCH_SET_TASK_STATUS(pTask, JOB_TASK_STATUS_SUCC);
       }
 
       SCH_TASK_DLOG("got fetch rsp, rows:%d, complete:%d", htonl(rsp->numOfRows), rsp->completed);
@@ -366,65 +378,24 @@ _return:
 
 int32_t schHandleCallback(void *param, SDataBuf *pMsg, int32_t rspCode) {
   int32_t                code = 0;
-  int32_t                msgType = pMsg->msgType;
   SSchTaskCallbackParam *pParam = (SSchTaskCallbackParam *)param;
   SSchTask              *pTask = NULL;
+  SSchJob               *pJob = NULL;
 
-  SSchJob *pJob = schAcquireJob(pParam->refId);
-  if (NULL == pJob) {
-    qWarn("QID:0x%" PRIx64 ",TID:0x%" PRIx64 "taosAcquireRef job failed, may be dropped, refId:0x%" PRIx64,
-          pParam->queryId, pParam->taskId, pParam->refId);
-    SCH_ERR_JRET(TSDB_CODE_QRY_JOB_FREED);
-  }
-
-  SCH_ERR_JRET(schGetTaskInJob(pJob, pParam->taskId, &pTask));
-
-  SCH_LOCK_TASK(pTask);
+  qDebug("begin to handle rsp msg, type:%s, handle:%p, code:%s", TMSG_INFO(pMsg->msgType), pMsg->handle, tstrerror(rspCode));
 
-  SCH_TASK_DLOG("rsp msg received, type:%s, handle:%p, code:%s", TMSG_INFO(msgType), pMsg->handle, tstrerror(rspCode));
+  SCH_ERR_RET(schProcessOnCbBegin(&pJob, &pTask, pParam->queryId, pParam->refId, pParam->taskId));
 
-  if (pParam->execId != pTask->execId) {
-    SCH_TASK_DLOG("execId %d mis-match current execId %d", pParam->execId, pTask->execId);
-    goto _return;
-  }
-
-  bool dropExecNode = (msgType == TDMT_SCH_LINK_BROKEN || SCH_NETWORK_ERR(rspCode));
-  SCH_ERR_JRET(schUpdateTaskHandle(pJob, pTask, dropExecNode, pMsg->handle, pParam->execId));
-
-  int8_t status = 0;
-  if (schJobNeedToStop(pJob, &status)) {
-    SCH_TASK_ELOG("rsp will not be processed cause of job status %s, rspCode:0x%x", jobTaskStatusStr(status), rspCode);
-    code = atomic_load_32(&pJob->errCode);
-    goto _return;
-  }
-
-  SCH_ERR_JRET(schValidateReceivedMsgType(pJob, pTask, msgType));
-
-  int32_t reqType = IsReq(pMsg) ? pMsg->msgType : (pMsg->msgType - 1);
-  if (SCH_NEED_REDIRECT(reqType, rspCode, pMsg->len)) {
-    code = schHandleRedirect(pJob, pTask, (SDataBuf *)pMsg, rspCode);
-    goto _return;
-  }
-  
-  schHandleResponseMsg(pJob, pTask, msgType, pMsg->pData, pMsg->len, rspCode);
+  code = schHandleResponseMsg(pJob, pTask, pParam->execId, pMsg, rspCode);
   pMsg->pData = NULL;
 
-_return:
-
-  if (pTask) {
-    if (code) {
-      schProcessOnTaskFailure(pJob, pTask, code);
-    }
-
-    SCH_UNLOCK_TASK(pTask);
-  }
-
-  if (pJob) {
-    schReleaseJob(pParam->refId);
-  }
+  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));
+
   SCH_RET(code);
 }
 
@@ -459,8 +430,38 @@ int32_t schHandleCommitCallback(void *param, SDataBuf *pMsg, int32_t code) {
   return schHandleCallback(param, pMsg, code);
 }
 
-int32_t schMakeCallbackParam(SSchJob *pJob, SSchTask *pTask, int32_t msgType, bool isHb, SSchTrans *trans,
-                             void **pParam) {
+int32_t schHandleHbCallback(void *param, SDataBuf *pMsg, int32_t code) {
+  SSchedulerHbRsp rsp = {0};
+  SSchTaskCallbackParam *pParam = (SSchTaskCallbackParam *)param;
+
+  if (code) {
+    qError("hb rsp error:%s", tstrerror(code));
+    SCH_ERR_JRET(code);
+  }
+
+  if (tDeserializeSSchedulerHbRsp(pMsg->pData, pMsg->len, &rsp)) {
+    qError("invalid hb rsp msg, size:%d", pMsg->len);
+    SCH_ERR_JRET(TSDB_CODE_QRY_INVALID_INPUT);
+  }
+
+  SSchTrans trans = {0};
+  trans.pTrans = pParam->pTrans;
+  trans.pHandle = pMsg->handle;
+
+  SCH_ERR_JRET(schUpdateHbConnection(&rsp.epId, &trans));
+
+  SCH_ERR_JRET(schProcessOnTaskStatusRsp(&rsp.epId, rsp.taskStatus));
+
+_return:
+
+  tFreeSSchedulerHbRsp(&rsp);
+  taosMemoryFree(param);
+
+  SCH_RET(code);
+}
+
+
+int32_t schMakeCallbackParam(SSchJob *pJob, SSchTask *pTask, int32_t msgType, bool isHb, SSchTrans *trans, void **pParam) {
   if (!isHb) {
     SSchTaskCallbackParam *param = taosMemoryCalloc(1, sizeof(SSchTaskCallbackParam));
     if (NULL == param) {
@@ -703,36 +704,6 @@ _return:
   SCH_RET(code);
 }
 
-int32_t schHandleHbCallback(void *param, SDataBuf *pMsg, int32_t code) {
-  SSchedulerHbRsp        rsp = {0};
-  SSchTaskCallbackParam *pParam = (SSchTaskCallbackParam *)param;
-
-  if (code) {
-    qError("hb rsp error:%s", tstrerror(code));
-    SCH_ERR_JRET(code);
-  }
-
-  if (tDeserializeSSchedulerHbRsp(pMsg->pData, pMsg->len, &rsp)) {
-    qError("invalid hb rsp msg, size:%d", pMsg->len);
-    SCH_ERR_JRET(TSDB_CODE_QRY_INVALID_INPUT);
-  }
-
-  SSchTrans trans = {0};
-  trans.pTrans = pParam->pTrans;
-  trans.pHandle = pMsg->handle;
-
-  SCH_ERR_JRET(schUpdateHbConnection(&rsp.epId, &trans));
-
-  SCH_ERR_JRET(schProcessOnTaskStatusRsp(&rsp.epId, rsp.taskStatus));
-
-_return:
-
-  tFreeSSchedulerHbRsp(&rsp);
-  taosMemoryFree(param);
-
-  SCH_RET(code);
-}
-
 int32_t schMakeBrokenLinkVal(SSchJob *pJob, SSchTask *pTask, SRpcBrokenlinkVal *brokenVal, bool isHb) {
   int32_t       code = 0;
   int32_t       msgType = TDMT_SCH_LINK_BROKEN;

source/libs/scheduler/src/schStatus.c

+/*
+ * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
+ *
+ * This program is free software: you can use, redistribute, and/or modify
+ * it under the terms of the GNU Affero General Public License, version 3
+ * or later ("AGPL"), as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "catalog.h"
+#include "command.h"
+#include "query.h"
+#include "schInt.h"
+#include "tmsg.h"
+#include "tref.h"
+#include "trpc.h"
+
+int32_t schSwitchJobStatus(SSchJob* pJob, int32_t status, void* param) {
+  int32_t code = 0;
+  SCH_ERR_JRET(schUpdateJobStatus(pJob, status));
+
+  switch (status) {
+    case JOB_TASK_STATUS_INIT:
+      break;
+    case JOB_TASK_STATUS_EXEC:
+      SCH_ERR_JRET(schExecJob(pJob, (SSchedulerReq*)param));    
+      break;
+    case JOB_TASK_STATUS_PART_SUCC:
+      SCH_ERR_JRET(schProcessOnJobPartialSuccess(pJob));
+      break;
+    case JOB_TASK_STATUS_SUCC:
+      break;
+    case JOB_TASK_STATUS_FAIL:      
+      SCH_RET(schProcessOnJobFailure(pJob, (param ? *(int32_t*)param : 0)));
+      break;
+    case JOB_TASK_STATUS_DROP:
+      SCH_ERR_JRET(schProcessOnJobDropped(pJob, *(int32_t*)param));
+      
+      if (taosRemoveRef(schMgmt.jobRef, pJob->refId)) {
+        SCH_JOB_ELOG("remove job from job list failed, refId:0x%" PRIx64, pJob->refId);
+      } else {
+        SCH_JOB_DLOG("job removed from jobRef list, refId:0x%" PRIx64, pJob->refId);
+      }
+      break;    
+    default: {
+      SCH_JOB_ELOG("unknown job status %d", status);
+      SCH_RET(TSDB_CODE_SCH_STATUS_ERROR);
+    }
+  }
+
+  return TSDB_CODE_SUCCESS;
+
+_return:
+
+  SCH_RET(schProcessOnJobFailure(pJob, code));
+}
+
+int32_t schHandleOpBeginEvent(int64_t jobId, SSchJob** job, SCH_OP_TYPE type, SSchedulerReq* pReq) {
+  SSchJob *pJob = schAcquireJob(jobId);
+  if (NULL == pJob) {
+    qError("Acquire sch job failed, may be dropped, jobId:0x%" PRIx64, jobId);
+    SCH_ERR_RET(TSDB_CODE_SCH_STATUS_ERROR);
+  }
+
+  *job = pJob;
+
+  SCH_RET(schProcessOnOpBegin(pJob, type, pReq));
+}
+
+int32_t schHandleOpEndEvent(SSchJob* pJob, SCH_OP_TYPE type, SSchedulerReq* pReq, int32_t errCode) {
+  int32_t code = errCode;
+  
+  if (NULL == pJob) {
+    SCH_RET(code);
+  }
+  
+  schProcessOnOpEnd(pJob, type, pReq, errCode);
+
+  if (TSDB_CODE_SCH_IGNORE_ERROR == errCode) {
+    code = pJob->errCode;
+  }
+
+  schReleaseJob(pJob->refId);
+
+  return code;
+}
+
+

source/libs/scheduler/src/schUtil.c

@@ -16,11 +16,25 @@
 #include "catalog.h"
 #include "command.h"
 #include "query.h"
-#include "schedulerInt.h"
+#include "schInt.h"
 #include "tmsg.h"
 #include "tref.h"
 #include "trpc.h"
 
+FORCE_INLINE SSchJob *schAcquireJob(int64_t refId) { 
+  qDebug("sch acquire jobId:0x%"PRIx64, refId); 
+  return (SSchJob *)taosAcquireRef(schMgmt.jobRef, refId); 
+}
+
+FORCE_INLINE int32_t schReleaseJob(int64_t refId) { 
+  if (0 == refId) {
+    return TSDB_CODE_SUCCESS;
+  }
+  
+  qDebug("sch release jobId:0x%"PRIx64, refId); 
+  return taosReleaseRef(schMgmt.jobRef, refId); 
+}
+
 char* schGetOpStr(SCH_OP_TYPE type) {
   switch (type) {
     case SCH_OP_NULL:
@@ -29,6 +43,8 @@ char* schGetOpStr(SCH_OP_TYPE type) {
       return "EXEC";
     case SCH_OP_FETCH:
       return "FETCH";
+    case SCH_OP_GET_STATUS:
+      return "GET STATUS";
     default:
       return "UNKNOWN";
   }
@@ -283,3 +299,20 @@ void schFreeSMsgSendInfo(SMsgSendInfo *msgSendInfo) {
   taosMemoryFree(msgSendInfo);
 }
 
+int32_t schGetTaskFromList(SHashObj *pTaskList, uint64_t taskId, SSchTask **pTask) {
+  int32_t s = taosHashGetSize(pTaskList);
+  if (s <= 0) {
+    return TSDB_CODE_SUCCESS;
+  }
+
+  SSchTask **task = taosHashGet(pTaskList, &taskId, sizeof(taskId));
+  if (NULL == task || NULL == (*task)) {
+    return TSDB_CODE_SUCCESS;
+  }
+
+  *pTask = *task;
+
+  return TSDB_CODE_SUCCESS;
+}
+
+

source/libs/scheduler/src/scheduler.c

@@ -16,7 +16,7 @@
 #include "catalog.h"
 #include "command.h"
 #include "query.h"
-#include "schedulerInt.h"
+#include "schInt.h"
 #include "tmsg.h"
 #include "tref.h"
 #include "trpc.h"
@@ -67,121 +67,45 @@ int32_t schedulerInit(SSchedulerCfg *cfg) {
   return TSDB_CODE_SUCCESS;
 }
 
-int32_t schedulerExecJob(SSchedulerReq *pReq, int64_t *pJobId, SQueryResult *pRes) {
-  qDebug("scheduler sync exec job start");
+int32_t schedulerExecJob(SSchedulerReq *pReq, int64_t *pJobId) {
+  qDebug("scheduler %s exec job start", pReq->syncReq ? "SYNC" : "ASYNC");
 
   int32_t code = 0;  
   SSchJob *pJob = NULL;
-  SCH_ERR_JRET(schInitJob(pReq, &pJob));
 
-  *pJobId = pJob->refId;
-  
-  SCH_ERR_JRET(schExecJobImpl(pReq, pJob, true));
+  SCH_ERR_JRET(schInitJob(pJobId, pReq));
 
-_return:
+  SCH_ERR_JRET(schHandleOpBeginEvent(*pJobId, &pJob, SCH_OP_EXEC, pReq));
 
-  if (code && NULL == pJob) {
-    qDestroyQueryPlan(pReq->pDag);
-  }
-  
-  if (pJob) {
-    schSetJobQueryRes(pJob, pRes);
-    schReleaseJob(pJob->refId);
-  }
-
-  return code;
-}
+  SCH_ERR_JRET(schSwitchJobStatus(pJob, JOB_TASK_STATUS_INIT, pReq));
 
-int32_t schedulerAsyncExecJob(SSchedulerReq *pReq, int64_t *pJobId) {
-  qDebug("scheduler async exec job start");
-
-  int32_t code = 0;  
-  SSchJob *pJob = NULL;
-  SCH_ERR_JRET(schInitJob(pReq, &pJob));
-
-  *pJobId = pJob->refId;
-  
-  SCH_ERR_JRET(schExecJobImpl(pReq, pJob, false));
+  SCH_ERR_JRET(schSwitchJobStatus(pJob, JOB_TASK_STATUS_EXEC, pReq));
 
 _return:
-
-  if (code && NULL == pJob) {
-    qDestroyQueryPlan(pReq->pDag);
-  }
   
-  if (pJob) {
-    schReleaseJob(pJob->refId);
-  }
-
-  return code;
+  SCH_RET(schHandleOpEndEvent(pJob, SCH_OP_EXEC, pReq, code));
 }
 
-int32_t schedulerFetchRows(int64_t job, void **pData) {
-  qDebug("scheduler sync fetch rows start");
-
-  if (NULL == pData) {
-    SCH_ERR_RET(TSDB_CODE_QRY_INVALID_INPUT);
-  }
+int32_t schedulerFetchRows(int64_t jobId, SSchedulerReq *pReq) {
+  qDebug("scheduler %s fetch rows start", pReq->syncReq ? "SYNC" : "ASYNC");
 
   int32_t  code = 0;
-  SSchJob *pJob = schAcquireJob(job);
-  if (NULL == pJob) {
-    qError("acquire job from jobRef list failed, may be dropped, jobId:0x%" PRIx64, job);
-    SCH_ERR_RET(TSDB_CODE_SCH_STATUS_ERROR);
-  }
-
-  SCH_ERR_RET(schBeginOperation(pJob, SCH_OP_FETCH, true));
-
-  pJob->userRes.fetchRes = pData;
-  code = schFetchRows(pJob);
-
-  schReleaseJob(job);
-
-  SCH_RET(code);
-}
-
-void schedulerAsyncFetchRows(int64_t job, schedulerFetchFp fp, void* param) {
-  qDebug("scheduler async fetch rows start");
+  SSchJob *pJob = NULL;
 
-  int32_t code = 0;
-  if (NULL == fp || NULL == param) {
-    SCH_ERR_JRET(TSDB_CODE_QRY_INVALID_INPUT);
-  }
+  SCH_ERR_JRET(schHandleOpBeginEvent(jobId, &pJob, SCH_OP_FETCH, pReq));
 
-  SSchJob *pJob = schAcquireJob(job);
-  if (NULL == pJob) {
-    qError("acquire sch job from job list failed, may be dropped, jobId:0x%" PRIx64, job);
-    SCH_ERR_JRET(TSDB_CODE_SCH_STATUS_ERROR);
-  }
-
-  SCH_ERR_JRET(schBeginOperation(pJob, SCH_OP_FETCH, false));
-  
-  pJob->userRes.fetchFp = fp;
-  pJob->userRes.userParam = param;
-  
-  SCH_ERR_JRET(schAsyncFetchRows(pJob));
+  SCH_ERR_JRET(schJobFetchRows(pJob));
 
 _return:
 
-  if (code) {
-    fp(NULL, param, code);
-  }
-  
-  schReleaseJob(job);
+  SCH_RET(schHandleOpEndEvent(pJob, SCH_OP_FETCH, pReq, code));
 }
 
-int32_t schedulerGetTasksStatus(int64_t job, SArray *pSub) {
+int32_t schedulerGetTasksStatus(int64_t jobId, SArray *pSub) {
   int32_t  code = 0;
-  SSchJob *pJob = schAcquireJob(job);
-  if (NULL == pJob) {
-    qDebug("acquire job from jobRef list failed, may not started or dropped, refId:0x%" PRIx64, job);
-    SCH_ERR_RET(TSDB_CODE_SCH_STATUS_ERROR);
-  }
+  SSchJob *pJob = NULL;
 
-  if (pJob->status < JOB_TASK_STATUS_NOT_START || pJob->levelNum <= 0 || NULL == pJob->levels) {
-    qDebug("job not initialized or not executable job, refId:0x%" PRIx64, job);
-    SCH_ERR_JRET(TSDB_CODE_SCH_STATUS_ERROR);
-  }
+  SCH_ERR_JRET(schHandleOpBeginEvent(jobId, &pJob, SCH_OP_GET_STATUS, NULL));
 
   for (int32_t i = pJob->levelNum - 1; i >= 0; --i) {
     SSchLevel *pLevel = taosArrayGet(pJob->levels, i);
@@ -198,23 +122,7 @@ int32_t schedulerGetTasksStatus(int64_t job, SArray *pSub) {
 
 _return:
 
-  schReleaseJob(job);
-
-  SCH_RET(code);
-}
-
-int32_t scheduleCancelJob(int64_t job) {
-  SSchJob *pJob = schAcquireJob(job);
-  if (NULL == pJob) {
-    qError("acquire job from jobRef list failed, may be dropped, jobId:0x%" PRIx64, job);
-    SCH_ERR_RET(TSDB_CODE_SCH_STATUS_ERROR);
-  }
-
-  int32_t code = schCancelJob(pJob);
-
-  schReleaseJob(job);
-
-  SCH_RET(code);
+  SCH_RET(schHandleOpEndEvent(pJob, SCH_OP_GET_STATUS, NULL, code));
 }
 
 void schedulerStopQueryHb(void *pTrans) {
@@ -225,33 +133,23 @@ void schedulerStopQueryHb(void *pTrans) {
   schCleanClusterHb(pTrans);
 }
 
-void schedulerFreeJob(int64_t* job, int32_t errCode) {
-  if (0 == *job) {
+void schedulerFreeJob(int64_t* jobId, int32_t errCode) {
+  if (0 == *jobId) {
     return;
   }
-  
-  SSchJob *pJob = schAcquireJob(*job);
+
+  SSchJob *pJob = schAcquireJob(*jobId);
   if (NULL == pJob) {
-    qError("acquire sch job failed, may be dropped, jobId:0x%" PRIx64, *job);
-    *job = 0;
+    qError("Acquire sch job failed, may be dropped, jobId:0x%" PRIx64, *jobId);
     return;
   }
 
-  int32_t code = schProcessOnJobDropped(pJob, errCode);
-  if (TSDB_CODE_SCH_JOB_IS_DROPPING == code) {
-    SCH_JOB_DLOG("sch job is already dropping, refId:0x%" PRIx64, *job);
-    *job = 0;
+  if (schJobDone(pJob)) {
     return;
   }
 
-  SCH_JOB_DLOG("start to remove job from jobRef list, refId:0x%" PRIx64, *job);
-
-  if (taosRemoveRef(schMgmt.jobRef, *job)) {
-    SCH_JOB_ELOG("remove job from job list failed, refId:0x%" PRIx64, *job);
-  }
-
-  schReleaseJob(*job);
-  *job = 0;
+  schSwitchJobStatus(pJob, JOB_TASK_STATUS_DROP, (void*)&errCode);
+  *jobId = 0;
 }
 
 void schedulerDestroy(void) {

source/libs/sync/inc/syncRaftCfg.h

@@ -36,7 +36,7 @@ typedef struct SRaftCfg {
   TdFilePtr pFile;
   char      path[TSDB_FILENAME_LEN * 2];
   int8_t    isStandBy;
-  int8_t    snapshotEnable;
+  int8_t    snapshotStrategy;
   SyncIndex lastConfigIndex;
 
   SyncIndex configIndexArr[MAX_CONFIG_INDEX_COUNT];
@@ -49,20 +49,20 @@ int32_t   raftCfgClose(SRaftCfg *pRaftCfg);
 int32_t   raftCfgPersist(SRaftCfg *pRaftCfg);
 int32_t   raftCfgAddConfigIndex(SRaftCfg *pRaftCfg, SyncIndex configIndex);
 
-cJSON * syncCfg2Json(SSyncCfg *pSyncCfg);
-char *  syncCfg2Str(SSyncCfg *pSyncCfg);
-char *  syncCfg2SimpleStr(SSyncCfg *pSyncCfg);
+cJSON  *syncCfg2Json(SSyncCfg *pSyncCfg);
+char   *syncCfg2Str(SSyncCfg *pSyncCfg);
+char   *syncCfg2SimpleStr(SSyncCfg *pSyncCfg);
 int32_t syncCfgFromJson(const cJSON *pRoot, SSyncCfg *pSyncCfg);
 int32_t syncCfgFromStr(const char *s, SSyncCfg *pSyncCfg);
 
-cJSON * raftCfg2Json(SRaftCfg *pRaftCfg);
-char *  raftCfg2Str(SRaftCfg *pRaftCfg);
+cJSON  *raftCfg2Json(SRaftCfg *pRaftCfg);
+char   *raftCfg2Str(SRaftCfg *pRaftCfg);
 int32_t raftCfgFromJson(const cJSON *pRoot, SRaftCfg *pRaftCfg);
 int32_t raftCfgFromStr(const char *s, SRaftCfg *pRaftCfg);
 
 typedef struct SRaftCfgMeta {
   int8_t    isStandBy;
-  int8_t    snapshotEnable;
+  int8_t    snapshotStrategy;
   SyncIndex lastConfigIndex;
 } SRaftCfgMeta;
 

source/libs/sync/src/syncElection.c

@@ -96,12 +96,20 @@ int32_t syncNodeElect(SSyncNode* pSyncNode) {
     return ret;
   }
 
-  if (pSyncNode->pRaftCfg->snapshotEnable) {
-    ret = syncNodeRequestVotePeersSnapshot(pSyncNode);
-  } else {
-    ret = syncNodeRequestVotePeers(pSyncNode);
+  switch (pSyncNode->pRaftCfg->snapshotStrategy) {
+    case SYNC_STRATEGY_NO_SNAPSHOT:
+      ret = syncNodeRequestVotePeers(pSyncNode);
+      break;
+
+    case SYNC_STRATEGY_STANDARD_SNAPSHOT:
+    case SYNC_STRATEGY_WAL_FIRST:
+      ret = syncNodeRequestVotePeersSnapshot(pSyncNode);
+      break;
+
+    default:
+      ret = syncNodeRequestVotePeers(pSyncNode);
+      break;
   }
-
   ASSERT(ret == 0);
   syncNodeResetElectTimer(pSyncNode);
 

source/libs/sync/src/syncRaftCfg.c

@@ -101,7 +101,7 @@ cJSON *syncCfg2Json(SSyncCfg *pSyncCfg) {
 
 char *syncCfg2Str(SSyncCfg *pSyncCfg) {
   cJSON *pJson = syncCfg2Json(pSyncCfg);
-  char * serialized = cJSON_Print(pJson);
+  char  *serialized = cJSON_Print(pJson);
   cJSON_Delete(pJson);
   return serialized;
 }
@@ -109,7 +109,7 @@ char *syncCfg2Str(SSyncCfg *pSyncCfg) {
 char *syncCfg2SimpleStr(SSyncCfg *pSyncCfg) {
   if (pSyncCfg != NULL) {
     int32_t len = 512;
-    char *  s = taosMemoryMalloc(len);
+    char   *s = taosMemoryMalloc(len);
     memset(s, 0, len);
 
     snprintf(s, len, "{replica-num:%d, my-index:%d, ", pSyncCfg->replicaNum, pSyncCfg->myIndex);
@@ -182,7 +182,7 @@ cJSON *raftCfg2Json(SRaftCfg *pRaftCfg) {
   cJSON *pRoot = cJSON_CreateObject();
   cJSON_AddItemToObject(pRoot, "SSyncCfg", syncCfg2Json(&(pRaftCfg->cfg)));
   cJSON_AddNumberToObject(pRoot, "isStandBy", pRaftCfg->isStandBy);
-  cJSON_AddNumberToObject(pRoot, "snapshotEnable", pRaftCfg->snapshotEnable);
+  cJSON_AddNumberToObject(pRoot, "snapshotStrategy", pRaftCfg->snapshotStrategy);
 
   char buf64[128];
   snprintf(buf64, sizeof(buf64), "%ld", pRaftCfg->lastConfigIndex);
@@ -205,7 +205,7 @@ cJSON *raftCfg2Json(SRaftCfg *pRaftCfg) {
 
 char *raftCfg2Str(SRaftCfg *pRaftCfg) {
   cJSON *pJson = raftCfg2Json(pRaftCfg);
-  char * serialized = cJSON_Print(pJson);
+  char  *serialized = cJSON_Print(pJson);
   cJSON_Delete(pJson);
   return serialized;
 }
@@ -228,7 +228,7 @@ int32_t raftCfgCreateFile(SSyncCfg *pCfg, SRaftCfgMeta meta, const char *path) {
   SRaftCfg raftCfg;
   raftCfg.cfg = *pCfg;
   raftCfg.isStandBy = meta.isStandBy;
-  raftCfg.snapshotEnable = meta.snapshotEnable;
+  raftCfg.snapshotStrategy = meta.snapshotStrategy;
   raftCfg.lastConfigIndex = meta.lastConfigIndex;
   raftCfg.configIndexCount = 1;
   memset(raftCfg.configIndexArr, 0, sizeof(raftCfg.configIndexArr));
@@ -257,8 +257,8 @@ int32_t raftCfgFromJson(const cJSON *pRoot, SRaftCfg *pRaftCfg) {
   cJSON *pJsonIsStandBy = cJSON_GetObjectItem(pJson, "isStandBy");
   pRaftCfg->isStandBy = cJSON_GetNumberValue(pJsonIsStandBy);
 
-  cJSON *pJsonSnapshotEnable = cJSON_GetObjectItem(pJson, "snapshotEnable");
-  pRaftCfg->snapshotEnable = cJSON_GetNumberValue(pJsonSnapshotEnable);
+  cJSON *pJsonSnapshotStrategy = cJSON_GetObjectItem(pJson, "snapshotStrategy");
+  pRaftCfg->snapshotStrategy = cJSON_GetNumberValue(pJsonSnapshotStrategy);
 
   cJSON *pJsonLastConfigIndex = cJSON_GetObjectItem(pJson, "lastConfigIndex");
   pRaftCfg->lastConfigIndex = atoll(cJSON_GetStringValue(pJsonLastConfigIndex));
@@ -280,7 +280,7 @@ int32_t raftCfgFromJson(const cJSON *pRoot, SRaftCfg *pRaftCfg) {
     (pRaftCfg->configIndexArr)[i] = atoll(pIndex->valuestring);
   }
 
-  cJSON * pJsonSyncCfg = cJSON_GetObjectItem(pJson, "SSyncCfg");
+  cJSON  *pJsonSyncCfg = cJSON_GetObjectItem(pJson, "SSyncCfg");
   int32_t code = syncCfgFromJson(pJsonSyncCfg, &(pRaftCfg->cfg));
   ASSERT(code == 0);
 

source/util/src/terror.c

@@ -395,6 +395,7 @@ TAOS_DEFINE_ERROR(TSDB_CODE_QRY_TASK_DROPPING,            "Task dropping")
 TAOS_DEFINE_ERROR(TSDB_CODE_QRY_DUPLICATTED_OPERATION,    "Duplicatted operation")
 TAOS_DEFINE_ERROR(TSDB_CODE_QRY_TASK_MSG_ERROR,           "Task message error")
 TAOS_DEFINE_ERROR(TSDB_CODE_QRY_JOB_FREED,                "Job already freed")
+TAOS_DEFINE_ERROR(TSDB_CODE_QRY_JOB_NOT_EXIST,            "Job not exist")
 TAOS_DEFINE_ERROR(TSDB_CODE_QRY_TASK_STATUS_ERROR,        "Task status error")
 TAOS_DEFINE_ERROR(TSDB_CODE_QRY_JSON_IN_ERROR,            "Json not support in in/notin operator")
 TAOS_DEFINE_ERROR(TSDB_CODE_QRY_JSON_NOT_SUPPORT_ERROR,   "Json not support in this place")