Merge branch 'develop' into feature/query

packaging/tools/post.sh

@@ -265,8 +265,14 @@ function install_config() {
         [ -f ${cfg_dir}/taos.cfg ] && ${csudo} cp ${cfg_dir}/taos.cfg ${cfg_install_dir}
         ${csudo} chmod 644 ${cfg_install_dir}/*
     fi
+    
+    # Save standard input to 6 and open / dev / TTY on standard input
+    exec 6<&0 0</dev/tty 
 
     local_fqdn_check
+    
+    # restore the backup standard input, and turn off 6
+    exec 0<&6 6<&-
 
     ${csudo} mv ${cfg_dir}/taos.cfg ${cfg_dir}/taos.cfg.org
     ${csudo} ln -s ${cfg_install_dir}/taos.cfg ${cfg_dir}
@@ -422,7 +428,7 @@ function install_service() {
 }
 
 function install_TDengine() {
-    echo -e "${GREEN}Start to install TDEngine...${NC}"
+    echo -e "${GREEN}Start to install TDengine...${NC}"
 
     #install log and data dir , then ln to /usr/local/taos
     ${csudo} mkdir -p ${log_dir} && ${csudo} chmod 777 ${log_dir}

packaging/tools/preun.sh

@@ -119,4 +119,4 @@ if ((${service_mod}==2)); then
     kill_taosd
 fi
 
-echo -e "${GREEN}TDEngine is removed successfully!${NC}"
+echo -e "${GREEN}TDengine is removed successfully!${NC}"

packaging/tools/set_core.sh

@@ -2,19 +2,39 @@
 #
 # This file is used to set config for core when taosd crash
 
+# Color setting
+RED='\033[0;31m'
+GREEN='\033[1;32m'
+GREEN_DARK='\033[0;32m'
+GREEN_UNDERLINE='\033[4;32m'
+NC='\033[0m'
+
 set -e
 # set -x
+corePath=$1
 
 csudo=""
 if command -v sudo > /dev/null; then
   csudo="sudo"
 fi
 
-#ulimit -c unlimited
+if [[ ! -n ${corePath} ]]; then
+  echo -e -n "${GREEN}Please enter a file directory to save the coredump file${NC}:"
+	read corePath
+  while true; do
+    if [[ ! -z "$corePath" ]]; then
+      break
+    else
+      read -p "Please enter a file directory to save the coredump file:" corePath
+    fi
+  done
+fi
+
+ulimit -c unlimited
 ${csudo} sed -i '/ulimit -c unlimited/d' /etc/profile ||:
 ${csudo} sed -i '$a\ulimit -c unlimited' /etc/profile ||:
 source /etc/profile
 
-${csudo} mkdir -p /coredump  ||:
+${csudo} mkdir -p ${corePath}  ||:
-${csudo} sysctl -w kernel.core_pattern='/coredump/core-%e-%p'  ||:
+${csudo} sysctl -w kernel.core_pattern=${corePath}/core-%e-%p  ||:
-${csudo} echo '/coredump/core-%e-%p' | ${csudo} tee /proc/sys/kernel/core_pattern  ||:
+${csudo} echo "${corePath}/core-%e-%p" | ${csudo} tee /proc/sys/kernel/core_pattern  ||:

src/balance/src/bnMain.c

@@ -330,7 +330,7 @@ void bnReset() {
   tsAccessSquence = 0;
 }
 
-static int32_t bnMonitorVgroups() {
+static bool bnMonitorVgroups() {
   void *  pIter = NULL;
   SVgObj *pVgroup = NULL;
   bool    hasUpdatingVgroup = false;
@@ -489,6 +489,7 @@ void bnCheckStatus() {
         mInfo("dnode:%d, set to offline state, access seq:%d last seq:%d laststat:%d", pDnode->dnodeId, tsAccessSquence,
               pDnode->lastAccess, pDnode->status);
         bnSetVgroupOffline(pDnode);
+        bnStartTimer(3000);
       }
     }
     mnodeDecDnodeRef(pDnode);

src/balance/src/bnThread.c

@@ -31,7 +31,10 @@ static void *bnThreadFunc(void *arg) {
     }
 
     pthread_cond_wait(&tsBnThread.cond, &tsBnThread.mutex);
+    mDebug("balance thread wakes up to work");
     bool updateSoon = bnStart();
+    mDebug("balance thread finished this poll, updateSoon:%d", updateSoon);
+    
     bnStartTimer(updateSoon ? 1000 : -1);
     pthread_mutex_unlock(&(tsBnThread.mutex));
   }
@@ -101,8 +104,8 @@ static void bnProcessTimer(void *handle, void *tmrId) {
   tsBnThread.timer = NULL;
   tsAccessSquence++;
 
-  bnCheckStatus();
   bnStartTimer(-1);
+  bnCheckStatus();
 
   if (handle == NULL) {
     if (tsAccessSquence % tsBalanceInterval == 0) {
@@ -121,6 +124,7 @@ void bnStartTimer(int64_t mseconds) {
 
   bool updateSoon = (mseconds != -1);
   if (updateSoon) {
+    mTrace("balance function will be called after %" PRId64 " ms", mseconds);
     taosTmrReset(bnProcessTimer, mseconds, (void *)mseconds, tsMnodeTmr, &tsBnThread.timer);
   } else {
     taosTmrReset(bnProcessTimer, tsStatusInterval * 1000, NULL, tsMnodeTmr, &tsBnThread.timer);

src/common/inc/tglobal.h

@@ -101,7 +101,8 @@ extern int32_t tsAlternativeRole;
 extern int32_t tsBalanceInterval;
 extern int32_t tsOfflineThreshold;
 extern int32_t tsMnodeEqualVnodeNum;
-extern int32_t tsFlowCtrl;
+extern int32_t tsEnableFlowCtrl;
+extern int32_t tsEnableSlaveQuery;
 
 // restful
 extern int32_t  tsEnableHttpModule;

src/common/src/tglobal.c

@@ -138,7 +138,8 @@ int32_t tsAlternativeRole = 0;
 int32_t tsBalanceInterval = 300;  // seconds
 int32_t tsOfflineThreshold = 86400*100;   // seconds 10days
 int32_t tsMnodeEqualVnodeNum = 4;
-int32_t tsFlowCtrl = 1;
+int32_t tsEnableFlowCtrl = 1;
+int32_t tsEnableSlaveQuery = 1;
 
 // restful
 int32_t  tsEnableHttpModule = 1;
@@ -542,7 +543,7 @@ static void doInitGlobalConfig(void) {
   cfg.ptr = &tsOfflineThreshold;
   cfg.valType = TAOS_CFG_VTYPE_INT32;
   cfg.cfgType = TSDB_CFG_CTYPE_B_CONFIG | TSDB_CFG_CTYPE_B_SHOW;
-  cfg.minValue = 5;
+  cfg.minValue = 3;
   cfg.maxValue = 7200000;
   cfg.ptrLength = 0;
   cfg.unitType = TAOS_CFG_UTYPE_SECOND;
@@ -1004,7 +1005,17 @@ static void doInitGlobalConfig(void) {
 
     // module configs
   cfg.option = "flowctrl";
-  cfg.ptr = &tsFlowCtrl;
+  cfg.ptr = &tsEnableFlowCtrl;
+  cfg.valType = TAOS_CFG_VTYPE_INT32;
+  cfg.cfgType = TSDB_CFG_CTYPE_B_CONFIG | TSDB_CFG_CTYPE_B_SHOW;
+  cfg.minValue = 0;
+  cfg.maxValue = 1;
+  cfg.ptrLength = 0;
+  cfg.unitType = TAOS_CFG_UTYPE_NONE;
+  taosInitConfigOption(cfg);
+
+  cfg.option = "slaveQuery";
+  cfg.ptr = &tsEnableSlaveQuery;
   cfg.valType = TAOS_CFG_VTYPE_INT32;
   cfg.cfgType = TSDB_CFG_CTYPE_B_CONFIG | TSDB_CFG_CTYPE_B_SHOW;
   cfg.minValue = 0;

src/dnode/src/dnodeMain.c

@@ -113,6 +113,7 @@ static void dnodeCleanupTmr() {
 int32_t dnodeInitSystem() {
   dnodeSetRunStatus(TSDB_RUN_STATUS_INITIALIZE);
   tscEmbedded  = 1;
+  taosIgnSIGPIPE();
   taosBlockSIGPIPE();
   taosResolveCRC();
   taosInitGlobalCfg();
@@ -120,7 +121,6 @@ int32_t dnodeInitSystem() {
   taosSetCoreDump();
   taosInitNotes();
   dnodeInitTmr();
-  signal(SIGPIPE, SIG_IGN);
 
   if (dnodeCreateDir(tsLogDir) < 0) {
    printf("failed to create dir: %s, reason: %s\n", tsLogDir, strerror(errno));

src/dnode/src/dnodeVRead.c

@@ -54,6 +54,7 @@ void dnodeCleanupVRead() {
 void dnodeDispatchToVReadQueue(SRpcMsg *pMsg) {
   int32_t queuedMsgNum = 0;
   int32_t leftLen = pMsg->contLen;
+  int32_t code = TSDB_CODE_VND_INVALID_VGROUP_ID;
   char *  pCont = pMsg->pCont;
 
   while (leftLen > 0) {
@@ -64,7 +65,7 @@ void dnodeDispatchToVReadQueue(SRpcMsg *pMsg) {
     assert(pHead->contLen > 0);
     void *pVnode = vnodeAcquire(pHead->vgId);
     if (pVnode != NULL) {
-      int32_t code = vnodeWriteToRQueue(pVnode, pCont, pHead->contLen, TAOS_QTYPE_RPC, pMsg);
+      code = vnodeWriteToRQueue(pVnode, pCont, pHead->contLen, TAOS_QTYPE_RPC, pMsg);
       if (code == TSDB_CODE_SUCCESS) queuedMsgNum++;
       vnodeRelease(pVnode);
     }
@@ -74,7 +75,7 @@ void dnodeDispatchToVReadQueue(SRpcMsg *pMsg) {
   }
 
   if (queuedMsgNum == 0) {
-    SRpcMsg rpcRsp = {.handle = pMsg->handle, .code = TSDB_CODE_VND_INVALID_VGROUP_ID};
+    SRpcMsg rpcRsp = {.handle = pMsg->handle, .code = code};
     rpcSendResponse(&rpcRsp);
   }
 

src/dnode/src/dnodeVWrite.c

@@ -188,6 +188,7 @@ static void *dnodeProcessVWriteQueue(void *wparam) {
   int32_t        numOfMsgs;
   int32_t        qtype;
 
+  taosBlockSIGPIPE();
   dDebug("dnode vwrite worker:%d is running", pWorker->workerId);
 
   while (1) {

src/kit/taosdump/taosdump.c

@@ -332,6 +332,9 @@ static error_t parse_opt(int key, char *arg, struct argp_state *state) {
       break;
     case 'N':
       arguments->data_batch = atoi(arg);
+      if (arguments->data_batch >= INT16_MAX) {
+        arguments->data_batch = INT16_MAX - 1;
+      } 
       break;
     case 'L': 
     {

src/mnode/src/mnodeMnode.c

@@ -377,6 +377,24 @@ static int32_t mnodeCreateMnodeCb(SMnodeMsg *pMsg, int32_t code) {
   return code;
 }
 
+static bool mnodeAllOnline() {
+  void *pIter = NULL;
+  bool  allOnline = true;
+
+  while (1) {
+    SMnodeObj *pMnode = NULL;
+    pIter = mnodeGetNextMnode(pIter, &pMnode);
+    if (pMnode == NULL) break;
+    if (pMnode->role != TAOS_SYNC_ROLE_MASTER && pMnode->role != TAOS_SYNC_ROLE_SLAVE) {
+      allOnline = false;
+      mnodeDecMnodeRef(pMnode);
+    }
+  }
+  mnodeCancelGetNextMnode(pIter);
+
+  return allOnline;
+}
+
 void mnodeCreateMnode(int32_t dnodeId, char *dnodeEp, bool needConfirm) {
   SMnodeObj *pMnode = calloc(1, sizeof(SMnodeObj));
   pMnode->mnodeId = dnodeId;
@@ -389,6 +407,11 @@ void mnodeCreateMnode(int32_t dnodeId, char *dnodeEp, bool needConfirm) {
     .fpRsp   = mnodeCreateMnodeCb
   };
 
+  if (needConfirm && !mnodeAllOnline()) {
+    mDebug("wait all mnode online then create new mnode");
+    return;
+  }
+
   int32_t code = TSDB_CODE_SUCCESS;
   if (needConfirm) {
     code = mnodeSendCreateMnodeMsg(dnodeId, dnodeEp);

src/mnode/src/mnodeSdb.c

@@ -1081,6 +1081,8 @@ static void *sdbWorkerFp(void *pWorker) {
   int32_t  qtype;
   void *   unUsed;
 
+  taosBlockSIGPIPE();
+
   while (1) {
     int32_t numOfMsgs = taosReadAllQitemsFromQset(tsSdbWQset, tsSdbWQall, &unUsed);
     if (numOfMsgs == 0) {

src/mnode/src/mnodeVgroup.c

@@ -659,7 +659,7 @@ static int32_t mnodeGetVgroupMeta(STableMetaMsg *pMeta, SShowObj *pShow, void *p
 
   pShow->bytes[cols] = 4;
   pSchema[cols].type = TSDB_DATA_TYPE_INT;
-  strcpy(pSchema[cols].name, "onlineVnodes");
+  strcpy(pSchema[cols].name, "onlines");
   pSchema[cols].bytes = htons(pShow->bytes[cols]);
   cols++;
 
@@ -674,13 +674,13 @@ static int32_t mnodeGetVgroupMeta(STableMetaMsg *pMeta, SShowObj *pShow, void *p
   for (int32_t i = 0; i < pShow->maxReplica; ++i) {
     pShow->bytes[cols] = 2;
     pSchema[cols].type = TSDB_DATA_TYPE_SMALLINT;
-    snprintf(pSchema[cols].name, TSDB_COL_NAME_LEN, "v%dDnode", i + 1);
+    snprintf(pSchema[cols].name, TSDB_COL_NAME_LEN, "v%d_dnode", i + 1);
     pSchema[cols].bytes = htons(pShow->bytes[cols]);
     cols++;
 
     pShow->bytes[cols] = 9 + VARSTR_HEADER_SIZE;
     pSchema[cols].type = TSDB_DATA_TYPE_BINARY;
-    snprintf(pSchema[cols].name, TSDB_COL_NAME_LEN, "v%dStatus", i + 1);
+    snprintf(pSchema[cols].name, TSDB_COL_NAME_LEN, "v%d_status", i + 1);
     pSchema[cols].bytes = htons(pShow->bytes[cols]);
     cols++;
   }

src/os/inc/osSocket.h

@@ -59,6 +59,7 @@ extern "C" {
 
 // TAOS_OS_FUNC_SOCKET
 int32_t taosSetNonblocking(SOCKET sock, int32_t on);
+void    taosIgnSIGPIPE();
 void    taosBlockSIGPIPE();
 
 // TAOS_OS_FUNC_SOCKET_SETSOCKETOPT

src/os/src/detail/osSocket.c

@@ -39,6 +39,10 @@ int32_t taosSetNonblocking(SOCKET sock, int32_t on) {
   return 0;
 }
 
+void taosIgnSIGPIPE() {
+  signal(SIGPIPE, SIG_IGN);
+}
+
 void taosBlockSIGPIPE() {
   sigset_t signal_mask;
   sigemptyset(&signal_mask);

src/os/src/windows/wSocket.c

@@ -46,6 +46,7 @@ int32_t taosSetNonblocking(SOCKET sock, int32_t on) {
   return 0;
 }
 
+void taosIgnSIGPIPE() {}
 void taosBlockSIGPIPE() {}
 
 int32_t taosSetSockOpt(SOCKET socketfd, int32_t level, int32_t optname, void *optval, int32_t optlen) {

src/sync/inc/syncInt.h

@@ -38,7 +38,7 @@ extern "C" {
 #define SYNC_MAX_FWDS 512
 #define SYNC_FWD_TIMER 300
 #define SYNC_ROLE_TIMER 15000             // ms
-#define SYNC_CHECK_INTERVAL 1             // ms
+#define SYNC_CHECK_INTERVAL 1000          // ms
 #define SYNC_WAIT_AFTER_CHOOSE_MASTER 10  // ms
 
 #define nodeRole    pNode->peerInfo[pNode->selfIndex]->role
@@ -86,9 +86,10 @@ typedef struct SsyncPeer {
   int32_t  peerFd;          // forward FD
   int32_t  numOfRetrieves;  // number of retrieves tried
   int32_t  fileChanged;     // a flag to indicate file is changed during retrieving process
+  int32_t  refCount;
+  int64_t  rid;
   void *   timer;
   void *   pConn;
-  int32_t  refCount;  // reference count
   struct   SSyncNode *pSyncNode;
 } SSyncPeer;
 
@@ -98,6 +99,7 @@ typedef struct SSyncNode {
   int8_t       quorum;
   int8_t       selfIndex;
   uint32_t     vgId;
+  int32_t      refCount;
   int64_t      rid;
   SSyncPeer *  peerInfo[TAOS_SYNC_MAX_REPLICA + 1];  // extra one for arbitrator
   SSyncPeer *  pMaster;
@@ -121,13 +123,13 @@ extern int32_t tsSyncNum;
 extern char    tsNodeFqdn[TSDB_FQDN_LEN];
 extern char *  syncStatus[];
 
-void *syncRetrieveData(void *param);
+void *     syncRetrieveData(void *param);
-void *syncRestoreData(void *param);
+void *     syncRestoreData(void *param);
-int32_t syncSaveIntoBuffer(SSyncPeer *pPeer, SWalHead *pHead);
+int32_t    syncSaveIntoBuffer(SSyncPeer *pPeer, SWalHead *pHead);
-void    syncRestartConnection(SSyncPeer *pPeer);
+void       syncRestartConnection(SSyncPeer *pPeer);
-void    syncBroadcastStatus(SSyncNode *pNode);
+void       syncBroadcastStatus(SSyncNode *pNode);
-void    syncAddPeerRef(SSyncPeer *pPeer);
+SSyncPeer *syncAcquirePeer(int64_t rid);
-int32_t syncDecPeerRef(SSyncPeer *pPeer);
+void       syncReleasePeer(SSyncPeer *pPeer);
 
 #ifdef __cplusplus
 }

src/sync/inc/syncTcp.h

@@ -25,14 +25,14 @@ typedef struct {
   uint32_t  serverIp;
   int16_t   port;
   int32_t   bufferSize;
-  void    (*processBrokenLink)(void *ahandle);
+  void    (*processBrokenLink)(int64_t handleId);
-  int32_t (*processIncomingMsg)(void *ahandle, void *buffer);
+  int32_t (*processIncomingMsg)(int64_t handleId, void *buffer);
   void    (*processIncomingConn)(int32_t fd, uint32_t ip);
 } SPoolInfo;
 
 void *syncOpenTcpThreadPool(SPoolInfo *pInfo);
 void  syncCloseTcpThreadPool(void *);
-void *syncAllocateTcpConn(void *, void *ahandle, int32_t connFd);
+void *syncAllocateTcpConn(void *, int64_t rid, int32_t connFd);
 void  syncFreeTcpConn(void *);
 
 #ifdef __cplusplus

src/sync/src/syncArbitrator.c

@@ -29,8 +29,8 @@
 
 static void    arbSignalHandler(int32_t signum, siginfo_t *sigInfo, void *context);
 static void    arbProcessIncommingConnection(int32_t connFd, uint32_t sourceIp);
-static void    arbProcessBrokenLink(void *param);
+static void    arbProcessBrokenLink(int64_t rid);
-static int32_t arbProcessPeerMsg(void *param, void *buffer);
+static int32_t arbProcessPeerMsg(int64_t rid, void *buffer);
 static tsem_t  tsArbSem;
 static void *  tsArbTcpPool;
 
@@ -138,20 +138,20 @@ static void arbProcessIncommingConnection(int32_t connFd, uint32_t sourceIp) {
 
   sDebug("%s, arbitrator request is accepted", pNode->id);
   pNode->nodeFd = connFd;
-  pNode->pConn = syncAllocateTcpConn(tsArbTcpPool, pNode, connFd);
+  pNode->pConn = syncAllocateTcpConn(tsArbTcpPool, (int64_t)pNode, connFd);
 
   return;
 }
 
-static void arbProcessBrokenLink(void *param) {
+static void arbProcessBrokenLink(int64_t rid) {
-  SNodeConn *pNode = param;
+  SNodeConn *pNode = (SNodeConn *)rid;
 
   sDebug("%s, TCP link is broken since %s, close connection", pNode->id, strerror(errno));
   tfree(pNode);
 }
 
-static int32_t arbProcessPeerMsg(void *param, void *buffer) {
+static int32_t arbProcessPeerMsg(int64_t rid, void *buffer) {
-  SNodeConn *pNode = param;
+  SNodeConn *pNode = (SNodeConn *)rid;
   SSyncHead  head;
   int32_t    bytes = 0;
   char *     cont = (char *)buffer;

src/sync/src/syncMain.c

@@ -35,19 +35,21 @@ char    tsNodeFqdn[TSDB_FQDN_LEN] = {0};
 static void *  tsTcpPool = NULL;
 static void *  tsSyncTmrCtrl = NULL;
 static void *  tsVgIdHash = NULL;
-static int32_t tsSyncRefId = -1;
+static int32_t tsNodeRefId = -1;
+static int32_t tsPeerRefId = -1;
 
 // local functions
 static void    syncProcessSyncRequest(char *pMsg, SSyncPeer *pPeer);
 static void    syncRecoverFromMaster(SSyncPeer *pPeer);
 static void    syncCheckPeerConnection(void *param, void *tmrId);
-static void    syncSendPeersStatusMsgToPeer(SSyncPeer *pPeer, char ack, int8_t type, uint16_t tranId);
+static int32_t syncSendPeersStatusMsgToPeer(SSyncPeer *pPeer, char ack, int8_t type, uint16_t tranId);
-static void    syncProcessBrokenLink(void *param);
+static void    syncProcessBrokenLink(int64_t rid);
-static int32_t syncProcessPeerMsg(void *param, void *buffer);
+static int32_t syncProcessPeerMsg(int64_t rid, void *buffer);
 static void    syncProcessIncommingConnection(int32_t connFd, uint32_t sourceIp);
 static void    syncRemovePeer(SSyncPeer *pPeer);
 static void    syncAddArbitrator(SSyncNode *pNode);
 static void    syncFreeNode(void *);
+static void    syncFreePeer(void *);
 static void    syncRemoveConfirmedFwdInfo(SSyncNode *pNode);
 static void    syncMonitorFwdInfos(void *param, void *tmrId);
 static void    syncMonitorNodeRole(void *param, void *tmrId);
@@ -55,7 +57,12 @@ static void    syncProcessFwdAck(SSyncNode *pNode, SFwdInfo *pFwdInfo, int32_t c
 static int32_t syncSaveFwdInfo(SSyncNode *pNode, uint64_t version, void *mhandle);
 static void    syncRestartPeer(SSyncPeer *pPeer);
 static int32_t syncForwardToPeerImpl(SSyncNode *pNode, void *data, void *mhandle, int32_t qtyp);
+
 static SSyncPeer *syncAddPeer(SSyncNode *pNode, const SNodeInfo *pInfo);
+static void       syncStartCheckPeerConn(SSyncPeer *pPeer);
+static void       syncStopCheckPeerConn(SSyncPeer *pPeer);
+static SSyncNode *syncAcquireNode(int64_t rid);
+static void       syncReleaseNode(SSyncNode *pNode);
 
 char* syncRole[] = {
   "offline",
@@ -87,29 +94,34 @@ int32_t syncInit() {
   tsTcpPool = syncOpenTcpThreadPool(&info);
   if (tsTcpPool == NULL) {
     sError("failed to init tcpPool");
+    syncCleanUp();
     return -1;
   }
 
   tsSyncTmrCtrl = taosTmrInit(1000, 50, 10000, "SYNC");
   if (tsSyncTmrCtrl == NULL) {
     sError("failed to init tmrCtrl");
-    syncCloseTcpThreadPool(tsTcpPool);
+    syncCleanUp();
-    tsTcpPool = NULL;
     return -1;
   }
 
   tsVgIdHash = taosHashInit(TSDB_MIN_VNODES, taosGetDefaultHashFunction(TSDB_DATA_TYPE_INT), true, HASH_ENTRY_LOCK);
   if (tsVgIdHash == NULL) {
     sError("failed to init vgIdHash");
-    taosTmrCleanUp(tsSyncTmrCtrl);
+    syncCleanUp();
-    syncCloseTcpThreadPool(tsTcpPool);
-    tsTcpPool = NULL;
-    tsSyncTmrCtrl = NULL;
     return -1;
   }
 
-  tsSyncRefId = taosOpenRef(200, syncFreeNode);
+  tsNodeRefId = taosOpenRef(200, syncFreeNode);
-  if (tsSyncRefId < 0) {
+  if (tsNodeRefId < 0) {
+    sError("failed to init node ref");
+    syncCleanUp();
+    return -1;
+  }
+
+  tsPeerRefId = taosOpenRef(1000, syncFreePeer);
+  if (tsPeerRefId < 0) {
+    sError("failed to init peer ref");
     syncCleanUp();
     return -1;
   }
@@ -121,12 +133,12 @@ int32_t syncInit() {
 }
 
 void syncCleanUp() {
-  if (tsTcpPool) {
+  if (tsTcpPool != NULL) {
     syncCloseTcpThreadPool(tsTcpPool);
     tsTcpPool = NULL;
   }
 
-  if (tsSyncTmrCtrl) {
+  if (tsSyncTmrCtrl != NULL) {
     taosTmrCleanUp(tsSyncTmrCtrl);
     tsSyncTmrCtrl = NULL;
   }
@@ -136,8 +148,15 @@ void syncCleanUp() {
     tsVgIdHash = NULL;
   }
 
-  taosCloseRef(tsSyncRefId);
+  if (tsNodeRefId != -1) {
-  tsSyncRefId = -1;
+    taosCloseRef(tsNodeRefId);
+    tsNodeRefId = -1;
+  }
+
+  if (tsPeerRefId != -1) {
+    taosCloseRef(tsPeerRefId);
+    tsPeerRefId = -1;
+  }
 
   sInfo("sync module is cleaned up");
 }
@@ -170,7 +189,8 @@ int64_t syncStart(const SSyncInfo *pInfo) {
   pNode->quorum = pCfg->quorum;
   if (pNode->quorum > pNode->replica) pNode->quorum = pNode->replica;
 
-  pNode->rid = taosAddRef(tsSyncRefId, pNode);
+  pNode->refCount = 1;
+  pNode->rid = taosAddRef(tsNodeRefId, pNode);
   if (pNode->rid < 0) {
     syncFreeNode(pNode);
     return -1;
@@ -232,13 +252,18 @@ int64_t syncStart(const SSyncInfo *pInfo) {
     (*pNode->notifyRole)(pNode->vgId, nodeRole);
   }
 
+  syncStartCheckPeerConn(pNode->peerInfo[TAOS_SYNC_MAX_REPLICA]);  // arb
+  for (int32_t index = 0; index < pNode->replica; ++index) {
+    syncStartCheckPeerConn(pNode->peerInfo[index]);
+  }
+
   return pNode->rid;
 }
 
 void syncStop(int64_t rid) {
   SSyncPeer *pPeer;
 
-  SSyncNode *pNode = taosAcquireRef(tsSyncRefId, rid);
+  SSyncNode *pNode = syncAcquireNode(rid);
   if (pNode == NULL) return;
 
   sInfo("vgId:%d, cleanup sync", pNode->vgId);
@@ -259,14 +284,14 @@ void syncStop(int64_t rid) {
 
   pthread_mutex_unlock(&pNode->mutex);
 
-  taosReleaseRef(tsSyncRefId, rid);
+  syncReleaseNode(pNode);
-  taosRemoveRef(tsSyncRefId, rid);
+  taosRemoveRef(tsNodeRefId, rid);
 }
 
 int32_t syncReconfig(int64_t rid, const SSyncCfg *pNewCfg) {
   int32_t i, j;
 
-  SSyncNode *pNode = taosAcquireRef(tsSyncRefId, rid);
+  SSyncNode *pNode = syncAcquireNode(rid);
   if (pNode == NULL) return TSDB_CODE_SYN_INVALID_CONFIG;
 
   sInfo("vgId:%d, reconfig, role:%s replica:%d old:%d", pNode->vgId, syncRole[nodeRole], pNewCfg->replica,
@@ -274,6 +299,11 @@ int32_t syncReconfig(int64_t rid, const SSyncCfg *pNewCfg) {
 
   pthread_mutex_lock(&pNode->mutex);
 
+  syncStopCheckPeerConn(pNode->peerInfo[TAOS_SYNC_MAX_REPLICA]);  // arb
+  for (int32_t index = 0; index < pNode->replica; ++index) {
+    syncStopCheckPeerConn(pNode->peerInfo[index]);
+  }
+
   for (i = 0; i < pNode->replica; ++i) {
     for (j = 0; j < pNewCfg->replica; ++j) {
       if ((strcmp(pNode->peerInfo[i]->fqdn, pNewCfg->nodeInfo[j].nodeFqdn) == 0) &&
@@ -330,28 +360,32 @@ int32_t syncReconfig(int64_t rid, const SSyncCfg *pNewCfg) {
     (*pNode->notifyRole)(pNode->vgId, nodeRole);
   }
 
+  syncStartCheckPeerConn(pNode->peerInfo[TAOS_SYNC_MAX_REPLICA]);  // arb
+  for (int32_t index = 0; index < pNode->replica; ++index) {
+    syncStartCheckPeerConn(pNode->peerInfo[index]);
+  }
+
   pthread_mutex_unlock(&pNode->mutex);
 
   sInfo("vgId:%d, %d replicas are configured, quorum:%d", pNode->vgId, pNode->replica, pNode->quorum);
   syncBroadcastStatus(pNode);
 
-  taosReleaseRef(tsSyncRefId, rid);
+  syncReleaseNode(pNode);
   return 0;
 }
 
 int32_t syncForwardToPeer(int64_t rid, void *data, void *mhandle, int32_t qtype) {
-  SSyncNode *pNode = taosAcquireRef(tsSyncRefId, rid);
+  SSyncNode *pNode = syncAcquireNode(rid);
-  if (pNode == NULL) return 0; 
+  if (pNode == NULL) return 0;
 
   int32_t code = syncForwardToPeerImpl(pNode, data, mhandle, qtype);
 
-  taosReleaseRef(tsSyncRefId, rid);
+  syncReleaseNode(pNode);
-
   return code;
 }
 
 void syncConfirmForward(int64_t rid, uint64_t version, int32_t code) {
-  SSyncNode *pNode = taosAcquireRef(tsSyncRefId, rid);
+  SSyncNode *pNode = syncAcquireNode(rid);
   if (pNode == NULL) return;
 
   SSyncPeer *pPeer = pNode->pMaster;
@@ -367,14 +401,14 @@ void syncConfirmForward(int64_t rid, uint64_t version, int32_t code) {
     }
   }
 
-  taosReleaseRef(tsSyncRefId, rid);
+  syncReleaseNode(pNode);
 }
 
 #if 0
 void syncRecover(int64_t rid) {
   SSyncPeer *pPeer;
 
-  SSyncNode *pNode = taosAcquireRef(tsSyncRefId, rid);
+  SSyncNode *pNode = syncAcquireNode(rid);
   if (pNode == NULL) return;
 
   // to do: add a few lines to check if recover is OK
@@ -395,12 +429,12 @@ void syncRecover(int64_t rid) {
 
   pthread_mutex_unlock(&pNode->mutex);
 
-  taosReleaseRef(tsSyncRefId, rid);
+  syncReleaseNode(pNode);
 }
 #endif
 
 int32_t syncGetNodesRole(int64_t rid, SNodesRole *pNodesRole) {
-  SSyncNode *pNode = taosAcquireRef(tsSyncRefId, rid);
+  SSyncNode *pNode = syncAcquireNode(rid);
   if (pNode == NULL) return -1;
 
   pNodesRole->selfIndex = pNode->selfIndex;
@@ -409,8 +443,7 @@ int32_t syncGetNodesRole(int64_t rid, SNodesRole *pNodesRole) {
     pNodesRole->role[i] = pNode->peerInfo[i]->role;
   }
 
-  taosReleaseRef(tsSyncRefId, rid);
+  syncReleaseNode(pNode);
-
   return 0;
 }
 
@@ -446,24 +479,61 @@ static void syncAddArbitrator(SSyncNode *pNode) {
 static void syncFreeNode(void *param) {
   SSyncNode *pNode = param;
 
+  int32_t refCount = atomic_sub_fetch_32(&pNode->refCount, 1);
+  sDebug("vgId:%d, syncnode is freed, refCount:%d", pNode->vgId, refCount);
+
   pthread_mutex_destroy(&pNode->mutex);
   tfree(pNode->pRecv);
   tfree(pNode->pSyncFwds);
   tfree(pNode);
 }
 
-void syncAddPeerRef(SSyncPeer *pPeer) { atomic_add_fetch_32(&pPeer->refCount, 1); }
+SSyncNode *syncAcquireNode(int64_t rid) {
-
+  SSyncNode *pNode = taosAcquireRef(tsNodeRefId, rid);
-int32_t syncDecPeerRef(SSyncPeer *pPeer) {
+  if (pNode == NULL) {
-  if (atomic_sub_fetch_32(&pPeer->refCount, 1) == 0) {
+    sDebug("failed to acquire syncnode from refId:%" PRId64, rid);
-    taosReleaseRef(tsSyncRefId, pPeer->pSyncNode->rid);
+  } else {
-
+    int32_t refCount = atomic_add_fetch_32(&pNode->refCount, 1);
-    sDebug("%s, resource is freed", pPeer->id);
+    sTrace("vgId:%d, acquire syncnode refId:%" PRId64 ", refCount:%d", pNode->vgId, rid, refCount);
-    tfree(pPeer);
-    return 0;
   }
 
-  return 1;
+  return pNode;
+}
+
+void syncReleaseNode(SSyncNode *pNode) {
+  int32_t refCount = atomic_sub_fetch_32(&pNode->refCount, 1);
+  sTrace("vgId:%d, dec syncnode refId:%" PRId64 " refCount:%d", pNode->vgId, pNode->rid, refCount);
+
+  taosReleaseRef(tsNodeRefId, pNode->rid);
+}
+
+static void syncFreePeer(void *param) {
+  SSyncPeer *pPeer = param;
+
+  int32_t refCount = atomic_sub_fetch_32(&pPeer->refCount, 1);
+  sDebug("%s, peer is freed, refCount:%d", pPeer->id, refCount);
+
+  syncReleaseNode(pPeer->pSyncNode);
+  tfree(pPeer);
+}
+
+SSyncPeer *syncAcquirePeer(int64_t rid) {
+  SSyncPeer *pPeer = taosAcquireRef(tsPeerRefId, rid);
+  if (pPeer == NULL) {
+    sDebug("failed to acquire peer from refId:%" PRId64, rid);
+  } else {
+    int32_t refCount = atomic_add_fetch_32(&pPeer->refCount, 1);
+    sTrace("%s, acquire peer refId:%" PRId64 ", refCount:%d", pPeer->id, rid, refCount);
+  }
+
+  return pPeer;
+}
+
+void syncReleasePeer(SSyncPeer *pPeer) {
+  int32_t refCount = atomic_sub_fetch_32(&pPeer->refCount, 1);
+  sTrace("%s, dec peer refId:%" PRId64 ", refCount:%d", pPeer->id, pPeer->rid, refCount);
+
+  taosReleaseRef(tsPeerRefId, pPeer->rid);
 }
 
 static void syncClosePeerConn(SSyncPeer *pPeer) {
@@ -473,7 +543,8 @@ static void syncClosePeerConn(SSyncPeer *pPeer) {
   taosClose(pPeer->syncFd);
   if (pPeer->peerFd >= 0) {
     pPeer->peerFd = -1;
-    syncFreeTcpConn(pPeer->pConn);
+    void *pConn = pPeer->pConn;
+    if (pConn != NULL) syncFreeTcpConn(pPeer->pConn);
   }
 }
 
@@ -482,7 +553,28 @@ static void syncRemovePeer(SSyncPeer *pPeer) {
 
   pPeer->ip = 0;
   syncClosePeerConn(pPeer);
-  syncDecPeerRef(pPeer);
+  //taosRemoveRef(tsPeerRefId, pPeer->rid);
+  syncReleasePeer(pPeer);
+}
+
+static void syncStartCheckPeerConn(SSyncPeer *pPeer) {
+  if (pPeer == NULL) return;
+  SSyncNode *pNode = pPeer->pSyncNode;
+
+  int32_t ret = strcmp(pPeer->fqdn, tsNodeFqdn);
+  if (pPeer->nodeId == 0 || (ret > 0) || (ret == 0 && pPeer->port > tsSyncPort)) {
+    int32_t checkMs = 100 + (pNode->vgId * 10) % 100;
+    if (pNode->vgId > 1) checkMs = tsStatusInterval * 1000 + checkMs;
+    sDebug("%s, check peer connection after %d ms", pPeer->id, checkMs);
+    taosTmrReset(syncCheckPeerConnection, checkMs, (void *)pPeer->rid, tsSyncTmrCtrl, &pPeer->timer);
+  }
+}
+
+static void syncStopCheckPeerConn(SSyncPeer *pPeer) {
+  if (pPeer == NULL) return;
+
+  taosTmrStopA(&pPeer->timer);
+  sDebug("%s, stop check peer connection", pPeer->id);
 }
 
 static SSyncPeer *syncAddPeer(SSyncNode *pNode, const SNodeInfo *pInfo) {
@@ -508,17 +600,11 @@ static SSyncPeer *syncAddPeer(SSyncNode *pNode, const SNodeInfo *pInfo) {
   pPeer->role = TAOS_SYNC_ROLE_OFFLINE;
   pPeer->pSyncNode = pNode;
   pPeer->refCount = 1;
+  pPeer->rid = taosAddRef(tsPeerRefId, pPeer);
 
-  sInfo("%s, it is configured, ep:%s:%u", pPeer->id, pPeer->fqdn, pPeer->port);
+  sInfo("%s, %p it is configured, ep:%s:%u rid:%" PRId64, pPeer->id, pPeer, pPeer->fqdn, pPeer->port, pPeer->rid);
-  int32_t ret = strcmp(pPeer->fqdn, tsNodeFqdn);
-  if (pPeer->nodeId == 0 || (ret > 0) || (ret == 0 && pPeer->port > tsSyncPort)) {
-    int32_t checkMs = 100 + (pNode->vgId * 10) % 100;
-    if (pNode->vgId > 1) checkMs = tsStatusInterval * 1000 + checkMs;
-    sDebug("%s, check peer connection after %d ms", pPeer->id, checkMs);
-    taosTmrReset(syncCheckPeerConnection, checkMs, pPeer, tsSyncTmrCtrl, &pPeer->timer);
-  }
 
-  taosAcquireRef(tsSyncRefId, pNode->rid);  
+  (void)syncAcquireNode(pNode->rid);
   return pPeer;
 }
 
@@ -560,6 +646,9 @@ static void syncChooseMaster(SSyncNode *pNode) {
         index = i;
       }
     }
+    sDebug("vgId:%d, master:%s may be choosed, index:%d", pNode->vgId, pNode->peerInfo[index]->id, index);
+  } else {
+    sDebug("vgId:%d, no master election since onlineNum:%d replica:%d", pNode->vgId, onlineNum, pNode->replica);
   }
 
   // add arbitrator connection
@@ -580,6 +669,11 @@ static void syncChooseMaster(SSyncNode *pNode) {
         }
       }
     }
+
+    if (index >= 0) {
+      sDebug("vgId:%d, master:%s may be choosed, index:%d onlineNum(arb):%d replica:%d", pNode->vgId,
+             pNode->peerInfo[index]->id, index, onlineNum, replica);
+    }
   }
 
   if (index >= 0) {
@@ -621,9 +715,13 @@ static SSyncPeer *syncCheckMaster(SSyncNode *pNode) {
 
   if (onlineNum <= replica * 0.5) {
     if (nodeRole != TAOS_SYNC_ROLE_UNSYNCED) {
-      nodeRole = TAOS_SYNC_ROLE_UNSYNCED;
+       if (nodeRole == TAOS_SYNC_ROLE_MASTER && onlineNum == replica * 0.5 && onlineNum >= 1) {
+         sInfo("vgId:%d, self keep work as master, online:%d replica:%d", pNode->vgId, onlineNum, replica);
+       } else {
+        nodeRole = TAOS_SYNC_ROLE_UNSYNCED;
+        sInfo("vgId:%d, self change to unsynced state, online:%d replica:%d", pNode->vgId, onlineNum, replica);
+       }
       (*pNode->notifyRole)(pNode->vgId, nodeRole);
-      sInfo("vgId:%d, self change to unsynced state, online:%d replica:%d", pNode->vgId, onlineNum, replica);
     }
   } else {
     for (int32_t index = 0; index < pNode->replica; ++index) {
@@ -678,7 +776,7 @@ static void syncCheckRole(SSyncPeer *pPeer, SPeerStatus* peersStatus, int8_t new
   if (pMaster) {
     // master is there
     pNode->pMaster = pMaster;
-    sDebug("%s, it is the master, sver:%" PRIu64, pMaster->id, pMaster->version);
+    sDebug("%s, it is the master, replica:%d sver:%" PRIu64, pMaster->id, pNode->replica, pMaster->version);
 
     if (syncValidateMaster(pPeer) < 0) return;
 
@@ -711,10 +809,10 @@ static void syncCheckRole(SSyncPeer *pPeer, SPeerStatus* peersStatus, int8_t new
     }
 
     if (consistent) {
-      sDebug("vgId:%d, choose master", pNode->vgId);
+      sDebug("vgId:%d, choose master, replica:%d", pNode->vgId, pNode->replica);
       syncChooseMaster(pNode);
     } else {
-      sDebug("vgId:%d, cannot choose master since roles inequality", pNode->vgId);
+      sDebug("vgId:%d, cannot choose master since roles inequality, replica:%d", pNode->vgId, pNode->replica);
     }
   }
 
@@ -743,7 +841,7 @@ static void syncRestartPeer(SSyncPeer *pPeer) {
   int32_t ret = strcmp(pPeer->fqdn, tsNodeFqdn);
   if (ret > 0 || (ret == 0 && pPeer->port > tsSyncPort)) {
     sDebug("%s, check peer connection in 1000 ms", pPeer->id);
-    taosTmrReset(syncCheckPeerConnection, SYNC_CHECK_INTERVAL, pPeer, tsSyncTmrCtrl, &pPeer->timer);
+    taosTmrReset(syncCheckPeerConnection, SYNC_CHECK_INTERVAL, (void *)pPeer->rid, tsSyncTmrCtrl, &pPeer->timer);
   }
 }
 
@@ -772,25 +870,30 @@ static void syncProcessSyncRequest(char *msg, SSyncPeer *pPeer) {
   }
 
   // start a new thread to retrieve the data
-  syncAddPeerRef(pPeer);
+  (void)syncAcquirePeer(pPeer->rid);
+
   pthread_attr_t thattr;
   pthread_t      thread;
   pthread_attr_init(&thattr);
   pthread_attr_setdetachstate(&thattr, PTHREAD_CREATE_DETACHED);
-  int32_t ret = pthread_create(&thread, &thattr, syncRetrieveData, pPeer);
+  int32_t ret = pthread_create(&thread, &thattr, syncRetrieveData, (void *)pPeer->rid);
   pthread_attr_destroy(&thattr);
 
   if (ret != 0) {
     sError("%s, failed to create sync thread since %s", pPeer->id, strerror(errno));
-    syncDecPeerRef(pPeer);
   } else {
     pPeer->sstatus = TAOS_SYNC_STATUS_START;
     sDebug("%s, thread is created to retrieve data, set sstatus:%s", pPeer->id, syncStatus[pPeer->sstatus]);
   }
+
+  syncReleasePeer(pPeer);
 }
 
 static void syncNotStarted(void *param, void *tmrId) {
-  SSyncPeer *pPeer = param;
+  int64_t    rid = (int64_t)param;
+  SSyncPeer *pPeer = syncAcquirePeer(rid);
+  if (pPeer == NULL) return;
+
   SSyncNode *pNode = pPeer->pSyncNode;
 
   pthread_mutex_lock(&pNode->mutex);
@@ -799,15 +902,22 @@ static void syncNotStarted(void *param, void *tmrId) {
   sInfo("%s, sync conn is still not up, restart and set sstatus:%s", pPeer->id, syncStatus[pPeer->sstatus]);
   syncRestartConnection(pPeer);
   pthread_mutex_unlock(&pNode->mutex);
+
+  syncReleasePeer(pPeer);
 }
 
 static void syncTryRecoverFromMaster(void *param, void *tmrId) {
-  SSyncPeer *pPeer = param;
+  int64_t    rid = (int64_t)param;
+  SSyncPeer *pPeer = syncAcquirePeer(rid);
+  if (pPeer == NULL) return;
+
   SSyncNode *pNode = pPeer->pSyncNode;
 
   pthread_mutex_lock(&pNode->mutex);
   syncRecoverFromMaster(pPeer);
   pthread_mutex_unlock(&pNode->mutex);
+
+  syncReleasePeer(pPeer);
 }
 
 static void syncRecoverFromMaster(SSyncPeer *pPeer) {
@@ -823,7 +933,7 @@ static void syncRecoverFromMaster(SSyncPeer *pPeer) {
   // Ensure the sync of mnode not interrupted
   if (pNode->vgId != 1 && tsSyncNum >= SYNC_MAX_NUM) {
     sInfo("%s, %d syncs are in process, try later", pPeer->id, tsSyncNum);
-    taosTmrReset(syncTryRecoverFromMaster, 500 + (pNode->vgId * 10) % 200, pPeer, tsSyncTmrCtrl, &pPeer->timer);
+    taosTmrReset(syncTryRecoverFromMaster, 500 + (pNode->vgId * 10) % 200, (void *)pPeer->rid, tsSyncTmrCtrl, &pPeer->timer);
     return;
   }
 
@@ -832,7 +942,7 @@ static void syncRecoverFromMaster(SSyncPeer *pPeer) {
   SSyncMsg msg;
   syncBuildSyncReqMsg(&msg, pNode->vgId);
 
-  taosTmrReset(syncNotStarted, SYNC_CHECK_INTERVAL, pPeer, tsSyncTmrCtrl, &pPeer->timer);
+  taosTmrReset(syncNotStarted, SYNC_CHECK_INTERVAL, (void *)pPeer->rid, tsSyncTmrCtrl, &pPeer->timer);
 
   if (taosWriteMsg(pPeer->peerFd, &msg, sizeof(SSyncMsg)) != sizeof(SSyncMsg)) {
     sError("%s, failed to send sync-req to peer", pPeer->id);
@@ -920,8 +1030,10 @@ static int32_t syncReadPeerMsg(SSyncPeer *pPeer, SSyncHead *pHead) {
   return 0;
 }
 
-static int32_t syncProcessPeerMsg(void *param, void *buffer) {
+static int32_t syncProcessPeerMsg(int64_t rid, void *buffer) {
-  SSyncPeer *pPeer = param;
+  SSyncPeer *pPeer = syncAcquirePeer(rid);
+  if (pPeer == NULL) return -1;
+
   SSyncHead *pHead = buffer;
   SSyncNode *pNode = pPeer->pSyncNode;
   
@@ -942,12 +1054,17 @@ static int32_t syncProcessPeerMsg(void *param, void *buffer) {
   }
 
   pthread_mutex_unlock(&pNode->mutex);
+  syncReleasePeer(pPeer);
 
   return code;
 }
 
-static void syncSendPeersStatusMsgToPeer(SSyncPeer *pPeer, char ack, int8_t type, uint16_t tranId) {
+static int32_t syncSendPeersStatusMsgToPeer(SSyncPeer *pPeer, char ack, int8_t type, uint16_t tranId) {
-  if (pPeer->peerFd < 0 || pPeer->ip == 0) return;
+  if (pPeer->peerFd < 0 || pPeer->ip == 0) {
+    sDebug("%s, failed to send status msg, restart fd:%d", pPeer->id, pPeer->peerFd);
+    syncRestartConnection(pPeer);
+    return -1;
+  }
 
   SSyncNode *  pNode = pPeer->pSyncNode;
   SPeersStatus msg;
@@ -970,9 +1087,11 @@ static void syncSendPeersStatusMsgToPeer(SSyncPeer *pPeer, char ack, int8_t type
     sDebug("%s, status is sent, self:%s:%s:%" PRIu64 ", peer:%s:%s:%" PRIu64 ", ack:%d tranId:%u type:%s pfd:%d",
            pPeer->id, syncRole[nodeRole], syncStatus[nodeSStatus], nodeVersion, syncRole[pPeer->role],
            syncStatus[pPeer->sstatus], pPeer->version, ack, tranId, statusType[type], pPeer->peerFd);
+    return 0;
   } else {
     sDebug("%s, failed to send status msg, restart", pPeer->id);
     syncRestartConnection(pPeer);
+    return -1;
   }
 }
 
@@ -989,7 +1108,7 @@ static void syncSetupPeerConnection(SSyncPeer *pPeer) {
   int32_t connFd = taosOpenTcpClientSocket(pPeer->ip, pPeer->port, 0);
   if (connFd < 0) {
     sDebug("%s, failed to open tcp socket since %s", pPeer->id, strerror(errno));
-    taosTmrReset(syncCheckPeerConnection, SYNC_CHECK_INTERVAL, pPeer, tsSyncTmrCtrl, &pPeer->timer);
+    taosTmrReset(syncCheckPeerConnection, SYNC_CHECK_INTERVAL, (void *)pPeer->rid, tsSyncTmrCtrl, &pPeer->timer);
     return;
   }
 
@@ -1000,17 +1119,19 @@ static void syncSetupPeerConnection(SSyncPeer *pPeer) {
     sDebug("%s, connection to peer server is setup, pfd:%d sfd:%d tranId:%u", pPeer->id, connFd, pPeer->syncFd, msg.tranId);
     pPeer->peerFd = connFd;
     pPeer->role = TAOS_SYNC_ROLE_UNSYNCED;
-    pPeer->pConn = syncAllocateTcpConn(tsTcpPool, pPeer, connFd);
+    pPeer->pConn = syncAllocateTcpConn(tsTcpPool, pPeer->rid, connFd);
-    syncAddPeerRef(pPeer);
   } else {
     sDebug("%s, failed to setup peer connection to server since %s, try later", pPeer->id, strerror(errno));
     taosClose(connFd);
-    taosTmrReset(syncCheckPeerConnection, SYNC_CHECK_INTERVAL, pPeer, tsSyncTmrCtrl, &pPeer->timer);
+    taosTmrReset(syncCheckPeerConnection, SYNC_CHECK_INTERVAL, (void *)pPeer->rid, tsSyncTmrCtrl, &pPeer->timer);
   }
 }
 
 static void syncCheckPeerConnection(void *param, void *tmrId) {
-  SSyncPeer *pPeer = param;
+  int64_t    rid = (int64_t)param;
+  SSyncPeer *pPeer = syncAcquirePeer(rid);
+  if (pPeer == NULL) return;
+
   SSyncNode *pNode = pPeer->pSyncNode;
 
   pthread_mutex_lock(&pNode->mutex);
@@ -1019,6 +1140,8 @@ static void syncCheckPeerConnection(void *param, void *tmrId) {
   syncSetupPeerConnection(pPeer);
 
   pthread_mutex_unlock(&pNode->mutex);
+
+  syncReleasePeer(pPeer);
 }
 
 static void syncCreateRestoreDataThread(SSyncPeer *pPeer) {
@@ -1029,8 +1152,9 @@ static void syncCreateRestoreDataThread(SSyncPeer *pPeer) {
   pthread_attr_init(&thattr);
   pthread_attr_setdetachstate(&thattr, PTHREAD_CREATE_DETACHED);
 
-  syncAddPeerRef(pPeer);
+  (void)syncAcquirePeer(pPeer->rid);
-  int32_t ret = pthread_create(&(thread), &thattr, (void *)syncRestoreData, pPeer);
+
+  int32_t ret = pthread_create(&(thread), &thattr, (void *)syncRestoreData, (void *)pPeer->rid);
   pthread_attr_destroy(&thattr);
 
   if (ret < 0) {
@@ -1038,10 +1162,11 @@ static void syncCreateRestoreDataThread(SSyncPeer *pPeer) {
     nodeSStatus = TAOS_SYNC_STATUS_INIT;
     sError("%s, failed to create sync thread, set sstatus:%s", pPeer->id, syncStatus[nodeSStatus]);
     taosClose(pPeer->syncFd);
-    syncDecPeerRef(pPeer);
   } else {
     sInfo("%s, sync connection is up", pPeer->id);
   }
+
+  syncReleasePeer(pPeer);
 }
 
 static void syncProcessIncommingConnection(int32_t connFd, uint32_t sourceIp) {
@@ -1073,7 +1198,7 @@ static void syncProcessIncommingConnection(int32_t connFd, uint32_t sourceIp) {
     return;
   }
 
-  sDebug("vgId:%d, sync msg is received, tranId:%u", vgId, msg.tranId);
+  sDebug("vgId:%d, sync connection is incomming, tranId:%u", vgId, msg.tranId);
 
   SSyncNode *pNode = *ppNode;
   pthread_mutex_lock(&pNode->mutex);
@@ -1101,8 +1226,7 @@ static void syncProcessIncommingConnection(int32_t connFd, uint32_t sourceIp) {
       sDebug("%s, TCP connection is up, pfd:%d sfd:%d, old pfd:%d", pPeer->id, connFd, pPeer->syncFd, pPeer->peerFd);
       syncClosePeerConn(pPeer);
       pPeer->peerFd = connFd;
-      pPeer->pConn = syncAllocateTcpConn(tsTcpPool, pPeer, connFd);
+      pPeer->pConn = syncAllocateTcpConn(tsTcpPool, pPeer->rid, connFd);
-      syncAddPeerRef(pPeer);
       sDebug("%s, ready to exchange data", pPeer->id);
       syncSendPeersStatusMsgToPeer(pPeer, 1, SYNC_STATUS_EXCHANGE_DATA, syncGenTranId());
     }
@@ -1111,23 +1235,21 @@ static void syncProcessIncommingConnection(int32_t connFd, uint32_t sourceIp) {
   pthread_mutex_unlock(&pNode->mutex);
 }
 
-static void syncProcessBrokenLink(void *param) {
+static void syncProcessBrokenLink(int64_t rid) {
-  if (param == NULL) return;  // the connection for arbitrator
+  SSyncPeer *pPeer = syncAcquirePeer(rid);
-  SSyncPeer *pPeer = param;
+  if (pPeer == NULL) return;
+
   SSyncNode *pNode = pPeer->pSyncNode;
 
-  if (taosAcquireRef(tsSyncRefId, pNode->rid) == NULL) return;
   pthread_mutex_lock(&pNode->mutex);
 
   sDebug("%s, TCP link is broken since %s, pfd:%d sfd:%d", pPeer->id, strerror(errno), pPeer->peerFd, pPeer->syncFd);
   pPeer->peerFd = -1;
 
-  if (syncDecPeerRef(pPeer) != 0) {
+  syncRestartConnection(pPeer);
-    syncRestartConnection(pPeer);
-  }
-
   pthread_mutex_unlock(&pNode->mutex);
-  taosReleaseRef(tsSyncRefId, pNode->rid);
+
+  syncReleasePeer(pPeer);
 }
 
 static int32_t syncSaveFwdInfo(SSyncNode *pNode, uint64_t version, void *mhandle) {
@@ -1198,7 +1320,7 @@ static void syncProcessFwdAck(SSyncNode *pNode, SFwdInfo *pFwdInfo, int32_t code
 
 static void syncMonitorNodeRole(void *param, void *tmrId) {
   int64_t    rid = (int64_t)param;
-  SSyncNode *pNode = taosAcquireRef(tsSyncRefId, rid);
+  SSyncNode *pNode = syncAcquireNode(rid);
   if (pNode == NULL) return;
 
   for (int32_t index = 0; index < pNode->replica; index++) {
@@ -1215,12 +1337,12 @@ static void syncMonitorNodeRole(void *param, void *tmrId) {
   }
 
   pNode->pRoleTimer = taosTmrStart(syncMonitorNodeRole, SYNC_ROLE_TIMER, (void *)pNode->rid, tsSyncTmrCtrl);
-  taosReleaseRef(tsSyncRefId, rid);
+  syncReleaseNode(pNode);
 }
 
 static void syncMonitorFwdInfos(void *param, void *tmrId) {
   int64_t    rid = (int64_t)param;
-  SSyncNode *pNode = taosAcquireRef(tsSyncRefId, rid);
+  SSyncNode *pNode = syncAcquireNode(rid);
   if (pNode == NULL) return;
 
   SSyncFwds *pSyncFwds = pNode->pSyncFwds;
@@ -1246,7 +1368,7 @@ static void syncMonitorFwdInfos(void *param, void *tmrId) {
     pNode->pFwdTimer = taosTmrStart(syncMonitorFwdInfos, SYNC_FWD_TIMER, (void *)pNode->rid, tsSyncTmrCtrl);
   }
 
-  taosReleaseRef(tsSyncRefId, rid);
+  syncReleaseNode(pNode);
 }
 
 static int32_t syncForwardToPeerImpl(SSyncNode *pNode, void *data, void *mhandle, int32_t qtype) {

src/sync/src/syncRestore.c

@@ -90,15 +90,18 @@ static int32_t syncRestoreFile(SSyncPeer *pPeer, uint64_t *fversion) {
       break;
     }
 
+    sDebug("%s, file:%s info is received from master, index:%d size:%" PRId64 " fver:%" PRIu64 " magic:%d", pPeer->id,
+           minfo.name, minfo.index, minfo.size, minfo.fversion, minfo.magic);
+
     // remove extra files on slave between the current and last index
     syncRemoveExtraFile(pPeer, pindex + 1, minfo.index - 1);
     pindex = minfo.index;
 
     // check the file info
     sinfo = minfo;
-    sDebug("%s, get file:%s info size:%" PRId64, pPeer->id, minfo.name, minfo.size);
+    sinfo.magic = (*pNode->getFileInfo)(pNode->vgId, sinfo.name, &sinfo.index, TAOS_SYNC_MAX_INDEX, &sinfo.size, &sinfo.fversion);
-    sinfo.magic = (*pNode->getFileInfo)(pNode->vgId, sinfo.name, &sinfo.index, TAOS_SYNC_MAX_INDEX, &sinfo.size,
+    sDebug("%s, local file:%s info, index:%d size:%" PRId64 " fver:%" PRIu64 " magic:%d", pPeer->id, sinfo.name,
-                                        &sinfo.fversion);
+           sinfo.index, sinfo.size, sinfo.fversion, sinfo.magic);
 
     // if file not there or magic is not the same, file shall be synced
     memset(&fileAck, 0, sizeof(SFileAck));
@@ -116,6 +119,8 @@ static int32_t syncRestoreFile(SSyncPeer *pPeer, uint64_t *fversion) {
     if (fileAck.sync == 0) {
       sDebug("%s, %s is the same", pPeer->id, minfo.name);
       continue;
+    } else {
+      sDebug("%s, %s will be received, size:%" PRId64, pPeer->id, minfo.name, minfo.size);
     }
 
     // if sync is required, open file, receive from master, and write to file
@@ -155,7 +160,7 @@ static int32_t syncRestoreFile(SSyncPeer *pPeer, uint64_t *fversion) {
   return code;
 }
 
-static int32_t syncRestoreWal(SSyncPeer *pPeer) {
+static int32_t syncRestoreWal(SSyncPeer *pPeer, uint64_t *wver) {
   SSyncNode *pNode = pPeer->pSyncNode;
   int32_t    ret, code = -1;
   uint64_t   lastVer = 0;
@@ -198,6 +203,7 @@ static int32_t syncRestoreWal(SSyncPeer *pPeer) {
   }
 
   free(pHead);
+  *wver = lastVer;
   return code;
 }
 
@@ -321,12 +327,19 @@ static int32_t syncRestoreDataStepByStep(SSyncPeer *pPeer) {
 
   nodeVersion = fversion;
 
-  sInfo("%s, start to restore wal", pPeer->id);
+  sInfo("%s, start to restore wal, fver:%" PRIu64, pPeer->id, nodeVersion);
-  if (syncRestoreWal(pPeer) < 0) {
+  uint64_t wver = 0;
-    sError("%s, failed to restore wal", pPeer->id);
+  code = syncRestoreWal(pPeer, &wver);  // lastwar
+  if (code < 0) {
+    sError("%s, failed to restore wal, code:%d", pPeer->id, code);
     return -1;
   }
 
+  if (wver != 0) {
+    nodeVersion = wver;
+    sDebug("%s, restore wal finished, set sver:%" PRIu64, pPeer->id, nodeVersion);
+  }
+
   nodeSStatus = TAOS_SYNC_STATUS_CACHE;
   sInfo("%s, start to insert buffered points, set sstatus:%s", pPeer->id, syncStatus[nodeSStatus]);
   if (syncProcessBufferedFwd(pPeer) < 0) {
@@ -338,7 +351,10 @@ static int32_t syncRestoreDataStepByStep(SSyncPeer *pPeer) {
 }
 
 void *syncRestoreData(void *param) {
-  SSyncPeer *pPeer = param;
+  int64_t    rid = (int64_t)param;
+  SSyncPeer *pPeer = syncAcquirePeer(rid);
+  if (pPeer == NULL) return NULL;
+
   SSyncNode *pNode = pPeer->pSyncNode;
 
   taosBlockSIGPIPE();
@@ -369,7 +385,7 @@ void *syncRestoreData(void *param) {
   taosClose(pPeer->syncFd);
   syncCloseRecvBuffer(pNode);
   __sync_fetch_and_sub(&tsSyncNum, 1);
-  syncDecPeerRef(pPeer);
+  syncReleasePeer(pPeer);
 
   return NULL;
 }

src/sync/src/syncRetrieve.c

@@ -104,7 +104,8 @@ static int32_t syncRetrieveFile(SSyncPeer *pPeer) {
     fileInfo.magic = (*pNode->getFileInfo)(pNode->vgId, fileInfo.name, &fileInfo.index, TAOS_SYNC_MAX_INDEX,
                                            &fileInfo.size, &fileInfo.fversion);
     syncBuildFileInfo(&fileInfo, pNode->vgId);
-    sDebug("%s, file:%s info is sent, size:%" PRId64, pPeer->id, fileInfo.name, fileInfo.size);
+    sDebug("%s, file:%s info is sent, index:%d size:%" PRId64 " fver:%" PRIu64 " magic:%d", pPeer->id, fileInfo.name,
+           fileInfo.index, fileInfo.size, fileInfo.fversion, fileInfo.magic);
 
     // send the file info
     int32_t ret = taosWriteMsg(pPeer->syncFd, &(fileInfo), sizeof(SFileInfo));
@@ -144,6 +145,8 @@ static int32_t syncRetrieveFile(SSyncPeer *pPeer) {
       fileInfo.index++;
       sDebug("%s, %s is the same", pPeer->id, fileInfo.name);
       continue;
+    } else {
+      sDebug("%s, %s will be sent", pPeer->id, fileInfo.name);
     }
 
     // get the full path to file
@@ -461,7 +464,10 @@ static int32_t syncRetrieveDataStepByStep(SSyncPeer *pPeer) {
 }
 
 void *syncRetrieveData(void *param) {
-  SSyncPeer *pPeer = (SSyncPeer *)param;
+  int64_t    rid = (int64_t)param;
+  SSyncPeer *pPeer = syncAcquirePeer(rid);
+  if (pPeer == NULL) return NULL;
+
   SSyncNode *pNode = pPeer->pSyncNode;
   taosBlockSIGPIPE();
 
@@ -490,7 +496,7 @@ void *syncRetrieveData(void *param) {
 
   pPeer->fileChanged = 0;
   taosClose(pPeer->syncFd);
-  syncDecPeerRef(pPeer);
+  syncReleasePeer(pPeer);
 
   return NULL;
 }

src/sync/src/syncTcp.c

@@ -42,7 +42,7 @@ typedef struct SPoolObj {
 
 typedef struct {
   SThreadObj *pThread;
-  void *      ahandle;
+  int64_t     handleId;
   int32_t     fd;
   int32_t     closedByApp;
 } SConnObj;
@@ -112,7 +112,7 @@ void syncCloseTcpThreadPool(void *param) {
   tfree(pPool);
 }
 
-void *syncAllocateTcpConn(void *param, void *pPeer, int32_t connFd) {
+void *syncAllocateTcpConn(void *param, int64_t rid, int32_t connFd) {
   struct epoll_event event;
   SPoolObj *pPool = param;
 
@@ -130,7 +130,7 @@ void *syncAllocateTcpConn(void *param, void *pPeer, int32_t connFd) {
 
   pConn->fd = connFd;
   pConn->pThread = pThread;
-  pConn->ahandle = pPeer;
+  pConn->handleId = rid;
   pConn->closedByApp = 0;
 
   event.events = EPOLLIN | EPOLLRDHUP;
@@ -164,7 +164,7 @@ static void taosProcessBrokenLink(SConnObj *pConn) {
   SPoolInfo * pInfo = &pPool->info;
 
   if (pConn->closedByApp == 0) shutdown(pConn->fd, SHUT_WR);
-  (*pInfo->processBrokenLink)(pConn->ahandle);
+  (*pInfo->processBrokenLink)(pConn->handleId);
 
   pThread->numOfFds--;
   epoll_ctl(pThread->pollFd, EPOLL_CTL_DEL, pConn->fd, NULL);
@@ -221,7 +221,7 @@ static void *syncProcessTcpData(void *param) {
       }
 
       if (pConn->closedByApp == 0) {
-        if ((*pInfo->processIncomingMsg)(pConn->ahandle, buffer) < 0) {
+        if ((*pInfo->processIncomingMsg)(pConn->handleId, buffer) < 0) {
           syncFreeTcpConn(pConn);
           continue;
         }

src/tsdb/src/tsdbCommit.c

@@ -161,6 +161,11 @@ _err:
 
 static void tsdbEndCommit(STsdbRepo *pRepo, int eno) {
   if (pRepo->appH.notifyStatus) pRepo->appH.notifyStatus(pRepo->appH.appH, TSDB_STATUS_COMMIT_OVER, eno);
+  SMemTable *pIMem = pRepo->imem;
+  tsdbLockRepo(pRepo);
+  pRepo->imem = NULL;
+  tsdbUnlockRepo(pRepo);
+  tsdbUnRefMemTable(pRepo, pIMem);
   sem_post(&(pRepo->readyToCommit));
 }
 

src/tsdb/src/tsdbMemTable.c

@@ -17,6 +17,7 @@
 #include "tsdbMain.h"
 
 #define TSDB_DATA_SKIPLIST_LEVEL 5
+#define TSDB_MAX_INSERT_BATCH 512
 
 static SMemTable * tsdbNewMemTable(STsdbRepo *pRepo);
 static void        tsdbFreeMemTable(SMemTable *pMemTable);
@@ -205,7 +206,7 @@ void *tsdbAllocBytes(STsdbRepo *pRepo, int bytes) {
 int tsdbAsyncCommit(STsdbRepo *pRepo) {
   if (pRepo->mem == NULL) return 0;
 
-  SMemTable *pIMem = pRepo->imem;
+  ASSERT(pRepo->imem == NULL);
 
   sem_wait(&(pRepo->readyToCommit));
 
@@ -220,8 +221,6 @@ int tsdbAsyncCommit(STsdbRepo *pRepo) {
   tsdbScheduleCommit(pRepo);
   if (tsdbUnlockRepo(pRepo) < 0) return -1;
 
-  if (tsdbUnRefMemTable(pRepo, pIMem) < 0) return -1;
-
   return 0;
 }
 
@@ -606,19 +605,13 @@ static int tsdbInsertDataToTable(STsdbRepo *pRepo, SSubmitBlk *pBlock, int32_t *
   STable *       pTable = NULL;
   SSubmitBlkIter blkIter = {0};
   SDataRow       row = NULL;
-  void **        rows = NULL;
+  void *         rows[TSDB_MAX_INSERT_BATCH] = {0};
   int            rowCounter = 0;
 
   ASSERT(pBlock->tid < pMeta->maxTables);
   pTable = pMeta->tables[pBlock->tid];
   ASSERT(pTable != NULL && TABLE_UID(pTable) == pBlock->uid);
 
-  rows = (void **)calloc(pBlock->numOfRows, sizeof(void *));
-  if (rows == NULL) {
-    terrno = TSDB_CODE_TDB_OUT_OF_MEMORY;
-    return -1;
-  }
-
   tsdbInitSubmitBlkIter(pBlock, &blkIter);
   while ((row = tsdbGetSubmitBlkNext(&blkIter)) != NULL) {
     if (tsdbCopyRowToMem(pRepo, row, pTable, &(rows[rowCounter])) < 0) {
@@ -632,9 +625,18 @@ static int tsdbInsertDataToTable(STsdbRepo *pRepo, SSubmitBlk *pBlock, int32_t *
     if (rows[rowCounter] != NULL) {
       rowCounter++;
     }
+
+    if (rowCounter == TSDB_MAX_INSERT_BATCH) {
+      if (tsdbInsertDataToTableImpl(pRepo, pTable, rows, rowCounter) < 0) {
+        goto _err;
+      }
+
+      rowCounter = 0;
+      memset(rows, 0, sizeof(rows));
+    }
   }
 
-  if (tsdbInsertDataToTableImpl(pRepo, pTable, rows, rowCounter) < 0) {
+  if (rowCounter > 0 && tsdbInsertDataToTableImpl(pRepo, pTable, rows, rowCounter) < 0) {
     goto _err;
   }
 
@@ -642,11 +644,9 @@ static int tsdbInsertDataToTable(STsdbRepo *pRepo, SSubmitBlk *pBlock, int32_t *
   pRepo->stat.pointsWritten += points * schemaNCols(pSchema);
   pRepo->stat.totalStorage += points * schemaVLen(pSchema);
 
-  free(rows);
   return 0;
 
 _err:
-  free(rows);
   return -1;
 }
 

src/util/src/tsocket.c

@@ -18,6 +18,10 @@
 #include "tsocket.h"
 #include "taoserror.h"
 
+#ifndef SIGPIPE
+  #define SIGPIPE EPIPE
+#endif
+
 int32_t taosGetFqdn(char *fqdn) {
   char hostname[1024];
   hostname[1023] = '\0';
@@ -115,6 +119,10 @@ int32_t taosWriteMsg(SOCKET fd, void *buf, int32_t nbytes) {
       nleft -= nwritten;
       ptr += nwritten;
     }
+
+    if (errno == SIGPIPE || errno == EPIPE) {
+      return -1;
+    }
   }
 
   return (nbytes - nleft);
@@ -142,6 +150,10 @@ int32_t taosReadMsg(SOCKET fd, void *buf, int32_t nbytes) {
       nleft -= nread;
       ptr += nread;
     }
+
+    if (errno == SIGPIPE || errno == EPIPE) {
+      return -1;
+    }
   }
 
   return (nbytes - nleft);

src/vnode/src/vnodeCfg.c

@@ -106,9 +106,10 @@ int32_t vnodeReadCfg(SVnodeObj *pVnode) {
   cJSON *vgCfgVersion = cJSON_GetObjectItem(root, "vgCfgVersion");
   if (!vgCfgVersion || vgCfgVersion->type != cJSON_Number) {
     vError("vgId:%d, failed to read %s, vgCfgVersion not found", pVnode->vgId, file);
-    goto PARSE_VCFG_ERROR;
+    vnodeMsg.cfg.vgCfgVersion = 0;
+  } else {
+    vnodeMsg.cfg.vgCfgVersion = vgCfgVersion->valueint;
   }
-  vnodeMsg.cfg.vgCfgVersion = vgCfgVersion->valueint;
 
   cJSON *cacheBlockSize = cJSON_GetObjectItem(root, "cacheBlockSize");
   if (!cacheBlockSize || cacheBlockSize->type != cJSON_Number) {

src/vnode/src/vnodeMgmt.c

@@ -89,7 +89,10 @@ static void vnodeIncRef(void *ptNode) {
 }
 
 void *vnodeAcquire(int32_t vgId) {
-  SVnodeObj **ppVnode = taosHashGetCB(tsVnodesHash, &vgId, sizeof(int32_t), vnodeIncRef, NULL, sizeof(void *));
+  SVnodeObj **ppVnode = NULL;
+  if (tsVnodesHash != NULL) {
+    ppVnode = taosHashGetCB(tsVnodesHash, &vgId, sizeof(int32_t), vnodeIncRef, NULL, sizeof(void *));
+  }
 
   if (ppVnode == NULL || *ppVnode == NULL) {
     terrno = TSDB_CODE_VND_INVALID_VGROUP_ID;

src/vnode/src/vnodeRead.c

@@ -65,13 +65,17 @@ static int32_t vnodeCheckRead(SVnodeObj *pVnode) {
     return TSDB_CODE_APP_NOT_READY;
   }
 
-  if (pVnode->role != TAOS_SYNC_ROLE_SLAVE && pVnode->role != TAOS_SYNC_ROLE_MASTER) {
+  if (pVnode->role == TAOS_SYNC_ROLE_MASTER) {
-    vDebug("vgId:%d, replica:%d role:%s, refCount:%d pVnode:%p", pVnode->vgId, pVnode->syncCfg.replica,
+    return TSDB_CODE_SUCCESS;
-           syncRole[pVnode->role], pVnode->refCount, pVnode);
-    return TSDB_CODE_APP_NOT_READY;
   }
 
-  return TSDB_CODE_SUCCESS;
+  if (tsEnableSlaveQuery && pVnode->role == TAOS_SYNC_ROLE_SLAVE) {
+    return TSDB_CODE_SUCCESS;
+  }
+
+  vDebug("vgId:%d, replica:%d role:%s, refCount:%d pVnode:%p, cant provide query service", pVnode->vgId, pVnode->syncCfg.replica,
+         syncRole[pVnode->role], pVnode->refCount, pVnode);
+  return TSDB_CODE_APP_NOT_READY;
 }
 
 void vnodeFreeFromRQueue(void *vparam, SVReadMsg *pRead) {

src/vnode/src/vnodeWrite.c

@@ -297,7 +297,7 @@ static int32_t vnodePerformFlowCtrl(SVWriteMsg *pWrite) {
   if (pWrite->qtype != TAOS_QTYPE_RPC) return 0;
   if (pVnode->queuedWMsg < MAX_QUEUED_MSG_NUM && pVnode->flowctrlLevel <= 0) return 0;
 
-  if (tsFlowCtrl == 0) {
+  if (tsEnableFlowCtrl == 0) {
     int32_t ms = pow(2, pVnode->flowctrlLevel + 2);
     if (ms > 100) ms = 100;
     vTrace("vgId:%d, msg:%p, app:%p, perform flowctrl for %d ms", pVnode->vgId, pWrite, pWrite->rpcMsg.ahandle, ms);

tests/Jenkinsfile

@@ -50,12 +50,7 @@ pipeline {
           agent{label 'master'}
           steps {
             pre_test()
-            catchError(buildResult: 'SUCCESS', stageResult: 'FAILURE') {
+            
-                sh '''
-                cd ${WKC}/tests/pytest
-                python3 concurrent_inquiry.py -c 1
-                '''
-            }
             sh '''
             cd ${WKC}/tests
             ./test-all.sh b1
@@ -82,7 +77,37 @@ pipeline {
                 ./handle_crash_gen_val_log.sh
                 '''
             }
+            sh'''
+            systemctl start taosd 
+            sleep 10
+            '''
+            catchError(buildResult: 'SUCCESS', stageResult: 'FAILURE') {
+                sh '''
+                cd ${WKC}/tests/gotest
+                bash batchtest.sh
+                '''
+            }
+            catchError(buildResult: 'SUCCESS', stageResult: 'FAILURE') {
+                sh '''
+                cd ${WKC}/tests/examples/python/PYTHONConnectorChecker
+                python3 PythonChecker.py
+                '''
+            }
+            catchError(buildResult: 'SUCCESS', stageResult: 'FAILURE') {
+                sh '''
+                cd ${WKC}/tests/examples/JDBC/JDBCDemo/
+                mvn clean package assembly:single -DskipTests >/dev/null 
+                java -jar target/jdbcChecker-SNAPSHOT-jar-with-dependencies.jar -host 127.0.0.1
+                '''
+            }
+            catchError(buildResult: 'SUCCESS', stageResult: 'FAILURE') {
+                sh '''
+                cd ${JENKINS_HOME}/workspace/C#NET/src/CheckC#
+                dotnet run
+                '''
+            }
             sh '''
+            systemctl stop taosd 
             cd ${WKC}/tests
             ./test-all.sh b2
             date
@@ -95,6 +120,15 @@ pipeline {
 
           steps {
             pre_test()
+            catchError(buildResult: 'SUCCESS', stageResult: 'FAILURE') {
+                sh '''
+                cd ${WKC}/tests/pytest
+                nohup taosd >/dev/null &
+                sleep 10
+                python3 concurrent_inquiry.py -c 1
+                
+                '''
+            }
             sh '''
             cd ${WKC}/tests/pytest
             ./valgrind-test.sh 2>&1 > mem-error-out.log
@@ -106,41 +140,7 @@ pipeline {
             date'''
           }
         }
-       stage('connector'){
+       
-         agent{label "release"}
-         steps{
-            sh'''
-            cd ${WORKSPACE}
-            git checkout develop
-            '''
-            catchError(buildResult: 'SUCCESS', stageResult: 'FAILURE') {
-                sh '''
-                cd ${WORKSPACE}/tests/gotest
-                bash batchtest.sh
-                '''
-            }
-            catchError(buildResult: 'SUCCESS', stageResult: 'FAILURE') {
-                sh '''
-                cd ${WORKSPACE}/tests/examples/python/PYTHONConnectorChecker
-                python3 PythonChecker.py
-                '''
-            }
-            catchError(buildResult: 'SUCCESS', stageResult: 'FAILURE') {
-                sh '''
-                cd ${WORKSPACE}/tests/examples/JDBC/JDBCDemo/
-                mvn clean package assembly:single >/dev/null 
-                java -jar target/jdbcChecker-SNAPSHOT-jar-with-dependencies.jar -host 127.0.0.1
-                '''
-            }
-            catchError(buildResult: 'SUCCESS', stageResult: 'FAILURE') {
-                sh '''
-                cd ${JENKINS_HOME}/workspace/C#NET/src/CheckC#
-                dotnet run
-                '''
-            }
-          
-         }
-       }
        stage('arm64_build'){
                 agent{label 'arm64'}
                 steps{

tests/pytest/concurrent_inquiry.py

@@ -388,7 +388,9 @@ class ConcurrentInquiry:
                     print(
                 "Failure thread%d, sql: %s \nexception: %s" %
                 (threadID, str(sql),str(e)))
-                    #exit(-1)
+                    err_uec='Unable to establish connection'
+                    if err_uec in str(e) and loop >0:
+                        exit(-1)
                 loop -= 1
                 if loop == 0: break
                     
@@ -415,7 +417,9 @@ class ConcurrentInquiry:
                 print(
             "Failure thread%d, sql: %s \nexception: %s" %
             (threadID, str(sql),str(e)))
-                #exit(-1)
+                err_uec='Unable to establish connection'
+                if err_uec in str(e) and loop >0:
+                    exit(-1)
             loop -= 1    
             if loop == 0: break
                 

tests/pytest/crash_gen/crash_gen_main.py

@@ -2224,22 +2224,25 @@ class ClientManager:
         if svcMgr: # gConfig.auto_start_service:
             svcMgr.stopTaosServices()
             svcMgr = None
+        
+
+        # Release global variables
+        gConfig = None
+        gSvcMgr = None
+        logger = None
+        
+        thPool = None
+        dbManager.cleanUp() # destructor wouldn't run in time
+        dbManager = None
+
         # Print exec status, etc., AFTER showing messages from the server
         self.conclude()
         # print("TC failed (2) = {}".format(self.tc.isFailed()))
         # Linux return code: ref https://shapeshed.com/unix-exit-codes/
         ret = 1 if self.tc.isFailed() else 0
         self.tc.cleanup()
-
-        # Release global variables
-        gConfig = None
-        gSvcMgr = None
-        logger = None
-
         # Release variables here
         self.tc = None
-        thPool = None
-        dbManager = None
 
         gc.collect() # force garbage collection
         # h = hpy()

tests/pytest/crash_gen/db.py

@@ -394,6 +394,7 @@ class DbManager():
             cType == 'native') else DbConn.createRest(dbTarget)
         try:
             self._dbConn.open()  # may throw taos.error.ProgrammingError: disconnected
+            Logging.debug("DbManager opened DB connection...")
         except taos.error.ProgrammingError as err:
             # print("Error type: {}, msg: {}, value: {}".format(type(err), err.msg, err))
             if (err.msg == 'client disconnected'):  # cannot open DB connection
@@ -412,6 +413,10 @@ class DbManager():
         # Moved to Database()
         # self._stateMachine = StateMechine(self._dbConn)
 
+    def __del__(self):
+        ''' Release the underlying DB connection upon deletion of DbManager '''
+        self.cleanUp()
+
     def getDbConn(self):
         return self._dbConn
 
@@ -438,5 +443,8 @@ class DbManager():
         return "table_{}".format(tblNum)
 
     def cleanUp(self):
-        self._dbConn.close()
+        if self._dbConn:
+            self._dbConn.close()
+            self._dbConn = None
+            Logging.debug("DbManager closed DB connection...")
 

tests/pytest/crash_gen/misc.py

@@ -2,6 +2,7 @@ import threading
 import random
 import logging
 import os
+import sys
 
 import taos
 
@@ -53,7 +54,7 @@ class Logging:
         # global misc.logger
         _logger = logging.getLogger('CrashGen')  # real logger
         _logger.addFilter(LoggingFilter())
-        ch = logging.StreamHandler()
+        ch = logging.StreamHandler(sys.stdout) # Ref: https://stackoverflow.com/questions/14058453/making-python-loggers-output-all-messages-to-stdout-in-addition-to-log-file
         _logger.addHandler(ch)
 
         # Logging adapter, to be used as a logger

tests/pytest/crash_gen_bootstrap.py

@@ -19,5 +19,5 @@ if __name__ == "__main__":
     mExec.init()
     exitCode = mExec.run()
 
-    print("Exiting with code: {}".format(exitCode))
+    print("\nCrash_Gen is now exiting with status code: {}".format(exitCode))
     sys.exit(exitCode)

tests/pytest/handle_crash_gen_val_log.sh

@@ -5,9 +5,10 @@ GREEN='\033[1;32m'
 GREEN_DARK='\033[0;32m'
 GREEN_UNDERLINE='\033[4;32m'
 NC='\033[0m'
-nohup /var/lib/jenkins/workspace/TDinternal/debug/build/bin/taosd -c /var/lib/jenkins/workspace/TDinternal/community/sim/dnode1/cfg >/dev/null &
+#nohup /var/lib/jenkins/workspace/TDinternal/debug/build/bin/taosd -c /var/lib/jenkins/workspace/TDinternal/community/sim/dnode1/cfg >/dev/null &
+nohup /root/TDinternal/debug/build/bin/taosd -c /root/TDinternal/community/sim/dnode1/cfg >/dev/null &
 ./crash_gen.sh --valgrind -p -t 10 -s 250 -b 4
-pidof taosd|xargs kill  
+pidof taosd|xargs kill -9
 grep 'start to execute\|ERROR SUMMARY' valgrind.err|grep -v 'grep'|uniq|tee crash_gen_mem_err.log
 
 for memError in `grep 'ERROR SUMMARY' crash_gen_mem_err.log | awk '{print $4}'`
@@ -31,4 +32,4 @@ if [ -n "$defiMemError" ]; then
     exit 8
     fi
 fi
-done
+done

tests/pytest/import_merge/importCacheFileH.py

@@ -55,7 +55,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importCacheFileHO.py

@@ -55,7 +55,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importCacheFileHPO.py

@@ -57,7 +57,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importCacheFileS.py

@@ -55,7 +55,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importCacheFileSub.py

@@ -55,7 +55,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importCacheFileT.py

@@ -55,7 +55,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importCacheFileTO.py

@@ -55,7 +55,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importCacheFileTPO.py

@@ -57,7 +57,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataH2.py

@@ -59,7 +59,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataHO.py

@@ -60,7 +60,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataHO2.py

@@ -60,7 +60,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataHPO.py

@@ -62,7 +62,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataLastH.py

@@ -59,7 +59,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataLastHO.py

@@ -59,7 +59,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataLastHPO.py

@@ -61,7 +61,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataLastS.py

@@ -59,7 +59,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataLastSub.py

@@ -32,7 +32,7 @@ class TDTestCase:
         tdDnodes.stop(1)
         tdDnodes.deploy(1)
         tdDnodes.start(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
 
         tdSql.execute('reset query cache')
         tdSql.execute('drop database if exists db')
@@ -60,9 +60,9 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
 
         tdLog.info("================= step5")
         tdLog.info("import 10 data totally repetitive")

tests/pytest/import_merge/importDataLastT.py

@@ -55,7 +55,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataLastTO.py

@@ -55,7 +55,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataLastTPO.py

@@ -57,7 +57,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataS.py

@@ -55,7 +55,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataSub.py

@@ -60,7 +60,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataT.py

@@ -55,7 +55,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataTO.py

@@ -55,7 +55,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importDataTPO.py

@@ -57,9 +57,9 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
 
         tdLog.info("================= step5")
         tdLog.info("import 20 data later with partly overlap")

tests/pytest/import_merge/importHORestart.py

@@ -57,7 +57,7 @@ class TDTestCase:
         tdLog.info("================= step5")
         tdDnodes.forcestop(1)
         tdDnodes.start(1)
-        tdLog.sleep(10)
+        #tdLog.sleep(10)
 
         tdLog.info("================= step6")
         tdSql.query('select * from tb1')

tests/pytest/import_merge/importHPORestart.py

@@ -62,7 +62,7 @@ class TDTestCase:
         tdLog.info("================= step5")
         tdDnodes.forcestop(1)
         tdDnodes.start(1)
-        tdLog.sleep(10)
+        #tdLog.sleep(10)
 
         tdLog.info("================= step6")
         tdSql.query('select * from tb1')

tests/pytest/import_merge/importHRestart.py

@@ -54,7 +54,7 @@ class TDTestCase:
         tdLog.info("================= step5")
         tdDnodes.forcestop(1)
         tdDnodes.start(1)
-        tdLog.sleep(10)
+        #tdLog.sleep(10)
 
         tdLog.info("================= step6")
         tdSql.query('select * from tb1')

tests/pytest/import_merge/importInsertThenImport.py

@@ -61,7 +61,7 @@ class TDTestCase:
         tdLog.info("================= step5")
         tdDnodes.stop(1)
         tdDnodes.start(1)
-        tdLog.sleep(10)
+        #tdLog.sleep(10)
 
         tdLog.info("================= step6")
         tdLog.info("import 100 sequential data again")

tests/pytest/import_merge/importLastH.py

@@ -53,7 +53,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importLastHO.py

@@ -53,7 +53,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importLastHPO.py

@@ -55,7 +55,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importLastS.py

@@ -53,7 +53,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importLastSub.py

@@ -53,7 +53,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importLastT.py

@@ -57,7 +57,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importLastTO.py

@@ -57,7 +57,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importLastTPO.py

@@ -59,7 +59,7 @@ class TDTestCase:
 
         tdLog.info("================= step4")
         tdDnodes.stop(1)
-        tdLog.sleep(5)
+        #tdLog.sleep(5)
         tdDnodes.start(1)
 
         tdLog.info("================= step5")

tests/pytest/import_merge/importSRestart.py

@@ -64,7 +64,7 @@ class TDTestCase:
         tdLog.info("================= step5")
         tdDnodes.forcestop(1)
         tdDnodes.start(1)
-        tdLog.sleep(10)
+        #tdLog.sleep(10)
 
         tdLog.info("================= step6")
         tdSql.query('select * from tb1')

tests/pytest/import_merge/importSubRestart.py

@@ -64,7 +64,7 @@ class TDTestCase:
         tdLog.info("================= step5")
         tdDnodes.forcestop(1)
         tdDnodes.start(1)
-        tdLog.sleep(10)
+        #tdLog.sleep(10)
 
         tdLog.info("================= step6")
         tdSql.query('select * from tb1')

tests/pytest/import_merge/importTORestart.py

@@ -64,7 +64,7 @@ class TDTestCase:
         tdLog.info("================= step5")
         tdDnodes.forcestop(1)
         tdDnodes.start(1)
-        tdLog.sleep(10)
+        #tdLog.sleep(10)
 
         tdLog.info("================= step6")
         tdSql.query('select * from tb1')

tests/pytest/import_merge/importTPORestart.py

@@ -68,7 +68,7 @@ class TDTestCase:
         tdLog.info("================= step5")
         tdDnodes.forcestop(1)
         tdDnodes.start(1)
-        tdLog.sleep(10)
+        #tdLog.sleep(10)
 
         tdLog.info("================= step6")
         tdSql.query('select * from tb1')

tests/pytest/import_merge/importTRestart.py

@@ -61,7 +61,7 @@ class TDTestCase:
         tdLog.info("================= step5")
         tdDnodes.forcestop(1)
         tdDnodes.start(1)
-        tdLog.sleep(10)
+        #tdLog.sleep(10)
 
         tdLog.info("================= step6")
         tdSql.query('select * from tb1')

tests/pytest/query/query.py

@@ -16,7 +16,7 @@ import taos
 from util.log import tdLog
 from util.cases import tdCases
 from util.sql import tdSql
-
+from util.dnodes import tdDnodes
 
 class TDTestCase:
     def init(self, conn, logSql):
@@ -44,6 +44,25 @@ class TDTestCase:
         tdSql.query("select * from db.st where ts='2020-05-13 10:00:00.000'")
         tdSql.checkRows(1)
 
+        ## test case for https://jira.taosdata.com:18080/browse/TD-2488
+        tdSql.execute("create table m1(ts timestamp, k int) tags(a int)")
+        tdSql.execute("create table t1 using m1 tags(1)")
+        tdSql.execute("create table t2 using m1 tags(2)")
+        tdSql.execute("insert into t1 values('2020-1-1 1:1:1', 1)")
+        tdSql.execute("insert into t1 values('2020-1-1 1:10:1', 2)")
+        tdSql.execute("insert into t2 values('2020-1-1 1:5:1', 99)")
+        
+        tdSql.query("select count(*) from m1 where ts = '2020-1-1 1:5:1' ")
+        tdSql.checkRows(1)
+        tdSql.checkData(0, 0, 1)
+
+        tdDnodes.stop(1)
+        tdDnodes.start(1)
+
+        tdSql.query("select count(*) from m1 where ts = '2020-1-1 1:5:1' ")
+        tdSql.checkRows(1)
+        tdSql.checkData(0, 0, 1)
+
     def stop(self):
         tdSql.close()
         tdLog.success("%s successfully executed" % __file__)

tests/pytest/table/alter_wal0.py

@@ -43,7 +43,8 @@ class TDTestCase:
 
         print("==============step2")
         tdDnodes.stopAll()
-        filename = '/var/lib/taos/mnode/wal/wal0'
+        path = tdDnodes.getDnodesRootDir()        
+        filename = path + '/dnode1/data/mnode/wal/wal0'
 
         with open(filename, 'rb') as f1:
             temp = f1.read()
@@ -57,12 +58,12 @@ class TDTestCase:
         print("==============step3")
         tdSql.execute("use demo;")
         tdSql.query('show tables;')
-        tdSql.checkRows(10)
+        tdSql.checkRows(9)
         for i in range(11,21):
             tdSql.execute("CREATE table if not exists test{num} using meters tags({num});".format(num=i))
 
         tdSql.query('show tables;')
-        tdSql.checkRows(20)
+        tdSql.checkRows(19)
         print("==============check table numbers and create 10 tables")
 
 

tests/script/general/db/alter_tables_d2.sim

@@ -11,11 +11,26 @@ system sh/cfg.sh -n dnode2 -c maxTablesPerVnode -v 5
 
 system sh/exec.sh -n dnode1 -s start
 system sh/exec.sh -n dnode2 -s start
-
-sleep 3000
 sql connect
 sql create dnode $hostname2
-sleep 1000
+
+$x = 0
+step1: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show dnodes
+print dnode1 $data4_1
+print dnode2 $data4_2
+if $data4_1 != ready then
+  goto step1
+endi
+if $data4_2 != ready then
+  goto step1
+endi
 
 print ============================ step1
 
@@ -80,10 +95,23 @@ system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/exec.sh -n dnode2 -s stop -x SIGINT
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 10
 system sh/cfg.sh -n dnode2 -c maxTablesPerVnode -v 10
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
 system sh/exec.sh -n dnode2 -s start
-sleep 5000
+$x = 0
+step2: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step1
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step2
+endi
 
 sql create table db.t100 using db.st tags(0)
 sql create table db.t101 using db.st tags(1)
@@ -110,7 +138,6 @@ if $rows != 40 then
   return -1
 endi
 
-
 sql insert into db.t100 values(now, 1)
 sql insert into db.t101 values(now, 1)
 sql insert into db.t102 values(now, 1)
@@ -144,10 +171,24 @@ system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/exec.sh -n dnode2 -s stop -x SIGINT
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 15
 system sh/cfg.sh -n dnode2 -c maxTablesPerVnode -v 15
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
 system sh/exec.sh -n dnode2 -s start
-sleep 5000
+
+$x = 0
+step3: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step3
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step3
+endi
 
 sql create table db.t200 using db.st tags(0)
 sql create table db.t201 using db.st tags(1)
@@ -198,7 +239,7 @@ sql insert into db.t219 values(now, 1)
 print ============================ step6
 
 sql reset query cache
-sleep 1000
+sleep 100
 
 sql select * from db.t000
 if $rows != 1 then 
@@ -268,10 +309,23 @@ system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/exec.sh -n dnode2 -s stop -x SIGINT
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 20
 system sh/cfg.sh -n dnode2 -c maxTablesPerVnode -v 20
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
 system sh/exec.sh -n dnode2 -s start
-sleep 5000
+$x = 0
+step9: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step9
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step9
+endi
 
 sql create table db.t300 using db.st tags(0)
 sql create table db.t301 using db.st tags(1)
@@ -321,7 +375,7 @@ if $rows != 80 then
 endi
 
 sql reset query cache
-sleep 1000
+sleep 100
 
 sql select * from db.t000
 if $rows != 1 then 
@@ -351,13 +405,26 @@ endi
 print ============================ step10
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/exec.sh -n dnode2 -s stop -x SIGINT
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
 system sh/exec.sh -n dnode2 -s start
-sleep 5000
+$x = 0
+step10: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step10
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step10
+endi
 
 sql reset query cache
-sleep 1000
+sleep 100
 
 sql show db.tables
 if $rows != 80 then 
@@ -401,10 +468,23 @@ system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/exec.sh -n dnode2 -s stop -x SIGINT
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 25
 system sh/cfg.sh -n dnode2 -c maxTablesPerVnode -v 25
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
 system sh/exec.sh -n dnode2 -s start
-sleep 5000
+$x = 0
+step1xx: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step1xx
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step1xx
+endi
 
 sql create table db.t400 using db.st tags(0)
 sql create table db.t401 using db.st tags(1)
@@ -454,7 +534,7 @@ if $rows != 100 then
 endi
 
 sql reset query cache
-sleep 1000
+sleep 100
 
 sql select * from db.t000
 if $rows != 1 then 

tests/script/general/db/alter_tables_v1.sim

@@ -5,8 +5,6 @@ system sh/cfg.sh -n dnode1 -c maxVgroupsPerDb -v 1
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 10
 
 system sh/exec.sh -n dnode1 -s start
-
-sleep 3000
 sql connect
 
 print ============================ step1
@@ -51,9 +49,22 @@ print ============================ step3
 
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 20
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
-sleep 5000
+$x = 0
+step2: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step2
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step2
+endi
 
 sql create table db.t10 using db.st tags(0)
 sql create table db.t11 using db.st tags(1)
@@ -91,9 +102,22 @@ endi
 print ============================ step5
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 30
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
-sleep 5000
+$x = 0
+step5: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step5
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step5
+endi
 
 sql create table db.t20 using db.st tags(0)
 sql create table db.t21 using db.st tags(1)
@@ -124,7 +148,7 @@ sql insert into db.t29 values(now, 1)
 print ============================ step6
 
 sql reset query cache
-sleep 1000
+sleep 100
 
 sql select * from db.t0
 if $rows != 1 then 
@@ -148,9 +172,22 @@ endi
 
 print ============================ step7
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
-sleep 5000
+$x = 0
+step7: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step7
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step7
+endi
 
 sql reset query cache
 sleep 1000
@@ -190,9 +227,22 @@ endi
 print ============================ step9
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 40
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
-sleep 5000
+$x = 0
+step9: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step9
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step9
+endi
 
 sql create table db.t30 using db.st tags(0)
 sql create table db.t31 using db.st tags(1)
@@ -249,12 +299,24 @@ if $rows != 40 then
   return -1
 endi
 
-
 print ============================ step10
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
-sleep 5000
+$x = 0
+step10: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step10
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step10
+endi
 
 sql reset query cache
 sleep 1000
@@ -294,9 +356,22 @@ endi
 print ============================ step12
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 50
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
-sleep 5000
+$x = 0
+step12: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step12
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step12
+endi
 
 sql create table db.t40 using db.st tags(0)
 sql create table db.t41 using db.st tags(1)

tests/script/general/db/alter_tables_v4.sim

@@ -5,8 +5,6 @@ system sh/cfg.sh -n dnode1 -c maxVgroupsPerDb -v 4
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 5
 
 system sh/exec.sh -n dnode1 -s start
-
-sleep 3000
 sql connect
 
 print ============================ step1
@@ -70,9 +68,22 @@ endi
 print ============================ step3
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 10
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
-sleep 5000
+$x = 0
+step3: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step3
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step3
+endi
 
 sql create table db.t100 using db.st tags(0)
 sql create table db.t101 using db.st tags(1)
@@ -131,9 +142,22 @@ endi
 print ============================ step5
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 15
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
-sleep 5000
+$x = 0
+step5: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step5
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step5
+endi
 
 sql create table db.t200 using db.st tags(0)
 sql create table db.t201 using db.st tags(1)
@@ -184,7 +208,7 @@ sql insert into db.t219 values(now, 1)
 print ============================ step6
 
 sql reset query cache
-sleep 1000
+sleep 100
 
 sql select * from db.t000
 if $rows != 1 then 
@@ -250,9 +274,22 @@ endi
 print ============================ step9
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 20
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
-sleep 5000
+$x = 0
+step9: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step9
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step9
+endi
 
 sql create table db.t300 using db.st tags(0)
 sql create table db.t301 using db.st tags(1)
@@ -331,9 +368,22 @@ endi
 
 print ============================ step10
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
-sleep 5000
+$x = 0
+step10: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step10
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step10
+endi
 
 sql reset query cache
 sleep 1000
@@ -378,9 +428,24 @@ endi
 print ============================ step12
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
 system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 25
-sleep 5000
 system sh/exec.sh -n dnode1 -s start
-sleep 5000
+$x = 0
+step12: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show mnodes -x step12
+print mnode1 $data2_1
+print mnode1 $data2_2
+print mnode1 $data2_3
+if $data2_1 != master then
+  goto step12
+endi
+
+
 sql create table db.t400 using db.st tags(0)
 sql create table db.t401 using db.st tags(1)
 sql create table db.t402 using db.st tags(2)

tests/script/general/db/delete_reusevnode.sim

@@ -5,7 +5,6 @@ system sh/deploy.sh -n dnode1 -i 1
 
 print ========= start dnodes
 system sh/exec.sh -n dnode1 -s start
-sleep 3000
 sql connect
 
 print ======== step1 
@@ -38,7 +37,6 @@ if $rows != 0 then
 endi 
 
 system sh/stop_dnodes.sh
-sleep 3000
 
 system sh/deploy.sh -n dnode1 -i 1
 print ========= start dnodes

tests/script/general/db/delete_writing1.sim

@@ -22,7 +22,6 @@ system sh/cfg.sh -n dnode4 -c mnodeEqualVnodeNum -v 4
 
 print ========= start dnodes
 system sh/exec.sh -n dnode1 -s start
-sleep 3000
 sql connect
 
 sql create database db 

tests/script/general/table/delete_writing.sim

@@ -21,8 +21,6 @@ system sh/cfg.sh -n dnode4 -c mnodeEqualVnodeNum -v 4
 
 print ========= start dnodes
 system sh/exec.sh -n dnode1 -s start
-
-sleep 3000
 sql connect
 
 sql create database db 

tests/script/general/wal/sync.sim

@@ -31,7 +31,6 @@ system sh/cfg.sh -n dnode3 -c maxSQLLength -v 940032
 print ============== deploy
 
 system sh/exec.sh -n dnode1 -s start 
-sleep 5001
 sql connect
 
 sql create dnode $hostname2
@@ -43,8 +42,8 @@ print  =============== step1
 $x = 0
 show1: 
 	$x = $x + 1
-	sleep 2000
+	sleep 1000
-	if $x == 5 then
+	if $x == 10 then
 	  return -1
 	endi
 sql show mnodes -x show1
@@ -82,7 +81,7 @@ restful d1 table_rest 1591772800 30000
 restful d1 table_rest 1591872800 30000
 restful d1 table_rest 1591972800 30000
 
-sleep 1000
+sleep 100
 sql select * from table_rest;
 print rows: $rows
 if $rows != 300000 then
@@ -91,29 +90,51 @@ endi
 
 print ===============  step3
 system sh/exec.sh -n dnode1 -s stop -x SIGINT
-sleep 5000
 sql select * from table_rest;
 print rows: $rows
 if $rows != 300000 then
   return -1
 endi
 system sh/exec.sh -n dnode1 -s start -x SIGINT
-sleep 5000
+
+$x = 0
+a1: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 40 then
+	  return -1
+	endi
+
+sql show vgroups 
+print online vnodes $data03
+if $data03 != 3 then
+	goto a1
+endi
 
 print ===============  step4
 system sh/exec.sh -n dnode2 -s stop -x SIGINT
-sleep 5000
 sql select * from table_rest;
 print rows: $rows
 if $rows != 300000 then
   return -1
 endi
 system sh/exec.sh -n dnode2 -s start -x SIGINT
-sleep 5000
+$x = 0
+a2: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 40 then
+	  return -1
+	endi
+
+sql show vgroups 
+print online vnodes $data03
+if $data03 != 3 then
+	goto a2
+endi
 
 print ===============  step5
 system sh/exec.sh -n dnode3 -s stop -x SIGINT
-sleep 5000
 sql select * from table_rest;
 print rows: $rows
 if $rows != 300000 then

tests/script/jenkins/basic.txt

@@ -288,6 +288,7 @@ cd ../../../debug; make
 ./test.sh -f unique/dnode/data1.sim
 ./test.sh -f unique/dnode/offline1.sim
 ./test.sh -f unique/dnode/offline2.sim
+./test.sh -f unique/dnode/offline3.sim
 ./test.sh -f unique/dnode/reason.sim
 ./test.sh -f unique/dnode/remove1.sim
 ./test.sh -f unique/dnode/remove2.sim

tests/script/jenkins/basic_3.txt

@@ -1,38 +1,4 @@
 
-./test.sh -f unique/stable/balance_replica1.sim
-./test.sh -f unique/stable/dnode2_stop.sim
-./test.sh -f unique/stable/dnode2.sim
-./test.sh -f unique/stable/dnode3.sim
-./test.sh -f unique/stable/replica2_dnode4.sim
-./test.sh -f unique/stable/replica2_vnode3.sim
-./test.sh -f unique/stable/replica3_dnode6.sim
-./test.sh -f unique/stable/replica3_vnode3.sim
-
-./test.sh -f unique/mnode/mgmt20.sim
-./test.sh -f unique/mnode/mgmt21.sim
-./test.sh -f unique/mnode/mgmt22.sim
-./test.sh -f unique/mnode/mgmt23.sim
-./test.sh -f unique/mnode/mgmt24.sim
-#./test.sh -f unique/mnode/mgmt25.sim
-#./test.sh -f unique/mnode/mgmt26.sim
-./test.sh -f unique/mnode/mgmt33.sim
-./test.sh -f unique/mnode/mgmt34.sim
-./test.sh -f unique/mnode/mgmtr2.sim
-
-
-
-./test.sh -f general/parser/stream_on_sys.sim
-./test.sh -f general/stream/metrics_del.sim
-./test.sh -f general/stream/metrics_n.sim
-./test.sh -f general/stream/metrics_replica1_vnoden.sim
-./test.sh -f general/stream/restart_stream.sim
-./test.sh -f general/stream/stream_3.sim
-./test.sh -f general/stream/stream_restart.sim
-./test.sh -f general/stream/table_1.sim
-./test.sh -f general/stream/table_del.sim
-./test.sh -f general/stream/table_n.sim
-./test.sh -f general/stream/table_replica1_vnoden.sim
-
 ./test.sh -f unique/arbitrator/check_cluster_cfg_para.sim
 #./test.sh -f unique/arbitrator/dn2_mn1_cache_file_sync.sim
 ./test.sh -f unique/arbitrator/dn3_mn1_full_createTableFail.sim
@@ -79,3 +45,35 @@
 ./test.sh -f unique/migrate/mn2_vn2_repl2_rmMnodeVnodeDir.sim
 ./test.sh -f unique/migrate/mn2_vn2_repl2_rmMnodeVnodeDir_stopAll_starAll.sim
 ./test.sh -f unique/migrate/mn2_vn2_repl2_rmVnodeDir.sim
+
+./test.sh -f unique/stable/balance_replica1.sim
+./test.sh -f unique/stable/dnode2_stop.sim
+./test.sh -f unique/stable/dnode2.sim
+./test.sh -f unique/stable/dnode3.sim
+./test.sh -f unique/stable/replica2_dnode4.sim
+./test.sh -f unique/stable/replica2_vnode3.sim
+./test.sh -f unique/stable/replica3_dnode6.sim
+./test.sh -f unique/stable/replica3_vnode3.sim
+
+./test.sh -f unique/mnode/mgmt20.sim
+./test.sh -f unique/mnode/mgmt21.sim
+./test.sh -f unique/mnode/mgmt22.sim
+./test.sh -f unique/mnode/mgmt23.sim
+./test.sh -f unique/mnode/mgmt24.sim
+#./test.sh -f unique/mnode/mgmt25.sim
+#./test.sh -f unique/mnode/mgmt26.sim
+./test.sh -f unique/mnode/mgmt33.sim
+./test.sh -f unique/mnode/mgmt34.sim
+./test.sh -f unique/mnode/mgmtr2.sim
+
+./test.sh -f general/parser/stream_on_sys.sim
+./test.sh -f general/stream/metrics_del.sim
+./test.sh -f general/stream/metrics_n.sim
+./test.sh -f general/stream/metrics_replica1_vnoden.sim
+./test.sh -f general/stream/restart_stream.sim
+./test.sh -f general/stream/stream_3.sim
+./test.sh -f general/stream/stream_restart.sim
+./test.sh -f general/stream/table_1.sim
+./test.sh -f general/stream/table_del.sim
+./test.sh -f general/stream/table_n.sim
+./test.sh -f general/stream/table_replica1_vnoden.sim

tests/script/jenkins/unique.txt

@@ -31,6 +31,8 @@ cd ../../../debug; make
 ./test.sh -f unique/dnode/balancex.sim
 ./test.sh -f unique/dnode/offline1.sim
 ./test.sh -f unique/dnode/offline2.sim
+./test.sh -f unique/dnode/offline3.sim
+./test.sh -f unique/dnode/reason.sim
 ./test.sh -f unique/dnode/remove1.sim
 ./test.sh -f unique/dnode/remove2.sim
 ./test.sh -f unique/dnode/vnode_clean.sim

tests/script/sh/deploy.sh

@@ -132,10 +132,12 @@ echo "cqdebugFlag            143"                >> $TAOS_CFG
 echo "monitor                0"                  >> $TAOS_CFG
 echo "monitorInterval        1"                  >> $TAOS_CFG
 echo "http                   0"                  >> $TAOS_CFG
+echo "slaveQuery             0"                  >> $TAOS_CFG
 echo "numOfThreadsPerCore    2.0"                >> $TAOS_CFG
 echo "defaultPass            taosdata"           >> $TAOS_CFG
 echo "numOfLogLines          20000000"           >> $TAOS_CFG
 echo "mnodeEqualVnodeNum     0"                  >> $TAOS_CFG
+echo "balanceInterval        1"                  >> $TAOS_CFG
 echo "clog                   2"                  >> $TAOS_CFG
 #echo "cache                 1"                  >> $TAOS_CFG
 echo "days                   10"                  >> $TAOS_CFG

tests/script/unique/arbitrator/dn3_mn1_r3_vnode_delDir.sim

@@ -360,12 +360,14 @@ print ============== step7: stop dnode3/dnode2, and cluster unable to provide se
 system sh/exec.sh -n dnode2 -s stop -x SIGINT
 system sh/exec.sh -n dnode3 -s stop -x SIGINT
 sleep 3000
-sql_error select count(*) from $stb
+sql select count(*) from $stb -x s71
+s71:
 
 print ============== step8: restart dnode2, and cluster Still unable to provide services
 system sh/exec.sh -n dnode2 -s start
 sleep 3000
-sql_error select count(*) from $stb
+sql select count(*) from $stb -x s81
+s81:
 
 print ============== step9: restart dnode3, and cluster Resume service delivery
 system sh/exec.sh -n dnode3 -s start

tests/script/unique/arbitrator/dn3_mn1_replica_change.sim

@@ -45,15 +45,30 @@ system sh/exec_tarbitrator.sh -s start
 
 print ============== step1: start dnode1, only deploy mnode
 system sh/exec.sh -n dnode1 -s start
-sleep 3000
 sql connect
 
 print ============== step2: start dnode2 and add into cluster , then create database with replica 1, and create table, insert data
 system sh/exec.sh -n dnode2 -s start
 sql create dnode $hostname2
-sleep 3000
 
-$sleepTimer = 10000
+$x = 0
+step1: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 40 then
+		return -1
+	endi
+
+sql show dnodes
+print dnode1 $data4_1
+print dnode2 $data4_2
+
+if $data4_1 != ready then
+  goto step1
+endi
+if $data4_2 != ready then
+  goto step1
+endi
 
 $db = db
 sql create database $db replica 1
@@ -84,7 +99,6 @@ while $i < $tblNum
 endw 
 
 sql select count(*) from $stb
-sleep 1000
 print data00 $data00
 if $data00 != $totalRows then
 	return -1
@@ -92,118 +106,65 @@ endi
 
 print ============== step2-1: stop dnode2 for falling disc, then restart dnode2, and check rows
 system sh/exec.sh -n dnode2 -s stop -x SIGINT
-sleep $sleepTimer
-
-
-$loopCnt = 0
-wait_dnode2_offline_0:
-$loopCnt = $loopCnt + 1
-if $loopCnt == 10 then 
-  return -1
-endi
-sql show dnodes
-if $rows != 2 then
-  sleep 2000
-  goto wait_dnode2_offline_0
-endi
-print $data0_1  $data1_1  $data2_1  $data3_1  $data4_1
-print $data0_2  $data1_2  $data2_2  $data3_2  $data4_2
-print $data0_3  $data1_3  $data2_3  $data3_3  $data4_3
-#print $data0_4  $data1_4  $data2_4  $data3_4  $data4_4
-#print $data0_5  $data1_5  $data2_5  $data3_5  $data4_5
-#print $data0_6  $data1_6  $data2_6  $data3_6  $data4_6
-#$dnode1Status = $data4_1  
-$dnode2Status = $data4_2
-#$dnode3Status = $data4_3  
-#$dnode4Status = $data4_4 
-#$dnode5Status = $data4_5
-
-if $dnode2Status != offline then
-  sleep 2000
-  goto wait_dnode2_offline_0
-endi
-
 system sh/exec.sh -n dnode2 -s start
-sleep $sleepTimer
 
+$x = 0
+a0: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 40 then
+	  return -1
+	endi
 
-$loopCnt = 0
+sql show vgroups 
-wait_dnode2_reready:
+print online vnodes $data03
-$loopCnt = $loopCnt + 1
+if $data03 != 1 then
-if $loopCnt == 10 then 
+	goto a0
-  return -1
-endi
-sql show dnodes
-if $rows != 2 then
-  sleep 2000
-  goto wait_dnode2_reready
-endi
-print $data0_1  $data1_1  $data2_1  $data3_1  $data4_1
-print $data0_2  $data1_2  $data2_2  $data3_2  $data4_2
-print $data0_3  $data1_3  $data2_3  $data3_3  $data4_3
-#print $data0_4  $data1_4  $data2_4  $data3_4  $data4_4
-#print $data0_5  $data1_5  $data2_5  $data3_5  $data4_5
-#print $data0_6  $data1_6  $data2_6  $data3_6  $data4_6
-#$dnode1Status = $data4_1  
-$dnode2Status = $data4_2
-#$dnode3Status = $data4_3  
-#$dnode4Status = $data4_4 
-#$dnode5Status = $data4_5
-
-if $dnode2Status != ready then
-  sleep 2000
-  goto wait_dnode2_reready
 endi
 
 sql select count(*) from $stb
-sleep 1000
 print data00 $data00
 if $data00 != $totalRows then
 	return -1
 endi
 
-
 print ============== step3: start dnode3 and add into cluster , then alter replica from 1 to 2, and waiting sync
 system sh/exec.sh -n dnode3 -s start
 sql create dnode $hostname3
-sleep 3000
 
-sql alter database $db replica 2
+$x = 0
-sleep $sleepTimer
+step2: 
-$loopCnt = 0
+	$x = $x + 1
-wait_dnode3_ready:
+	sleep 1000
-$loopCnt = $loopCnt + 1
+	if $x == 40 then
-if $loopCnt == 10 then 
+		return -1
-  return -1
+	endi
-endi
 
 sql show dnodes
-if $rows != 3 then
+print dnode1 $data4_1
-  sleep 2000
+print dnode2 $data4_2
-  goto wait_dnode3_ready
+print dnode3 $data4_3
-endi
-print $data0_1  $data1_1  $data2_1  $data3_1  $data4_1
-print $data0_2  $data1_2  $data2_2  $data3_2  $data4_2
-print $data0_3  $data1_3  $data2_3  $data3_3  $data4_3
-#print $data0_4  $data1_4  $data2_4  $data3_4  $data4_4
-#print $data0_5  $data1_5  $data2_5  $data3_5  $data4_5
-#print $data0_6  $data1_6  $data2_6  $data3_6  $data4_6
-#$dnode1Status = $data4_1  
-$dnode2Status = $data4_2
-$dnode3Status = $data4_3  
-#$dnode4Status = $data4_4 
-#$dnode5Status = $data4_5
 
-if $dnode2Status != ready then
+if $data4_3 != ready then
-  sleep 2000
+  goto step2
-  goto wait_dnode3_ready
+endi
-endi
+
-if $dnode3Status != ready then
+sql alter database $db replica 2
-  sleep 2000
+
-  goto wait_dnode3_ready
+$x = 0
+a1: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 40 then
+	  return -1
+	endi
+
+sql show vgroups 
+print online vnodes $data03
+if $data03 != 2 then
+	goto a1
 endi
 
-sleep $sleepTimer
 # check using select
 sql select count(*) from $stb
 print data00 $data00
@@ -211,44 +172,22 @@ if $data00 != $totalRows then
 	return -1
 endi
 
-
 print ============== step4: stop dnode2 for checking if sync ok
 system sh/exec.sh -n dnode2 -s stop
-sleep $sleepTimer
+$x = 0
-$loopCnt = 0
+a2: 
-wait_dnode2_offline:
+	$x = $x + 1
-$loopCnt = $loopCnt + 1
+	sleep 1000
-if $loopCnt == 10 then 
+	if $x == 40 then
-  return -1
+	  return -1
+	endi
+
+sql show vgroups 
+print online vnodes $data03
+if $data03 != 1 then
+	goto a2
 endi
 
-sql show dnodes
-if $rows != 3 then
-  sleep 2000
-  goto wait_dnode2_offline
-endi
-print $data0_1  $data1_1  $data2_1  $data3_1  $data4_1
-print $data0_2  $data1_2  $data2_2  $data3_2  $data4_2
-print $data0_3  $data1_3  $data2_3  $data3_3  $data4_3
-#print $data0_4  $data1_4  $data2_4  $data3_4  $data4_4
-#print $data0_5  $data1_5  $data2_5  $data3_5  $data4_5
-#print $data0_6  $data1_6  $data2_6  $data3_6  $data4_6
-#$dnode1Status = $data4_1  
-$dnode2Status = $data4_2
-$dnode3Status = $data4_3  
-#$dnode4Status = $data4_4 
-#$dnode5Status = $data4_5
-
-if $dnode2Status != offline then
-  sleep 2000
-  goto wait_dnode2_offline
-endi
-if $dnode3Status != ready then
-  sleep 2000
-  goto wait_dnode2_offline
-endi
-
-sleep $sleepTimer   # waitting for move master vnode of dnode2 to dnode3
 # check using select
 sql select count(*) from $stb
 print data00 $data00
@@ -258,38 +197,20 @@ endi
 
 print ============== step5: restart dnode2
 system sh/exec.sh -n dnode2 -s start
-sleep 3000
+$x = 0
+a3: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 40 then
+	  return -1
+	endi
 
-$loopCnt = 0
+sql show vgroups 
-wait_dnode2_ready:
+print online vnodes $data03
-$loopCnt = $loopCnt + 1
+if $data03 != 2 then
-if $loopCnt == 10 then 
+	goto a3
-  return -1
 endi
 
-sql show dnodes
-if $rows != 3 then
-  sleep 2000
-  goto wait_dnode2_ready
-endi
-print $data0_1  $data1_1  $data2_1  $data3_1  $data4_1
-print $data0_2  $data1_2  $data2_2  $data3_2  $data4_2
-print $data0_3  $data1_3  $data2_3  $data3_3  $data4_3
-#print $data0_4  $data1_4  $data2_4  $data3_4  $data4_4
-#print $data0_5  $data1_5  $data2_5  $data3_5  $data4_5
-#print $data0_6  $data1_6  $data2_6  $data3_6  $data4_6
-#$dnode1Status = $data4_1  
-$dnode2Status = $data4_2
-$dnode3Status = $data4_3  
-#$dnode4Status = $data4_4 
-#$dnode5Status = $data4_5
-
-if $dnode2Status != ready then
-  sleep 2000
-  goto wait_dnode2_ready
-endi
-
-sleep $sleepTimer
 # check using select
 sql select count(*) from $stb
 print data00 $data00
@@ -300,40 +221,39 @@ endi
 print ============== step6: start dnode4 and add into cluster , then alter replica from 2 to 3, and waiting sync
 system sh/exec.sh -n dnode4 -s start
 sql create dnode $hostname4
-sleep 3000
+$x = 0
-
+step6: 
-sql alter database $db replica 3
+	$x = $x + 1
-sleep $sleepTimer
+	sleep 1000
-$loopCnt = 0
+	if $x == 10 then
-wait_dnode4_ready:
+		return -1
-$loopCnt = $loopCnt + 1
+	endi
-if $loopCnt == 10 then 
-  return -1
-endi
 
 sql show dnodes
-if $rows != 4 then
+print dnode1 $data4_1
-  sleep 2000
+print dnode2 $data4_2
-  goto wait_dnode4_ready
+print dnode3 $data4_3
-endi
+print dnode4 $data4_4
-print $data0_1  $data1_1  $data2_1  $data3_1  $data4_1
-print $data0_2  $data1_2  $data2_2  $data3_2  $data4_2
-print $data0_3  $data1_3  $data2_3  $data3_3  $data4_3
-print $data0_4  $data1_4  $data2_4  $data3_4  $data4_4
-#print $data0_5  $data1_5  $data2_5  $data3_5  $data4_5
-#print $data0_6  $data1_6  $data2_6  $data3_6  $data4_6
-#$dnode1Status = $data4_1  
-#$dnode2Status = $data4_2
-$dnode3Status = $data4_3  
-$dnode4Status = $data4_4 
-#$dnode5Status = $data4_5
 
-if $dnode4Status != ready then
+if $data4_4 != ready then
-  sleep 2000
+  goto step6
-  goto wait_dnode4_ready
+endi
+
+sql alter database $db replica 3
+$x = 0
+a4: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 40 then
+	  return -1
+	endi
+
+sql show vgroups 
+print online vnodes $data03
+if $data03 != 3 then
+	goto a4
 endi
 
-sleep $sleepTimer
 # check using select
 sql select count(*) from $stb
 print data00 $data00
@@ -343,44 +263,20 @@ endi
 
 print ============== step7: alter replica from 3 to 2, and waiting sync
 sql alter database $db replica 2
-sleep $sleepTimer
+$x = 0
-$loopCnt = 0
+a5: 
-wait_vgroups_replic_to_2:
+	$x = $x + 1
-$loopCnt = $loopCnt + 1
+	sleep 1000
-if $loopCnt == 10 then 
+	if $x == 40 then
-  return -1
+	  return -1
+	endi
+
+sql show vgroups 
+print online vnodes $data03
+if $data03 != 2 then
+	goto a5
 endi
 
-sql show vgroups
-print $data0_1  $data1_1  $data2_1  $data3_1  $data4_1  $data5_1  $data6_1  $data7_1  $data8_1  $data9_1  $data10_1
-print $data0_2  $data1_2  $data2_2  $data3_2  $data4_2  $data5_2  $data6_2  $data7_2  $data8_2  $data9_2  $data10_2
-print $data0_3  $data1_3  $data2_3  $data3_3  $data4_3  $data5_3  $data6_3  $data7_3  $data8_3  $data9_3  $data10_3
-print $data0_4  $data1_4  $data2_4  $data3_4  $data4_4  $data5_4  $data6_4  $data7_4  $data8_4  $data9_4  $data10_4
-#print $data0_5  $data1_5  $data2_5  $data3_5  $data4_5
-#print $data0_6  $data1_6  $data2_6  $data3_6  $data4_6
-$thirdDnode_2 = $data8_1 
-$thirdDnode_3 = $data8_2 
-$thirdDnode_4 = $data8_3 
-$thirdDnode_5 = $data8_4 
-
-if $thirdDnode_2 != null then
-  sleep 2000
-  goto wait_vgroups_replic_to_2
-endi
-if $thirdDnode_3 != null then
-  sleep 2000
-  goto wait_vgroups_replic_to_2
-endi
-if $thirdDnode_4 != null then
-  sleep 2000
-  goto wait_vgroups_replic_to_2
-endi
-if $thirdDnode_5 != null then
-  sleep 2000
-  goto wait_vgroups_replic_to_2
-endi
-
-sleep $sleepTimer  #waiting del one replica data
 # check using select
 sql select count(*) from $stb
 print data00 $data00
@@ -390,85 +286,20 @@ endi
 
 print ============== step8: alter replica from 2 to 1, and waiting sync
 sql alter database $db replica 1
-sleep $sleepTimer
+$x = 0
-$loopCnt = 0
+a6: 
-wait_vgroups_replic_to_1:
+	$x = $x + 1
-$loopCnt = $loopCnt + 1
+	sleep 1000
-if $loopCnt == 10 then 
+	if $x == 40 then
-  return -1
+	  return -1
+	endi
+
+sql show vgroups 
+print online vnodes $data03
+if $data03 != 1 then
+	goto a6
 endi
 
-sql show vgroups
-print $data0_1  $data1_1  $data2_1  $data3_1  $data4_1  $data5_1  $data6_1  $data7_1  $data8_1
-print $data0_2  $data1_2  $data2_2  $data3_2  $data4_2  $data5_2  $data6_2  $data7_2  $data8_2
-print $data0_3  $data1_3  $data2_3  $data3_3  $data4_3  $data5_3  $data6_3  $data7_3  $data8_3
-print $data0_4  $data1_4  $data2_4  $data3_4  $data4_4  $data5_4  $data6_4  $data7_4  $data8_4
-#print $data0_5  $data1_5  $data2_5  $data3_5  $data4_5
-#print $data0_6  $data1_6  $data2_6  $data3_6  $data4_6
-$sencodDnode_2 = $data6_1 
-$sencodDnode_3 = $data6_2 
-$sencodDnode_4 = $data6_3 
-$sencodDnode_5 = $data6_4 
-
-if $sencodDnode_2 != null then
-  sleep 2000
-  goto wait_vgroups_replic_to_1
-endi
-if $sencodDnode_3 != null then
-  sleep 2000
-  goto wait_vgroups_replic_to_1
-endi
-if $sencodDnode_4 != null then
-  sleep 2000
-  goto wait_vgroups_replic_to_1
-endi
-if $sencodDnode_5 != null then
-  sleep 2000
-  goto wait_vgroups_replic_to_1
-endi
-
-$loopCnt = 0
-all_dnodes_ready:
-$loopCnt = $loopCnt + 1
-if $loopCnt == 10 then 
-  return -1
-endi
-
-sql show dnodes
-if $rows != 4 then
-  sleep 2000
-  goto all_dnodes_ready
-endi
-print $data0_1  $data1_1  $data2_1  $data3_1  $data4_1
-print $data0_2  $data1_2  $data2_2  $data3_2  $data4_2
-print $data0_3  $data1_3  $data2_3  $data3_3  $data4_3
-print $data0_4  $data1_4  $data2_4  $data3_4  $data4_4
-#print $data0_5  $data1_5  $data2_5  $data3_5  $data4_5
-#print $data0_6  $data1_6  $data2_6  $data3_6  $data4_6
-$dnode1Status = $data4_1  
-$dnode2Status = $data4_2
-$dnode3Status = $data4_3  
-$dnode4Status = $data4_4 
-#$dnode5Status = $data4_5
-
-if $dnode1Status != ready then
-  sleep 2000
-  goto all_dnodes_ready
-endi
-if $dnode2Status != ready then
-  sleep 2000
-  goto all_dnodes_ready
-endi
-if $dnode3Status != ready then
-  sleep 2000
-  goto all_dnodes_ready
-endi
-if $dnode4Status != ready then
-  sleep 2000
-  goto all_dnodes_ready
-endi
-
-sleep $sleepTimer  #waiting del one replica data
 # check using select
 sql select count(*) from $stb
 print data00 $data00
@@ -481,39 +312,31 @@ sql drop dnode $hostname2
 sql drop dnode $hostname3
 sleep $sleepTimer
 
-$loopCnt = 0
+$x = 0
-wait_dnode23_dropped:
+step9: 
-$loopCnt = $loopCnt + 1
+	$x = $x + 1
-if $loopCnt == 10 then 
+	sleep 1000
-  return -1
+	if $x == 40 then
-endi
+	  return -1
+	endi
 
 sql show dnodes
 if $rows != 2 then
-  sleep 2000
+  goto step9
-  goto wait_dnode23_dropped
 endi
-print $data0_1  $data1_1  $data2_1  $data3_1  $data4_1
-print $data0_2  $data1_2  $data2_2  $data3_2  $data4_2
-print $data0_3  $data1_3  $data2_3  $data3_3  $data4_3
-print $data0_4  $data1_4  $data2_4  $data3_4  $data4_4
-#print $data0_5  $data1_5  $data2_5  $data3_5  $data4_5
-#print $data0_6  $data1_6  $data2_6  $data3_6  $data4_6
-$dnode1Status = $data4_1  
-$dnode2Status = $data4_2
-$dnode3Status = $data4_3  
-$dnode4Status = $data4_4 
 
-if $dnode2Status != null then
+$x = 0
-  sleep 2000
+a7: 
-  goto wait_dnode23_dropped
+	$x = $x + 1
-endi
+	sleep 1000
-if $dnode3Status != null then
+	if $x == 40 then
-  sleep 2000
+	  return -1
-  goto wait_dnode23_dropped
+	endi
-endi
+
-if $dnode4Status != ready then
+sql show vgroups 
-  return -1
+print online vnodes $data03
+if $data03 != 1 then
+	goto a7
 endi
 
 sleep $sleepTimer  #waiting move vnode from dnode3/dnode3 to dnode4
@@ -523,3 +346,13 @@ print data00 $data00
 if $data00 != $totalRows then
 	return -1
 endi
+
+
+system sh/exec.sh -n dnode1 -s stop  -x SIGINT
+system sh/exec.sh -n dnode2 -s stop  -x SIGINT
+system sh/exec.sh -n dnode3 -s stop  -x SIGINT
+system sh/exec.sh -n dnode4 -s stop  -x SIGINT
+system sh/exec.sh -n dnode5 -s stop  -x SIGINT
+system sh/exec.sh -n dnode6 -s stop  -x SIGINT
+system sh/exec.sh -n dnode7 -s stop  -x SIGINT
+system sh/exec.sh -n dnode8 -s stop  -x SIGINT

tests/script/unique/arbitrator/dn3_mn1_replica_change_dropDnod.sim

@@ -140,3 +140,11 @@ if $data00 != $totalRows then
 	return -1
 endi
 
+system sh/exec.sh -n dnode1 -s stop  -x SIGINT
+system sh/exec.sh -n dnode2 -s stop  -x SIGINT
+system sh/exec.sh -n dnode3 -s stop  -x SIGINT
+system sh/exec.sh -n dnode4 -s stop  -x SIGINT
+system sh/exec.sh -n dnode5 -s stop  -x SIGINT
+system sh/exec.sh -n dnode6 -s stop  -x SIGINT
+system sh/exec.sh -n dnode7 -s stop  -x SIGINT
+system sh/exec.sh -n dnode8 -s stop  -x SIGINT

tests/script/unique/arbitrator/dn3_mn1_vnode_nomaster.sim

@@ -158,13 +158,15 @@ if $dnode4Vtatus != offline then
   sleep 2000
   goto wait_dnode4_vgroup_offline
 endi
-if $dnode3Vtatus != unsynced then
+if $dnode3Vtatus != master then
   sleep 2000
   goto wait_dnode4_vgroup_offline
 endi
 
-sql_error select count(*) from $stb
+sql select count(*) from $stb -x s31
-sql_error insert into $tb values (now, 9988)
+s31:
+#sql_error insert into $tb values (now, 9988) -x s32
+#s32:
 
 print ============== step4: restart dnode2, then create database with replica 2, and create table, insert data
 system sh/exec.sh -n dnode2 -s start

tests/script/unique/arbitrator/sync_replica3_alterTable_add.sim

@@ -140,11 +140,9 @@ if $rows != 1 then
   goto wait_dnode4_vgroup_offline
 endi
 print show vgroups:
-print $data0_1  $data1_1  $data2_1  $data3_1  $data4_1  $data5_1  $data6_1  $data7_1  $data8_1  $data9_1
+print $data00  $data01  $data02  $data03  $data04  $data05  $data06  $data07  $data08  $data09
-print $data0_2  $data1_2  $data2_2  $data3_2  $data4_2  $data5_2  $data6_2  $data7_2  $data8_2  $data9_2
+$dnode4Vtatus = $data05
-print $data0_3  $data1_3  $data2_3  $data3_3  $data4_3  $data5_3  $data6_3  $data7_3  $data8_3  $data9_3
+$dnode3Vtatus = $data07
-$dnode4Vtatus = $data5_2
-$dnode3Vtatus = $data7_2
 
 if $dnode4Vtatus != offline then
   sleep 2000
@@ -198,7 +196,3 @@ if $data00 != $totalRows then
 	sleep 2000
 	goto wait_table_altered
 endi
-
-
-
-

tests/script/unique/arbitrator/sync_replica_alterTable_background_add.sim

@@ -14,6 +14,9 @@ sql alter table $stb add column f1 float
 $tblNum = $totalTableNum
 $alterTblNum = 10
 
+sql reset query cache 
+sleep 100
+
 $i = 1
 while $i < $alterTblNum
   $tb = tb . $i

tests/script/unique/big/balance.sim

@@ -21,7 +21,6 @@ system sh/cfg.sh -n dnode5 -c maxTablesPerVnode -v 1000
 
 print =============== prepare data
 system sh/exec.sh -n dnode1 -s start
-sleep 3000
 sql connect
 
 $i = 0
@@ -92,8 +91,8 @@ system sh/exec.sh -n dnode2 -s start
 $x = 0
 show1: 
 	$x = $x + 1
-	sleep 3000
+	sleep 1000
-	if $x == 10 then 
+	if $x == 30 then 
 		return -1
 	endi
 	
@@ -108,7 +107,7 @@ if $data2_2 != 2 then
 endi
 
 sql reset query cache
-sleep 1000
+sleep 100
 
 sql select count(*) from t10
 print select count(*) from t10 $data00 expect $rowNum 
@@ -143,7 +142,6 @@ endi
 print ========== step2
 sql create dnode $hostname3
 system sh/exec.sh -n dnode3 -s start
-sleep 10000
 
 print ========== step3
 sql drop dnode $hostname2
@@ -151,8 +149,8 @@ sql drop dnode $hostname2
 $x = 0
 show3: 
 	$x = $x + 1
-	sleep 3000
+	sleep 1000
-	if $x == 10 then 
+	if $x == 30 then 
 		return -1
 	endi
 	
@@ -171,10 +169,9 @@ if $data2_3 != 2 then
 endi
 
 system sh/exec.sh -n dnode2 -s stop -x SIGINT
-sleep 5000
 
 sql reset query cache
-sleep 1000
+sleep 100
 
 sql select count(*) from t10
 print select count(*) from t10 $data00 expect $rowNum 
@@ -212,8 +209,8 @@ sql drop dnode $hostname3
 $x = 0
 show4: 
 	$x = $x + 1
-	sleep 3000
+	sleep 1000
-	if $x == 10 then 
+	if $x == 30 then 
 		return -1
 	endi
 	
@@ -228,10 +225,9 @@ if $data2_3 != null then
 endi
 
 system sh/exec.sh -n dnode3 -s stop -x SIGINT
-sleep 5000
 
 sql reset query cache
-sleep 1000
+sleep 100
 
 sql select count(*) from t10
 print select count(*) from t10 $data00 expect $rowNum 
@@ -267,14 +263,31 @@ print ========== step5
 system sh/exec.sh -n dnode4 -s start
 sql create dnode $hostname4
 
-sleep 3000
+$x = 0
+step5: 
+	$x = $x + 1
+	sleep 1000
+	if $x == 10 then
+		return -1
+	endi
+
+sql show dnodes
+print dnode1 $data4_1
+print dnode2 $data4_2
+print dnode3 $data4_3
+print dnode4 $data4_4
+
+if $data4_4 != ready then
+  goto step5
+endi
+
 sql alter database db replica 2
 
 $x = 0
 show5: 
 	$x = $x + 1
-	sleep 3000
+	sleep 1000
-	if $x == 10 then 
+	if $x == 30 then 
 		return -1
 	endi
 	
@@ -289,7 +302,7 @@ if $data2_4 != 4 then
 endi
 
 sql reset query cache
-sleep 1000
+sleep 100
 
 sql select count(*) from t10
 print select count(*) from t10 $data00 expect $rowNum 
@@ -321,16 +334,14 @@ if $data00 != $totalNum then
   goto show5
 endi
 
-
 print ========== step6
-sleep 3000
 sql alter database db replica 1
 
 $x = 0
 show6: 
 	$x = $x + 1
-	sleep 3000
+	sleep 1000
-	if $x == 10 then 
+	if $x == 30 then 
 		return -1
 	endi
 	
@@ -345,7 +356,7 @@ if $data2_4 != 2 then
 endi
 
 sql reset query cache
-sleep 1000
+sleep 100
 
 sql select count(*) from t10
 print select count(*) from t10 $data00 expect $rowNum