Merge pull request #746 from taosdata/feature/slguan

Fix crashes and other exceptions that may occur when deleting a database

src/inc/taoserror.h

@@ -134,6 +134,8 @@ extern "C" {
 #define TSDB_CODE_INVALID_SUBMIT_MSG         113
 #define TSDB_CODE_NOT_ACTIVE_TABLE           114
 #define TSDB_CODE_INVALID_TABLE_ID           115
+#define TSDB_CODE_INVALID_VNODE_STATUS       116
+#define TSDB_CODE_FAILED_TO_LOCK_RESOURCES   117
 
 #ifdef __cplusplus
 }

src/inc/tsdb.h

@@ -44,22 +44,40 @@ extern "C" {
 #define TSDB_TIME_PRECISION_MILLI_STR "ms"
 #define TSDB_TIME_PRECISION_MICRO_STR "us"
 
-enum _status {
-  TSDB_STATUS_OFFLINE,
-  TSDB_STATUS_CREATING,
-  TSDB_STATUS_UNSYNCED,
-  TSDB_STATUS_SLAVE,
-  TSDB_STATUS_MASTER,
-  TSDB_STATUS_READY,
+  enum _vnode_status {
+  TSDB_VNODE_STATUS_OFFLINE,
+  TSDB_VNODE_STATUS_CREATING,
+  TSDB_VNODE_STATUS_UNSYNCED,
+  TSDB_VNODE_STATUS_SLAVE,
+  TSDB_VNODE_STATUS_MASTER,
+  TSDB_VNODE_STATUS_CLOSING,
+  TSDB_VNODE_STATUS_DELETING,
 };
 
-enum _syncstatus {
+enum _vnode_sync_status {
   STDB_SSTATUS_INIT,
   TSDB_SSTATUS_SYNCING,
   TSDB_SSTATUS_SYNC_CACHE,
   TSDB_SSTATUS_SYNC_FILE,
 };
 
+enum _dnode_status {
+    TSDB_DNODE_STATUS_OFFLINE,
+    TSDB_DNODE_STATUS_READY
+};
+
+enum _dnode_balance_status {
+    LB_DNODE_STATE_BALANCED,
+    LB_DNODE_STATE_BALANCING,
+    LB_DNODE_STATE_OFFLINE_REMOVING,
+    LB_DNODE_STATE_SHELL_REMOVING
+};
+
+enum _vgroup_status {
+    LB_VGROUP_STATE_READY,
+    LB_VGROUP_STATE_UPDATE
+};
+
 #define TSDB_DATA_TYPE_BOOL       1     // 1 bytes
 #define TSDB_DATA_TYPE_TINYINT    2     // 1 byte
 #define TSDB_DATA_TYPE_SMALLINT   3     // 2 bytes

src/inc/tstatus.h

@@ -20,10 +20,11 @@
 extern "C" {
 #endif
 
-extern char *sdbDnodeStatusStr[];
-extern char *sdbDnodeBalanceStateStr[];
-extern char *sdbVnodeDropStateStr[];
-extern char *sdbVnodeSyncStatusStr[];
+const char* taosGetVnodeStatusStr(int vnodeStatus);
+const char* taosGetDnodeStatusStr(int dnodeStatus);
+const char* taosGetDnodeBalanceStateStr(int dnodeBalanceStatus);
+const char* taosGetVnodeSyncStatusStr(int vnodeSyncStatus);
+const char* taosGetVnodeDropStatusStr(int dropping);
 
 #ifdef __cplusplus
 }

src/rpc/src/trpc.c

@@ -354,6 +354,8 @@ int taosOpenRpcChannWithQ(void *handle, int cid, int sessions, void *qhandle) {
   STaosRpc * pServer = (STaosRpc *)handle;
   SRpcChann *pChann;
 
+  tTrace("cid:%d, handle:%p open rpc chann", cid, handle);
+
   if (pServer == NULL) return -1;
   if (cid >= pServer->numOfChanns || cid < 0) {
     tError("%s: cid:%d, chann is out of range, max:%d", pServer->label, cid, pServer->numOfChanns);
@@ -402,6 +404,8 @@ void taosCloseRpcChann(void *handle, int cid) {
   STaosRpc * pServer = (STaosRpc *)handle;
   SRpcChann *pChann;
 
+  tTrace("cid:%d, handle:%p close rpc chann", cid, handle);
+
   if (pServer == NULL) return;
   if (cid >= pServer->numOfChanns || cid < 0) {
     tError("%s cid:%d, chann is out of range, max:%d", pServer->label, cid, pServer->numOfChanns);

src/rpc/src/tstring.c

@@ -240,4 +240,6 @@ char *tsError[] = {"success",
                    "invalid submit message",
                    "not active table(not created yet or deleted already)",  //114
                    "invalid table id",
+                   "invalid vnode status",    //116
+                   "failed to lock resources",
 };

src/system/detail/inc/mgmtBalance.h

@@ -28,15 +28,6 @@ extern "C" {
 #include "tstatus.h"
 #include "ttime.h"
 
-enum {
-  LB_DNODE_STATE_BALANCED,
-  LB_DNODE_STATE_BALANCING,
-  LB_DNODE_STATE_OFFLINE_REMOVING,
-  LB_DNODE_STATE_SHELL_REMOVING
-};
-
-enum { LB_VGROUP_STATE_READY, LB_VGROUP_STATE_UPDATE };
-
 void mgmtCreateDnodeOrderList();
 
 void mgmtReleaseDnodeOrderList();

src/system/detail/inc/vnode.h

@@ -92,7 +92,7 @@ typedef struct {
   SVPeerDesc          vpeers[TSDB_VNODES_SUPPORT];
   SVnodePeer *        peerInfo[TSDB_VNODES_SUPPORT];
   char                selfIndex;
-  char                status;
+  char                vnodeStatus;
   char                accessState;  // Vnode access state, Readable/Writable
   char                syncStatus;
   char                commitInProcess;

src/system/detail/src/dnodeMgmt.c

@@ -26,6 +26,7 @@
 #include "vnodeMgmt.h"
 #include "vnodeSystem.h"
 #include "vnodeUtil.h"
+#include "tstatus.h"
 
 SMgmtObj mgmtObj;
 extern uint64_t tsCreatedTime;
@@ -330,7 +331,7 @@ int vnodeProcessVPeerCfg(char *msg, int msgLen, SMgmtObj *pMgmtObj) {
     return -1;
   }
 
-  if (vnodeList[vnode].status == TSDB_STATUS_CREATING) {
+  if (vnodeList[vnode].vnodeStatus == TSDB_VNODE_STATUS_CREATING) {
     dTrace("vid:%d, vnode is still under creating", vnode);
     return 0;
   }
@@ -359,11 +360,25 @@ int vnodeProcessVPeerCfg(char *msg, int msgLen, SMgmtObj *pMgmtObj) {
   }
 
   if (vnodeList[vnode].cfg.maxSessions == 0) {
+    dTrace("vid:%d, vnode is empty", vnode);
     if (pCfg->maxSessions > 0) {
+      if (vnodeList[vnode].vnodeStatus == TSDB_VNODE_STATUS_OFFLINE) {
+        dTrace("vid:%d, status:%s, start to create vnode", vnode, taosGetVnodeStatusStr(vnodeList[vnode].vnodeStatus));
         return vnodeCreateVnode(vnode, pCfg, pMsg->vpeerDesc);
+      } else {
+        dTrace("vid:%d, status:%s, cannot preform create vnode operation", vnode, taosGetVnodeStatusStr(vnodeList[vnode].vnodeStatus));
+        return TSDB_CODE_INVALID_VNODE_STATUS;
+      }
     }
   } else {
+    dTrace("vid:%d, vnode is not empty", vnode);
     if (pCfg->maxSessions > 0) {
+      if (vnodeList[vnode].vnodeStatus == TSDB_VNODE_STATUS_DELETING) {
+        dTrace("vid:%d, status:%s, wait vnode delete finished", vnode, taosGetVnodeStatusStr(vnodeList[vnode].vnodeStatus));
+      } else {
+        dTrace("vid:%d, status:%s, start to update vnode", vnode, taosGetVnodeStatusStr(vnodeList[vnode].vnodeStatus));
+      }
+      /*
       if (pCfg->maxSessions != vnodeList[vnode].cfg.maxSessions) {
         vnodeCleanUpOneVnode(vnode);
       }
@@ -376,7 +391,10 @@ int vnodeProcessVPeerCfg(char *msg, int msgLen, SMgmtObj *pMgmtObj) {
         vnodeList[vnode].cfg.maxSessions = pCfg->maxSessions;
         vnodeOpenVnode(vnode);
       }
+      */
+      return 0;
     } else {
+      dTrace("vid:%d, status:%s, start to delete vnode", vnode, taosGetVnodeStatusStr(vnodeList[vnode].vnodeStatus));
       vnodeRemoveVnode(vnode);
     }
   }
@@ -434,11 +452,11 @@ int vnodeProcessFreeVnodeRequest(char *pMsg, int msgLen, SMgmtObj *pMgmtObj) {
   pFree->vnode = htons(pFree->vnode);
 
   if (pFree->vnode < 0 || pFree->vnode >= TSDB_MAX_VNODES) {
-    dWarn("vid:%d out of range", pFree->vnode);
+    dWarn("vid:%d, out of range", pFree->vnode);
     return -1;
   }
 
-  dTrace("vid:%d receive free vnode message", pFree->vnode);
+  dTrace("vid:%d, receive free vnode message", pFree->vnode);
   int32_t code = vnodeRemoveVnode(pFree->vnode);
   assert(code == TSDB_CODE_SUCCESS || code == TSDB_CODE_ACTION_IN_PROGRESS);
 

src/system/detail/src/mgmtDb.c

@@ -141,11 +141,10 @@ int mgmtCheckDbParams(SCreateDbMsg *pCreate) {
   if (pCreate->cacheNumOfBlocks.fraction < 0) pCreate->cacheNumOfBlocks.fraction = tsAverageCacheBlocks;  //
   //-1 for balance
 
-#ifdef CLUSTER
-  if (pCreate->replications > TSDB_VNODES_SUPPORT - 1) pCreate->replications = TSDB_VNODES_SUPPORT - 1;
-#else
-  pCreate->replications = 1;
-#endif
+  if (pCreate->replications <= 0 || pCreate->replications > TSDB_REPLICA_MAX_NUM) {
+    mTrace("invalid db option replications: %d", pCreate->replications);
+    return TSDB_CODE_INVALID_OPTION;
+  }
 
   if (pCreate->commitLog < 0 || pCreate->commitLog > 1) {
     mTrace("invalid db option commitLog: %d", pCreate->commitLog);
@@ -316,7 +315,7 @@ bool mgmtCheckDropDbFinished(SDbObj *pDb) {
       SDnodeObj *pDnode = mgmtGetDnode(pVnodeGid->ip);
 
       if (pDnode == NULL) continue;
-      if (pDnode->status == TSDB_STATUS_OFFLINE) continue;
+      if (pDnode->status == TSDB_DNODE_STATUS_OFFLINE) continue;
 
       SVnodeLoad *pVload = &pDnode->vload[pVnodeGid->vnode];
       if (pVload->dropStatus == TSDB_VN_STATUS_DROPPING) {

src/system/detail/src/mgmtDnode.c

@@ -21,6 +21,7 @@
 #include "mgmt.h"
 #include "tschemautil.h"
 #include "tstatus.h"
+#include "tstatus.h"
 
 bool mgmtCheckModuleInDnode(SDnodeObj *pDnode, int moduleType);
 int  mgmtGetDnodesNum();
@@ -43,9 +44,9 @@ void mgmtSetDnodeMaxVnodes(SDnodeObj *pDnode) {
   pDnode->openVnodes = 0;
 
 #ifdef CLUSTER
-  pDnode->status = TSDB_STATUS_OFFLINE;
+  pDnode->status = TSDB_DNODE_STATUS_OFFLINE;
 #else
-  pDnode->status = TSDB_STATUS_READY;
+  pDnode->status = TSDB_DNODE_STATUS_READY;
 #endif
 }
 
@@ -57,9 +58,9 @@ void mgmtCalcNumOfFreeVnodes(SDnodeObj *pDnode) {
     if (pVload->vgId != 0) {
       mTrace("dnode:%s, calc free vnodes, exist vnode:%d, vgroup:%d, state:%d %s, dropstate:%d %s, syncstatus:%d %s",
              taosIpStr(pDnode->privateIp), i, pVload->vgId,
-             pVload->status, sdbDnodeStatusStr[pVload->status],
-             pVload->dropStatus, sdbVnodeDropStateStr[pVload->dropStatus],
-             pVload->syncStatus, sdbVnodeSyncStatusStr[pVload->syncStatus]);
+             pVload->status, taosGetDnodeStatusStr(pVload->status),
+             pVload->dropStatus, taosGetVnodeDropStatusStr(pVload->dropStatus),
+             pVload->syncStatus, taosGetVnodeSyncStatusStr(pVload->syncStatus));
       totalVnodes++;
     }
   }
@@ -196,11 +197,11 @@ int mgmtRetrieveDnodes(SShowObj *pShow, char *data, int rows, SConnObj *pConn) {
     cols++;
 
     pWrite = data + pShow->offset[cols] * rows + pShow->bytes[cols] * numOfRows;
-    strcpy(pWrite, sdbDnodeStatusStr[pDnode->status]);
+    strcpy(pWrite, taosGetDnodeStatusStr(pDnode->status) );
     cols++;
 
     pWrite = data + pShow->offset[cols] * rows + pShow->bytes[cols] * numOfRows;
-    strcpy(pWrite, sdbDnodeBalanceStateStr[pDnode->lbState]);
+    strcpy(pWrite, taosGetDnodeBalanceStateStr(pDnode->lbState));
     cols++;
 
     tinet_ntoa(ipstr, pDnode->publicIp);
@@ -292,7 +293,7 @@ int mgmtRetrieveModules(SShowObj *pShow, char *data, int rows, SConnObj *pConn)
       cols++;
 
       pWrite = data + pShow->offset[cols] * rows + pShow->bytes[cols] * numOfRows;
-      strcpy(pWrite, sdbDnodeStatusStr[pDnode->status]);
+      strcpy(pWrite, taosGetDnodeStatusStr(pDnode->status) );
       cols++;
 
       numOfRows++;

src/system/detail/src/mgmtDnodeInt.c

@@ -128,7 +128,7 @@ int mgmtProcessVPeersRsp(char *msg, int msgLen, SDnodeObj *pObj) {
   }
 
   if (pDb->vgStatus != TSDB_VG_STATUS_IN_PROGRESS) {
-    mTrace("dnode:%s, db:%s vpeer rsp already disposed, code:%d", taosIpStr(pObj->privateIp), pRsp->more, pRsp->code);
+    mTrace("dnode:%s, db:%s vpeer rsp already disposed, vgroup status:%d code:%d", taosIpStr(pObj->privateIp), pRsp->more, pDb->vgStatus, pRsp->code);
     return 0;
   }
 
@@ -140,10 +140,11 @@ int mgmtProcessVPeersRsp(char *msg, int msgLen, SDnodeObj *pObj) {
 
   if (pRsp->code == TSDB_CODE_VG_COMMITLOG_INIT_FAILED) {
     pDb->vgStatus = TSDB_VG_STATUS_COMMITLOG_INIT_FAILED;
+    mError("dnode:%s, db:%s vgroup commit log init failed, code:%d", taosIpStr(pObj->privateIp), pRsp->more, pRsp->code);
   } else {
     pDb->vgStatus = TSDB_VG_STATUS_INIT_FAILED;
+    mError("dnode:%s, db:%s vgroup init failed, code:%d", taosIpStr(pObj->privateIp), pRsp->more, pRsp->code);
   }
-  mError("dnode:%s, db:%s vgroup create failed, code:%d", taosIpStr(pObj->privateIp), pRsp->more, pRsp->code);
 
   return 0;
 }
@@ -331,7 +332,6 @@ char *mgmtBuildVpeersIe(char *pMsg, SVgObj *pVgroup, int vnode) {
   pCfg->replications = (char)pVgroup->numOfVnodes;
   pCfg->rowsInFileBlock = htonl(pCfg->rowsInFileBlock);
 
-#ifdef CLUSTER
   SVPeerDesc *vpeerDesc = pVPeers->vpeerDesc;
 
   pMsg = (char *)(pVPeers->vpeerDesc);
@@ -341,7 +341,6 @@ char *mgmtBuildVpeersIe(char *pMsg, SVgObj *pVgroup, int vnode) {
     vpeerDesc[j].vnode = htonl(pVgroup->vnodeGid[j].vnode);
     pMsg += sizeof(SVPeerDesc);
   }
-#endif
 
   return pMsg;
 }

src/system/detail/src/mgmtMeter.c

@@ -502,7 +502,7 @@ int mgmtCreateMeter(SDbObj *pDb, SCreateTableMsg *pCreate) {
 
   int numOfTables = sdbGetNumOfRows(meterSdb);
   if (numOfTables >= tsMaxTables) {
-    mWarn("numOfTables:%d, exceed tsMaxTables:%d", numOfTables, tsMaxTables);
+    mError("table:%s, numOfTables:%d exceed maxTables:%d", pCreate->meterId, numOfTables, tsMaxTables);
     return TSDB_CODE_TOO_MANY_TABLES;
   }
 
@@ -510,6 +510,7 @@ int mgmtCreateMeter(SDbObj *pDb, SCreateTableMsg *pCreate) {
   assert(pAcct != NULL);
   int code = mgmtCheckMeterLimit(pAcct, pCreate);
   if (code != 0) {
+    mError("table:%s, exceed the limit", pCreate->meterId);
     return code;
   }
 
@@ -517,8 +518,10 @@ int mgmtCreateMeter(SDbObj *pDb, SCreateTableMsg *pCreate) {
   pMeter = mgmtGetMeter(pCreate->meterId);
   if (pMeter) {
     if (pCreate->igExists) {
+      mError("table:%s, igExists is true", pCreate->meterId);
       return TSDB_CODE_SUCCESS;
     } else {
+      mError("table:%s, table is already exist", pCreate->meterId);
       return TSDB_CODE_TABLE_ALREADY_EXIST;
     }
   }
@@ -533,6 +536,7 @@ int mgmtCreateMeter(SDbObj *pDb, SCreateTableMsg *pCreate) {
     char *pTagData = (char *)pCreate->schema;  // it is a tag key
     pMetric = mgmtGetMeter(pTagData);
     if (pMetric == NULL) {
+      mError("table:%s, corresponding super table does not exist", pCreate->meterId);
       return TSDB_CODE_INVALID_TABLE;
     }
 
@@ -545,6 +549,7 @@ int mgmtCreateMeter(SDbObj *pDb, SCreateTableMsg *pCreate) {
     pMeter->schema = (char *)malloc(size);
     if (pMeter->schema == NULL) {
       mgmtDestroyMeter(pMeter);
+      mError("table:%s, corresponding super table schema is null", pCreate->meterId);
       return TSDB_CODE_INVALID_TABLE;
     }
     memset(pMeter->schema, 0, size);
@@ -556,13 +561,13 @@ int mgmtCreateMeter(SDbObj *pDb, SCreateTableMsg *pCreate) {
     pMeter->pTagData = pMeter->schema;
     pMeter->nextColId = pMetric->nextColId;
     memcpy(pMeter->pTagData, pTagData, size);
-
   } else {
     int numOfCols = pCreate->numOfColumns + pCreate->numOfTags;
     size = numOfCols * sizeof(SSchema) + pCreate->sqlLen;
     pMeter->schema = (char *)malloc(size);
     if (pMeter->schema == NULL) {
       mgmtDestroyMeter(pMeter);
+      mError("table:%s, no schema input", pCreate->meterId);
       return TSDB_CODE_SERV_OUT_OF_MEMORY;
     }
     memset(pMeter->schema, 0, size);
@@ -583,7 +588,7 @@ int mgmtCreateMeter(SDbObj *pDb, SCreateTableMsg *pCreate) {
       pMeter->pSql = pMeter->schema + numOfCols * sizeof(SSchema);
       memcpy(pMeter->pSql, (char *)(pCreate->schema) + numOfCols * sizeof(SSchema), pCreate->sqlLen);
       pMeter->pSql[pCreate->sqlLen - 1] = 0;
-      mTrace("stream sql len:%d, sql:%s", pCreate->sqlLen, pMeter->pSql);
+      mTrace("table:%s, stream sql len:%d sql:%s", pCreate->meterId, pCreate->sqlLen, pMeter->pSql);
     } else {
       if (pCreate->numOfTags > 0) {
         pMeter->meterType = TSDB_METER_METRIC;
@@ -596,13 +601,14 @@ int mgmtCreateMeter(SDbObj *pDb, SCreateTableMsg *pCreate) {
   pMeter->createdTime = taosGetTimestampMs();
   strcpy(pMeter->meterId, pCreate->meterId);
   if (pthread_rwlock_init(&pMeter->rwLock, NULL)) {
-    mError("Failed to init meter lock");
+    mError("table:%s, failed to init meter lock", pCreate->meterId);
     mgmtDestroyMeter(pMeter);
-    return TSDB_CODE_OTHERS;
+    return TSDB_CODE_FAILED_TO_LOCK_RESOURCES;
   }
 
   code = mgmtCheckMeterGrant(pCreate, pMeter);
   if (code != 0) {
+    mError("table:%s, grant expired", pCreate->meterId);
     return code;
   }
 
@@ -611,21 +617,25 @@ int mgmtCreateMeter(SDbObj *pDb, SCreateTableMsg *pCreate) {
 
     if (pDb->vgStatus == TSDB_VG_STATUS_IN_PROGRESS) {
       mgmtDestroyMeter(pMeter);
+      mTrace("table:%s, vgroup in creating progress", pCreate->meterId);
       return TSDB_CODE_ACTION_IN_PROGRESS;
     }
 
     if (pDb->vgStatus == TSDB_VG_STATUS_FULL) {
       mgmtDestroyMeter(pMeter);
+      mError("table:%s, vgroup is full", pCreate->meterId);
       return TSDB_CODE_NO_ENOUGH_DNODES;
     }
 
     if (pDb->vgStatus == TSDB_VG_STATUS_COMMITLOG_INIT_FAILED) {
       mgmtDestroyMeter(pMeter);
+      mError("table:%s, commit log init failed", pCreate->meterId);
       return TSDB_CODE_VG_COMMITLOG_INIT_FAILED;
     }
 
     if (pDb->vgStatus == TSDB_VG_STATUS_INIT_FAILED) {
       mgmtDestroyMeter(pMeter);
+      mError("table:%s, vgroup init failed", pCreate->meterId);
       return TSDB_CODE_VG_INIT_FAILED;
     }
 
@@ -633,12 +643,13 @@ int mgmtCreateMeter(SDbObj *pDb, SCreateTableMsg *pCreate) {
       pDb->vgStatus = TSDB_VG_STATUS_IN_PROGRESS;
       mgmtCreateVgroup(pDb);
       mgmtDestroyMeter(pMeter);
+      mTrace("table:%s, vgroup malloced, wait for create progress finished", pCreate->meterId);
       return TSDB_CODE_ACTION_IN_PROGRESS;
     }
 
     int sid = taosAllocateId(pVgroup->idPool);
     if (sid < 0) {
-      mWarn("db:%s, vgroup:%d, run out of ID, num:%d", pDb->name, pVgroup->vgId, taosIdPoolNumOfUsed(pVgroup->idPool));
+      mWarn("table:%s, vgroup:%d run out of ID, num:%d", pCreate->meterId, pVgroup->vgId, taosIdPoolNumOfUsed(pVgroup->idPool));
       pDb->vgStatus = TSDB_VG_STATUS_IN_PROGRESS;
       mgmtCreateVgroup(pDb);
       mgmtDestroyMeter(pMeter);
@@ -650,18 +661,21 @@ int mgmtCreateMeter(SDbObj *pDb, SCreateTableMsg *pCreate) {
     pMeter->uid = (((uint64_t)pMeter->gid.vgId) << 40) + ((((uint64_t)pMeter->gid.sid) & ((1ul << 24) - 1ul)) << 16) +
                   ((uint64_t)sdbVersion & ((1ul << 16) - 1ul));
 
-    mTrace("meter:%s, create meter in vgroup, vgId:%d, sid:%d, vnode:%d, uid:%d",
-           pMeter->meterId, pVgroup->vgId, sid, pVgroup->vnodeGid[0].vnode, pMeter->uid);
+    mTrace("table:%s, create table in vgroup, vgId:%d sid:%d vnode:%d uid:%d db:%s",
+           pMeter->meterId, pVgroup->vgId, sid, pVgroup->vnodeGid[0].vnode, pMeter->uid, pDb->name);
   } else {
     pMeter->uid = (((uint64_t)pMeter->createdTime) << 16) + ((uint64_t)sdbVersion & ((1ul << 16) - 1ul));
   }
 
-  if (sdbInsertRow(meterSdb, pMeter, 0) < 0) return TSDB_CODE_SDB_ERROR;
+  if (sdbInsertRow(meterSdb, pMeter, 0) < 0) {
+    mError("table:%s, update sdb error", pCreate->meterId);
+    return TSDB_CODE_SDB_ERROR;
+  }
 
   // send create message to the selected vnode servers
   if (pCreate->numOfTags == 0) {
-    mTrace("meter:%s, send msg to dnode, vgId:%d, sid:%d, vnode:%d, dbname:%s",
-           pMeter->meterId, pMeter->gid.vgId, pMeter->gid.sid, pVgroup->vnodeGid[0].vnode, pDb->name);
+    mTrace("table:%s, send create msg to dnode, vgId:%d, sid:%d, vnode:%d",
+           pMeter->meterId, pMeter->gid.vgId, pMeter->gid.sid, pVgroup->vnodeGid[0].vnode);
 
     grantAddTimeSeries(pMeter->numOfColumns - 1);
     mgmtSendCreateMsgToVgroup(pMeter, pVgroup);
@@ -881,7 +895,10 @@ void mgmtCleanUpMeters() { sdbCloseTable(meterSdb); }
 int mgmtGetMeterMeta(SMeterMeta *pMeta, SShowObj *pShow, SConnObj *pConn) {
   int cols = 0;
 
-  if (pConn->pDb == NULL) return TSDB_CODE_DB_NOT_SELECTED;
+  SDbObj *pDb = NULL;
+  if (pConn->pDb != NULL) pDb = mgmtGetDb(pConn->pDb->name);
+
+  if (pDb == NULL) return TSDB_CODE_DB_NOT_SELECTED;
 
   SSchema *pSchema = tsGetSchema(pMeta);
 
@@ -916,7 +933,7 @@ int mgmtGetMeterMeta(SMeterMeta *pMeta, SShowObj *pShow, SConnObj *pConn) {
   for (int i = 1; i < cols; ++i) pShow->offset[i] = pShow->offset[i - 1] + pShow->bytes[i - 1];
 
   //  pShow->numOfRows = sdbGetNumOfRows (meterSdb);
-  pShow->numOfRows = pConn->pDb->numOfTables;
+  pShow->numOfRows = pDb->numOfTables;
   pShow->rowSize = pShow->offset[cols - 1] + pShow->bytes[cols - 1];
 
   return 0;
@@ -1208,8 +1225,12 @@ int mgmtRetrieveMeters(SShowObj *pShow, char *data, int rows, SConnObj *pConn) {
   int      numOfRead = 0;
   char     prefix[20] = {0};
 
-  if (pConn->pDb == NULL) return 0;
-  strcpy(prefix, pConn->pDb->name);
+  SDbObj *pDb = NULL;
+  if (pConn->pDb != NULL) pDb = mgmtGetDb(pConn->pDb->name);
+
+  if (pDb == NULL) return 0;
+
+  strcpy(prefix, pDb->name);
   strcat(prefix, TS_PATH_DELIMITER);
   prefixLen = strlen(prefix);
 
@@ -1269,7 +1290,10 @@ int mgmtRetrieveMeters(SShowObj *pShow, char *data, int rows, SConnObj *pConn) {
 int mgmtGetMetricMeta(SMeterMeta *pMeta, SShowObj *pShow, SConnObj *pConn) {
   int cols = 0;
 
-  if (pConn->pDb == NULL) return TSDB_CODE_DB_NOT_SELECTED;
+  SDbObj *pDb = NULL;
+  if (pConn->pDb != NULL) pDb = mgmtGetDb(pConn->pDb->name);
+
+  if (pDb == NULL) return TSDB_CODE_DB_NOT_SELECTED;
 
   SSchema *pSchema = tsGetSchema(pMeta);
 
@@ -1309,8 +1333,8 @@ int mgmtGetMetricMeta(SMeterMeta *pMeta, SShowObj *pShow, SConnObj *pConn) {
   pShow->offset[0] = 0;
   for (int i = 1; i < cols; ++i) pShow->offset[i] = pShow->offset[i - 1] + pShow->bytes[i - 1];
 
-  pShow->numOfRows = pConn->pDb->numOfMetrics;
-  pShow->pNode = pConn->pDb->pMetric;
+  pShow->numOfRows = pDb->numOfMetrics;
+  pShow->pNode = pDb->pMetric;
   pShow->rowSize = pShow->offset[cols - 1] + pShow->bytes[cols - 1];
 
   return 0;

src/system/detail/src/mgmtShell.c

@@ -189,8 +189,11 @@ int mgmtProcessMeterMetaMsg(char *pMsg, int msgLen, SConnObj *pConn) {
   int size = sizeof(STaosHeader) + sizeof(STaosRsp) + sizeof(SMeterMeta) + sizeof(SSchema) * TSDB_MAX_COLUMNS +
              sizeof(SSchema) * TSDB_MAX_TAGS + TSDB_MAX_TAGS_LEN + TSDB_EXTRA_PAYLOAD_SIZE;
 
+  SDbObj *pDb = NULL;
+  if (pConn->pDb != NULL) pDb = mgmtGetDb(pConn->pDb->name);
+
   // todo db check should be extracted
-  if (pConn->pDb == NULL || (pConn->pDb != NULL && pConn->pDb->dropStatus != TSDB_DB_STATUS_READY)) {
+  if (pDb == NULL || (pDb != NULL && pDb->dropStatus != TSDB_DB_STATUS_READY)) {
 
     if ((pStart = mgmtAllocMsg(pConn, size, &pMsg, &pRsp)) == NULL) {
       taosSendSimpleRsp(pConn->thandle, TSDB_MSG_TYPE_METERINFO_RSP, TSDB_CODE_SERV_OUT_OF_MEMORY);
@@ -223,10 +226,10 @@ int mgmtProcessMeterMetaMsg(char *pMsg, int msgLen, SConnObj *pConn) {
 
     SDbObj* pMeterDb = mgmtGetDbByMeterId(pCreateMsg->meterId);
     mTrace("meter:%s, pConnDb:%p, pConnDbName:%s, pMeterDb:%p, pMeterDbName:%s",
-           pCreateMsg->meterId, pConn->pDb, pConn->pDb->name, pMeterDb, pMeterDb->name);
-    assert(pConn->pDb == pMeterDb);
+           pCreateMsg->meterId, pDb, pDb->name, pMeterDb, pMeterDb->name);
+    assert(pDb == pMeterDb);
 
-    int32_t code = mgmtCreateMeter(pConn->pDb, pCreateMsg);
+    int32_t code = mgmtCreateMeter(pDb, pCreateMsg);
 
     char stableName[TSDB_METER_ID_LEN] = {0};
     strncpy(stableName, pInfo->tags, TSDB_METER_ID_LEN);
@@ -256,7 +259,7 @@ int mgmtProcessMeterMetaMsg(char *pMsg, int msgLen, SConnObj *pConn) {
   }
 
   if (pMeterObj == NULL) {
-    if (pConn->pDb)
+    if (pDb)
       pRsp->code = TSDB_CODE_INVALID_TABLE;
     else
       pRsp->code = TSDB_CODE_DB_NOT_SELECTED;
@@ -274,7 +277,7 @@ int mgmtProcessMeterMetaMsg(char *pMsg, int msgLen, SConnObj *pConn) {
     pMeta->vgid = htonl(pMeterObj->gid.vgId);
     pMeta->sversion = htons(pMeterObj->sversion);
 
-    pMeta->precision = pConn->pDb->cfg.precision;
+    pMeta->precision = pDb->cfg.precision;
 
     pMeta->numOfTags = pMeterObj->numOfTags;
     pMeta->numOfColumns = htons(pMeterObj->numOfColumns);
@@ -505,7 +508,10 @@ int mgmtProcessMetricMetaMsg(char *pMsg, int msgLen, SConnObj *pConn) {
   SMetricMetaElemMsg *pElem = (SMetricMetaElemMsg *)(((char *)pMetricMetaMsg) + pMetricMetaMsg->metaElem[0]);
   pMetric = mgmtGetMeter(pElem->meterId);
 
-  if (pMetric == NULL || (pConn->pDb != NULL && pConn->pDb->dropStatus != TSDB_DB_STATUS_READY)) {
+  SDbObj *pDb = NULL;
+  if (pConn->pDb != NULL) pDb = mgmtGetDb(pConn->pDb->name);
+
+  if (pMetric == NULL || (pDb != NULL && pDb->dropStatus != TSDB_DB_STATUS_READY)) {
     pStart = taosBuildRspMsg(pConn->thandle, TSDB_MSG_TYPE_METRIC_META_RSP);
     if (pStart == NULL) {
       taosSendSimpleRsp(pConn->thandle, TSDB_MSG_TYPE_METRIC_META_RSP, TSDB_CODE_SERV_OUT_OF_MEMORY);
@@ -514,7 +520,7 @@ int mgmtProcessMetricMetaMsg(char *pMsg, int msgLen, SConnObj *pConn) {
 
     pMsg = pStart;
     pRsp = (STaosRsp *)pMsg;
-    if (pConn->pDb)
+    if (pDb)
       pRsp->code = TSDB_CODE_INVALID_TABLE;
     else
       pRsp->code = TSDB_CODE_DB_NOT_SELECTED;
@@ -957,17 +963,23 @@ int mgmtProcessCreateTableMsg(char *pMsg, int msgLen, SConnObj *pConn) {
       pSchema++;
     }
 
-    if (pConn->pDb) {
-      code = mgmtCreateMeter(pConn->pDb, pCreate);
-      if (code == 0) {
-        mTrace("meter:%s is created by %s", pCreate->meterId, pConn->pUser->user);
-        // mLPrint("meter:%s is created by %s", pCreate->meterId, pConn->pUser->user);
-      }
+    SDbObj *pDb = NULL;
+    if (pConn->pDb != NULL) pDb = mgmtGetDb(pConn->pDb->name);
+
+    if (pDb) {
+      code = mgmtCreateMeter(pDb, pCreate);
     } else {
       code = TSDB_CODE_DB_NOT_SELECTED;
     }
   }
 
+  if (code != 0) {
+    mError("table:%s, failed to create table, code:%d", pCreate->meterId, code);
+  } else {
+    mTrace("table:%s, table is created by %s", pCreate->meterId, pConn->pUser->user);
+    //mLPrint("meter:%s is created by %s", pCreate->meterId, pConn->pUser->user);
+  }
+
   taosSendSimpleRsp(pConn->thandle, TSDB_MSG_TYPE_CREATE_TABLE_RSP, code);
 
   return 0;
@@ -984,7 +996,10 @@ int mgmtProcessDropTableMsg(char *pMsg, int msgLen, SConnObj *pConn) {
   if (!pConn->writeAuth) {
     code = TSDB_CODE_NO_RIGHTS;
   } else {
-    code = mgmtDropMeter(pConn->pDb, pDrop->meterId, pDrop->igNotExists);
+    SDbObj *pDb = NULL;
+    if (pConn->pDb != NULL) pDb = mgmtGetDb(pConn->pDb->name);
+
+    code = mgmtDropMeter(pDb, pDrop->meterId, pDrop->igNotExists);
     if (code == 0) {
       mTrace("meter:%s is dropped by user:%s", pDrop->meterId, pConn->pUser->user);
       // mLPrint("meter:%s is dropped by user:%s", pDrop->meterId, pConn->pUser->user);
@@ -1014,12 +1029,15 @@ int mgmtProcessAlterTableMsg(char *pMsg, int msgLen, SConnObj *pConn) {
       mError("meter:%s error numOfCols:%d in alter table", pAlter->meterId, pAlter->numOfCols);
       code = TSDB_CODE_APP_ERROR;
     } else {
-      if (pConn->pDb) {
+      SDbObj *pDb = NULL;
+      if (pConn->pDb != NULL) pDb = mgmtGetDb(pConn->pDb->name);
+
+      if (pDb) {
         for (int32_t i = 0; i < pAlter->numOfCols; ++i) {
           pAlter->schema[i].bytes = htons(pAlter->schema[i].bytes);
         }
 
-        code = mgmtAlterMeter(pConn->pDb, pAlter);
+        code = mgmtAlterMeter(pDb, pAlter);
         if (code == 0) {
           mLPrint("meter:%s is altered by %s", pAlter->meterId, pConn->pUser->user);
         }
@@ -1263,8 +1281,7 @@ void *mgmtProcessMsgFromShell(char *msg, void *ahandle, void *thandle) {
     }
 
     if (pConn->pAcct) {
-      if (pConn->pDb == NULL ||
-          strncmp(pConn->pDb->name, pHead->db, tListLen(pConn->pDb->name)) != 0) {
+      if (pConn->pDb == NULL || strncmp(pConn->pDb->name, pHead->db, tListLen(pConn->pDb->name)) != 0) {
         pConn->pDb = mgmtGetDb(pHead->db);
       }
 

src/system/detail/src/mgmtVgroup.c

@@ -123,7 +123,7 @@ void mgmtProcessVgTimer(void *handle, void *tmrId) {
   if (pDb == NULL) return;
 
   if (pDb->vgStatus > TSDB_VG_STATUS_IN_PROGRESS) {
-    mTrace("db:%s, set vgstatus from %d to %d", pDb->name, pDb->vgStatus, TSDB_VG_STATUS_READY);
+    mTrace("db:%s, set vgroup status from %d to ready", pDb->name, pDb->vgStatus);
     pDb->vgStatus = TSDB_VG_STATUS_READY;
   }
 
@@ -143,7 +143,7 @@ SVgObj *mgmtCreateVgroup(SDbObj *pDb) {
 
   // based on load balance, create a new one
   if (mgmtAllocVnodes(pVgroup) != 0) {
-    mError("no enough free dnode");
+    mError("db:%s, no enough free dnode to alloc %d vnodes", pDb->name, pVgroup->numOfVnodes);
     free(pVgroup);
     pDb->vgStatus = TSDB_VG_STATUS_FULL;
     taosTmrReset(mgmtProcessVgTimer, 5000, pDb, mgmtTmr, &pDb->vgTimer);
@@ -152,9 +152,9 @@ SVgObj *mgmtCreateVgroup(SDbObj *pDb) {
 
   sdbInsertRow(vgSdb, pVgroup, 0);
 
-  mTrace("vgroup:%d, db:%s replica:%d is created", pVgroup->vgId, pDb->name, pVgroup->numOfVnodes);
+  mTrace("vgroup:%d, vgroup is created, db:%s replica:%d", pVgroup->vgId, pDb->name, pVgroup->numOfVnodes);
   for (int i = 0; i < pVgroup->numOfVnodes; ++i)
-    mTrace("dnode:%s, vgroup:%d, vnode:%d is created", taosIpStr(pVgroup->vnodeGid[i].ip), pVgroup->vgId, pVgroup->vnodeGid[i].vnode);
+    mTrace("vgroup:%d, dnode:%s vnode:%d is created", pVgroup->vgId, taosIpStr(pVgroup->vnodeGid[i].ip), pVgroup->vnodeGid[i].vnode);
 
   mgmtSendVPeersMsg(pVgroup);
 
@@ -206,7 +206,10 @@ void mgmtCleanUpVgroups() { sdbCloseTable(vgSdb); }
 int mgmtGetVgroupMeta(SMeterMeta *pMeta, SShowObj *pShow, SConnObj *pConn) {
   int cols = 0;
 
-  if (pConn->pDb == NULL) return TSDB_CODE_DB_NOT_SELECTED;
+  SDbObj *pDb = NULL;
+  if (pConn->pDb != NULL) pDb = mgmtGetDb(pConn->pDb->name);
+
+  if (pDb == NULL) return TSDB_CODE_DB_NOT_SELECTED;
 
   SSchema *pSchema = tsGetSchema(pMeta);
 
@@ -229,7 +232,7 @@ int mgmtGetVgroupMeta(SMeterMeta *pMeta, SShowObj *pShow, SConnObj *pConn) {
   cols++;
 
   int     maxReplica = 0;
-  SVgObj *pVgroup = pConn->pDb->pHead;
+  SVgObj *pVgroup = pDb->pHead;
   while (pVgroup != NULL) {
     maxReplica = pVgroup->numOfVnodes > maxReplica ? pVgroup->numOfVnodes : maxReplica;
     pVgroup = pVgroup->next;
@@ -267,8 +270,8 @@ int mgmtGetVgroupMeta(SMeterMeta *pMeta, SShowObj *pShow, SConnObj *pConn) {
   pShow->offset[0] = 0;
   for (int i = 1; i < cols; ++i) pShow->offset[i] = pShow->offset[i - 1] + pShow->bytes[i - 1];
 
-  pShow->numOfRows = pConn->pDb->numOfVgroups;
-  pShow->pNode = pConn->pDb->pHead;
+  pShow->numOfRows = pDb->numOfVgroups;
+  pShow->pNode = pDb->pHead;
   pShow->rowSize = pShow->offset[cols - 1] + pShow->bytes[cols - 1];
 
   return 0;
@@ -282,7 +285,11 @@ int mgmtRetrieveVgroups(SShowObj *pShow, char *data, int rows, SConnObj *pConn)
   char    ipstr[20];
 
   int maxReplica = 0;
-  pVgroup = pConn->pDb->pHead;
+
+  SDbObj *pDb = NULL;
+  if (pConn->pDb != NULL) pDb = mgmtGetDb(pConn->pDb->name);
+
+  pVgroup = pDb->pHead;
   while (pVgroup != NULL) {
     maxReplica = pVgroup->numOfVnodes > maxReplica ? pVgroup->numOfVnodes : maxReplica;
     pVgroup = pVgroup->next;

src/system/detail/src/vnodeCache.c

@@ -297,7 +297,7 @@ pthread_t vnodeCreateCommitThread(SVnodeObj *pVnode) {
 
   taosTmrStopA(&pVnode->commitTimer);
 
-  if (pVnode->status == TSDB_STATUS_UNSYNCED) {
+  if (pVnode->vnodeStatus == TSDB_VNODE_STATUS_UNSYNCED) {
     taosTmrReset(vnodeProcessCommitTimer, pVnode->cfg.commitTime * 1000, pVnode, vnodeTmrCtrl, &pVnode->commitTimer);
     dTrace("vid:%d, it is in unsyc state, commit later", pVnode->vnode);
     return pVnode->commitThread;

src/system/detail/src/vnodeFile.c

@@ -1290,7 +1290,7 @@ int vnodeWriteBlockToFile(SMeterObj *pObj, SCompBlock *pCompBlock, SData *data[]
     pCompBlock->len += wlen;
   }
 
-  dTrace("vid: %d vnode compStorage size is: %ld", pObj->vnode, pVnode->vnodeStatistic.compStorage);
+  dTrace("vid:%d, vnode compStorage size is: %ld", pObj->vnode, pVnode->vnodeStatistic.compStorage);
 
   pCompBlock->algorithm = pCfg->compression;
   pCompBlock->numOfPoints = points;

src/system/detail/src/vnodeShell.c

@@ -27,6 +27,7 @@
 #include "vnode.h"
 #include "vnodeRead.h"
 #include "vnodeUtil.h"
+#include "vnodeStore.h"
 
 #pragma GCC diagnostic ignored "-Wint-conversion"
 extern int tsMaxQueues;
@@ -89,9 +90,9 @@ void *vnodeProcessMsgFromShell(char *msg, void *ahandle, void *thandle) {
   // if ( vnodeList[vnode].status != TSDB_STATUS_MASTER && pMsg->msgType != TSDB_MSG_TYPE_RETRIEVE ) {
 
 #ifdef CLUSTER
-  if (vnodeList[vnode].status != TSDB_STATUS_MASTER) {
+  if (vnodeList[vnode].vnodeStatus != TSDB_VNODE_STATUS_MASTER) {
     taosSendSimpleRsp(thandle, pMsg->msgType + 1, TSDB_CODE_NOT_READY);
-    dTrace("vid:%d sid:%d, shell msg is ignored since in state:%d", vnode, sid, vnodeList[vnode].status);
+    dTrace("vid:%d sid:%d, shell msg is ignored since in state:%d", vnode, sid, vnodeList[vnode].vnodeStatus);
   } else {
 #endif
     dTrace("vid:%d sid:%d, msg:%s is received pConn:%p", vnode, sid, taosMsg[pMsg->msgType], thandle);
@@ -154,6 +155,11 @@ int vnodeInitShell() {
 }
 
 int vnodeOpenShellVnode(int vnode) {
+  if (shellList[vnode] != NULL) {
+    dError("vid:%d, shell is already opened", vnode);
+    return -1;
+  }
+
   const int32_t MIN_NUM_OF_SESSIONS = 300;
 
   SVnodeCfg *pCfg = &vnodeList[vnode].cfg;
@@ -162,23 +168,29 @@ int vnodeOpenShellVnode(int vnode) {
   size_t size = sessions * sizeof(SShellObj);
   shellList[vnode] = (SShellObj *)calloc(1, size);
   if (shellList[vnode] == NULL) {
-    dError("vid:%d failed to allocate shellObj, size:%d", vnode, size);
+    dError("vid:%d, sessions:%d, failed to allocate shellObj, size:%d", vnode, pCfg->maxSessions, size);
     return -1;
   }
 
   if(taosOpenRpcChannWithQ(pShellServer, vnode, sessions, rpcQhandle[(vnode+1)%tsMaxQueues]) != TSDB_CODE_SUCCESS) {
+    dError("vid:%d, sessions:%d, failed to open shell", vnode, pCfg->maxSessions);
     return -1;
   }
 
+  dTrace("vid:%d, sessions:%d, shell is opened", vnode, pCfg->maxSessions);
   return TSDB_CODE_SUCCESS;
 }
 
 static void vnodeDelayedFreeResource(void *param, void *tmrId) {
   int32_t vnode = *(int32_t*) param;
-  taosCloseRpcChann(pShellServer, vnode); // close connection
-  tfree (shellList[vnode]);  //free SShellObj
+  dTrace("vid:%d, start to free resources", vnode);
 
+  taosCloseRpcChann(pShellServer, vnode); // close connection
+  tfree(shellList[vnode]);  //free SShellObj
   tfree(param);
+
+  memset(vnodeList + vnode, 0, sizeof(SVnodeObj));
+  vnodeCalcOpenVnodes();
 }
 
 void vnodeCloseShellVnode(int vnode) {

src/system/detail/src/vnodeStore.c

@@ -22,6 +22,7 @@
 #include "vnode.h"
 #include "vnodeStore.h"
 #include "vnodeUtil.h"
+#include "tstatus.h"
 
 #pragma GCC diagnostic push
 #pragma GCC diagnostic warning "-Woverflow"
@@ -30,12 +31,14 @@ int        tsMaxVnode = -1;
 int        tsOpenVnodes = 0;
 SVnodeObj *vnodeList = NULL;
 
-int vnodeInitStoreVnode(int vnode) {
+static int vnodeInitStoreVnode(int vnode) {
   SVnodeObj *pVnode = vnodeList + vnode;
 
   pVnode->vnode = vnode;
   vnodeOpenMetersVnode(vnode);
-  if (pVnode->cfg.maxSessions == 0) return 0;
+  if (pVnode->cfg.maxSessions <= 0) {
+    return TSDB_CODE_SUCCESS;
+  }
 
   pVnode->firstKey = taosGetTimestamp(pVnode->cfg.precision);
 
@@ -45,9 +48,10 @@ int vnodeInitStoreVnode(int vnode) {
     return -1;
   }
 
-  if (vnodeInitFile(vnode) < 0) return -1;
-
-  // vnodeOpenMeterMgmtStoreVnode(vnode);
+  if (vnodeInitFile(vnode) < 0) {
+    dError("vid:%d, files init failed.", pVnode->vnode);
+    return -1;
+  }
 
   if (vnodeInitCommit(vnode) < 0) {
     dError("vid:%d, commit init failed.", pVnode->vnode);
@@ -70,10 +74,17 @@ int vnodeOpenVnode(int vnode) {
   pVnode->accessState = TSDB_VN_ALL_ACCCESS;
 
   // vnode is empty
-  if (pVnode->cfg.maxSessions == 0) return 0;
+  if (pVnode->cfg.maxSessions <= 0) {
+    return TSDB_CODE_SUCCESS;
+  }
 
+  if (!(pVnode->vnodeStatus == TSDB_VNODE_STATUS_OFFLINE || pVnode->vnodeStatus == TSDB_VNODE_STATUS_CREATING)) {
+    dError("vid:%d, status:%s, cannot enter open operation", vnode, taosGetVnodeStatusStr(pVnode->vnodeStatus));
+    return TSDB_CODE_INVALID_VNODE_STATUS;
+  }
+
+  dTrace("vid:%d, status:%s, start to open", vnode, taosGetVnodeStatusStr(pVnode->vnodeStatus));
   pthread_mutex_lock(&dmutex);
-  // vnodeOpenMeterMgmtVnode(vnode);
 
   // not enough memory, abort
   if ((code = vnodeOpenShellVnode(vnode)) != TSDB_CODE_SUCCESS) {
@@ -93,14 +104,13 @@ int vnodeOpenVnode(int vnode) {
   vnodeOpenStreams(pVnode, NULL);
 #endif
 
-  dTrace("vid:%d, vnode is opened, openVnodes:%d", vnode, tsOpenVnodes);
+  dTrace("vid:%d, vnode is opened, openVnodes:%d, status:%s", vnode, tsOpenVnodes, taosGetVnodeStatusStr(pVnode->vnodeStatus));
 
-  return 0;
+  return TSDB_CODE_SUCCESS;
 }
 
 static int32_t vnodeMarkAllMetersDropped(SVnodeObj* pVnode) {
   if (pVnode->meterList == NULL) {
-    assert(pVnode->cfg.maxSessions == 0);
     return TSDB_CODE_SUCCESS;
   }
 
@@ -119,7 +129,7 @@ static int32_t vnodeMarkAllMetersDropped(SVnodeObj* pVnode) {
   return ready? TSDB_CODE_SUCCESS:TSDB_CODE_ACTION_IN_PROGRESS;
 }
 
-int vnodeCloseVnode(int vnode) {
+static int vnodeCloseVnode(int vnode) {
   if (vnodeList == NULL) return TSDB_CODE_SUCCESS;
 
   SVnodeObj* pVnode = &vnodeList[vnode];
@@ -130,12 +140,23 @@ int vnodeCloseVnode(int vnode) {
     return TSDB_CODE_SUCCESS;
   }
 
+  if (pVnode->vnodeStatus == TSDB_VNODE_STATUS_DELETING) {
+    dTrace("vid:%d, status:%s, another thread performed delete operation", vnode, taosGetVnodeStatusStr(pVnode->vnodeStatus));
+    return TSDB_CODE_SUCCESS;
+  } else {
+    dTrace("vid:%d, status:%s, enter close operation", vnode, taosGetVnodeStatusStr(pVnode->vnodeStatus));
+    pVnode->vnodeStatus = TSDB_VNODE_STATUS_CLOSING;
+  }
+
   // set the meter is dropped flag 
   if (vnodeMarkAllMetersDropped(pVnode) != TSDB_CODE_SUCCESS) {
     pthread_mutex_unlock(&dmutex);
     return TSDB_CODE_ACTION_IN_PROGRESS;
   }
 
+  dTrace("vid:%d, status:%s, enter delete operation", vnode, taosGetVnodeStatusStr(pVnode->vnodeStatus));
+  pVnode->vnodeStatus = TSDB_VNODE_STATUS_DELETING;
+
   vnodeCloseStream(vnodeList + vnode);
   vnodeCancelCommit(vnodeList + vnode);
   vnodeClosePeerVnode(vnode);
@@ -149,9 +170,6 @@ int vnodeCloseVnode(int vnode) {
   if (tsMaxVnode == vnode) tsMaxVnode = vnode - 1;
 
   tfree(vnodeList[vnode].meterIndex);
-  memset(vnodeList + vnode, 0, sizeof(SVnodeObj));
-
-  vnodeCalcOpenVnodes();
 
   pthread_mutex_unlock(&dmutex);
   return TSDB_CODE_SUCCESS;
@@ -160,7 +178,12 @@ int vnodeCloseVnode(int vnode) {
 int vnodeCreateVnode(int vnode, SVnodeCfg *pCfg, SVPeerDesc *pDesc) {
   char fileName[128];
 
-  vnodeList[vnode].status = TSDB_STATUS_CREATING;
+  if (vnodeList[vnode].vnodeStatus != TSDB_VNODE_STATUS_OFFLINE) {
+    dError("vid:%d, status:%s, cannot enter create operation", vnode, taosGetVnodeStatusStr(vnodeList[vnode].vnodeStatus));
+    return TSDB_CODE_INVALID_VNODE_STATUS;
+  }
+
+  vnodeList[vnode].vnodeStatus = TSDB_VNODE_STATUS_CREATING;
 
   sprintf(fileName, "%s/vnode%d", tsDirectory, vnode);
   mkdir(fileName, 0755);
@@ -177,14 +200,14 @@ int vnodeCreateVnode(int vnode, SVnodeCfg *pCfg, SVPeerDesc *pDesc) {
     return TSDB_CODE_VG_INIT_FAILED;
   }
 
-  if (vnodeInitStoreVnode(vnode) != 0) {
+  if (vnodeInitStoreVnode(vnode) < 0) {
     return TSDB_CODE_VG_COMMITLOG_INIT_FAILED;
   }
 
   return vnodeOpenVnode(vnode);
 }
 
-void vnodeRemoveDataFiles(int vnode) {
+static void vnodeRemoveDataFiles(int vnode) {
   char           vnodeDir[TSDB_FILENAME_LEN];
   char           dfilePath[TSDB_FILENAME_LEN];
   char           linkFile[TSDB_FILENAME_LEN];
@@ -227,19 +250,28 @@ void vnodeRemoveDataFiles(int vnode) {
 
   sprintf(vnodeDir, "%s/vnode%d", tsDirectory, vnode);
   rmdir(vnodeDir);
-  dTrace("vnode %d is removed!", vnode);
+  dTrace("vid:%d, vnode is removed!", vnode);
 }
 
 int vnodeRemoveVnode(int vnode) {
   if (vnodeList == NULL) return TSDB_CODE_SUCCESS;
 
   if (vnodeList[vnode].cfg.maxSessions > 0) {
+    SVnodeObj* pVnode = &vnodeList[vnode];
+    if (pVnode->vnodeStatus == TSDB_VNODE_STATUS_CREATING
+        || pVnode->vnodeStatus == TSDB_VNODE_STATUS_OFFLINE
+        || pVnode->vnodeStatus == TSDB_VNODE_STATUS_DELETING) {
+      dError("vid:%d, status:%s, cannot enter close/delete operation", vnode, taosGetVnodeStatusStr(pVnode->vnodeStatus));
+      return TSDB_CODE_ACTION_IN_PROGRESS;
+    } else {
       int32_t ret = vnodeCloseVnode(vnode);
       if (ret != TSDB_CODE_SUCCESS) {
         return ret;
       }
 
       vnodeRemoveDataFiles(vnode);
+    }
+
   } else {
     dTrace("vid:%d, max sessions:%d, this vnode already dropped!!!", vnode, vnodeList[vnode].cfg.maxSessions);
     vnodeList[vnode].cfg.maxSessions = 0;  //reset value
@@ -293,7 +325,7 @@ void vnodeCleanUpOneVnode(int vnode) {
   again = 1;
 
   if (vnodeList[vnode].pCachePool) {
-    vnodeList[vnode].status = TSDB_STATUS_OFFLINE;
+    vnodeList[vnode].vnodeStatus = TSDB_VNODE_STATUS_OFFLINE;
     vnodeClosePeerVnode(vnode);
   }
 
@@ -322,7 +354,7 @@ void vnodeCleanUpVnodes() {
 
   for (int vnode = 0; vnode < TSDB_MAX_VNODES; ++vnode) {
     if (vnodeList[vnode].pCachePool) {
-      vnodeList[vnode].status = TSDB_STATUS_OFFLINE;
+      vnodeList[vnode].vnodeStatus = TSDB_VNODE_STATUS_OFFLINE;
       vnodeClosePeerVnode(vnode);
     }
   }

src/system/detail/src/vnodeStream.c

@@ -171,7 +171,7 @@ void vnodeCloseStream(SVnodeObj *pVnode) {
 void vnodeUpdateStreamRole(SVnodeObj *pVnode) {
   /* SMeterObj *pObj; */
 
-  int newRole = (pVnode->status == TSDB_STATUS_MASTER) ? 1 : 0;
+  int newRole = (pVnode->vnodeStatus == TSDB_VNODE_STATUS_MASTER) ? 1 : 0;
   if (newRole != pVnode->streamRole) {
     dTrace("vid:%d, stream role is changed to:%d", pVnode->vnode, newRole);
     pVnode->streamRole = newRole;

src/system/lite/src/dnodeMgmt.spec.c

@@ -50,7 +50,7 @@ char *taosBuildReqMsgToMnode(SMgmtObj *pObj, char type) {
 }
 
 int taosSendMsgToMnode(SMgmtObj *pObj, char *msg, int msgLen) {
-  mTrace("msg:%s is sent to mnode", taosMsg[*(msg-1)]);
+  dTrace("msg:%s is sent to mnode", taosMsg[*(msg-1)]);
 
   /*
    * Lite version has no message header, so minus one

src/system/lite/src/mgmtDnode.spec.c

@@ -30,7 +30,7 @@ int mgmtInitDnodes() {
   dnodeObj.createdTime = (int64_t)tsRebootTime * 1000;
   dnodeObj.lastReboot = tsRebootTime;
   dnodeObj.numOfCores = (uint16_t)tsNumOfCores;
-  dnodeObj.status = TSDB_STATUS_READY;
+  dnodeObj.status = TSDB_DNODE_STATUS_READY;
   dnodeObj.alternativeRole = TSDB_DNODE_ROLE_ANY;
   dnodeObj.numOfTotalVnodes = tsNumOfTotalVnodes;
   dnodeObj.thandle = (void*)(1);  //hack way

src/system/lite/src/mgmtDnodeInt.spec.c

@@ -82,7 +82,7 @@ void mgmtCleanUpDnodeInt() {}
 void mgmtProcessDnodeStatus(void *handle, void *tmrId) {
   SDnodeObj *pObj = &dnodeObj;
   pObj->openVnodes = tsOpenVnodes;
-  pObj->status = TSDB_STATUS_READY;
+  pObj->status = TSDB_DNODE_STATUS_READY;
 
   float memoryUsedMB = 0;
   taosGetSysMemory(&memoryUsedMB);
@@ -97,7 +97,7 @@ void mgmtProcessDnodeStatus(void *handle, void *tmrId) {
       if (vnodeList[vnode].cfg.maxSessions <= 0) {
         pVload->dropStatus = TSDB_VN_STATUS_READY;
         pVload->status = TSDB_VN_STATUS_READY;
-        mPrint("vid:%d, drop finished", pObj->privateIp, vnode);
+        mPrint("dnode:%s, vid:%d, drop finished", taosIpStr(pObj->privateIp), vnode);
         taosTmrStart(mgmtMonitorDbDrop, 10000, NULL, mgmtTmr);
       }
     }

src/system/lite/src/vnodePeer.spec.c

@@ -15,6 +15,7 @@
 
 #define _DEFAULT_SOURCE
 #include "vnode.h"
+#include "tstatus.h"
 
 int vnodeInitPeer(int numOfThreads) { return 0; }
 
@@ -30,8 +31,8 @@ void vnodeBroadcastStatusToUnsyncedPeer(SVnodeObj *pVnode) {}
 
 int vnodeOpenPeerVnode(int vnode) {
   SVnodeObj *pVnode = vnodeList + vnode;
-  pVnode->status = (pVnode->cfg.replications > 1) ? TSDB_STATUS_UNSYNCED : TSDB_STATUS_MASTER;
-  dTrace("vid:%d, vnode status:%d numOfPeers:%d", vnode, pVnode->status, pVnode->cfg.replications-1);
+  pVnode->vnodeStatus = (pVnode->cfg.replications > 1) ? TSDB_VNODE_STATUS_UNSYNCED : TSDB_VNODE_STATUS_MASTER;
+  dTrace("vid:%d, status:%s numOfPeers:%d", vnode, taosGetVnodeStatusStr(pVnode->vnodeStatus), pVnode->cfg.replications - 1);
   vnodeUpdateStreamRole(pVnode);
   return 0;
 }

src/util/src/tstatus.c

@@ -13,10 +13,54 @@
  * along with this program. If not, see <http://www.gnu.org/licenses/>.
  */
 
-char* sdbDnodeStatusStr[] = {"offline", "creating", "unsynced", "slave", "master", "ready"};
+#include "taosmsg.h"
+#include "tsdb.h"
 
-char* sdbDnodeBalanceStateStr[] = {"balanced", "balancing", "offline removing", "shell removing"};
+const char* taosGetVnodeStatusStr(int vnodeStatus) {
+  switch (vnodeStatus) {
+    case TSDB_VNODE_STATUS_OFFLINE:return "offline";
+    case TSDB_VNODE_STATUS_CREATING: return "creating";
+    case TSDB_VNODE_STATUS_UNSYNCED: return "unsynced";
+    case TSDB_VNODE_STATUS_SLAVE: return "slave";
+    case TSDB_VNODE_STATUS_MASTER: return "master";
+    case TSDB_VNODE_STATUS_CLOSING: return "closing";
+    case TSDB_VNODE_STATUS_DELETING: return "deleting";
+    default: return "undefined";
+  }
+}
 
-char* sdbVnodeSyncStatusStr[] = {"init", "syncing", "sync_cache", "sync_file"};
+const char* taosGetDnodeStatusStr(int dnodeStatus) {
+  switch (dnodeStatus) {
+    case TSDB_DNODE_STATUS_OFFLINE: return "offline";
+    case TSDB_DNODE_STATUS_READY: return "ready";
+    default: return "undefined";
+  }
+}
 
-char* sdbVnodeDropStateStr[] = {"ready", "dropping"};
+const char* taosGetDnodeBalanceStateStr(int dnodeBalanceStatus) {
+  switch (dnodeBalanceStatus) {
+    case LB_DNODE_STATE_BALANCED: return "balanced";
+    case LB_DNODE_STATE_BALANCING: return "balancing";
+    case LB_DNODE_STATE_OFFLINE_REMOVING: return "offline removing";
+    case LB_DNODE_STATE_SHELL_REMOVING: return "removing";
+    default: return "undefined";
+  }
+}
+
+const char* taosGetVnodeSyncStatusStr(int vnodeSyncStatus) {
+  switch (vnodeSyncStatus) {
+    case STDB_SSTATUS_INIT: return "init";
+    case TSDB_SSTATUS_SYNCING: return "syncing";
+    case TSDB_SSTATUS_SYNC_CACHE: return "sync_cache";
+    case TSDB_SSTATUS_SYNC_FILE: return "sync_file";
+    default: return "undefined";
+  }
+}
+
+const char* taosGetVnodeDropStatusStr(int dropping) {
+  switch (dropping) {
+    case 0: return "ready";
+    case 1: return "dropping";
+    default: return "undefined";
+  }
+}
\ No newline at end of file