diff --git a/include/libs/transport/trpc.h b/include/libs/transport/trpc.h
index 7f02eafe3f..8341b2f0bb 100644
--- a/include/libs/transport/trpc.h
+++ b/include/libs/transport/trpc.h
@@ -51,6 +51,7 @@ typedef struct {
   SRpcMsg rpcMsg;
   int32_t rspLen;
   void   *pRsp;
+  void   *pNode;
 } SNodeMsg;
 
 typedef struct SRpcInit {
diff --git a/source/dnode/mgmt/container/inc/dndInt.h b/source/dnode/mgmt/container/inc/dndInt.h
index 90f9e86c25..788f615553 100644
--- a/source/dnode/mgmt/container/inc/dndInt.h
+++ b/source/dnode/mgmt/container/inc/dndInt.h
@@ -56,7 +56,7 @@ extern "C" {
 #define dDebug(...) { if (dDebugFlag & DEBUG_DEBUG) { taosPrintLog("DND ", DEBUG_DEBUG, dDebugFlag, __VA_ARGS__); }}
 #define dTrace(...) { if (dDebugFlag & DEBUG_TRACE) { taosPrintLog("DND ", DEBUG_TRACE, dDebugFlag, __VA_ARGS__); }}
 
-typedef enum { MNODE, VNODES, QNODE, SNODE, BNODE, DNODE, NODE_MAX } ENodeType;
+typedef enum { VNODES, QNODE, SNODE, MNODE, BNODE, DNODE, NODE_MAX } ENodeType;
 typedef enum { PROC_SINGLE, PROC_CHILD, PROC_PARENT } EProcType;
 typedef enum { DND_STAT_INIT, DND_STAT_RUNNING, DND_STAT_STOPPED } EDndStatus;
 typedef enum { DND_WORKER_SINGLE, DND_WORKER_MULTI } EWorkerType;
@@ -77,29 +77,6 @@ typedef int32_t (*OpenNodeFp)(SMgmtWrapper *pWrapper);
 typedef void (*CloseNodeFp)(SMgmtWrapper *pWrapper);
 typedef bool (*RequireNodeFp)(SMgmtWrapper *pWrapper);
 
-typedef struct SMsgHandle {
-  NodeMsgFp     msgFp;
-  SMgmtWrapper *pWrapper;
-} SMsgHandle;
-
-typedef struct SMgmtFp {
-  OpenNodeFp     openFp;
-  CloseNodeFp    closeFp;
-  RequireNodeFp  requiredFp;
-} SMgmtFp;
-
-typedef struct SMgmtWrapper {
-  const char *name;
-  char       *path;
-  bool        required;
-  EProcType   procType;
-  SProcObj   *pProc;
-  void       *pMgmt;
-  SDnode     *pDnode;
-  NodeMsgFp   msgFps[TDMT_MAX];
-  SMgmtFp     fp;
-} SMgmtWrapper;
-
 typedef struct {
   EWorkerType type;
   const char *name;
@@ -114,6 +91,29 @@ typedef struct {
   };
 } SDnodeWorker;
 
+typedef struct SMsgHandle {
+  NodeMsgFp     msgFp;
+  SMgmtWrapper *pWrapper;
+} SMsgHandle;
+
+typedef struct SMgmtFp {
+  OpenNodeFp    openFp;
+  CloseNodeFp   closeFp;
+  RequireNodeFp requiredFp;
+} SMgmtFp;
+
+typedef struct SMgmtWrapper {
+  const char *name;
+  char       *path;
+  bool        required;
+  EProcType   procType;
+  SProcObj   *pProc;
+  void       *pMgmt;
+  SDnode     *pDnode;
+  NodeMsgFp   msgFps[TDMT_MAX];
+  SMgmtFp     fp;
+} SMgmtWrapper;
+
 typedef struct {
   void      *serverRpc;
   void      *clientRpc;
@@ -143,6 +143,7 @@ void          dndGetStartup(SDnode *pDnode, SStartupReq *pStartup);
 TdFilePtr     dndCheckRunning(char *dataDir);
 SMgmtWrapper *dndGetWrapper(SDnode *pDnode, ENodeType nodeType);
 void          dndSetMsgHandle(SMgmtWrapper *pWrapper, int32_t msgType, NodeMsgFp nodeMsgFp);
+void          dndProcessStartupReq(SDnode *pDnode, SRpcMsg *pMsg);
 
 // dndMonitor.h
 void dndSendMonitorReport(SDnode *pDnode);
@@ -153,6 +154,7 @@ void    dndClose(SDnode *pDnode);
 int32_t dndRun(SDnode *pDnode);
 void    dndeHandleEvent(SDnode *pDnode, EDndEvent event);
 void    dndProcessRpcMsg(SMgmtWrapper *pWrapper, SRpcMsg *pMsg, SEpSet *pEpSet);
+void    dndSendRsp(SMgmtWrapper *pWrapper, SRpcMsg *pRsp);
 
 // dndTransport.h
 int32_t dndSendReqToMnode(SDnode *pDnode, SRpcMsg *pRpcMsg);
diff --git a/source/dnode/mgmt/container/inc/dndNode.h b/source/dnode/mgmt/container/inc/dndNode.h
index 7a92248021..0fbae69d89 100644
--- a/source/dnode/mgmt/container/inc/dndNode.h
+++ b/source/dnode/mgmt/container/inc/dndNode.h
@@ -27,6 +27,7 @@ void    dndClose(SDnode *pDnode);
 int32_t dndRun(SDnode *pDnode);
 void    dndeHandleEvent(SDnode *pDnode, EDndEvent event);
 void    dndProcessRpcMsg(SMgmtWrapper *pWrapper, SRpcMsg *pMsg, SEpSet *pEpSet);
+void    dndSendRsp(SMgmtWrapper *pWrapper, SRpcMsg *pRsp);
 
 #ifdef __cplusplus
 }
diff --git a/source/dnode/mgmt/container/src/dndInt.c b/source/dnode/mgmt/container/src/dndInt.c
index 1306c862bc..7f4b000ed5 100644
--- a/source/dnode/mgmt/container/src/dndInt.c
+++ b/source/dnode/mgmt/container/src/dndInt.c
@@ -126,3 +126,15 @@ TdFilePtr dndCheckRunning(char *dataDir) {
   dDebug("file:%s is locked", filepath);
   return pFile;
 }
+
+void dndProcessStartupReq(SDnode *pDnode, SRpcMsg *pReq) {
+  dDebug("startup req is received");
+
+  SStartupReq *pStartup = rpcMallocCont(sizeof(SStartupReq));
+  dndGetStartup(pDnode, pStartup);
+
+  dDebug("startup req is sent, step:%s desc:%s finished:%d", pStartup->name, pStartup->desc, pStartup->finished);
+
+  SRpcMsg rpcRsp = {.handle = pReq->handle, .pCont = pStartup, .contLen = sizeof(SStartupReq)};
+  rpcSendResponse(&rpcRsp);
+}
diff --git a/source/dnode/mgmt/container/src/dndMonitor.c b/source/dnode/mgmt/container/src/dndMonitor.c
index 4362049a5a..8b7ff389f6 100644
--- a/source/dnode/mgmt/container/src/dndMonitor.c
+++ b/source/dnode/mgmt/container/src/dndMonitor.c
@@ -66,7 +66,7 @@ void dndSendMonitorReport(SDnode *pDnode) {
   SMonClusterInfo clusterInfo = {0};
   SMonVgroupInfo  vgroupInfo = {0};
   SMonGrantInfo   grantInfo = {0};
-  if (mmGetMonitorInfo(pDnode, &clusterInfo, &vgroupInfo, &grantInfo) == 0) {
+  if (mmGetMonitorInfo(dndGetWrapper(pDnode, MNODE), &clusterInfo, &vgroupInfo, &grantInfo) == 0) {
     monSetClusterInfo(pMonitor, &clusterInfo);
     monSetVgroupInfo(pMonitor, &vgroupInfo);
     monSetGrantInfo(pMonitor, &grantInfo);
diff --git a/source/dnode/mgmt/container/src/dndNode.c b/source/dnode/mgmt/container/src/dndNode.c
index 82130660ab..a64cb4f2b8 100644
--- a/source/dnode/mgmt/container/src/dndNode.c
+++ b/source/dnode/mgmt/container/src/dndNode.c
@@ -49,6 +49,10 @@ static void dndCloseNode(SMgmtWrapper *pWrapper) {
   if (pWrapper->required) {
     (*pWrapper->fp.closeFp)(pWrapper);
   }
+  if (pWrapper->pProc) {
+    taosProcCleanup(pWrapper->pProc);
+    pWrapper->pProc = NULL;
+  }
 }
 
 static void dndClearMemory(SDnode *pDnode) {
@@ -171,6 +175,7 @@ static int32_t dndRunInSingleProcess(SDnode *pDnode) {
     if (!pWrapper->required) continue;
 
     dInfo("node:%s, will start in single process", pWrapper->name);
+    pWrapper->procType = PROC_SINGLE;
     if (dndOpenNode(pWrapper) != 0) {
       dError("node:%s, failed to start since %s", pWrapper->name, terrstr());
       return -1;
@@ -189,14 +194,67 @@ static void dndClearNodesExecpt(SDnode *pDnode, ENodeType except) {
   }
 }
 
+static void dndSendRpcRsp(SMgmtWrapper *pWrapper, SRpcMsg *pRsp) {
+  if (pRsp->code == TSDB_CODE_DND_MNODE_NOT_DEPLOYED || pRsp->code == TSDB_CODE_APP_NOT_READY) {
+    dmSendRedirectRsp(pWrapper->pDnode, pRsp);
+  } else {
+    rpcSendResponse(pRsp);
+  }
+}
+
+void dndSendRsp(SMgmtWrapper *pWrapper, SRpcMsg *pRsp) {
+  int32_t code = -1;
+
+  if (pWrapper->procType != PROC_CHILD) {
+    dndSendRpcRsp(pWrapper, pRsp);
+  } else {
+    do {
+      code = taosProcPutToParentQueue(pWrapper->pProc, pRsp, sizeof(SRpcMsg), pRsp->pCont, pRsp->contLen);
+      if (code != 0) {
+        taosMsleep(10);
+      }
+    } while (code != 0);
+  }
+}
+
+static void dndConsumeChildQueue(SMgmtWrapper *pWrapper, SNodeMsg *pMsg, int32_t msgLen, void *pCont, int32_t contLen) {
+  dTrace("msg:%p, get from child queue", pMsg);
+  SRpcMsg *pRpc = &pMsg->rpcMsg;
+  pRpc->pCont = pCont;
+
+  NodeMsgFp msgFp = pWrapper->msgFps[TMSG_INDEX(pRpc->msgType)];
+  int32_t   code = (*msgFp)(pWrapper, pMsg);
+
+  if (code != 0) {
+    bool isReq = (pRpc->msgType & 1U);
+    if (isReq) {
+      SRpcMsg rsp = {.handle = pRpc->handle, .ahandle = pRpc->ahandle, .code = terrno};
+      dndSendRsp(pWrapper, &rsp);
+    }
+
+    dTrace("msg:%p, is freed", pMsg);
+    taosFreeQitem(pMsg);
+    rpcFreeCont(pCont);
+  }
+}
+
+static void dndConsumeParentQueue(SMgmtWrapper *pWrapper, SRpcMsg *pRsp, int32_t msgLen, void *pCont, int32_t contLen) {
+  dTrace("msg:%p, get from parent queue", pRsp);
+  pRsp->pCont = pCont;
+  dndSendRpcRsp(pWrapper, pRsp);
+  free(pRsp);
+}
+
 static int32_t dndRunInMultiProcess(SDnode *pDnode) {
+  dInfo("dnode run in multi process mode");
+
   for (ENodeType n = 0; n < NODE_MAX; ++n) {
     SMgmtWrapper *pWrapper = &pDnode->wrappers[n];
     if (!pWrapper->required) continue;
 
     if (n == DNODE) {
       dInfo("node:%s, will start in parent process", pWrapper->name);
-      pWrapper->procType = PROC_PARENT;
+      pWrapper->procType = PROC_SINGLE;
       if (dndOpenNode(pWrapper) != 0) {
         dError("node:%s, failed to start since %s", pWrapper->name, terrstr());
         return -1;
@@ -204,7 +262,21 @@ static int32_t dndRunInMultiProcess(SDnode *pDnode) {
       continue;
     }
 
-    SProcCfg  cfg = {0};
+    SProcCfg  cfg = {.childQueueSize = 1024 * 1024 * 2,  // size will be a configuration item
+                     .childConsumeFp = (ProcConsumeFp)dndConsumeChildQueue,
+                     .childMallocHeadFp = (ProcMallocFp)taosAllocateQitem,
+                     .childFreeHeadFp = (ProcFreeFp)taosFreeQitem,
+                     .childMallocBodyFp = (ProcMallocFp)rpcMallocCont,
+                     .childFreeBodyFp = (ProcFreeFp)rpcFreeCont,
+                     .parentQueueSize = 1024 * 1024 * 2,  // size will be a configuration item
+                     .parentConsumeFp = (ProcConsumeFp)dndConsumeParentQueue,
+                     .parentdMallocHeadFp = (ProcMallocFp)malloc,
+                     .parentFreeHeadFp = (ProcFreeFp)free,
+                     .parentMallocBodyFp = (ProcMallocFp)rpcMallocCont,
+                     .parentFreeBodyFp = (ProcFreeFp)rpcFreeCont,
+                     .testFlag = 0,
+                     .pParent = pWrapper,
+                     .name = pWrapper->name};
     SProcObj *pProc = taosProcInit(&cfg);
     if (pProc == NULL) {
       dError("node:%s, failed to fork since %s", pWrapper->name, terrstr());
@@ -227,6 +299,11 @@ static int32_t dndRunInMultiProcess(SDnode *pDnode) {
       dInfo("node:%s, will not start in parent process", pWrapper->name);
       pWrapper->procType = PROC_PARENT;
     }
+
+    if (taosProcRun(pProc) != 0) {
+      dError("node:%s, failed to run proc since %s", pWrapper->name, terrstr());
+      return -1;
+    }
   }
 
   return 0;
@@ -275,14 +352,6 @@ static int32_t dndBuildMsg(SNodeMsg *pMsg, SRpcMsg *pRpc, SEpSet *pEpSet) {
   return 0;
 }
 
-static void dndSendRpcRsp(SDnode *pDnode, SRpcMsg *pRpc) {
-  if (pRpc->code == TSDB_CODE_APP_NOT_READY) {
-    dmSendRedirectRsp(pDnode, pRpc);
-  } else {
-    rpcSendResponse(pRpc);
-  }
-}
-
 void dndProcessRpcMsg(SMgmtWrapper *pWrapper, SRpcMsg *pRpc, SEpSet *pEpSet) {
   if (pEpSet && pEpSet->numOfEps > 0 && pRpc->msgType == TDMT_MND_STATUS_RSP) {
     dmUpdateMnodeEpSet(pWrapper->pDnode, pEpSet);
@@ -307,19 +376,33 @@ void dndProcessRpcMsg(SMgmtWrapper *pWrapper, SRpcMsg *pRpc, SEpSet *pEpSet) {
   }
 
   dTrace("msg:%p, is created, app:%p user:%s", pMsg, pRpc->ahandle, pMsg->user);
-  code = (*msgFp)(pWrapper, pMsg);
+
+  if (pWrapper->procType == PROC_SINGLE) {
+    code = (*msgFp)(pWrapper, pMsg);
+  } else if (pWrapper->procType == PROC_PARENT) {
+    code = taosProcPutToChildQueue(pWrapper->pProc, pMsg, sizeof(SNodeMsg), pRpc->pCont, pRpc->contLen);
+  } else {
+    terrno = TSDB_CODE_MEMORY_CORRUPTED;
+    dError("msg:%p, won't be processed for it is child process", pMsg);
+  }
 
 _OVER:
 
-  if (code != 0) {
+  if (code == 0) {
+    if (pWrapper->procType == PROC_PARENT) {
+      dTrace("msg:%p, is freed", pMsg);
+      taosFreeQitem(pMsg);
+      rpcFreeCont(pRpc->pCont);
+    }
+  } else {
     dError("msg:%p, failed to process since %s", pMsg, terrstr());
     bool isReq = (pRpc->msgType & 1U);
     if (isReq) {
       SRpcMsg rsp = {.handle = pRpc->handle, .ahandle = pRpc->ahandle, .code = terrno};
-      dndSendRpcRsp(pWrapper->pDnode, &rsp);
+      dndSendRsp(pWrapper, &rsp);
     }
     dTrace("msg:%p, is freed", pMsg);
     taosFreeQitem(pMsg);
     rpcFreeCont(pRpc->pCont);
   }
-}
\ No newline at end of file
+}
diff --git a/source/dnode/mgmt/container/src/dndTransport.c b/source/dnode/mgmt/container/src/dndTransport.c
index 196bdb165b..002ce73460 100644
--- a/source/dnode/mgmt/container/src/dndTransport.c
+++ b/source/dnode/mgmt/container/src/dndTransport.c
@@ -25,8 +25,7 @@
 static void dndProcessResponse(void *parent, SRpcMsg *pRsp, SEpSet *pEpSet) {
   SDnode     *pDnode = parent;
   STransMgmt *pMgmt = &pDnode->trans;
-
-  tmsg_t msgType = pRsp->msgType;
+  tmsg_t      msgType = pRsp->msgType;
 
   if (dndGetStatus(pDnode) == DND_STAT_STOPPED) {
     if (pRsp == NULL || pRsp->pCont == NULL) return;
@@ -88,11 +87,11 @@ void dndCleanupClient(SDnode *pDnode) {
 static void dndProcessRequest(void *param, SRpcMsg *pReq, SEpSet *pEpSet) {
   SDnode     *pDnode = param;
   STransMgmt *pMgmt = &pDnode->trans;
+  tmsg_t      msgType = pReq->msgType;
 
-  tmsg_t msgType = pReq->msgType;
   if (msgType == TDMT_DND_NETWORK_TEST) {
     dTrace("RPC %p, network test req will be processed, app:%p", pReq->handle, pReq->ahandle);
-    dmProcessStartupReq(pDnode, pReq);
+    dndProcessStartupReq(pDnode, pReq);
     return;
   }
 
@@ -249,8 +248,8 @@ void dndCleanupServer(SDnode *pDnode) {
 int32_t dndInitMsgHandle(SDnode *pDnode) {
   STransMgmt *pMgmt = &pDnode->trans;
 
-  for (ENodeType nodeType = 0; nodeType < NODE_MAX; ++nodeType) {
-    SMgmtWrapper *pWrapper = &pDnode->wrappers[nodeType];
+  for (ENodeType n = 0; n < NODE_MAX; ++n) {
+    SMgmtWrapper *pWrapper = &pDnode->wrappers[n];
 
     for (int32_t msgIndex = 0; msgIndex < TDMT_MAX; ++msgIndex) {
       NodeMsgFp msgFp = pWrapper->msgFps[msgIndex];
diff --git a/source/dnode/mgmt/dnode/inc/dmInt.h b/source/dnode/mgmt/dnode/inc/dmInt.h
index 6a3e205f7c..466c124866 100644
--- a/source/dnode/mgmt/dnode/inc/dmInt.h
+++ b/source/dnode/mgmt/dnode/inc/dmInt.h
@@ -51,9 +51,6 @@ void    dmUpdateMnodeEpSet(SDnode *pDnode, SEpSet *pEpSet);
 void    dmGetDnodeEp(SDnode *pDnode, int32_t dnodeId, char *pEp, char *pFqdn, uint16_t *pPort);
 void    dmSendRedirectRsp(SDnode *pDnode, SRpcMsg *pMsg);
 
-// dmHandle.h
-void dmProcessStartupReq(SDnode *pDnode, SRpcMsg *pMsg);
-
 #ifdef __cplusplus
 }
 #endif
diff --git a/source/dnode/mgmt/dnode/inc/dmMsg.h b/source/dnode/mgmt/dnode/inc/dmMsg.h
index 3877d03c96..036471f17f 100644
--- a/source/dnode/mgmt/dnode/inc/dmMsg.h
+++ b/source/dnode/mgmt/dnode/inc/dmMsg.h
@@ -24,7 +24,6 @@ extern "C" {
 
 void    dmInitMsgHandles(SMgmtWrapper *pWrapper);
 void    dmSendStatusReq(SDnodeMgmt *pMgmt);
-void    dmProcessStartupReq(SDnode *pDnode, SRpcMsg *pMsg);
 int32_t dmProcessConfigReq(SDnode *pDnode, SRpcMsg *pReq);
 void    dmProcessStatusRsp(SDnode *pDnode, SRpcMsg *pRsp);
 void    dmProcessAuthRsp(SDnode *pDnode, SRpcMsg *pRsp);
diff --git a/source/dnode/mgmt/dnode/inc/dmWorker.h b/source/dnode/mgmt/dnode/inc/dmWorker.h
index 992dca1ea7..611a5e7c67 100644
--- a/source/dnode/mgmt/dnode/inc/dmWorker.h
+++ b/source/dnode/mgmt/dnode/inc/dmWorker.h
@@ -22,8 +22,8 @@
 extern "C" {
 #endif
 
-int32_t dmStartWorker();
-void    dmStopWorker();
+int32_t dmStartWorker(SDnodeMgmt *pMgmt);
+void    dmStopWorker(SDnodeMgmt *pMgmt);
 int32_t dmProcessMgmtMsg(SMgmtWrapper *pWrapper, SNodeMsg *pMsg);
 
 #ifdef __cplusplus
diff --git a/source/dnode/mgmt/dnode/src/dmFile.c b/source/dnode/mgmt/dnode/src/dmFile.c
index 0794824ac6..03fc47ae27 100644
--- a/source/dnode/mgmt/dnode/src/dmFile.c
+++ b/source/dnode/mgmt/dnode/src/dmFile.c
@@ -151,8 +151,8 @@ PRASE_DNODE_OVER:
 
   dmResetDnodes(pMgmt, pMgmt->pDnodeEps);
 
-  terrno = 0;
-  return 0;
+  terrno = code;
+  return code;
 }
 
 int32_t dmWriteFile(SDnodeMgmt *pMgmt) {
diff --git a/source/dnode/mgmt/dnode/src/dmInt.c b/source/dnode/mgmt/dnode/src/dmInt.c
index 27d4c5a3f9..23251a2ff2 100644
--- a/source/dnode/mgmt/dnode/src/dmInt.c
+++ b/source/dnode/mgmt/dnode/src/dmInt.c
@@ -108,6 +108,7 @@ void dmSendRedirectRsp(SDnode *pDnode, SRpcMsg *pReq) {
 int32_t dmInit(SMgmtWrapper *pWrapper) {
   SDnode     *pDnode = pWrapper->pDnode;
   SDnodeMgmt *pMgmt = calloc(1, sizeof(SDnodeMgmt));
+  dInfo("dnode-mgmt is initialized");
 
   pMgmt->dnodeId = 0;
   pMgmt->dropped = 0;
@@ -120,23 +121,23 @@ int32_t dmInit(SMgmtWrapper *pWrapper) {
 
   pMgmt->dnodeHash = taosHashInit(4, taosGetDefaultHashFunction(TSDB_DATA_TYPE_INT), true, HASH_NO_LOCK);
   if (pMgmt->dnodeHash == NULL) {
-    dError("node:%s, failed to init dnode hash", pWrapper->name);
+    dError("failed to init dnode hash");
     terrno = TSDB_CODE_OUT_OF_MEMORY;
     return -1;
   }
 
   if (dmReadFile(pMgmt) != 0) {
-    dError("node:%s, failed to read file since %s", pWrapper->name, terrstr());
+    dError("failed to read file since %s", terrstr());
     return -1;
   }
 
   if (pMgmt->dropped) {
-    dError("node:%s, will not start since its already dropped", pWrapper->name);
+    dError("dnode will not start since its already dropped");
     return -1;
   }
 
   if (dmStartWorker(pMgmt) != 0) {
-    dError("node:%s, failed to start worker since %s", pWrapper->name, terrstr());
+    dError("failed to start dnode worker since %s", terrstr());
     return -1;
   }
 
@@ -153,6 +154,7 @@ void dmCleanup(SMgmtWrapper *pWrapper) {
   SDnodeMgmt *pMgmt = pWrapper->pMgmt;
   if (pMgmt == NULL) return;
 
+  dInfo("dnode-mgmt start to clean up");
   dmStopWorker(pMgmt);
 
   taosWLockLatch(&pMgmt->latch);
@@ -168,6 +170,9 @@ void dmCleanup(SMgmtWrapper *pWrapper) {
   }
 
   taosWUnLockLatch(&pMgmt->latch);
+
+  free(pMgmt);
+  pWrapper->pMgmt = NULL;
   dInfo("dnode-mgmt is cleaned up");
 }
 
diff --git a/source/dnode/mgmt/dnode/src/dmMsg.c b/source/dnode/mgmt/dnode/src/dmMsg.c
index 5879e29b56..28d7eb0d29 100644
--- a/source/dnode/mgmt/dnode/src/dmMsg.c
+++ b/source/dnode/mgmt/dnode/src/dmMsg.c
@@ -102,18 +102,6 @@ int32_t dmProcessConfigReq(SDnode *pDnode, SRpcMsg *pReq) {
   return TSDB_CODE_OPS_NOT_SUPPORT;
 }
 
-void dmProcessStartupReq(SDnode *pDnode, SRpcMsg *pReq) {
-  dDebug("startup req is received");
-
-  SStartupReq *pStartup = rpcMallocCont(sizeof(SStartupReq));
-  dndGetStartup(pDnode, pStartup);
-
-  dDebug("startup req is sent, step:%s desc:%s finished:%d", pStartup->name, pStartup->desc, pStartup->finished);
-
-  SRpcMsg rpcRsp = {.handle = pReq->handle, .pCont = pStartup, .contLen = sizeof(SStartupReq)};
-  rpcSendResponse(&rpcRsp);
-}
-
 void dmInitMsgHandles(SMgmtWrapper *pWrapper) {
   // Requests handled by DNODE
   dndSetMsgHandle(pWrapper, TDMT_DND_CREATE_MNODE, dmProcessMgmtMsg);
diff --git a/source/dnode/mgmt/dnode/src/dmWorker.c b/source/dnode/mgmt/dnode/src/dmWorker.c
index 80fa34839e..161b740ac9 100644
--- a/source/dnode/mgmt/dnode/src/dmWorker.c
+++ b/source/dnode/mgmt/dnode/src/dmWorker.c
@@ -57,7 +57,7 @@ static void *dmThreadRoutine(void *param) {
 static void dmProcessMgmtQueue(SDnode *pDnode, SNodeMsg *pNodeMsg) {
   int32_t  code = 0;
   SRpcMsg *pMsg = &pNodeMsg->rpcMsg;
-  dTrace("msg:%p, will be processed in mgmt queue", pNodeMsg);
+  dTrace("msg:%p, will be processed", pNodeMsg);
 
   switch (pMsg->msgType) {
     case TDMT_DND_CREATE_MNODE:
diff --git a/source/dnode/mgmt/mnode/inc/mmFile.h b/source/dnode/mgmt/mnode/inc/mmFile.h
index 0aae15f077..8beceea305 100644
--- a/source/dnode/mgmt/mnode/inc/mmFile.h
+++ b/source/dnode/mgmt/mnode/inc/mmFile.h
@@ -22,8 +22,8 @@
 extern "C" {
 #endif
 
-int32_t mmReadFile(SDnode *pDnode);
-int32_t mmWriteFile(SDnode *pDnode);
+int32_t mmReadFile(SMnodeMgmt *pMgmt);
+int32_t mmWriteFile(SMnodeMgmt *pMgmt);
 
 #ifdef __cplusplus
 }
diff --git a/source/dnode/mgmt/mnode/inc/mmInt.h b/source/dnode/mgmt/mnode/inc/mmInt.h
index 1f7d94aa27..ac1b79ee62 100644
--- a/source/dnode/mgmt/mnode/inc/mmInt.h
+++ b/source/dnode/mgmt/mnode/inc/mmInt.h
@@ -30,31 +30,35 @@ typedef struct SMnodeMgmt {
   int8_t       selfIndex;
   SReplica     replicas[TSDB_MAX_REPLICA];
   SMnode      *pMnode;
+  SDnode      *pDnode;
   SProcObj    *pProcess;
   bool         singleProc;
   SRWLatch     latch;
+  const char  *path;
   SDnodeWorker readWorker;
   SDnodeWorker writeWorker;
   SDnodeWorker syncWorker;
 } SMnodeMgmt;
 
-// mmInt.h
-void mmGetMgmtFp(SMgmtWrapper *pMgmt);
+// interface
+void    mmGetMgmtFp(SMgmtWrapper *pMgmt);
+int32_t mmGetUserAuth(SMgmtWrapper *pWrapper, char *user, char *spi, char *encrypt, char *secret, char *ckey);
+int32_t mmGetMonitorInfo(SMgmtWrapper *pWrapper, SMonClusterInfo *pClusterInfo, SMonVgroupInfo *pVgroupInfo,
+                         SMonGrantInfo *pGrantInfo);
 
-// mmMgmt.h
-int32_t mmInit(SDnode *pDnode);
-void    mmCleanup(SDnode *pDnode);
+// mmInt.h
+SMnode *mmAcquire(SMnodeMgmt *pMgmt);
+void    mmRelease(SMnodeMgmt *pMgmt, SMnode *pMnode);
+int32_t mmOpen(SMnodeMgmt *pMgmt, SMnodeOpt *pOption);
+int32_t mmAlter(SMnodeMgmt *pMgmt, SMnodeOpt *pOption);
+int32_t mmDrop(SMnodeMgmt *pMgmt);
+int32_t mmBuildOptionFromReq(SMnodeMgmt *pMgmt, SMnodeOpt *pOption, SDCreateMnodeReq *pCreate);
 
 // mmHandle.h
 int32_t mmProcessCreateReq(SDnode *pDnode, SRpcMsg *pRpcMsg);
 int32_t mmProcessAlterReq(SDnode *pDnode, SRpcMsg *pRpcMsg);
 int32_t mmProcessDropReq(SDnode *pDnode, SRpcMsg *pRpcMsg);
 
-int32_t mmGetUserAuth(SMgmtWrapper *pWrapper, char *user, char *spi, char *encrypt, char *secret, char *ckey);
-int32_t mmGetMonitorInfo(SDnode *pDnode, SMonClusterInfo *pClusterInfo, SMonVgroupInfo *pVgroupInfo,
-                         SMonGrantInfo *pGrantInfo);
-
-
 #ifdef __cplusplus
 }
 #endif
diff --git a/source/dnode/mgmt/mnode/inc/mmMgmt.h b/source/dnode/mgmt/mnode/inc/mmMgmt.h
deleted file mode 100644
index d2075e40b6..0000000000
--- a/source/dnode/mgmt/mnode/inc/mmMgmt.h
+++ /dev/null
@@ -1,36 +0,0 @@
-/*
- * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
- *
- * This program is free software: you can use, redistribute, and/or modify
- * it under the terms of the GNU Affero General Public License, version 3
- * or later ("AGPL"), as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.
- *
- * You should have received a copy of the GNU Affero General Public License
- * along with this program. If not, see <http://www.gnu.org/licenses/>.
- */
-
-#ifndef _TD_DND_MNODE_MGMT_H_
-#define _TD_DND_MNODE_MGMT_H_
-
-#include "mmInt.h"
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-SMnode *mmAcquire(SDnode *pDnode);
-void    mmRelease(SDnode *pDnode, SMnode *pMnode);
-int32_t mmOpen(SDnode *pDnode, SMnodeOpt *pOption);
-int32_t mmAlter(SDnode *pDnode, SMnodeOpt *pOption);
-int32_t mmDrop(SDnode *pDnode);
-int32_t mmBuildOptionFromReq(SDnode *pDnode, SMnodeOpt *pOption, SDCreateMnodeReq *pCreate);
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif /*_TD_DND_MNODE_MGMT_H_*/
\ No newline at end of file
diff --git a/source/dnode/mgmt/mnode/inc/mmMsg.h b/source/dnode/mgmt/mnode/inc/mmMsg.h
index 620d2903e4..fdafff901f 100644
--- a/source/dnode/mgmt/mnode/inc/mmMsg.h
+++ b/source/dnode/mgmt/mnode/inc/mmMsg.h
@@ -27,11 +27,6 @@ int32_t mmProcessCreateReq(SDnode *pDnode, SRpcMsg *pRpcMsg);
 int32_t mmProcessAlterReq(SDnode *pDnode, SRpcMsg *pRpcMsg);
 int32_t mmProcessDropReq(SDnode *pDnode, SRpcMsg *pRpcMsg);
 
-int32_t mmGetUserAuth(SMgmtWrapper *pWrapper, char *user, char *spi, char *encrypt, char *secret, char *ckey);
-int32_t mmGetMonitorInfo(SDnode *pDnode, SMonClusterInfo *pClusterInfo, SMonVgroupInfo *pVgroupInfo,
-                         SMonGrantInfo *pGrantInfo);
-
-
 #ifdef __cplusplus
 }
 #endif
diff --git a/source/dnode/mgmt/mnode/inc/mmWorker.h b/source/dnode/mgmt/mnode/inc/mmWorker.h
index a1dbd5c267..f86b484699 100644
--- a/source/dnode/mgmt/mnode/inc/mmWorker.h
+++ b/source/dnode/mgmt/mnode/inc/mmWorker.h
@@ -22,19 +22,15 @@
 extern "C" {
 #endif
 
-int32_t mmStartWorker(SDnode *pDnode);
-void    mmStopWorker(SDnode *pDnode);
-void    mmInitMsgFp(SMnodeMgmt *pMgmt);
-void    mmProcessRpcMsg(SDnode *pDnode, SRpcMsg *pMsg, SEpSet *pEpSet);
-int32_t mmPutMsgToWriteQueue(SDnode *pDnode, SRpcMsg *pRpcMsg);
-int32_t mmPutMsgToReadQueue(SDnode *pDnode, SRpcMsg *pRpcMsg);
-void    mmConsumeChildQueue(SDnode *pDnode, SMndMsg *pMsg, int32_t msgLen, void *pCont, int32_t contLen);
-void    mmConsumeParentQueue(SDnode *pDnode, SRpcMsg *pMsg, int32_t msgLen, void *pCont, int32_t contLen);
-
+int32_t mmStartWorker(SMnodeMgmt *pMgmt);
+void    mmStopWorker(SMnodeMgmt *pMgmt);
 int32_t mmProcessWriteMsg(SMgmtWrapper *pWrapper, SNodeMsg *pMsg);
 int32_t mmProcessSyncMsg(SMgmtWrapper *pWrapper, SNodeMsg *pMsg);
 int32_t mmProcessReadMsg(SMgmtWrapper *pWrapper, SNodeMsg *pMsg);
 
+int32_t mmPutMsgToWriteQueue(SDnode *pDnode, SRpcMsg *pRpcMsg);
+int32_t mmPutMsgToReadQueue(SDnode *pDnode, SRpcMsg *pRpcMsg);
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/source/dnode/mgmt/mnode/src/mmFile.c b/source/dnode/mgmt/mnode/src/mmFile.c
index e6edfe1dce..a232d9a3ad 100644
--- a/source/dnode/mgmt/mnode/src/mmFile.c
+++ b/source/dnode/mgmt/mnode/src/mmFile.c
@@ -14,22 +14,19 @@
  */
 
 #define _DEFAULT_SOURCE
-#include "mmInt.h"
+#include "mmFile.h"
 
-#if 0
-int32_t mmReadFile(SDnode *pDnode) {
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
+int32_t mmReadFile(SMnodeMgmt *pMgmt) {
+  int32_t   code = TSDB_CODE_DND_MNODE_READ_FILE_ERROR;
+  int32_t   len = 0;
+  int32_t   maxLen = 4096;
+  char     *content = calloc(1, maxLen + 1);
+  cJSON    *root = NULL;
+  char      file[PATH_MAX];
+  TdFilePtr pFile = NULL;
 
-  int32_t code = TSDB_CODE_DND_MNODE_READ_FILE_ERROR;
-  int32_t len = 0;
-  int32_t maxLen = 4096;
-  char   *content = calloc(1, maxLen + 1);
-  cJSON  *root = NULL;
-  char    file[PATH_MAX + 20];
-
-  snprintf(file, sizeof(file), "%s%smnode.json", pDnode->dir.dnode, TD_DIRSEP);
-
-  TdFilePtr pFile = taosOpenFile(file, TD_FILE_READ);
+  snprintf(file, sizeof(file), "%s%smnode.json", pMgmt->path, TD_DIRSEP);
+  pFile = taosOpenFile(file, TD_FILE_READ);
   if (pFile == NULL) {
     dDebug("file %s not exist", file);
     code = 0;
@@ -115,11 +112,9 @@ PRASE_MNODE_OVER:
   return code;
 }
 
-int32_t mmWriteFile(SDnode *pDnode) {
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-
+int32_t mmWriteFile(SMnodeMgmt *pMgmt) {
   char file[PATH_MAX];
-  snprintf(file, sizeof(file), "%s%smnode.json.bak", pDnode->dir.dnode, TD_DIRSEP);
+  snprintf(file, sizeof(file), "%s%smnode.json.bak", pMgmt->path, TD_DIRSEP);
 
   TdFilePtr pFile = taosOpenFile(file, TD_FILE_CTEATE | TD_FILE_WRITE | TD_FILE_TRUNC);
   if (pFile == NULL) {
@@ -156,7 +151,7 @@ int32_t mmWriteFile(SDnode *pDnode) {
   free(content);
 
   char realfile[PATH_MAX + 20];
-  snprintf(realfile, sizeof(realfile), "%s%smnode.json", pDnode->dir.dnode, TD_DIRSEP);
+  snprintf(realfile, sizeof(realfile), "%s%smnode.json", pMgmt->path, TD_DIRSEP);
 
   if (taosRenameFile(file, realfile) != 0) {
     terrno = TSDB_CODE_DND_MNODE_WRITE_FILE_ERROR;
@@ -167,5 +162,3 @@ int32_t mmWriteFile(SDnode *pDnode) {
   dInfo("successed to write %s, deployed:%d dropped:%d", realfile, pMgmt->deployed, pMgmt->dropped);
   return 0;
 }
-
-#endif
\ No newline at end of file
diff --git a/source/dnode/mgmt/mnode/src/mmInt.c b/source/dnode/mgmt/mnode/src/mmInt.c
index 77d94ffef5..b0acd9def1 100644
--- a/source/dnode/mgmt/mnode/src/mmInt.c
+++ b/source/dnode/mgmt/mnode/src/mmInt.c
@@ -15,15 +15,286 @@
 
 #define _DEFAULT_SOURCE
 #include "mmInt.h"
+#include "dmInt.h"
+#include "mmFile.h"
 #include "mmMsg.h"
+#include "mmWorker.h"
 
-bool mmRequireNode(SMgmtWrapper *pWrapper) { return false; }
+SMnode *mmAcquire(SMnodeMgmt *pMgmt) {
+  SMnode *pMnode = NULL;
+  int32_t refCount = 0;
+
+  taosRLockLatch(&pMgmt->latch);
+  if (pMgmt->deployed && !pMgmt->dropped) {
+    refCount = atomic_add_fetch_32(&pMgmt->refCount, 1);
+    pMnode = pMgmt->pMnode;
+  } else {
+    terrno = TSDB_CODE_DND_MNODE_NOT_DEPLOYED;
+  }
+  taosRUnLockLatch(&pMgmt->latch);
+
+  if (pMnode != NULL) {
+    dTrace("acquire mnode, refCount:%d", refCount);
+  }
+  return pMnode;
+}
+
+void mmRelease(SMnodeMgmt *pMgmt, SMnode *pMnode) {
+  if (pMnode == NULL) return;
+
+  taosRLockLatch(&pMgmt->latch);
+  int32_t refCount = atomic_sub_fetch_32(&pMgmt->refCount, 1);
+  taosRUnLockLatch(&pMgmt->latch);
+  dTrace("release mnode, refCount:%d", refCount);
+}
+
+int32_t mmOpen(SMnodeMgmt *pMgmt, SMnodeOpt *pOption) {
+  SMnode *pMnode = mndOpen(pMgmt->path, pOption);
+  if (pMnode == NULL) {
+    dError("failed to open mnode since %s", terrstr());
+    return -1;
+  }
+
+  if (mmStartWorker(pMgmt) != 0) {
+    dError("failed to start mnode worker since %s", terrstr());
+    mndClose(pMnode);
+    mndDestroy(pMgmt->path);
+    return -1;
+  }
+
+  pMgmt->deployed = 1;
+  if (mmWriteFile(pMgmt) != 0) {
+    dError("failed to write mnode file since %s", terrstr());
+    pMgmt->deployed = 0;
+    mmStopWorker(pMgmt);
+    mndClose(pMnode);
+    mndDestroy(pMgmt->path);
+    return -1;
+  }
+
+  taosWLockLatch(&pMgmt->latch);
+  pMgmt->pMnode = pMnode;
+  pMgmt->deployed = 1;
+  taosWUnLockLatch(&pMgmt->latch);
+
+  dInfo("mnode open successfully");
+  return 0;
+}
+
+int32_t mmAlter(SMnodeMgmt *pMgmt, SMnodeOpt *pOption) {
+  SMnode *pMnode = mmAcquire(pMgmt);
+  if (pMnode == NULL) {
+    dError("failed to alter mnode since %s", terrstr());
+    return -1;
+  }
+
+  if (mndAlter(pMnode, pOption) != 0) {
+    dError("failed to alter mnode since %s", terrstr());
+    mmRelease(pMgmt, pMnode);
+    return -1;
+  }
+
+  mmRelease(pMgmt, pMnode);
+  return 0;
+}
+
+int32_t mmDrop(SMnodeMgmt *pMgmt) {
+  SMnode *pMnode = mmAcquire(pMgmt);
+  if (pMnode == NULL) {
+    dError("failed to drop mnode since %s", terrstr());
+    return -1;
+  }
+
+  taosRLockLatch(&pMgmt->latch);
+  pMgmt->dropped = 1;
+  taosRUnLockLatch(&pMgmt->latch);
+
+  if (mmWriteFile(pMgmt) != 0) {
+    taosRLockLatch(&pMgmt->latch);
+    pMgmt->dropped = 0;
+    taosRUnLockLatch(&pMgmt->latch);
+
+    mmRelease(pMgmt, pMnode);
+    dError("failed to drop mnode since %s", terrstr());
+    return -1;
+  }
+
+  mmRelease(pMgmt, pMnode);
+  mmStopWorker(pMgmt);
+  pMgmt->deployed = 0;
+  mmWriteFile(pMgmt);
+  mndClose(pMnode);
+  pMgmt->pMnode = NULL;
+  mndDestroy(pMgmt->path);
+
+  return 0;
+}
+
+static void mmInitOption(SMnodeMgmt *pMgmt, SMnodeOpt *pOption) {
+  SDnode *pDnode = pMgmt->pDnode;
+
+  pOption->pDnode = pDnode;
+  pOption->sendReqToDnodeFp = dndSendReqToDnode;
+  pOption->sendReqToMnodeFp = dndSendReqToMnode;
+  pOption->sendRedirectRspFp = dmSendRedirectRsp;
+  pOption->putReqToMWriteQFp = mmPutMsgToWriteQueue;
+  pOption->putReqToMReadQFp = mmPutMsgToReadQueue;
+  pOption->dnodeId = dmGetDnodeId(pDnode);
+  pOption->clusterId = dmGetClusterId(pDnode);
+}
+
+static void mmBuildOptionForDeploy(SMnodeMgmt *pMgmt, SMnodeOpt *pOption) {
+  SDnode *pDnode = pMgmt->pDnode;
+
+  mmInitOption(pMgmt, pOption);
+  pOption->replica = 1;
+  pOption->selfIndex = 0;
+  SReplica *pReplica = &pOption->replicas[0];
+  pReplica->id = 1;
+  pReplica->port = pDnode->cfg.serverPort;
+  memcpy(pReplica->fqdn, pDnode->cfg.localFqdn, TSDB_FQDN_LEN);
+
+  pMgmt->selfIndex = pOption->selfIndex;
+  pMgmt->replica = pOption->replica;
+  memcpy(&pMgmt->replicas, pOption->replicas, sizeof(SReplica) * TSDB_MAX_REPLICA);
+}
+
+static void mmBuildOptionForOpen(SMnodeMgmt *pMgmt, SMnodeOpt *pOption) {
+  mmInitOption(pMgmt, pOption);
+  pOption->selfIndex = pMgmt->selfIndex;
+  pOption->replica = pMgmt->replica;
+  memcpy(&pOption->replicas, pMgmt->replicas, sizeof(SReplica) * TSDB_MAX_REPLICA);
+}
+
+int32_t mmBuildOptionFromReq(SMnodeMgmt *pMgmt, SMnodeOpt *pOption, SDCreateMnodeReq *pCreate) {
+  SDnode *pDnode = pMgmt->pDnode;
+
+  mmInitOption(pMgmt, pOption);
+  pOption->dnodeId = dmGetDnodeId(pDnode);
+  pOption->clusterId = dmGetClusterId(pDnode);
+
+  pOption->replica = pCreate->replica;
+  pOption->selfIndex = -1;
+  for (int32_t i = 0; i < pCreate->replica; ++i) {
+    SReplica *pReplica = &pOption->replicas[i];
+    pReplica->id = pCreate->replicas[i].id;
+    pReplica->port = pCreate->replicas[i].port;
+    memcpy(pReplica->fqdn, pCreate->replicas[i].fqdn, TSDB_FQDN_LEN);
+    if (pReplica->id == pOption->dnodeId) {
+      pOption->selfIndex = i;
+    }
+  }
+
+  if (pOption->selfIndex == -1) {
+    dError("failed to build mnode options since %s", terrstr());
+    return -1;
+  }
+
+  pMgmt->selfIndex = pOption->selfIndex;
+  pMgmt->replica = pOption->replica;
+  memcpy(&pMgmt->replicas, pOption->replicas, sizeof(SReplica) * TSDB_MAX_REPLICA);
+  return 0;
+}
+
+static void mmCleanup(SMgmtWrapper *pWrapper) {
+  SMnodeMgmt *pMgmt = pWrapper->pMgmt;
+  if (pMgmt == NULL) return;
+
+  dInfo("mnode-mgmt start to clean up");
+  if (pMgmt->pMnode) {
+    mmStopWorker(pMgmt);
+    mndClose(pMgmt->pMnode);
+    pMgmt->pMnode = NULL;
+  }
+  free(pMgmt);
+  pWrapper->pMgmt = NULL;
+  dInfo("mnode-mgmt is cleaned up");
+}
+
+static int32_t mmInit(SMgmtWrapper *pWrapper) {
+  SDnode     *pDnode = pWrapper->pDnode;
+  SMnodeMgmt *pMgmt = calloc(1, sizeof(SMnodeMgmt));
+  int32_t     code = -1;
+  SMnodeOpt   option = {0};
+
+  dInfo("mnode-mgmt is initialized");
+  pMgmt->path = pWrapper->path;
+  pMgmt->pDnode = pWrapper->pDnode;
+  taosInitRWLatch(&pMgmt->latch);
+
+  if (mmReadFile(pMgmt) != 0) {
+    dError("failed to read file since %s", terrstr());
+    goto _OVER;
+  }
+
+  if (!pMgmt->deployed) {
+    dInfo("mnode start to deploy");
+    mmBuildOptionForDeploy(pMgmt, &option);
+    code = mmOpen(pMgmt, &option);
+  } else {
+    dInfo("mnode start to open");
+    mmBuildOptionForOpen(pMgmt, &option);
+    code = mmOpen(pMgmt, &option);
+  }
+
+_OVER:
+  if (code == 0) {
+    dInfo("mnode-mgmt is initialized");
+  } else {
+    dError("failed to init mnode-mgmtsince %s", terrstr());
+    mmCleanup(pWrapper);
+  }
+
+  return code;
+}
+
+static bool mmDeployRequired(SDnode *pDnode) {
+  if (dmGetDnodeId(pDnode) > 0) {
+    return false;
+  }
+
+  if (dmGetClusterId(pDnode) > 0) {
+    return false;
+  }
+
+  if (strcmp(pDnode->cfg.localEp, pDnode->cfg.firstEp) != 0) {
+    return false;
+  }
+
+  return true;
+}
+
+static bool mmRequire(SMgmtWrapper *pWrapper) {
+  SMnodeMgmt mgmt = {0};
+  mgmt.path = pWrapper->path;
+  if (mmReadFile(&mgmt) != 0) {
+    return false;
+  }
+
+  if (mgmt.dropped) {
+    dInfo("mnode has been dropped and needs to be deleted");
+    mndDestroy(mgmt.path);
+    return false;
+  }
+
+  if (mgmt.deployed) {
+    dInfo("mnode has been deployed");
+    return true;
+  }
+
+  bool required = mmDeployRequired(pWrapper->pDnode);
+  if (required) {
+    dInfo("mnode need to be deployed");
+  }
+
+  return required;
+}
 
 void mmGetMgmtFp(SMgmtWrapper *pWrapper) {
   SMgmtFp mgmtFp = {0};
-  mgmtFp.openFp = NULL;
+  mgmtFp.openFp = mmInit;
   mgmtFp.closeFp = NULL;
-  mgmtFp.requiredFp = mmRequireNode;
+  mgmtFp.requiredFp = mmRequire;
 
   mmInitMsgHandles(pWrapper);
   pWrapper->name = "mnode";
@@ -31,10 +302,29 @@ void mmGetMgmtFp(SMgmtWrapper *pWrapper) {
 }
 
 int32_t mmGetUserAuth(SMgmtWrapper *pWrapper, char *user, char *spi, char *encrypt, char *secret, char *ckey) {
-  return 0;
+  SMnodeMgmt *pMgmt = pWrapper->pMgmt;
+
+  SMnode *pMnode = mmAcquire(pMgmt);
+  if (pMnode == NULL) {
+    terrno = TSDB_CODE_APP_NOT_READY;
+    dTrace("failed to get user auth since %s", terrstr());
+    return -1;
+  }
+
+  int32_t code = mndRetriveAuth(pMnode, user, spi, encrypt, secret, ckey);
+  mmRelease(pMgmt, pMnode);
+
+  dTrace("user:%s, retrieve auth spi:%d encrypt:%d", user, *spi, *encrypt);
+  return code;
 }
 
-int32_t mmGetMonitorInfo(SDnode *pDnode, SMonClusterInfo *pClusterInfo, SMonVgroupInfo *pVgroupInfo,
+int32_t mmGetMonitorInfo(SMgmtWrapper *pWrapper, SMonClusterInfo *pClusterInfo, SMonVgroupInfo *pVgroupInfo,
                          SMonGrantInfo *pGrantInfo) {
-  return 0;
+  SMnodeMgmt *pMgmt = pWrapper->pMgmt;
+  SMnode     *pMnode = mmAcquire(pMgmt);
+  if (pMnode == NULL) return -1;
+
+  int32_t code = mndGetMonitorInfo(pMnode, pClusterInfo, pVgroupInfo, pGrantInfo);
+  mmRelease(pMgmt, pMnode);
+  return code;
 }
\ No newline at end of file
diff --git a/source/dnode/mgmt/mnode/src/mmMgmt.c b/source/dnode/mgmt/mnode/src/mmMgmt.c
deleted file mode 100644
index c7e3f39814..0000000000
--- a/source/dnode/mgmt/mnode/src/mmMgmt.c
+++ /dev/null
@@ -1,321 +0,0 @@
-/*
- * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
- *
- * This program is free software: you can use, redistribute, and/or modify
- * it under the terms of the GNU Affero General Public License, version 3
- * or later ("AGPL"), as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.
- *
- * You should have received a copy of the GNU Affero General Public License
- * along with this program. If not, see <http:www.gnu.org/licenses/>.
- */
-
-#define _DEFAULT_SOURCE
-#include "mmInt.h"
-
-#include "dmInt.h"
-#include "dndTransport.h"
-
-#if 0
-static void    mmInitOption(SDnode *pDnode, SMnodeOpt *pOption);
-static void    mmBuildOptionForDeploy(SDnode *pDnode, SMnodeOpt *pOption);
-static void    mmBuildOptionForOpen(SDnode *pDnode, SMnodeOpt *pOption);
-static bool    mmDeployRequired(SDnode *pDnode);
-static int32_t mmOpenImp(SDnode *pDnode, SMnodeOpt *pOption);
-
-int32_t mmInit(SDnode *pDnode) {
-  dInfo("mnode mgmt start to init");
-  int32_t code = -1;
-
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-  taosInitRWLatch(&pMgmt->latch);
-  mmInitMsgFp(pMgmt);
-
-  if (mmReadFile(pDnode) != 0) {
-    goto _OVER;
-  }
-
-  if (pMgmt->dropped) {
-    dInfo("mnode has been dropped and needs to be deleted");
-    mndDestroy(pDnode->dir.mnode);
-    code = 0;
-    goto _OVER;
-  }
-
-  if (!pMgmt->deployed) {
-    bool required = mmDeployRequired(pDnode);
-    if (!required) {
-      dInfo("mnode does not need to be deployed");
-      code = 0;
-      goto _OVER;
-    }
-
-    dInfo("mnode start to deploy");
-    SMnodeOpt option = {0};
-    mmBuildOptionForDeploy(pDnode, &option);
-    code = mmOpen(pDnode, &option);
-  } else {
-    dInfo("mnode start to open");
-    SMnodeOpt option = {0};
-    mmBuildOptionForOpen(pDnode, &option);
-    code = mmOpen(pDnode, &option);
-  }
-
-_OVER:
-  if (code == 0) {
-    dInfo("mnode mgmt init success");
-  } else {
-    dError("failed to init mnode mgmt since %s", terrstr());
-    mmCleanup(pDnode);
-  }
-
-  return code;
-}
-
-void mmCleanup(SDnode *pDnode) {
-  dInfo("mnode mgmt start to clean up");
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-  if (pMgmt->pMnode) {
-    mmStopWorker(pDnode);
-    mndClose(pMgmt->pMnode);
-    pMgmt->pMnode = NULL;
-  }
-  dInfo("mnode mgmt is cleaned up");
-}
-
-SMnode *mmAcquire(SDnode *pDnode) {
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-  SMnode     *pMnode = NULL;
-  int32_t     refCount = 0;
-
-  taosRLockLatch(&pMgmt->latch);
-  if (pMgmt->deployed && !pMgmt->dropped) {
-    refCount = atomic_add_fetch_32(&pMgmt->refCount, 1);
-    pMnode = pMgmt->pMnode;
-  } else {
-    terrno = TSDB_CODE_DND_MNODE_NOT_DEPLOYED;
-  }
-  taosRUnLockLatch(&pMgmt->latch);
-
-  if (pMnode != NULL) {
-    dTrace("acquire mnode, refCount:%d", refCount);
-  }
-  return pMnode;
-}
-
-void mmRelease(SDnode *pDnode, SMnode *pMnode) {
-  if (pMnode == NULL) return;
-
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-  taosRLockLatch(&pMgmt->latch);
-  int32_t refCount = atomic_sub_fetch_32(&pMgmt->refCount, 1);
-  taosRUnLockLatch(&pMgmt->latch);
-  dTrace("release mnode, refCount:%d", refCount);
-}
-
-int32_t mmOpen(SDnode *pDnode, SMnodeOpt *pOption) {
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-  pMgmt->singleProc = true;
-
-  int32_t code = mmOpenImp(pDnode, pOption);
-
-  if (code == 0 && !pMgmt->singleProc) {
-    SProcCfg cfg = {.childQueueSize = 1024 * 1024,
-                    .childConsumeFp = (ProcConsumeFp)mmConsumeChildQueue,
-                    .childMallocHeadFp = (ProcMallocFp)taosAllocateQitem,
-                    .childFreeHeadFp = (ProcFreeFp)taosFreeQitem,
-                    .childMallocBodyFp = (ProcMallocFp)rpcMallocCont,
-                    .childFreeBodyFp = (ProcFreeFp)rpcFreeCont,
-                    .parentQueueSize = 1024 * 1024,
-                    .parentConsumeFp = (ProcConsumeFp)mmConsumeParentQueue,
-                    .parentdMallocHeadFp = (ProcMallocFp)malloc,
-                    .parentFreeHeadFp = (ProcFreeFp)free,
-                    .parentMallocBodyFp = (ProcMallocFp)rpcMallocCont,
-                    .parentFreeBodyFp = (ProcFreeFp)rpcFreeCont,
-                    .testFlag = 0,
-                    .pParent = pDnode,
-                    .name = "mnode"};
-
-    pMgmt->pProcess = taosProcInit(&cfg);
-    if (pMgmt->pProcess == NULL) {
-      return -1;
-    }
-
-    return taosProcRun(pMgmt->pProcess);
-  }
-
-  return code;
-}
-
-int32_t mmAlter(SDnode *pDnode, SMnodeOpt *pOption) {
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-
-  SMnode *pMnode = mmAcquire(pDnode);
-  if (pMnode == NULL) {
-    dError("failed to alter mnode since %s", terrstr());
-    return -1;
-  }
-
-  if (mndAlter(pMnode, pOption) != 0) {
-    dError("failed to alter mnode since %s", terrstr());
-    mmRelease(pDnode, pMnode);
-    return -1;
-  }
-
-  mmRelease(pDnode, pMnode);
-  return 0;
-}
-
-int32_t mmDrop(SDnode *pDnode) {
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-
-  SMnode *pMnode = mmAcquire(pDnode);
-  if (pMnode == NULL) {
-    dError("failed to drop mnode since %s", terrstr());
-    return -1;
-  }
-
-  taosRLockLatch(&pMgmt->latch);
-  pMgmt->dropped = 1;
-  taosRUnLockLatch(&pMgmt->latch);
-
-  if (mmWriteFile(pDnode) != 0) {
-    taosRLockLatch(&pMgmt->latch);
-    pMgmt->dropped = 0;
-    taosRUnLockLatch(&pMgmt->latch);
-
-    mmRelease(pDnode, pMnode);
-    dError("failed to drop mnode since %s", terrstr());
-    return -1;
-  }
-
-  mmRelease(pDnode, pMnode);
-  mmStopWorker(pDnode);
-  pMgmt->deployed = 0;
-  mmWriteFile(pDnode);
-  mndClose(pMnode);
-  pMgmt->pMnode = NULL;
-  mndDestroy(pDnode->dir.mnode);
-
-  return 0;
-}
-
-static bool mmDeployRequired(SDnode *pDnode) {
-  if (dmGetDnodeId(pDnode) > 0) {
-    return false;
-  }
-
-  if (dmGetClusterId(pDnode) > 0) {
-    return false;
-  }
-
-  if (strcmp(pDnode->cfg.localEp, pDnode->cfg.firstEp) != 0) {
-    return false;
-  }
-
-  return true;
-}
-
-static void mmInitOption(SDnode *pDnode, SMnodeOpt *pOption) {
-  pOption->pDnode = pDnode;
-  pOption->sendReqToDnodeFp = dndSendReqToDnode;
-  pOption->sendReqToMnodeFp = dndSendReqToMnode;
-  pOption->sendRedirectRspFp = dmSendRedirectRsp;
-  pOption->putReqToMWriteQFp = mmPutMsgToWriteQueue;
-  pOption->putReqToMReadQFp = mmPutMsgToReadQueue;
-  pOption->dnodeId = dmGetDnodeId(pDnode);
-  pOption->clusterId = dmGetClusterId(pDnode);
-}
-
-static void mmBuildOptionForDeploy(SDnode *pDnode, SMnodeOpt *pOption) {
-  mmInitOption(pDnode, pOption);
-  pOption->replica = 1;
-  pOption->selfIndex = 0;
-  SReplica *pReplica = &pOption->replicas[0];
-  pReplica->id = 1;
-  pReplica->port = pDnode->cfg.serverPort;
-  memcpy(pReplica->fqdn, pDnode->cfg.localFqdn, TSDB_FQDN_LEN);
-
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-  pMgmt->selfIndex = pOption->selfIndex;
-  pMgmt->replica = pOption->replica;
-  memcpy(&pMgmt->replicas, pOption->replicas, sizeof(SReplica) * TSDB_MAX_REPLICA);
-}
-
-static void mmBuildOptionForOpen(SDnode *pDnode, SMnodeOpt *pOption) {
-  mmInitOption(pDnode, pOption);
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-  pOption->selfIndex = pMgmt->selfIndex;
-  pOption->replica = pMgmt->replica;
-  memcpy(&pOption->replicas, pMgmt->replicas, sizeof(SReplica) * TSDB_MAX_REPLICA);
-}
-
-int32_t mmBuildOptionFromReq(SDnode *pDnode, SMnodeOpt *pOption, SDCreateMnodeReq *pCreate) {
-  mmInitOption(pDnode, pOption);
-  pOption->dnodeId = dmGetDnodeId(pDnode);
-  pOption->clusterId = dmGetClusterId(pDnode);
-
-  pOption->replica = pCreate->replica;
-  pOption->selfIndex = -1;
-  for (int32_t i = 0; i < pCreate->replica; ++i) {
-    SReplica *pReplica = &pOption->replicas[i];
-    pReplica->id = pCreate->replicas[i].id;
-    pReplica->port = pCreate->replicas[i].port;
-    memcpy(pReplica->fqdn, pCreate->replicas[i].fqdn, TSDB_FQDN_LEN);
-    if (pReplica->id == pOption->dnodeId) {
-      pOption->selfIndex = i;
-    }
-  }
-
-  if (pOption->selfIndex == -1) {
-    dError("failed to build mnode options since %s", terrstr());
-    return -1;
-  }
-
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-  pMgmt->selfIndex = pOption->selfIndex;
-  pMgmt->replica = pOption->replica;
-  memcpy(&pMgmt->replicas, pOption->replicas, sizeof(SReplica) * TSDB_MAX_REPLICA);
-  return 0;
-}
-
-static int32_t mmOpenImp(SDnode *pDnode, SMnodeOpt *pOption) {
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-
-  SMnode *pMnode = mndOpen(pDnode->dir.mnode, pOption);
-  if (pMnode == NULL) {
-    dError("failed to open mnode since %s", terrstr());
-    return -1;
-  }
-
-  if (mmStartWorker(pDnode) != 0) {
-    dError("failed to start mnode worker since %s", terrstr());
-    mndClose(pMnode);
-    mndDestroy(pDnode->dir.mnode);
-    return -1;
-  }
-
-  pMgmt->deployed = 1;
-  if (mmWriteFile(pDnode) != 0) {
-    dError("failed to write mnode file since %s", terrstr());
-    pMgmt->deployed = 0;
-    mmStopWorker(pDnode);
-    mndClose(pMnode);
-    mndDestroy(pDnode->dir.mnode);
-    return -1;
-  }
-
-  taosWLockLatch(&pMgmt->latch);
-  pMgmt->pMnode = pMnode;
-  pMgmt->deployed = 1;
-  taosWUnLockLatch(&pMgmt->latch);
-
-  dInfo("mnode open successfully");
-  return 0;
-}
-
-#endif
-
diff --git a/source/dnode/mgmt/mnode/src/mmMsg.c b/source/dnode/mgmt/mnode/src/mmMsg.c
index 389ebfe853..bbb49848cc 100644
--- a/source/dnode/mgmt/mnode/src/mmMsg.c
+++ b/source/dnode/mgmt/mnode/src/mmMsg.c
@@ -15,12 +15,13 @@
 
 #define _DEFAULT_SOURCE
 #include "mmMsg.h"
+#include "dmInt.h"
 #include "mmWorker.h"
 
-#if 0
-#include "dmInt.h"
-
 int32_t mmProcessCreateReq(SDnode *pDnode, SRpcMsg *pReq) {
+  SMgmtWrapper *pWrapper = dndGetWrapper(pDnode, MNODE);
+  SMnodeMgmt   *pMgmt = pWrapper->pMgmt;
+
   SDCreateMnodeReq createReq = {0};
   if (tDeserializeSDCreateMnodeReq(pReq->pCont, pReq->contLen, &createReq) != 0) {
     terrno = TSDB_CODE_INVALID_MSG;
@@ -34,25 +35,28 @@ int32_t mmProcessCreateReq(SDnode *pDnode, SRpcMsg *pReq) {
   }
 
   SMnodeOpt option = {0};
-  if (mmBuildOptionFromReq(pDnode, &option, &createReq) != 0) {
+  if (mmBuildOptionFromReq(pMgmt, &option, &createReq) != 0) {
     terrno = TSDB_CODE_DND_MNODE_INVALID_OPTION;
     dError("failed to create mnode since %s", terrstr());
     return -1;
   }
 
-  SMnode *pMnode = mmAcquire(pDnode);
+  SMnode *pMnode = mmAcquire(pMgmt);
   if (pMnode != NULL) {
-    mmRelease(pDnode, pMnode);
+    mmRelease(pMgmt, pMnode);
     terrno = TSDB_CODE_DND_MNODE_ALREADY_DEPLOYED;
     dError("failed to create mnode since %s", terrstr());
     return -1;
   }
 
   dDebug("start to create mnode");
-  return mmOpen(pDnode, &option);
+  return mmOpen(pMgmt, &option);
 }
 
 int32_t mmProcessAlterReq(SDnode *pDnode, SRpcMsg *pReq) {
+  SMgmtWrapper *pWrapper = dndGetWrapper(pDnode, MNODE);
+  SMnodeMgmt   *pMgmt = pWrapper->pMgmt;
+
   SDAlterMnodeReq alterReq = {0};
   if (tDeserializeSDCreateMnodeReq(pReq->pCont, pReq->contLen, &alterReq) != 0) {
     terrno = TSDB_CODE_INVALID_MSG;
@@ -66,13 +70,13 @@ int32_t mmProcessAlterReq(SDnode *pDnode, SRpcMsg *pReq) {
   }
 
   SMnodeOpt option = {0};
-  if (mmBuildOptionFromReq(pDnode, &option, &alterReq) != 0) {
+  if (mmBuildOptionFromReq(pMgmt, &option, &alterReq) != 0) {
     terrno = TSDB_CODE_DND_MNODE_INVALID_OPTION;
     dError("failed to alter mnode since %s", terrstr());
     return -1;
   }
 
-  SMnode *pMnode = mmAcquire(pDnode);
+  SMnode *pMnode = mmAcquire(pMgmt);
   if (pMnode == NULL) {
     terrno = TSDB_CODE_DND_MNODE_NOT_DEPLOYED;
     dError("failed to alter mnode since %s", terrstr());
@@ -80,13 +84,16 @@ int32_t mmProcessAlterReq(SDnode *pDnode, SRpcMsg *pReq) {
   }
 
   dDebug("start to alter mnode");
-  int32_t code = mmAlter(pDnode, &option);
-  mmRelease(pDnode, pMnode);
+  int32_t code = mmAlter(pMgmt, &option);
+  mmRelease(pMgmt, pMnode);
 
   return code;
 }
 
 int32_t mmProcessDropReq(SDnode *pDnode, SRpcMsg *pReq) {
+  SMgmtWrapper *pWrapper = dndGetWrapper(pDnode, MNODE);
+  SMnodeMgmt   *pMgmt = pWrapper->pMgmt;
+
   SDDropMnodeReq dropReq = {0};
   if (tDeserializeSMCreateDropMnodeReq(pReq->pCont, pReq->contLen, &dropReq) != 0) {
     terrno = TSDB_CODE_INVALID_MSG;
@@ -99,7 +106,7 @@ int32_t mmProcessDropReq(SDnode *pDnode, SRpcMsg *pReq) {
     return -1;
   }
 
-  SMnode *pMnode = mmAcquire(pDnode);
+  SMnode *pMnode = mmAcquire(pMgmt);
   if (pMnode == NULL) {
     terrno = TSDB_CODE_DND_MNODE_NOT_DEPLOYED;
     dError("failed to drop mnode since %s", terrstr());
@@ -107,45 +114,12 @@ int32_t mmProcessDropReq(SDnode *pDnode, SRpcMsg *pReq) {
   }
 
   dDebug("start to drop mnode");
-  int32_t code = mmDrop(pDnode);
-  mmRelease(pDnode, pMnode);
+  int32_t code = mmDrop(pMgmt);
+  mmRelease(pMgmt, pMnode);
 
   return code;
 }
 
-int32_t mmGetMonitorInfo(SDnode *pDnode, SMonClusterInfo *pClusterInfo, SMonVgroupInfo *pVgroupInfo,
-                         SMonGrantInfo *pGrantInfo) {
-  SMnode *pMnode = mmAcquire(pDnode);
-  if (pMnode == NULL) return -1;
-
-  int32_t code = mndGetMonitorInfo(pMnode, pClusterInfo, pVgroupInfo, pGrantInfo);
-  mmRelease(pDnode, pMnode);
-  return code;
-}
-
-int32_t mmGetUserAuth(SDnode *pDnode, char *user, char *spi, char *encrypt, char *secret, char *ckey) {
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-
-  SMnode *pMnode = mmAcquire(pDnode);
-  if (pMnode == NULL) {
-    terrno = TSDB_CODE_APP_NOT_READY;
-    dTrace("failed to get user auth since %s", terrstr());
-    return -1;
-  }
-
-  int32_t code = mndRetriveAuth(pMnode, user, spi, encrypt, secret, ckey);
-  mmRelease(pDnode, pMnode);
-
-  dTrace("user:%s, retrieve auth spi:%d encrypt:%d", user, *spi, *encrypt);
-  return code;
-}
-
-#endif
-
-int32_t mmProcessCreateReq(SDnode *pDnode, SRpcMsg *pRpcMsg) {return 0;}
-int32_t mmProcessAlterReq(SDnode *pDnode, SRpcMsg *pRpcMsg) {return 0;}
-int32_t mmProcessDropReq(SDnode *pDnode, SRpcMsg *pRpcMsg) {return 0;}
-
 void mmInitMsgHandles(SMgmtWrapper *pWrapper) {
   // Requests handled by DNODE
   dndSetMsgHandle(pWrapper, TDMT_DND_CREATE_MNODE_RSP, mmProcessWriteMsg);
diff --git a/source/dnode/mgmt/mnode/src/mmWorker.c b/source/dnode/mgmt/mnode/src/mmWorker.c
index 3ca112782f..de8809ab17 100644
--- a/source/dnode/mgmt/mnode/src/mmWorker.c
+++ b/source/dnode/mgmt/mnode/src/mmWorker.c
@@ -20,16 +20,63 @@
 #include "dndTransport.h"
 #include "dndWorker.h"
 
-#if 0
-static int32_t mmProcessWriteMsg(SDnode *pDnode, SMndMsg *pMsg);
-static int32_t mmProcessSyncMsg(SDnode *pDnode, SMndMsg *pMsg);
-static int32_t mmProcessReadMsg(SDnode *pDnode, SMndMsg *pMsg);
-static int32_t mmPutMndMsgToWorker(SDnode *pDnode, SDnodeWorker *pWorker, SMndMsg *pMsg);
-static int32_t mmPutRpcMsgToWorker(SDnode *pDnode, SDnodeWorker *pWorker, SRpcMsg *pRpc);
-static void    mmConsumeMsgQueue(SDnode *pDnode, SMndMsg *pMsg);
+static void mmSendRpcRsp(SMnodeMgmt *pMgmt, SRpcMsg *pRpc) {
+  if (pRpc->code == TSDB_CODE_DND_MNODE_NOT_DEPLOYED || pRpc->code == TSDB_CODE_APP_NOT_READY) {
+    dmSendRedirectRsp(pMgmt->pDnode, pRpc);
+  } else {
+    rpcSendResponse(pRpc);
+  }
+}
 
-int32_t mmStartWorker(SDnode *pDnode) {
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
+void mmPutRpcRspToWorker(SMnodeMgmt *pMgmt, SRpcMsg *pRpc) {
+  int32_t code = -1;
+
+  if (pMgmt->singleProc) {
+    mmSendRpcRsp(pMgmt, pRpc);
+  } else {
+    do {
+      code = taosProcPutToParentQueue(pMgmt->pProcess, pRpc, sizeof(SRpcMsg), pRpc->pCont, pRpc->contLen);
+      if (code != 0) {
+        taosMsleep(10);
+      }
+    } while (code != 0);
+  }
+}
+
+static void mmConsumeMsgQueue(SMnodeMgmt *pMgmt, SNodeMsg *pMsg) {
+  dTrace("msg:%p, will be processed", pMsg);
+
+  SMnode  *pMnode = mmAcquire(pMgmt);
+  SRpcMsg *pRpc = &pMsg->rpcMsg;
+  bool     isReq = (pRpc->msgType & 1U);
+  int32_t  code = -1;
+
+  if (pMnode != NULL) {
+    pMsg->pNode = pMnode;
+    code = mndProcessMsg((SMndMsg*)pMsg);
+    mmRelease(pMgmt, pMnode);
+  }
+
+  if (isReq) {
+    if (pMsg->rpcMsg.handle == NULL) return;
+    if (code == 0) {
+      SRpcMsg rsp = {.handle = pRpc->handle, .contLen = pMsg->rspLen, .pCont = pMsg->pRsp};
+      mmPutRpcRspToWorker(pMgmt, &rsp);
+    } else {
+      if (terrno != TSDB_CODE_MND_ACTION_IN_PROGRESS) {
+        SRpcMsg rsp = {.handle = pRpc->handle, .contLen = pMsg->rspLen, .pCont = pMsg->pRsp, .code = terrno};
+        mmPutRpcRspToWorker(pMgmt, &rsp);
+      }
+    }
+  }
+
+  dTrace("msg:%p, is freed", pMsg);
+  rpcFreeCont(pRpc->pCont);
+  taosFreeQitem(pMsg);
+}
+
+int32_t mmStartWorker(SMnodeMgmt *pMgmt) {
+  SDnode *pDnode = pMgmt->pDnode;
   if (dndInitWorker(pDnode, &pMgmt->readWorker, DND_WORKER_SINGLE, "mnode-read", 0, 1, mmConsumeMsgQueue) != 0) {
     dError("failed to start mnode read worker since %s", terrstr());
     return -1;
@@ -48,9 +95,7 @@ int32_t mmStartWorker(SDnode *pDnode) {
   return 0;
 }
 
-void mmStopWorker(SDnode *pDnode) {
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-
+void mmStopWorker(SMnodeMgmt *pMgmt) {
   taosWLockLatch(&pMgmt->latch);
   pMgmt->deployed = 0;
   taosWUnLockLatch(&pMgmt->latch);
@@ -64,122 +109,41 @@ void mmStopWorker(SDnode *pDnode) {
   dndCleanupWorker(&pMgmt->syncWorker);
 }
 
-void mmInitMsgFp(SMnodeMgmt *pMgmt) {
-  
-}
-
-static void mmSendRpcRsp(SDnode *pDnode, SRpcMsg *pRpc) {
-  if (pRpc->code == TSDB_CODE_DND_MNODE_NOT_DEPLOYED || pRpc->code == TSDB_CODE_APP_NOT_READY) {
-    dmSendRedirectRsp(pDnode, pRpc);
-  } else {
-    rpcSendResponse(pRpc);
-  }
-}
-
-static int32_t mmBuildMsg(SMndMsg *pMsg, SRpcMsg *pRpc) {
-  SRpcConnInfo connInfo = {0};
-  if ((pRpc->msgType & 1U) && rpcGetConnInfo(pRpc->handle, &connInfo) != 0) {
-    terrno = TSDB_CODE_MND_NO_USER_FROM_CONN;
-    dError("failed to create msg since %s, app:%p RPC:%p", terrstr(), pRpc->ahandle, pRpc->handle);
-    return -1;
-  }
-
-  memcpy(pMsg->user, connInfo.user, TSDB_USER_LEN);
-  pMsg->rpcMsg = *pRpc;
-  pMsg->createdTime = taosGetTimestampSec();
-
-  return 0;
-}
-
-void mmProcessRpcMsg(SDnode *pDnode, SRpcMsg *pRpc, SEpSet *pEpSet) {
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-  int32_t   code = -1;
-  SMndMsg  *pMsg = NULL;
-
-  MndMsgFp msgFp = pMgmt->msgFp[TMSG_INDEX(pRpc->msgType)];
-  if (msgFp == NULL) {
-    terrno = TSDB_CODE_MSG_NOT_PROCESSED;
-    goto _OVER;
-  }
-
-  pMsg = taosAllocateQitem(sizeof(SMndMsg));
-  if (pMsg == NULL) {
-    goto _OVER;
-  }
-
-  if (mmBuildMsg(pMsg, pRpc) != 0) {
-    goto _OVER;
-  }
-
-  dTrace("msg:%p, is created, app:%p RPC:%p user:%s", pMsg, pRpc->ahandle, pRpc->handle, pMsg->user);
-
-  if (pMgmt->singleProc) {
-    code = (*msgFp)(pDnode, pMsg);
-  } else {
-    code = taosProcPutToChildQueue(pMgmt->pProcess, pMsg, sizeof(SMndMsg), pRpc->pCont, pRpc->contLen);
-  }
-
-_OVER:
-
-  if (code == 0) {
-    if (!pMgmt->singleProc) {
-      dTrace("msg:%p, is freed", pMsg);
-      taosFreeQitem(pMsg);
-      rpcFreeCont(pRpc->pCont);
-    }
-  } else {
-    bool isReq = (pRpc->msgType & 1U);
-    if (isReq) {
-      SRpcMsg rsp = {.handle = pRpc->handle, .ahandle = pRpc->ahandle, .code = terrno};
-      mmSendRpcRsp(pDnode, &rsp);
-    }
-    dTrace("msg:%p, is freed", pMsg);
-    taosFreeQitem(pMsg);
-    rpcFreeCont(pRpc->pCont);
-  }
-}
-
-int32_t mmProcessWriteMsg(SDnode *pDnode, SMndMsg *pMsg) {
-  return mmPutMndMsgToWorker(pDnode, &pDnode->mmgmt.writeWorker, pMsg);
-}
-
-int32_t mmProcessSyncMsg(SDnode *pDnode, SMndMsg *pMsg) {
-  return mmPutMndMsgToWorker(pDnode, &pDnode->mmgmt.syncWorker, pMsg);
-}
-
-int32_t mmProcessReadMsg(SDnode *pDnode, SMndMsg *pMsg) {
-  return mmPutMndMsgToWorker(pDnode, &pDnode->mmgmt.readWorker, pMsg);
-}
-
-int32_t mmPutMsgToWriteQueue(SDnode *pDnode, SRpcMsg *pRpc) {
-  return mmPutRpcMsgToWorker(pDnode, &pDnode->mmgmt.writeWorker, pRpc);
-}
-
-int32_t mmPutMsgToReadQueue(SDnode *pDnode, SRpcMsg *pRpc) {
-  return mmPutRpcMsgToWorker(pDnode, &pDnode->mmgmt.readWorker, pRpc);
-}
-
-static int32_t mmPutMndMsgToWorker(SDnode *pDnode, SDnodeWorker *pWorker, SMndMsg *pMsg) {
-  SMnode *pMnode = mmAcquire(pDnode);
+static int32_t mmPutMsgToWorker(SMnodeMgmt *pMgmt, SDnodeWorker *pWorker, SNodeMsg *pMsg) {
+  SMnode *pMnode = mmAcquire(pMgmt);
   if (pMnode == NULL) return -1;
 
   dTrace("msg:%p, put into worker %s", pMsg, pWorker->name);
   int32_t code = dndWriteMsgToWorker(pWorker, pMsg, 0);
-  mmRelease(pDnode, pMnode);
+  mmRelease(pMgmt, pMnode);
   return code;
 }
 
-static int32_t mmPutRpcMsgToWorker(SDnode *pDnode, SDnodeWorker *pWorker, SRpcMsg *pRpc) {
-  SMndMsg *pMsg = taosAllocateQitem(sizeof(SMndMsg));
+int32_t mmProcessWriteMsg(SMgmtWrapper *pWrapper, SNodeMsg *pMsg) {
+  SMnodeMgmt *pMgmt = pWrapper->pMgmt;
+  return mmPutMsgToWorker(pMgmt, &pMgmt->writeWorker, pMsg);
+}
+
+int32_t mmProcessSyncMsg(SMgmtWrapper *pWrapper, SNodeMsg *pMsg) {
+  SMnodeMgmt *pMgmt = pWrapper->pMgmt;
+  return mmPutMsgToWorker(pMgmt, &pMgmt->syncWorker, pMsg);
+}
+
+int32_t mmProcessReadMsg(SMgmtWrapper *pWrapper, SNodeMsg *pMsg) {
+  SMnodeMgmt *pMgmt = pWrapper->pMgmt;
+  return mmPutMsgToWorker(pMgmt, &pMgmt->readWorker, pMsg);
+}
+
+static int32_t mmPutRpcMsgToWorker(SMgmtWrapper *pWrapper, SDnodeWorker *pWorker, SRpcMsg *pRpc) {
+  SNodeMsg *pMsg = taosAllocateQitem(sizeof(SNodeMsg));
   if (pMsg == NULL) {
     return -1;
   }
 
   dTrace("msg:%p, is created", pMsg);
   pMsg->rpcMsg = *pRpc;
-  pMsg->createdTime = taosGetTimestampSec();
 
-  int32_t code = mmPutMndMsgToWorker(pDnode, pWorker, pMsg);
+  int32_t code = mmPutMsgToWorker(pWrapper->pMgmt, pWorker, pMsg);
   if (code != 0) {
     dTrace("msg:%p, is freed", pMsg);
     taosFreeQitem(pMsg);
@@ -189,88 +153,14 @@ static int32_t mmPutRpcMsgToWorker(SDnode *pDnode, SDnodeWorker *pWorker, SRpcMs
   return code;
 }
 
-void mmPutRpcRspToWorker(SDnode *pDnode, SRpcMsg *pRpc) {
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-  int32_t   code = -1;
-
-  if (pMgmt->singleProc) {
-    mmSendRpcRsp(pDnode, pRpc);
-  } else {
-    do {
-      code = taosProcPutToParentQueue(pMgmt->pProcess, pRpc, sizeof(SRpcMsg), pRpc->pCont, pRpc->contLen);
-      if (code != 0) {
-        taosMsleep(10);
-      }
-    } while (code != 0);
-  }
+int32_t mmPutMsgToWriteQueue(SDnode *pDnode, SRpcMsg *pRpc) {
+  SMgmtWrapper *pWrapper = dndGetWrapper(pDnode, MNODE);
+  SMnodeMgmt   *pMgmt = pWrapper->pMgmt;
+  return mmPutRpcMsgToWorker(pWrapper, &pMgmt->writeWorker, pRpc);
 }
 
-void mmConsumeChildQueue(SDnode *pDnode, SMndMsg *pMsg, int32_t msgLen, void *pCont, int32_t contLen) {
-  dTrace("msg:%p, get from child queue", pMsg);
-  SMnodeMgmt *pMgmt = &pDnode->mmgmt;
-
-  SRpcMsg *pRpc = &pMsg->rpcMsg;
-  pRpc->pCont = pCont;
-
-  MndMsgFp msgFp = pMgmt->msgFp[TMSG_INDEX(pRpc->msgType)];
-  int32_t  code = (*msgFp)(pDnode, pMsg);
-
-  if (code != 0) {
-    bool isReq = (pRpc->msgType & 1U);
-    if (isReq) {
-      SRpcMsg rsp = {.handle = pRpc->handle, .ahandle = pRpc->ahandle, .code = terrno};
-      mmPutRpcRspToWorker(pDnode, &rsp);
-    }
-
-    dTrace("msg:%p, is freed", pMsg);
-    taosFreeQitem(pMsg);
-    rpcFreeCont(pCont);
-  }
+int32_t mmPutMsgToReadQueue(SDnode *pDnode, SRpcMsg *pRpc) {
+  SMgmtWrapper *pWrapper = dndGetWrapper(pDnode, MNODE);
+  SMnodeMgmt   *pMgmt = pWrapper->pMgmt;
+  return mmPutRpcMsgToWorker(pWrapper, &pMgmt->readWorker, pRpc);
 }
-
-void mmConsumeParentQueue(SDnode *pDnode, SRpcMsg *pMsg, int32_t msgLen, void *pCont, int32_t contLen) {
-  dTrace("msg:%p, get from parent queue", pMsg);
-  pMsg->pCont = pCont;
-  mmSendRpcRsp(pDnode, pMsg);
-  free(pMsg);
-}
-
-static void mmConsumeMsgQueue(SDnode *pDnode, SMndMsg *pMsg) {
-  dTrace("msg:%p, get from msg queue", pMsg);
-  SMnode  *pMnode = mmAcquire(pDnode);
-  SRpcMsg *pRpc = &pMsg->rpcMsg;
-  bool     isReq = (pRpc->msgType & 1U);
-  int32_t  code = -1;
-
-  if (pMnode != NULL) {
-    pMsg->pMnode = pMnode;
-    code = mndProcessMsg(pMsg);
-    mmRelease(pDnode, pMnode);
-  }
-
-  if (isReq) {
-    if (pMsg->rpcMsg.handle == NULL) return;
-    if (code == 0) {
-      SRpcMsg rsp = {.handle = pRpc->handle, .contLen = pMsg->contLen, .pCont = pMsg->pCont};
-      mmPutRpcRspToWorker(pDnode, &rsp);
-    } else {
-      if (terrno != TSDB_CODE_MND_ACTION_IN_PROGRESS) {
-        SRpcMsg rsp = {.handle = pRpc->handle, .contLen = pMsg->contLen, .pCont = pMsg->pCont, .code = terrno};
-        mmPutRpcRspToWorker(pDnode, &rsp);
-      }
-    }
-  }
-
-  dTrace("msg:%p, is freed", pMsg);
-  rpcFreeCont(pRpc->pCont);
-  taosFreeQitem(pMsg);
-}
-
-#endif
-
-int32_t mmProcessWriteMsg( SMgmtWrapper *pWrapper, SNodeMsg *pMsg) {return 0;}
-int32_t mmProcessSyncMsg( SMgmtWrapper *pWrapper, SNodeMsg *pMsg) {return 0;}
-int32_t mmProcessReadMsg(SMgmtWrapper *pWrapper, SNodeMsg *pMsg) {
-  terrno = TSDB_CODE_MSG_NOT_PROCESSED;
-  return -1;
-}
\ No newline at end of file
diff --git a/source/util/src/tprocess.c b/source/util/src/tprocess.c
index 54cd894692..0286bdf972 100644
--- a/source/util/src/tprocess.c
+++ b/source/util/src/tprocess.c
@@ -348,13 +348,13 @@ SProcObj *taosProcInit(const SProcCfg *pCfg) {
 
   pProc->pChildQueue = taosProcQueueInit(pCfg->childQueueSize);
   pProc->pParentQueue = taosProcQueueInit(pCfg->parentQueueSize);
-    if (pProc->pChildQueue == NULL || pProc->pParentQueue == NULL) {
+  if (pProc->pChildQueue == NULL || pProc->pParentQueue == NULL) {
     taosProcQueueCleanup(pProc->pChildQueue);
     free(pProc);
     return NULL;
   }
 
-    pProc->pChildQueue->name = pCfg->name;
+  pProc->pChildQueue->name = pCfg->name;
   pProc->pChildQueue->pParent = pCfg->pParent;
   pProc->pChildQueue->mallocHeadFp = pCfg->childMallocHeadFp;
   pProc->pChildQueue->freeHeadFp = pCfg->childFreeHeadFp;
@@ -436,8 +436,7 @@ int32_t taosProcRun(SProcObj *pProc) {
 
 void taosProcStop(SProcObj *pProc) {
   pProc->stopFlag = true;
-  // todo
-  // join
+  // todo join
 }
 
 bool taosProcIsChild(SProcObj *pProc) { return pProc->isChild; }
@@ -445,6 +444,7 @@ bool taosProcIsChild(SProcObj *pProc) { return pProc->isChild; }
 void taosProcCleanup(SProcObj *pProc) {
   if (pProc != NULL) {
     uDebug("proc:%s, clean up", pProc->name);
+    taosProcStop(pProc);
     taosProcQueueCleanup(pProc->pChildQueue);
     taosProcQueueCleanup(pProc->pParentQueue);
     free(pProc);