Merge remote-tracking branch 'origin/3.0' into feature/3.0_debug_wxy

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/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

@@ -389,10 +389,10 @@ 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,
+  SSchedulerReq    req = {
+    .syncReq = true,
+    .pConn = &conn,
     .pNodeList = pNodeList,
     .pDag = pDag,
     .sql = pRequest->sqlstr,
     .startTs = pRequest->metric.start,
     .execFp = NULL,
-                       .execParam = NULL,
+    .cbParam = NULL,
     .chkKillFp = chkRequestKilled,
-                       .chkKillParam = (void*)pRequest->self};
+    .chkKillParam = (void*)pRequest->self,
+    .pExecRes = &res,
+  };
 
-  int32_t code = schedulerExecJob(&req, &pRequest->body.queryJob, &res);
-  pRequest->body.resInfo.execRes = res.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,
+        SSchedulerReq req = {
+          .syncReq = false,
+          .pConn = &conn,
           .pNodeList = pNodeList,
           .pDag = pDag,
           .sql = pRequest->sqlstr,
           .startTs = pRequest->metric.start,
           .execFp = schedulerExecCb,
-                             .execParam = pRequest,
+          .cbParam = pRequest,
           .chkKillFp = chkRequestKilled,
-                             .chkKillParam = (void*)pRequest->self};
-        code = schedulerAsyncExecJob(&req, &pRequest->body.queryJob);
+          .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/dnode/mgmt/mgmt_vnode/src/vmWorker.c

@@ -107,7 +107,7 @@ static void vmProcessSyncQueue(SQueueInfo *pInfo, STaosQall *qall, int32_t numOf
     const STraceId *trace = &pMsg->info.traceId;
     dGTrace("vgId:%d, msg:%p get from vnode-sync queue", pVnode->vgId, pMsg);
 
-    int32_t code = vnodeProcessSyncReq(pVnode->pImpl, pMsg, NULL);  // no response here
+    int32_t code = vnodeProcessSyncMsg(pVnode->pImpl, pMsg, NULL);  // no response here
     dGTrace("vgId:%d, msg:%p is freed, code:0x%x", pVnode->vgId, pMsg, code);
     rpcFreeCont(pMsg->pCont);
     taosFreeQitem(pMsg);

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);
+
   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/inc/vnode.h

@@ -52,10 +52,10 @@ int32_t vnodeCreate(const char *path, SVnodeCfg *pCfg, STfs *pTfs);
 void    vnodeDestroy(const char *path, STfs *pTfs);
 SVnode *vnodeOpen(const char *path, STfs *pTfs, SMsgCb msgCb);
 void    vnodeClose(SVnode *pVnode);
-int32_t vnodePreprocessReq(SVnode *pVnode, SRpcMsg *pMsg);
+int32_t vnodePreProcessReq(SVnode *pVnode, SRpcMsg *pMsg);
 int32_t vnodeProcessWriteReq(SVnode *pVnode, SRpcMsg *pMsg, int64_t version, SRpcMsg *pRsp);
 int32_t vnodeProcessCMsg(SVnode *pVnode, SRpcMsg *pMsg, SRpcMsg **pRsp);
-int32_t vnodeProcessSyncReq(SVnode *pVnode, SRpcMsg *pMsg, SRpcMsg **pRsp);
+int32_t vnodeProcessSyncMsg(SVnode *pVnode, SRpcMsg *pMsg, SRpcMsg **pRsp);
 int32_t vnodePreprocessQueryMsg(SVnode *pVnode, SRpcMsg *pMsg);
 int32_t vnodeProcessQueryMsg(SVnode *pVnode, SRpcMsg *pMsg);
 int32_t vnodeProcessFetchMsg(SVnode *pVnode, SRpcMsg *pMsg, SQueueInfo *pInfo);

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

@@ -240,7 +240,7 @@ struct SVnode {
   SSink*     pSink;
   tsem_t     canCommit;
   int64_t    sync;
-  int32_t    syncCount;
+  int32_t    blockCount;
   tsem_t     syncSem;
   SQHandle*  pQuery;
 };

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

@@ -85,6 +85,7 @@ STqOffsetStore* tqOffsetOpen(STQ* pTq) {
 void tqOffsetClose(STqOffsetStore* pStore) {
   tqOffsetSnapshot(pStore);
   taosHashCleanup(pStore->pHash);
+  taosMemoryFree(pStore);
 }
 
 STqOffset* tqOffsetRead(STqOffsetStore* pStore, const char* subscribeKey) {

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

@@ -81,7 +81,7 @@ SVnode *vnodeOpen(const char *path, STfs *pTfs, SMsgCb msgCb) {
   pVnode->state.applied = info.state.committed;
   pVnode->pTfs = pTfs;
   pVnode->msgCb = msgCb;
-  pVnode->syncCount = 0;
+  pVnode->blockCount = 0;
 
   tsem_init(&pVnode->syncSem, 0, 0);
   tsem_init(&(pVnode->canCommit), 0, 1);

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

@@ -28,7 +28,7 @@ static int32_t vnodeProcessAlterHasnRangeReq(SVnode *pVnode, int64_t version, vo
 static int32_t vnodeProcessWriteMsg(SVnode *pVnode, int64_t version, SRpcMsg *pMsg, SRpcMsg *pRsp);
 static int32_t vnodeProcessDropTtlTbReq(SVnode *pVnode, int64_t version, void *pReq, int32_t len, SRpcMsg *pRsp);
 
-int32_t vnodePreprocessReq(SVnode *pVnode, SRpcMsg *pMsg) {
+int32_t vnodePreProcessReq(SVnode *pVnode, SRpcMsg *pMsg) {
   int32_t  code = 0;
   SDecoder dc = {0};
 

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

@@ -25,12 +25,12 @@ static inline bool vnodeIsMsgWeak(tmsg_t type) { return false; }
 static inline void vnodeAccumBlockMsg(SVnode *pVnode, tmsg_t type) {
   if (!vnodeIsMsgBlock(type)) return;
 
-  int32_t count = atomic_add_fetch_32(&pVnode->syncCount, 1);
+  int32_t count = atomic_add_fetch_32(&pVnode->blockCount, 1);
   vTrace("vgId:%d, accum block, count:%d type:%s", pVnode->config.vgId, count, TMSG_INFO(type));
 }
 
 static inline void vnodeWaitBlockMsg(SVnode *pVnode) {
-  int32_t count = atomic_load_32(&pVnode->syncCount);
+  int32_t count = atomic_load_32(&pVnode->blockCount);
   if (count <= 0) return;
 
   vTrace("vgId:%d, wait block finish, count:%d", pVnode->config.vgId, count);
@@ -40,10 +40,10 @@ static inline void vnodeWaitBlockMsg(SVnode *pVnode) {
 static inline void vnodePostBlockMsg(SVnode *pVnode, tmsg_t type) {
   if (!vnodeIsMsgBlock(type)) return;
 
-  int32_t count = atomic_load_32(&pVnode->syncCount);
+  int32_t count = atomic_load_32(&pVnode->blockCount);
   if (count <= 0) return;
 
-  count = atomic_sub_fetch_32(&pVnode->syncCount, 1);
+  count = atomic_sub_fetch_32(&pVnode->blockCount, 1);
   vTrace("vgId:%d, post block, count:%d type:%s", pVnode->config.vgId, count, TMSG_INFO(type));
   if (count <= 0) {
     tsem_post(&pVnode->syncSem);
@@ -84,8 +84,10 @@ static int32_t vnodeProcessAlterReplicaReq(SVnode *pVnode, SRpcMsg *pMsg) {
     terrno = TSDB_CODE_INVALID_MSG;
     return TSDB_CODE_INVALID_MSG;
   }
-  STraceId *trace = &pMsg->info.traceId;
+
+  const STraceId *trace = &pMsg->info.traceId;
   vGTrace("vgId:%d, start to alter vnode replica to %d, handle:%p", TD_VID(pVnode), req.replica, pMsg->info.handle);
+
   SSyncCfg cfg = {.replicaNum = req.replica, .myIndex = req.selfIndex};
   for (int32_t r = 0; r < req.replica; ++r) {
     SNodeInfo *pNode = &cfg.nodeInfo[r];
@@ -126,32 +128,23 @@ void vnodeProposeMsg(SQueueInfo *pInfo, STaosQall *qall, int32_t numOfMsgs) {
 
   for (int32_t m = 0; m < numOfMsgs; m++) {
     if (taosGetQitem(qall, (void **)&pMsg) == 0) continue;
-    STraceId *trace = &pMsg->info.traceId;
+    const STraceId *trace = &pMsg->info.traceId;
     vGTrace("vgId:%d, msg:%p get from vnode-write queue handle:%p", vgId, pMsg, pMsg->info.handle);
 
+    code = vnodePreProcessReq(pVnode, pMsg);
+    if (code != 0) {
+      vError("vgId:%d, msg:%p failed to pre-process since %s", vgId, pMsg, terrstr());
+    } else {
       if (pMsg->msgType == TDMT_VND_ALTER_REPLICA) {
         code = vnodeProcessAlterReplicaReq(pVnode, pMsg);
-    } else {
-      code = vnodePreprocessReq(pVnode, pMsg);
-      if (code != 0) {
-        vError("vgId:%d, failed to pre-process msg:%p since %s", vgId, pMsg, terrstr());
       } else {
         code = syncPropose(pVnode->sync, pMsg, vnodeIsMsgWeak(pMsg->msgType));
-        if (code == 1) {
-          do {
-            static int32_t cnt = 0;
-            if (cnt++ % 1000 == 1) {
-              vInfo("vgId:%d, msg:%p apply right now, apply index:%ld, msgtype:%s,%d", vgId, pMsg,
-                    pMsg->info.conn.applyIndex, TMSG_INFO(pMsg->msgType), pMsg->msgType);
-            }
-          } while (0);
-
+        if (code > 0) {
           SRpcMsg rsp = {.code = pMsg->code, .info = pMsg->info};
           if (vnodeProcessWriteReq(pVnode, pMsg, pMsg->info.conn.applyIndex, &rsp) < 0) {
             rsp.code = terrno;
-            vInfo("vgId:%d, msg:%p failed to apply right now since %s", vgId, pMsg, terrstr());
+            vError("vgId:%d, msg:%p failed to apply right now since %s", vgId, pMsg, terrstr());
           }
-
           if (rsp.info.handle != NULL) {
             tmsgSendRsp(&rsp);
           }
@@ -161,18 +154,10 @@ void vnodeProposeMsg(SQueueInfo *pInfo, STaosQall *qall, int32_t numOfMsgs) {
 
     if (code == 0) {
       vnodeAccumBlockMsg(pVnode, pMsg->msgType);
-    } else if (code == -1 && terrno == TSDB_CODE_SYN_NOT_LEADER) {
+    } else if (code < 0) {
+      if (terrno == TSDB_CODE_SYN_NOT_LEADER) {
         SEpSet newEpSet = {0};
         syncGetRetryEpSet(pVnode->sync, &newEpSet);
-
-      /*
-      syncGetEpSet(pVnode->sync, &newEpSet);
-      SEp *pEp = &newEpSet.eps[newEpSet.inUse];
-      if (pEp->port == tsServerPort && strcmp(pEp->fqdn, tsLocalFqdn) == 0) {
-        newEpSet.inUse = (newEpSet.inUse + 1) % newEpSet.numOfEps;
-      }
-      */
-
         vGTrace("vgId:%d, msg:%p is redirect since not leader, numOfEps:%d inUse:%d", vgId, pMsg, newEpSet.numOfEps,
                 newEpSet.inUse);
         for (int32_t i = 0; i < newEpSet.numOfEps; ++i) {
@@ -182,13 +167,15 @@ void vnodeProposeMsg(SQueueInfo *pInfo, STaosQall *qall, int32_t numOfMsgs) {
         SRpcMsg rsp = {.code = TSDB_CODE_RPC_REDIRECT, .info = pMsg->info};
         tmsgSendRedirectRsp(&rsp, &newEpSet);
       } else {
-      if (code != 1) {
         if (terrno != 0) code = terrno;
         vError("vgId:%d, msg:%p failed to propose since %s, code:0x%x", vgId, pMsg, tstrerror(code), code);
         SRpcMsg rsp = {.code = code, .info = pMsg->info};
+        if (rsp.info.handle != NULL) {
           tmsgSendRsp(&rsp);
         }
       }
+    } else {
+    }
 
     vGTrace("vgId:%d, msg:%p is freed, code:0x%x", vgId, pMsg, code);
     rpcFreeCont(pMsg->pCont);
@@ -206,7 +193,7 @@ void vnodeApplyMsg(SQueueInfo *pInfo, STaosQall *qall, int32_t numOfMsgs) {
 
   for (int32_t i = 0; i < numOfMsgs; ++i) {
     if (taosGetQitem(qall, (void **)&pMsg) == 0) continue;
-    STraceId *trace = &pMsg->info.traceId;
+    const STraceId *trace = &pMsg->info.traceId;
     vGTrace("vgId:%d, msg:%p get from vnode-apply queue, type:%s handle:%p", vgId, pMsg, TMSG_INFO(pMsg->msgType),
             pMsg->info.handle);
 
@@ -229,17 +216,24 @@ void vnodeApplyMsg(SQueueInfo *pInfo, STaosQall *qall, int32_t numOfMsgs) {
   }
 }
 
-int32_t vnodeProcessSyncReq(SVnode *pVnode, SRpcMsg *pMsg, SRpcMsg **pRsp) {
-  int32_t ret = 0;
+int32_t vnodeProcessSyncMsg(SVnode *pVnode, SRpcMsg *pMsg, SRpcMsg **pRsp) {
+  int32_t         code = 0;
+  const STraceId *trace = &pMsg->info.traceId;
 
-  if (syncEnvIsStart()) {
-    SSyncNode *pSyncNode = syncNodeAcquire(pVnode->sync);
-    assert(pSyncNode != NULL);
+  if (!syncEnvIsStart()) {
+    vGError("vgId:%d, msg:%p failed to process since sync env not start", pVnode->config.vgId);
+    terrno = TSDB_CODE_APP_ERROR;
+    return -1;
+  }
 
-    SMsgHead *pHead = pMsg->pCont;
-    STraceId *trace = &pMsg->info.traceId;
+  SSyncNode *pSyncNode = syncNodeAcquire(pVnode->sync);
+  if (pSyncNode == NULL) {
+    vGError("vgId:%d, msg:%p failed to process since invalid sync node", pVnode->config.vgId);
+    terrno = TSDB_CODE_SYN_INTERNAL_ERROR;
+    return -1;
+  }
 
-    do {
+#if 1
   char          *syncNodeStr = sync2SimpleStr(pVnode->sync);
   static int64_t vndTick = 0;
   if (++vndTick % 10 == 1) {
@@ -247,154 +241,125 @@ int32_t vnodeProcessSyncReq(SVnode *pVnode, SRpcMsg *pMsg, SRpcMsg **pRsp) {
   }
   if (gRaftDetailLog) {
     char logBuf[512] = {0};
-        snprintf(logBuf, sizeof(logBuf), "==vnodeProcessSyncReq== msgType:%d, syncNode: %s", pMsg->msgType,
-                 syncNodeStr);
+    snprintf(logBuf, sizeof(logBuf), "vnode process syncmsg, msgType:%d, syncNode:%s", pMsg->msgType, syncNodeStr);
     syncRpcMsgLog2(logBuf, pMsg);
   }
   taosMemoryFree(syncNodeStr);
-    } while (0);
-
-    SRpcMsg *pRpcMsg = pMsg;
+#endif
 
-    // ToDo: ugly! use function pointer
-    // use different strategy
   if (syncNodeStrategy(pSyncNode) == SYNC_STRATEGY_NO_SNAPSHOT) {
-      if (pRpcMsg->msgType == TDMT_SYNC_TIMEOUT) {
-        SyncTimeout *pSyncMsg = syncTimeoutFromRpcMsg2(pRpcMsg);
+    if (pMsg->msgType == TDMT_SYNC_TIMEOUT) {
+      SyncTimeout *pSyncMsg = syncTimeoutFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnTimeoutCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnTimeoutCb(pSyncNode, pSyncMsg);
       syncTimeoutDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_PING) {
-        SyncPing *pSyncMsg = syncPingFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_PING) {
+      SyncPing *pSyncMsg = syncPingFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnPingCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnPingCb(pSyncNode, pSyncMsg);
       syncPingDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_PING_REPLY) {
-        SyncPingReply *pSyncMsg = syncPingReplyFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_PING_REPLY) {
+      SyncPingReply *pSyncMsg = syncPingReplyFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnPingReplyCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnPingReplyCb(pSyncNode, pSyncMsg);
       syncPingReplyDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_CLIENT_REQUEST) {
-        SyncClientRequest *pSyncMsg = syncClientRequestFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_CLIENT_REQUEST) {
+      SyncClientRequest *pSyncMsg = syncClientRequestFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnClientRequestCb(pSyncNode, pSyncMsg, NULL);
+      code = syncNodeOnClientRequestCb(pSyncNode, pSyncMsg, NULL);
       syncClientRequestDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_REQUEST_VOTE) {
-        SyncRequestVote *pSyncMsg = syncRequestVoteFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_REQUEST_VOTE) {
+      SyncRequestVote *pSyncMsg = syncRequestVoteFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnRequestVoteCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnRequestVoteCb(pSyncNode, pSyncMsg);
       syncRequestVoteDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_REQUEST_VOTE_REPLY) {
-        SyncRequestVoteReply *pSyncMsg = syncRequestVoteReplyFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_REQUEST_VOTE_REPLY) {
+      SyncRequestVoteReply *pSyncMsg = syncRequestVoteReplyFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnRequestVoteReplyCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnRequestVoteReplyCb(pSyncNode, pSyncMsg);
       syncRequestVoteReplyDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_APPEND_ENTRIES) {
-        SyncAppendEntries *pSyncMsg = syncAppendEntriesFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_APPEND_ENTRIES) {
+      SyncAppendEntries *pSyncMsg = syncAppendEntriesFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnAppendEntriesCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnAppendEntriesCb(pSyncNode, pSyncMsg);
       syncAppendEntriesDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_APPEND_ENTRIES_REPLY) {
-        SyncAppendEntriesReply *pSyncMsg = syncAppendEntriesReplyFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_APPEND_ENTRIES_REPLY) {
+      SyncAppendEntriesReply *pSyncMsg = syncAppendEntriesReplyFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnAppendEntriesReplyCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnAppendEntriesReplyCb(pSyncNode, pSyncMsg);
       syncAppendEntriesReplyDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_SET_VNODE_STANDBY) {
-        ret = vnodeSetStandBy(pVnode);
-        if (ret != 0 && terrno != 0) ret = terrno;
-        SRpcMsg rsp = {.code = ret, .info = pMsg->info};
+    } else if (pMsg->msgType == TDMT_SYNC_SET_VNODE_STANDBY) {
+      code = vnodeSetStandBy(pVnode);
+      if (code != 0 && terrno != 0) code = terrno;
+      SRpcMsg rsp = {.code = code, .info = pMsg->info};
       tmsgSendRsp(&rsp);
     } else {
-        vError("==vnodeProcessSyncReq== error msg type:%d", pRpcMsg->msgType);
-        ret = -1;
+      vGError("vgId:%d, msg:%p failed to process since error msg type:%d", pVnode->config.vgId, pMsg->msgType);
+      code = -1;
     }
-
   } else {
     // use wal first strategy
-
-      if (pRpcMsg->msgType == TDMT_SYNC_TIMEOUT) {
-        SyncTimeout *pSyncMsg = syncTimeoutFromRpcMsg2(pRpcMsg);
+    if (pMsg->msgType == TDMT_SYNC_TIMEOUT) {
+      SyncTimeout *pSyncMsg = syncTimeoutFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnTimeoutCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnTimeoutCb(pSyncNode, pSyncMsg);
       syncTimeoutDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_PING) {
-        SyncPing *pSyncMsg = syncPingFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_PING) {
+      SyncPing *pSyncMsg = syncPingFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnPingCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnPingCb(pSyncNode, pSyncMsg);
       syncPingDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_PING_REPLY) {
-        SyncPingReply *pSyncMsg = syncPingReplyFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_PING_REPLY) {
+      SyncPingReply *pSyncMsg = syncPingReplyFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnPingReplyCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnPingReplyCb(pSyncNode, pSyncMsg);
       syncPingReplyDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_CLIENT_REQUEST) {
-        SyncClientRequest *pSyncMsg = syncClientRequestFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_CLIENT_REQUEST) {
+      SyncClientRequest *pSyncMsg = syncClientRequestFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnClientRequestCb(pSyncNode, pSyncMsg, NULL);
+      code = syncNodeOnClientRequestCb(pSyncNode, pSyncMsg, NULL);
       syncClientRequestDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_CLIENT_REQUEST_BATCH) {
-        SyncClientRequestBatch *pSyncMsg = syncClientRequestBatchFromRpcMsg(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_CLIENT_REQUEST_BATCH) {
+      SyncClientRequestBatch *pSyncMsg = syncClientRequestBatchFromRpcMsg(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnClientRequestBatchCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnClientRequestBatchCb(pSyncNode, pSyncMsg);
       syncClientRequestBatchDestroyDeep(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_REQUEST_VOTE) {
-        SyncRequestVote *pSyncMsg = syncRequestVoteFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_REQUEST_VOTE) {
+      SyncRequestVote *pSyncMsg = syncRequestVoteFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnRequestVoteCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnRequestVoteCb(pSyncNode, pSyncMsg);
       syncRequestVoteDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_REQUEST_VOTE_REPLY) {
-        SyncRequestVoteReply *pSyncMsg = syncRequestVoteReplyFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_REQUEST_VOTE_REPLY) {
+      SyncRequestVoteReply *pSyncMsg = syncRequestVoteReplyFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnRequestVoteReplyCb(pSyncNode, pSyncMsg);
+      code = syncNodeOnRequestVoteReplyCb(pSyncNode, pSyncMsg);
       syncRequestVoteReplyDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_APPEND_ENTRIES_BATCH) {
-        SyncAppendEntriesBatch *pSyncMsg = syncAppendEntriesBatchFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_APPEND_ENTRIES_BATCH) {
+      SyncAppendEntriesBatch *pSyncMsg = syncAppendEntriesBatchFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnAppendEntriesSnapshot2Cb(pSyncNode, pSyncMsg);
+      code = syncNodeOnAppendEntriesSnapshot2Cb(pSyncNode, pSyncMsg);
       syncAppendEntriesBatchDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_APPEND_ENTRIES_REPLY) {
-        SyncAppendEntriesReply *pSyncMsg = syncAppendEntriesReplyFromRpcMsg2(pRpcMsg);
+    } else if (pMsg->msgType == TDMT_SYNC_APPEND_ENTRIES_REPLY) {
+      SyncAppendEntriesReply *pSyncMsg = syncAppendEntriesReplyFromRpcMsg2(pMsg);
       ASSERT(pSyncMsg != NULL);
-        ret = syncNodeOnAppendEntriesReplySnapshot2Cb(pSyncNode, pSyncMsg);
+      code = syncNodeOnAppendEntriesReplySnapshot2Cb(pSyncNode, pSyncMsg);
       syncAppendEntriesReplyDestroy(pSyncMsg);
-
-      } else if (pRpcMsg->msgType == TDMT_SYNC_SET_VNODE_STANDBY) {
-        ret = vnodeSetStandBy(pVnode);
-        if (ret != 0 && terrno != 0) ret = terrno;
-        SRpcMsg rsp = {.code = ret, .info = pMsg->info};
+    } else if (pMsg->msgType == TDMT_SYNC_SET_VNODE_STANDBY) {
+      code = vnodeSetStandBy(pVnode);
+      if (code != 0 && terrno != 0) code = terrno;
+      SRpcMsg rsp = {.code = code, .info = pMsg->info};
       tmsgSendRsp(&rsp);
     } else {
-        vError("==vnodeProcessSyncReq== error msg type:%d", pRpcMsg->msgType);
-        ret = -1;
+      vGError("vgId:%d, msg:%p failed to process since error msg type:%d", pVnode->config.vgId, pMsg->msgType);
+      code = -1;
     }
   }
 
   syncNodeRelease(pSyncNode);
-  } else {
-    vError("==vnodeProcessSyncReq== error syncEnv stop");
-    ret = -1;
-  }
-
-  if (ret != 0 && terrno == 0) {
+  if (code != 0 && terrno == 0) {
     terrno = TSDB_CODE_SYN_INTERNAL_ERROR;
   }
-  return ret;
+  return code;
 }
 
 static int32_t vnodeSyncEqMsg(const SMsgCb *msgcb, SRpcMsg *pMsg) {
@@ -427,7 +392,7 @@ static void vnodeSyncReconfig(struct SSyncFSM *pFsm, const SRpcMsg *pMsg, SReCon
   syncGetAndDelRespRpc(pVnode->sync, cbMeta.newCfgSeqNum, &rpcMsg.info);
   rpcMsg.info.conn.applyIndex = cbMeta.index;
 
-  STraceId *trace = (STraceId *)&pMsg->info.traceId;
+  const STraceId *trace = (STraceId *)&pMsg->info.traceId;
   vGTrace("vgId:%d, alter vnode replica is confirmed, type:%s contLen:%d seq:%" PRIu64 " handle:%p", TD_VID(pVnode),
           TMSG_INFO(pMsg->msgType), pMsg->contLen, cbMeta.seqNum, rpcMsg.info.handle);
   if (rpcMsg.info.handle != NULL) {
@@ -444,9 +409,8 @@ static void vnodeSyncCommitMsg(SSyncFSM *pFsm, const SRpcMsg *pMsg, SFsmCbMeta c
   char      logBuf[256] = {0};
 
   snprintf(logBuf, sizeof(logBuf),
-           "==callback== ==CommitCb== execute, pFsm:%p, index:%ld, isWeak:%d, code:%d, state:%d %s, beginIndex :%ld\n",
-           pFsm, cbMeta.index, cbMeta.isWeak, cbMeta.code, cbMeta.state, syncUtilState2String(cbMeta.state),
-           beginIndex);
+           "commitCb execute, pFsm:%p, index:%ld, isWeak:%d, code:%d, state:%d %s, beginIndex :%ld\n", pFsm,
+           cbMeta.index, cbMeta.isWeak, cbMeta.code, cbMeta.state, syncUtilState2String(cbMeta.state), beginIndex);
   syncRpcMsgLog2(logBuf, (SRpcMsg *)pMsg);
 
   SRpcMsg rpcMsg = {.msgType = pMsg->msgType, .contLen = pMsg->contLen};
@@ -459,16 +423,15 @@ static void vnodeSyncCommitMsg(SSyncFSM *pFsm, const SRpcMsg *pMsg, SFsmCbMeta c
 
 static void vnodeSyncPreCommitMsg(SSyncFSM *pFsm, const SRpcMsg *pMsg, SFsmCbMeta cbMeta) {
   char logBuf[256] = {0};
-  snprintf(logBuf, sizeof(logBuf),
-           "==callback== ==PreCommitCb== pFsm:%p, index:%ld, isWeak:%d, code:%d, state:%d %s \n", pFsm, cbMeta.index,
-           cbMeta.isWeak, cbMeta.code, cbMeta.state, syncUtilState2String(cbMeta.state));
+  snprintf(logBuf, sizeof(logBuf), "preCommitCb== pFsm:%p, index:%ld, isWeak:%d, code:%d, state:%d %s \n", pFsm,
+           cbMeta.index, cbMeta.isWeak, cbMeta.code, cbMeta.state, syncUtilState2String(cbMeta.state));
   syncRpcMsgLog2(logBuf, (SRpcMsg *)pMsg);
 }
 
 static void vnodeSyncRollBackMsg(SSyncFSM *pFsm, const SRpcMsg *pMsg, SFsmCbMeta cbMeta) {
   char logBuf[256] = {0};
-  snprintf(logBuf, sizeof(logBuf), "==callback== ==RollBackCb== pFsm:%p, index:%ld, isWeak:%d, code:%d, state:%d %s \n",
-           pFsm, cbMeta.index, cbMeta.isWeak, cbMeta.code, cbMeta.state, syncUtilState2String(cbMeta.state));
+  snprintf(logBuf, sizeof(logBuf), "rollBackCb== pFsm:%p, index:%ld, isWeak:%d, code:%d, state:%d %s \n", pFsm,
+           cbMeta.index, cbMeta.isWeak, cbMeta.code, cbMeta.state, syncUtilState2String(cbMeta.state));
   syncRpcMsgLog2(logBuf, (SRpcMsg *)pMsg);
 }
 

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/executorMain.c

@@ -27,6 +27,10 @@ static TdThreadOnce initPoolOnce = PTHREAD_ONCE_INIT;
 int32_t             exchangeObjRefPool = -1;
 
 static void initRefPool() { exchangeObjRefPool = taosOpenRef(1024, doDestroyExchangeOperatorInfo); }
+static void cleanupRefPool() {
+  int32_t ref = atomic_val_compare_exchange_32(&exchangeObjRefPool, exchangeObjRefPool, 0);
+  taosCloseRef(ref);
+}
 
 int32_t qCreateExecTask(SReadHandle* readHandle, int32_t vgId, uint64_t taskId, SSubplan* pSubplan,
                         qTaskInfo_t* pTaskInfo, DataSinkHandle* handle, const char* sql, EOPTR_EXEC_MODEL model) {
@@ -34,7 +38,7 @@ int32_t qCreateExecTask(SReadHandle* readHandle, int32_t vgId, uint64_t taskId,
   SExecTaskInfo** pTask = (SExecTaskInfo**)pTaskInfo;
 
   taosThreadOnce(&initPoolOnce, initRefPool);
-
+  atexit(cleanupRefPool);
   int32_t code = createExecTaskInfoImpl(pSubplan, pTask, readHandle, taskId, sql, model);
   if (code != TSDB_CODE_SUCCESS) {
     goto _error;

source/libs/executor/src/scanoperator.c

@@ -1375,6 +1375,7 @@ static void destroySysScanOperator(void* param, int32_t numOfOutput) {
   }
 
   taosArrayDestroy(pInfo->scanCols);
+  taosMemoryFreeClear(pInfo->pUser);
 }
 
 static int32_t getSysTableDbNameColId(const char* pTable) {

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);
   }
 

source/libs/function/src/builtinsimpl.c

@@ -92,10 +92,12 @@ typedef struct SStddevRes {
   union {
     double   quadraticDSum;
     int64_t  quadraticISum;
+    uint64_t quadraticUSum;
   };
   union {
     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/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/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;
@@ -330,7 +330,7 @@ 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: {
@@ -447,7 +447,7 @@ 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) {
@@ -467,7 +467,7 @@ _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));
@@ -499,7 +499,7 @@ 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:
 
@@ -698,7 +698,7 @@ 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);
 
@@ -749,7 +749,7 @@ 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;
@@ -770,7 +770,7 @@ _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"
@@ -37,7 +37,7 @@ 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));
       }
@@ -51,7 +51,7 @@ 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_CREATE_TABLE_RSP: {
@@ -308,7 +320,7 @@ 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);              
 
@@ -325,7 +337,7 @@ 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);
@@ -362,65 +374,24 @@ _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));
-
-  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;
-  }
+  SCH_ERR_RET(schProcessOnCbBegin(&pJob, &pTask, pParam->queryId, pParam->refId, pParam->taskId));
   
-  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);
 }
 
@@ -451,6 +422,37 @@ int32_t schHandleLinkBrokenCallback(void *param, SDataBuf *pMsg, int32_t 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 schMakeCallbackParam(SSchJob *pJob, SSchTask *pTask, int32_t msgType, bool isHb, SSchTrans *trans, void **pParam) {
   if (!isHb) {
     SSchTaskCallbackParam *param = taosMemoryCalloc(1, sizeof(SSchTaskCallbackParam));
@@ -692,36 +694,6 @@ _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(schInitJob(pJobId, pReq));
 
-  SCH_ERR_JRET(schExecJobImpl(pReq, pJob, true));
+  SCH_ERR_JRET(schHandleOpBeginEvent(*pJobId, &pJob, SCH_OP_EXEC, pReq));
 
-_return:
-
-  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");
-
-  int32_t code = 0;
-  if (NULL == fp || NULL == param) {
-    SCH_ERR_JRET(TSDB_CODE_QRY_INVALID_INPUT);
-  }
-
-  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));
+  SSchJob *pJob = NULL;
 
-  pJob->userRes.fetchFp = fp;
-  pJob->userRes.userParam = param;
+  SCH_ERR_JRET(schHandleOpBeginEvent(jobId, &pJob, SCH_OP_FETCH, pReq));
 
-  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/os/src/osFile.c

@@ -58,6 +58,15 @@ typedef struct TdFile {
 
 #define FILE_WITH_LOCK 1
 
+typedef struct AutoDelFile * AutoDelFilePtr;
+typedef struct AutoDelFile {
+  char           *name;
+  AutoDelFilePtr  lastAutoDelFilePtr;
+} AutoDelFile;
+static TdThreadMutex autoDelFileLock;
+static AutoDelFilePtr nowAutoDelFilePtr = NULL;
+static TdThreadOnce autoDelFileInit = PTHREAD_ONCE_INIT;
+
 void taosGetTmpfilePath(const char *inputTmpDir, const char *fileNamePrefix, char *dstPath) {
 #ifdef WINDOWS
   const char *tdengineTmpFileNamePrefix = "tdengine-";
@@ -238,7 +247,33 @@ int32_t taosDevInoFile(TdFilePtr pFile, int64_t *stDev, int64_t *stIno) {
   return 0;
 }
 
-void autoDelFileListAdd(const char *path) { return; }
+void autoDelFileList() {
+  taosThreadMutexLock(&autoDelFileLock);
+  while (nowAutoDelFilePtr != NULL) {
+    taosRemoveFile(nowAutoDelFilePtr->name);
+    AutoDelFilePtr tmp = nowAutoDelFilePtr->lastAutoDelFilePtr;
+    taosMemoryFree(nowAutoDelFilePtr->name);
+    taosMemoryFree(nowAutoDelFilePtr);
+    nowAutoDelFilePtr = tmp;
+  }
+  taosThreadMutexUnlock(&autoDelFileLock);
+  taosThreadMutexDestroy(&autoDelFileLock);
+}
+
+void autoDelFileListInit() {
+  taosThreadMutexInit(&autoDelFileLock, NULL);
+  atexit(autoDelFileList);
+}
+
+void autoDelFileListAdd(const char *path) {
+  taosThreadOnce(&autoDelFileInit, autoDelFileListInit);
+  taosThreadMutexLock(&autoDelFileLock);
+  AutoDelFilePtr tmp = taosMemoryMalloc(sizeof(AutoDelFile));
+  tmp->lastAutoDelFilePtr = nowAutoDelFilePtr;
+  tmp->name = taosMemoryStrDup(path);
+  nowAutoDelFilePtr = tmp;
+  taosThreadMutexUnlock(&autoDelFileLock);
+}
 
 TdFilePtr taosOpenFile(const char *path, int32_t tdFileOptions) {
   int   fd = -1;
@@ -283,10 +318,6 @@ TdFilePtr taosOpenFile(const char *path, int32_t tdFileOptions) {
     }
   }
 
-  if (tdFileOptions & TD_FILE_AUTO_DEL) {
-    autoDelFileListAdd(path);
-  }
-
   TdFilePtr pFile = (TdFilePtr)taosMemoryMalloc(sizeof(TdFile));
   if (pFile == NULL) {
     if (fd >= 0) close(fd);
@@ -299,6 +330,9 @@ TdFilePtr taosOpenFile(const char *path, int32_t tdFileOptions) {
   pFile->fd = fd;
   pFile->fp = fp;
   pFile->refId = 0;
+  if (tdFileOptions & TD_FILE_AUTO_DEL) {
+    autoDelFileListAdd(path);
+  }
   return pFile;
 }
 

source/os/src/osMemory.c

@@ -282,14 +282,14 @@ void *taosMemoryRealloc(void *ptr, int32_t size) {
 #endif
 }
 
-void *taosMemoryStrDup(void *ptr) {
+void *taosMemoryStrDup(const char *ptr) {
 #ifdef USE_TD_MEMORY
   if (ptr == NULL) return NULL;
 
   TdMemoryInfoPtr pTdMemoryInfo = (TdMemoryInfoPtr)((char *)ptr - sizeof(TdMemoryInfo));
   assert(pTdMemoryInfo->symbol == TD_MEMORY_SYMBOL);
 
-  void *tmp = tstrdup((const char *)pTdMemoryInfo);
+  void *tmp = tstrdup(pTdMemoryInfo);
   if (tmp == NULL) return NULL;
 
   memcpy(tmp, pTdMemoryInfo, sizeof(TdMemoryInfo));
@@ -297,7 +297,7 @@ void *taosMemoryStrDup(void *ptr) {
 
   return (char *)tmp + sizeof(TdMemoryInfo);
 #else
-  return tstrdup((const char *)ptr);
+  return tstrdup(ptr);
 #endif
 }
 

source/util/src/terror.c

@@ -394,6 +394,7 @@ 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")