Merge branch 'develop' into feature/query

src/client/inc/tsclient.h

@@ -52,12 +52,20 @@ typedef struct STableComInfo {
   int32_t rowSize;
 } STableComInfo;
 
+typedef struct SCMCorVgroupInfo {
+  int32_t version;
+  int8_t inUse;
+  int8_t  numOfEps;
+  SEpAddr epAddr[TSDB_MAX_REPLICA];
+} SCMCorVgroupInfo;
+
 typedef struct STableMeta {
   STableComInfo tableInfo;
   uint8_t       tableType;
   int16_t       sversion;
   int16_t       tversion;
-  SCMVgroupInfo vgroupInfo;
+  SCMVgroupInfo  vgroupInfo;
+  SCMCorVgroupInfo  corVgroupInfo;
   int32_t       sid;       // the index of one table in a virtual node
   uint64_t      uid;       // unique id of a table
   SSchema       schema[];  // if the table is TSDB_CHILD_TABLE, schema is acquired by super table meta info
@@ -456,7 +464,8 @@ extern void *    tscQhandle;
 extern int       tscKeepConn[];
 extern int       tsInsertHeadSize;
 extern int       tscNumOfThreads;
-extern SRpcEpSet tscMgmtEpSet;
+  
+extern SRpcCorEpSet tscMgmtEpSet;
 
 extern int (*tscBuildMsg[TSDB_SQL_MAX])(SSqlObj *pSql, SSqlInfo *pInfo);
 

src/client/src/tscSchemaUtil.c

@@ -140,7 +140,15 @@ struct SSchema tscGetTbnameColumnSchema() {
   strcpy(s.name, TSQL_TBNAME_L);
   return s;
 }
-
+static void tscInitCorVgroupInfo(SCMCorVgroupInfo *corVgroupInfo, SCMVgroupInfo *vgroupInfo) {
+  corVgroupInfo->version = 0;
+  corVgroupInfo->inUse = 0;
+  corVgroupInfo->numOfEps = vgroupInfo->numOfEps;
+  for (int32_t i = 0; i < corVgroupInfo->numOfEps; i++) {
+    strncpy(corVgroupInfo->epAddr[i].fqdn, vgroupInfo->epAddr[i].fqdn, TSDB_FQDN_LEN);
+    corVgroupInfo->epAddr[i].port = vgroupInfo->epAddr[i].port;
+  }
+}
 STableMeta* tscCreateTableMetaFromMsg(STableMetaMsg* pTableMetaMsg, size_t* size) {
   assert(pTableMetaMsg != NULL);
   
@@ -157,6 +165,9 @@ STableMeta* tscCreateTableMetaFromMsg(STableMetaMsg* pTableMetaMsg, size_t* size
   pTableMeta->sid = pTableMetaMsg->sid;
   pTableMeta->uid = pTableMetaMsg->uid;
   pTableMeta->vgroupInfo = pTableMetaMsg->vgroup;
+
+  tscInitCorVgroupInfo(&pTableMeta->corVgroupInfo, &pTableMeta->vgroupInfo);
+
   pTableMeta->sversion = pTableMetaMsg->sversion;
   pTableMeta->tversion = pTableMetaMsg->tversion;
   

src/client/src/tscServer.c

@@ -26,10 +26,11 @@
 #include "ttime.h"
 #include "ttimer.h"
 #include "tutil.h"
+#include "tlockfree.h"
 
 #define TSC_MGMT_VNODE 999
 
-SRpcEpSet  tscMgmtEpSet;
+SRpcCorEpSet  tscMgmtEpSet;
 SRpcEpSet  tscDnodeEpSet;
 
 int (*tscBuildMsg[TSDB_SQL_MAX])(SSqlObj *pSql, SSqlInfo *pInfo) = {0};
@@ -51,37 +52,78 @@ static void tscSetDnodeEpSet(SSqlObj* pSql, SCMVgroupInfo* pVgroupInfo) {
     pEpSet->numOfEps = 0;
     return;
   }
-  
+
   pEpSet->numOfEps = pVgroupInfo->numOfEps;
   for(int32_t i = 0; i < pVgroupInfo->numOfEps; ++i) {
     strcpy(pEpSet->fqdn[i], pVgroupInfo->epAddr[i].fqdn);
     pEpSet->port[i] = pVgroupInfo->epAddr[i].port;
   }
 }
-
-void tscPrintMgmtEp() {
-  if (tscMgmtEpSet.numOfEps <= 0) {
-    tscError("invalid mnode EP list:%d", tscMgmtEpSet.numOfEps);
-  } else {
-    for (int i = 0; i < tscMgmtEpSet.numOfEps; ++i) {
-      tscDebug("mnode index:%d %s:%d", i, tscMgmtEpSet.fqdn[i], tscMgmtEpSet.port[i]);
-    }
+static void tscDumpMgmtEpSet(SRpcEpSet *epSet) {
+  taosCorBeginRead(&tscMgmtEpSet.version);
+  *epSet = tscMgmtEpSet.epSet;
+  taosCorEndRead(&tscMgmtEpSet.version);
+}  
+static void tscEpSetHtons(SRpcEpSet *s) {
+   for (int32_t i = 0; i < s->numOfEps; i++) {
+      s->port[i] = htons(s->port[i]);    
+   }
+} 
+bool tscEpSetIsEqual(SRpcEpSet *s1, SRpcEpSet *s2) {
+   if (s1->numOfEps != s2->numOfEps || s1->inUse != s2->inUse) {
+     return false;
+   } 
+   for (int32_t i = 0; i < s1->numOfEps; i++) {
+     if (s1->port[i] != s2->port[i] 
+        || strncmp(s1->fqdn[i], s2->fqdn[i], TSDB_FQDN_LEN) != 0)
+        return false;
+   }
+   return true;
+}
+void tscUpdateMgmtEpSet(SRpcEpSet *pEpSet) {
+  // no need to update if equal
+  taosCorBeginWrite(&tscMgmtEpSet.version);
+  tscMgmtEpSet.epSet = *pEpSet;
+  taosCorEndWrite(&tscMgmtEpSet.version);
+}
+static void tscDumpEpSetFromVgroupInfo(SCMCorVgroupInfo *pVgroupInfo, SRpcEpSet *pEpSet) {
+  if (pVgroupInfo == NULL) { return;}
+  taosCorBeginRead(&pVgroupInfo->version);
+  int8_t inUse = pVgroupInfo->inUse;
+  pEpSet->inUse = (inUse >= 0 && inUse < TSDB_MAX_REPLICA) ? inUse: 0; 
+  pEpSet->numOfEps = pVgroupInfo->numOfEps;  
+  for (int32_t i = 0; i < pVgroupInfo->numOfEps; ++i) {
+    strncpy(pEpSet->fqdn[i], pVgroupInfo->epAddr[i].fqdn, TSDB_FQDN_LEN);
+    pEpSet->port[i] = pVgroupInfo->epAddr[i].port;
   }
+  taosCorEndRead(&pVgroupInfo->version);
 }
 
-void tscSetMgmtEpSet(SRpcEpSet *pEpSet) {
-  tscMgmtEpSet.numOfEps = pEpSet->numOfEps;
-  tscMgmtEpSet.inUse = pEpSet->inUse;
-  for (int32_t i = 0; i < tscMgmtEpSet.numOfEps; ++i) {
-    tscMgmtEpSet.port[i] = htons(pEpSet->port[i]);
+static void tscUpdateVgroupInfo(SSqlObj *pObj, SRpcEpSet *pEpSet) {
+  SSqlCmd *pCmd = &pObj->cmd;
+  STableMetaInfo *pTableMetaInfo = tscGetTableMetaInfoFromCmd(pCmd, pCmd->clauseIndex, 0);
+  if (pTableMetaInfo == NULL || pTableMetaInfo->pTableMeta == NULL) { return;}
+  SCMCorVgroupInfo *pVgroupInfo = &pTableMetaInfo->pTableMeta->corVgroupInfo;
+
+  taosCorBeginWrite(&pVgroupInfo->version);
+  //TODO(dengyihao), dont care vgid 
+  pVgroupInfo->inUse = pEpSet->inUse;
+  pVgroupInfo->numOfEps = pEpSet->numOfEps;
+  for (int32_t i = 0; pVgroupInfo->numOfEps; i++) {
+    strncpy(pVgroupInfo->epAddr[i].fqdn, pEpSet->fqdn[i], TSDB_FQDN_LEN);
+    pVgroupInfo->epAddr[i].port = pEpSet->port[i];
   }
+  taosCorEndWrite(&pVgroupInfo->version);
 }
-
-void tscUpdateEpSet(void *ahandle, SRpcEpSet *pEpSet) {
-  tscMgmtEpSet = *pEpSet;
-  tscDebug("mnode EP list is changed for ufp is called, numOfEps:%d inUse:%d", tscMgmtEpSet.numOfEps, tscMgmtEpSet.inUse);
-  for (int32_t i = 0; i < tscMgmtEpSet.numOfEps; ++i) {
-    tscDebug("index:%d fqdn:%s port:%d", i, tscMgmtEpSet.fqdn[i], tscMgmtEpSet.port[i]);
+void tscPrintMgmtEp() {
+  SRpcEpSet dump;
+  tscDumpMgmtEpSet(&dump);
+  if (dump.numOfEps <= 0) {
+    tscError("invalid mnode EP list:%d", dump.numOfEps);
+  } else {
+    for (int i = 0; i < dump.numOfEps; ++i) {
+      tscDebug("mnode index:%d %s:%d", i, dump.fqdn[i], dump.port[i]);
+    }
   }
 }
 
@@ -95,7 +137,9 @@ void tscUpdateEpSet(void *ahandle, SRpcEpSet *pEpSet) {
 UNUSED_FUNC
 static int32_t tscGetMgmtConnMaxRetryTimes() {
   int32_t factor = 2;
-  return tscMgmtEpSet.numOfEps * factor;
+  SRpcEpSet dump;
+  tscDumpMgmtEpSet(&dump);
+  return dump.numOfEps * factor;
 }
 
 void tscProcessHeartBeatRsp(void *param, TAOS_RES *tres, int code) {
@@ -111,9 +155,11 @@ void tscProcessHeartBeatRsp(void *param, TAOS_RES *tres, int code) {
 
   if (code == 0) {
     SCMHeartBeatRsp *pRsp = (SCMHeartBeatRsp *)pRes->pRsp;
-    SRpcEpSet *      pEpSet = &pRsp->epSet;
-    if (pEpSet->numOfEps > 0) 
-      tscSetMgmtEpSet(pEpSet);
+    SRpcEpSet *      epSet = &pRsp->epSet;
+    if (epSet->numOfEps > 0) {
+      tscEpSetHtons(epSet);
+      tscUpdateMgmtEpSet(epSet);
+    } 
 
     pSql->pTscObj->connId = htonl(pRsp->connId);
 
@@ -185,7 +231,7 @@ int tscSendMsgToServer(SSqlObj *pSql) {
 
   // set the mgmt ip list
   if (pSql->cmd.command >= TSDB_SQL_MGMT) {
-    pSql->epSet = tscMgmtEpSet;
+    tscDumpMgmtEpSet(&pSql->epSet);
   }
 
   memcpy(pMsg, pSql->cmd.payload, pSql->cmd.payloadLen);
@@ -236,10 +282,16 @@ void tscProcessMsgFromServer(SRpcMsg *rpcMsg, SRpcEpSet *pEpSet) {
     return;
   }
 
-  if (pCmd->command < TSDB_SQL_MGMT) {
-    if (pEpSet) pSql->epSet = *pEpSet;
-  } else {
-    if (pEpSet) tscMgmtEpSet = *pEpSet;
+  if (pEpSet) { 
+    //SRpcEpSet dump; 
+    tscEpSetHtons(pEpSet); 
+    if (tscEpSetIsEqual(&pSql->epSet, pEpSet)) {
+      if(pCmd->command < TSDB_SQL_MGMT)  { 
+        tscUpdateVgroupInfo(pSql, pEpSet); 
+      } else {
+        tscUpdateMgmtEpSet(pEpSet);
+    }
+    }
   }
 
   STableMetaInfo *pTableMetaInfo = tscGetTableMetaInfoFromCmd(pCmd, pCmd->clauseIndex, 0);
@@ -400,7 +452,8 @@ int tscProcessSql(SSqlObj *pSql) {
       return pSql->res.code;
     }
   } else if (pCmd->command < TSDB_SQL_LOCAL) {
-    pSql->epSet = tscMgmtEpSet;
+
+    //pSql->epSet = tscMgmtEpSet;
   } else {  // local handler
     return (*tscProcessMsgRsp[pCmd->command])(pSql);
   }
@@ -504,8 +557,8 @@ int tscBuildSubmitMsg(SSqlObj *pSql, SSqlInfo *pInfo) {
 
   // pSql->cmd.payloadLen is set during copying data into payload
   pSql->cmd.msgType = TSDB_MSG_TYPE_SUBMIT;
-  tscSetDnodeEpSet(pSql, &pTableMeta->vgroupInfo);
-  
+  tscDumpEpSetFromVgroupInfo(&pTableMeta->corVgroupInfo, &pSql->epSet);
+
   tscDebug("%p build submit msg, vgId:%d numOfTables:%d numberOfEP:%d", pSql, vgId, pSql->cmd.numOfTablesInSubmit,
       pSql->epSet.numOfEps);
   return TSDB_CODE_SUCCESS;
@@ -546,11 +599,11 @@ static char *doSerializeTableInfo(SQueryTableMsg* pQueryMsg, SSqlObj *pSql, char
     } else {
       pVgroupInfo = &pTableMeta->vgroupInfo;
     }
-
     tscSetDnodeEpSet(pSql, pVgroupInfo);
+
     if (pVgroupInfo != NULL) {
       pQueryMsg->head.vgId = htonl(pVgroupInfo->vgId);
-    }
+    } 
 
     STableIdInfo *pTableIdInfo = (STableIdInfo *)pMsg;
     pTableIdInfo->tid = htonl(pTableMeta->sid);
@@ -567,8 +620,8 @@ static char *doSerializeTableInfo(SQueryTableMsg* pQueryMsg, SSqlObj *pSql, char
     tscDebug("%p query on stable, vgIndex:%d, numOfVgroups:%d", pSql, index, numOfVgroups);
 
     SVgroupTableInfo* pTableIdList = taosArrayGet(pTableMetaInfo->pVgroupTables, index);
-    
-    // set the vgroup info
+
+    // set the vgroup info 
     tscSetDnodeEpSet(pSql, &pTableIdList->vgInfo);
     pQueryMsg->head.vgId = htonl(pTableIdList->vgInfo.vgId);
     
@@ -1115,11 +1168,11 @@ int32_t tscBuildShowMsg(SSqlObj *pSql, SSqlInfo *pInfo) {
       pShowMsg->payloadLen = htons(pPattern->n);
     }
   } else {
-    SSQLToken *pIpAddr = &pShowInfo->prefix;
-    assert(pIpAddr->n > 0 && pIpAddr->type > 0);
+    SSQLToken *pEpAddr = &pShowInfo->prefix;
+    assert(pEpAddr->n > 0 && pEpAddr->type > 0);
 
-    strncpy(pShowMsg->payload, pIpAddr->z, pIpAddr->n);
-    pShowMsg->payloadLen = htons(pIpAddr->n);
+    strncpy(pShowMsg->payload, pEpAddr->z, pEpAddr->n);
+    pShowMsg->payloadLen = htons(pEpAddr->n);
   }
 
   pCmd->payloadLen = sizeof(SCMShowMsg) + pShowMsg->payloadLen;
@@ -1302,7 +1355,7 @@ int tscBuildUpdateTagMsg(SSqlObj* pSql, SSqlInfo *pInfo) {
   SQueryInfo *    pQueryInfo = tscGetQueryInfoDetail(pCmd, 0);
   STableMetaInfo *pTableMetaInfo = tscGetMetaInfo(pQueryInfo, 0);
 
-  tscSetDnodeEpSet(pSql, &pTableMetaInfo->pTableMeta->vgroupInfo);
+  tscDumpEpSetFromVgroupInfo(&pTableMetaInfo->pTableMeta->corVgroupInfo, &pSql->epSet);
 
   return TSDB_CODE_SUCCESS;
 }
@@ -1826,13 +1879,14 @@ int tscProcessSTableVgroupRsp(SSqlObj *pSql) {
 
     memcpy(pInfo->vgroupList, pVgroupInfo, size);
     for (int32_t j = 0; j < pInfo->vgroupList->numOfVgroups; ++j) {
+      //just init, no need to lock
       SCMVgroupInfo *pVgroups = &pInfo->vgroupList->vgroups[j];
-
       pVgroups->vgId = htonl(pVgroups->vgId);
       assert(pVgroups->numOfEps >= 1);
 
       for (int32_t k = 0; k < pVgroups->numOfEps; ++k) {
         pVgroups->epAddr[k].port = htons(pVgroups->epAddr[k].port);
+
       }
 
       pMsg += size;
@@ -1925,8 +1979,10 @@ int tscProcessConnectRsp(SSqlObj *pSql) {
   assert(len <= sizeof(pObj->db));
   tstrncpy(pObj->db, temp, sizeof(pObj->db));
   
-  if (pConnect->epSet.numOfEps > 0) 
-    tscSetMgmtEpSet(&pConnect->epSet);
+  if (pConnect->epSet.numOfEps > 0) {
+    tscEpSetHtons(&pConnect->epSet);
+    tscUpdateMgmtEpSet(&pConnect->epSet);
+  } 
 
   strcpy(pObj->sversion, pConnect->serverVersion);
   pObj->writeAuth = pConnect->writeAuth;

src/client/src/tscSql.c

@@ -63,7 +63,7 @@ SSqlObj *taosConnectImpl(const char *ip, const char *user, const char *pass, con
 
   if (ip) {
     if (tscSetMgmtEpSetFromCfg(ip, NULL) < 0) return NULL;
-    if (port) tscMgmtEpSet.port[0] = port;
+    if (port) tscMgmtEpSet.epSet.port[0] = port;
   } 
  
   void *pDnodeConn = NULL;

src/client/src/tscSystem.c

@@ -41,7 +41,7 @@ int tscNumOfThreads;
 
 static pthread_once_t tscinit = PTHREAD_ONCE_INIT;
 void taosInitNote(int numOfNoteLines, int maxNotes, char* lable);
-void tscUpdateEpSet(void *ahandle, SRpcEpSet *pEpSet);
+//void tscUpdateEpSet(void *ahandle, SRpcEpSet *pEpSet);
 
 void tscCheckDiskUsage(void *UNUSED_PARAM(para), void* UNUSED_PARAM(param)) {
   taosGetDisk();

src/client/src/tscUtil.c

@@ -2146,16 +2146,19 @@ char* strdup_throw(const char* str) {
 }
 
 int tscSetMgmtEpSetFromCfg(const char *first, const char *second) {
-  tscMgmtEpSet.numOfEps = 0;
-  tscMgmtEpSet.inUse = 0;
+  // init mgmt ip set 
+  tscMgmtEpSet.version = 0;
+  SRpcEpSet *mgmtEpSet = &(tscMgmtEpSet.epSet);
+  mgmtEpSet->numOfEps = 0;
+  mgmtEpSet->inUse = 0;
 
   if (first && first[0] != 0) {
     if (strlen(first) >= TSDB_EP_LEN) {
       terrno = TSDB_CODE_TSC_INVALID_FQDN;
       return -1;
     }
-    taosGetFqdnPortFromEp(first, tscMgmtEpSet.fqdn[tscMgmtEpSet.numOfEps], &tscMgmtEpSet.port[tscMgmtEpSet.numOfEps]);
-    tscMgmtEpSet.numOfEps++;
+    taosGetFqdnPortFromEp(first, mgmtEpSet->fqdn[mgmtEpSet->numOfEps], &(mgmtEpSet->port[mgmtEpSet->numOfEps]));
+    mgmtEpSet->numOfEps++;
   }
 
   if (second && second[0] != 0) {
@@ -2163,11 +2166,11 @@ int tscSetMgmtEpSetFromCfg(const char *first, const char *second) {
       terrno = TSDB_CODE_TSC_INVALID_FQDN;
       return -1;
     }
-    taosGetFqdnPortFromEp(second, tscMgmtEpSet.fqdn[tscMgmtEpSet.numOfEps], &tscMgmtEpSet.port[tscMgmtEpSet.numOfEps]);
-    tscMgmtEpSet.numOfEps++;
+    taosGetFqdnPortFromEp(second, mgmtEpSet->fqdn[mgmtEpSet->numOfEps], &(mgmtEpSet->port[mgmtEpSet->numOfEps]));
+    mgmtEpSet->numOfEps++;
   }
 
-  if ( tscMgmtEpSet.numOfEps == 0) {
+  if (mgmtEpSet->numOfEps == 0) {
     terrno = TSDB_CODE_TSC_INVALID_FQDN;
     return -1;
   }

src/inc/trpc.h

@@ -35,6 +35,11 @@ typedef struct SRpcEpSet {
   char      fqdn[TSDB_MAX_REPLICA][TSDB_FQDN_LEN];
 } SRpcEpSet;
 
+typedef struct SRpcCorEpSet {
+  int32_t version; 
+  SRpcEpSet epSet; 
+} SRpcCorEpSet;
+
 typedef struct SRpcConnInfo {
   uint32_t  clientIp;
   uint16_t  clientPort;

src/mnode/src/mnodeSdb.c

@@ -77,11 +77,6 @@ typedef struct {
   pthread_mutex_t mutex;
 } SSdbObject;
 
-typedef struct {
-  int32_t rowSize;
-  void *  row;
-} SSdbRow;
-
 typedef struct {
   pthread_t thread;
   int32_t   workerId;
@@ -419,32 +414,28 @@ void sdbDecRef(void *handle, void *pObj) {
   }
 }
 
-static SSdbRow *sdbGetRowMeta(SSdbTable *pTable, void *key) {
+static void *sdbGetRowMeta(SSdbTable *pTable, void *key) {
   if (pTable == NULL) return NULL;
 
   int32_t keySize = sizeof(int32_t);
   if (pTable->keyType == SDB_KEY_STRING || pTable->keyType == SDB_KEY_VAR_STRING) {
     keySize = strlen((char *)key);
   }
-  
-  return taosHashGet(pTable->iHandle, key, keySize);
+
+  void **ppRow = (void **)taosHashGet(pTable->iHandle, key, keySize);
+  if (ppRow == NULL) return NULL;
+  return *ppRow;
 }
 
-static SSdbRow *sdbGetRowMetaFromObj(SSdbTable *pTable, void *key) {
+static void *sdbGetRowMetaFromObj(SSdbTable *pTable, void *key) {
   return sdbGetRowMeta(pTable, sdbGetObjKey(pTable, key));
 }
 
 void *sdbGetRow(void *handle, void *key) {
-  SSdbTable *pTable = (SSdbTable *)handle;
-  int32_t keySize = sizeof(int32_t);
-  if (pTable->keyType == SDB_KEY_STRING || pTable->keyType == SDB_KEY_VAR_STRING) {
-    keySize = strlen((char *)key);
-  }
-  
-  SSdbRow *pMeta = taosHashGet(pTable->iHandle, key, keySize);
-  if (pMeta) {
-    sdbIncRef(pTable, pMeta->row);
-    return pMeta->row;
+  void *pRow = sdbGetRowMeta(handle, key);
+  if (pRow) {
+    sdbIncRef(handle, pRow);
+    return pRow;
   } else {
     return NULL;
   }
@@ -455,10 +446,6 @@ static void *sdbGetRowFromObj(SSdbTable *pTable, void *key) {
 }
 
 static int32_t sdbInsertHash(SSdbTable *pTable, SSdbOper *pOper) {
-  SSdbRow rowMeta;
-  rowMeta.rowSize = pOper->rowSize;
-  rowMeta.row = pOper->pObj;
-
   void *  key = sdbGetObjKey(pTable, pOper->pObj);
   int32_t keySize = sizeof(int32_t);
 
@@ -466,7 +453,7 @@ static int32_t sdbInsertHash(SSdbTable *pTable, SSdbOper *pOper) {
     keySize = strlen((char *)key);
   }
 
-  taosHashPut(pTable->iHandle, key, keySize, &rowMeta, sizeof(SSdbRow));
+  taosHashPut(pTable->iHandle, key, keySize, &pOper->pObj, sizeof(void **));
 
   sdbIncRef(pTable, pOper->pObj);
   atomic_add_fetch_32(&pTable->numOfRows, 1);
@@ -586,17 +573,17 @@ static int sdbWrite(void *param, void *data, int type) {
     code = (*pTable->decodeFp)(&oper);
     return sdbInsertHash(pTable, &oper);
   } else if (action == SDB_ACTION_DELETE) {
-    SSdbRow *rowMeta = sdbGetRowMeta(pTable, pHead->cont);
-    if (rowMeta == NULL || rowMeta->row == NULL) {
+    void *pRow = sdbGetRowMeta(pTable, pHead->cont);
+    if (pRow == NULL) {
       sdbError("table:%s, failed to get object:%s from wal while dispose delete action", pTable->tableName,
                pHead->cont);
       return TSDB_CODE_SUCCESS;
     }
-    SSdbOper oper = {.table = pTable, .pObj = rowMeta->row};
+    SSdbOper oper = {.table = pTable, .pObj = pRow};
     return sdbDeleteHash(pTable, &oper);
   } else if (action == SDB_ACTION_UPDATE) {
-    SSdbRow *rowMeta = sdbGetRowMeta(pTable, pHead->cont);
-    if (rowMeta == NULL || rowMeta->row == NULL) {
+    void *pRow = sdbGetRowMeta(pTable, pHead->cont);
+    if (pRow == NULL) {
       sdbError("table:%s, failed to get object:%s from wal while dispose update action", pTable->tableName,
                pHead->cont);
       return TSDB_CODE_SUCCESS;
@@ -675,18 +662,12 @@ int32_t sdbDeleteRow(SSdbOper *pOper) {
   SSdbTable *pTable = (SSdbTable *)pOper->table;
   if (pTable == NULL) return TSDB_CODE_MND_SDB_INVALID_TABLE_TYPE;
 
-  SSdbRow *pMeta = sdbGetRowMetaFromObj(pTable, pOper->pObj);
-  if (pMeta == NULL) {
+  void *pRow = sdbGetRowMetaFromObj(pTable, pOper->pObj);
+  if (pRow == NULL) {
     sdbDebug("table:%s, record is not there, delete failed", pTable->tableName);
     return TSDB_CODE_MND_SDB_OBJ_NOT_THERE;
   }
 
-  void *pMetaRow = pMeta->row;
-  if (pMetaRow == NULL) {
-    sdbError("table:%s, record meta is null", pTable->tableName);
-    return TSDB_CODE_MND_SDB_INVAID_META_ROW;
-  }
-
   sdbIncRef(pTable, pOper->pObj);
 
   int32_t code = sdbDeleteHash(pTable, pOper);
@@ -728,18 +709,12 @@ int32_t sdbUpdateRow(SSdbOper *pOper) {
   SSdbTable *pTable = (SSdbTable *)pOper->table;
   if (pTable == NULL) return TSDB_CODE_MND_SDB_INVALID_TABLE_TYPE;
 
-  SSdbRow *pMeta = sdbGetRowMetaFromObj(pTable, pOper->pObj);
-  if (pMeta == NULL) {
+  void *pRow = sdbGetRowMetaFromObj(pTable, pOper->pObj);
+  if (pRow == NULL) {
     sdbDebug("table:%s, record is not there, update failed", pTable->tableName);
     return TSDB_CODE_MND_SDB_OBJ_NOT_THERE;
   }
 
-  void *pMetaRow = pMeta->row;
-  if (pMetaRow == NULL) {
-    sdbError("table:%s, record meta is null", pTable->tableName);
-    return TSDB_CODE_MND_SDB_INVAID_META_ROW;
-  }
-
   int32_t code = sdbUpdateHash(pTable, pOper);
   if (code != TSDB_CODE_SUCCESS) {
     sdbError("table:%s, failed to update hash", pTable->tableName);
@@ -789,14 +764,14 @@ void *sdbFetchRow(void *handle, void *pNode, void **ppRow) {
     return NULL;
   }
 
-  SSdbRow *pMeta = taosHashIterGet(pIter);
-  if (pMeta == NULL) {
+  void **ppMetaRow = taosHashIterGet(pIter);
+  if (ppMetaRow == NULL) {
     taosHashDestroyIter(pIter);
     return NULL;
   }
 
-  *ppRow = pMeta->row;
-  sdbIncRef(handle, pMeta->row);
+  *ppRow = *ppMetaRow;
+  sdbIncRef(handle, *ppMetaRow);
 
   return pIter;
 }
@@ -846,11 +821,11 @@ void sdbCloseTable(void *handle) {
 
   SHashMutableIterator *pIter = taosHashCreateIter(pTable->iHandle);
   while (taosHashIterNext(pIter)) {
-    SSdbRow *pMeta = taosHashIterGet(pIter);
-    if (pMeta == NULL) continue;
+    void **ppRow = taosHashIterGet(pIter);
+    if (ppRow == NULL) continue;
 
     SSdbOper oper = {
-      .pObj = pMeta->row,
+      .pObj = *ppRow,
       .table = pTable,
     };
 

src/mnode/src/mnodeTable.c

@@ -107,42 +107,41 @@ static int32_t mnodeChildTableActionInsert(SSdbOper *pOper) {
   SVgObj *pVgroup = mnodeGetVgroup(pTable->vgId);
   if (pVgroup == NULL) {
     mError("ctable:%s, not in vgId:%d", pTable->info.tableId, pTable->vgId);
-    return TSDB_CODE_MND_VGROUP_NOT_EXIST;
   }
-  mnodeDecVgroupRef(pVgroup);
 
-  SDbObj *pDb = mnodeGetDb(pVgroup->dbName);
-  if (pDb == NULL) {
-    mError("ctable:%s, vgId:%d not in db:%s", pTable->info.tableId, pVgroup->vgId, pVgroup->dbName);
-    return TSDB_CODE_MND_INVALID_DB;
-  }
-  
-  if (pDb->status != TSDB_DB_STATUS_READY) {
-    mError("db:%s, status:%d, in dropping", pDb->name, pDb->status);
-    return TSDB_CODE_MND_DB_IN_DROPPING;
+  SDbObj *pDb = NULL;
+  if (pVgroup != NULL) {
+    pDb = mnodeGetDb(pVgroup->dbName);
+    if (pDb == NULL) {
+      mError("ctable:%s, vgId:%d not in db:%s", pTable->info.tableId, pVgroup->vgId, pVgroup->dbName);
+    }
   }
-  mnodeDecDbRef(pDb);
 
-  SAcctObj *pAcct = mnodeGetAcct(pDb->acct);
-  if (pAcct == NULL) {
-    mError("ctable:%s, acct:%s not exists", pTable->info.tableId, pDb->acct);
-    return TSDB_CODE_MND_INVALID_ACCT;
+  SAcctObj *pAcct = NULL;
+  if (pDb != NULL) {
+    pAcct = mnodeGetAcct(pDb->acct);
+    if (pAcct == NULL) {
+      mError("ctable:%s, acct:%s not exists", pTable->info.tableId, pDb->acct);
+    }
   }
-  mnodeDecAcctRef(pAcct);
 
   if (pTable->info.type == TSDB_CHILD_TABLE) {
     // add ref
     pTable->superTable = mnodeGetSuperTableByUid(pTable->suid);
     mnodeAddTableIntoStable(pTable->superTable, pTable);
     grantAdd(TSDB_GRANT_TIMESERIES, pTable->superTable->numOfColumns - 1);
-    pAcct->acctInfo.numOfTimeSeries += (pTable->superTable->numOfColumns - 1);
+    if (pAcct) pAcct->acctInfo.numOfTimeSeries += (pTable->superTable->numOfColumns - 1);
   } else {
     grantAdd(TSDB_GRANT_TIMESERIES, pTable->numOfColumns - 1);
-    pAcct->acctInfo.numOfTimeSeries += (pTable->numOfColumns - 1);
+    if (pAcct) pAcct->acctInfo.numOfTimeSeries += (pTable->numOfColumns - 1);
   }
 
-  mnodeAddTableIntoDb(pDb);
-  mnodeAddTableIntoVgroup(pVgroup, pTable);
+  if (pDb) mnodeAddTableIntoDb(pDb);
+  if (pVgroup) mnodeAddTableIntoVgroup(pVgroup, pTable);
+
+  mnodeDecVgroupRef(pVgroup);
+  mnodeDecDbRef(pDb);
+  mnodeDecAcctRef(pAcct);
 
   return TSDB_CODE_SUCCESS;
 }

src/util/inc/tlockfree.h

@@ -75,7 +75,7 @@ void taosRUnLockLatch(SRWLatch *pLatch);
 
 // copy on read
 #define taosCorBeginRead(x) for (uint32_t i_ = 1; 1; ++i_) { \
-    int32_t old_ = atomic_load_32(x); \
+    int32_t old_ = atomic_add_fetch_32((x), 0); \
     if (old_ & 0x00000001) { \
       if (i_ % 1000 == 0) { \
         sched_yield(); \
@@ -84,7 +84,7 @@ void taosRUnLockLatch(SRWLatch *pLatch);
     }
 
 #define taosCorEndRead(x) \
-    if (atomic_load_32(x) == old_) { \
+    if (atomic_add_fetch_32((x), 0) == old_) { \
       break; \
     } \
   }