diff --git a/include/dnode/mnode/mnode.h b/include/dnode/mnode/mnode.h
index 5b88a9d6aff34bf3a151650f42c48e4ef325e81d..08ab63e55abb4498e058c8e014387c1046a76621 100644
--- a/include/dnode/mnode/mnode.h
+++ b/include/dnode/mnode/mnode.h
@@ -29,8 +29,7 @@ extern "C" {
 typedef struct SMnode SMnode;
 
 typedef struct {
-  int32_t  dnodeId;
-  int64_t  clusterId;
+  bool     deploy;
   int8_t   replica;
   int8_t   selfIndex;
   SReplica replicas[TSDB_MAX_REPLICA];
diff --git a/source/dnode/mgmt/mm/src/mmFile.c b/source/dnode/mgmt/mm/src/mmFile.c
index e5cc0ce0871bfa398bd7c46f4c9260bb16e8f0de..76aba771cbd08271a8995f6d4a79ebb7b4ba235d 100644
--- a/source/dnode/mgmt/mm/src/mmFile.c
+++ b/source/dnode/mgmt/mm/src/mmFile.c
@@ -111,7 +111,7 @@ int32_t mmWriteFile(SMnodeMgmt *pMgmt, bool deployed) {
 
   TdFilePtr pFile = taosOpenFile(file, TD_FILE_CTEATE | TD_FILE_WRITE | TD_FILE_TRUNC);
   if (pFile == NULL) {
-    terrno = TAOS_SYSTEM_ERROR(errno);;
+    terrno = TAOS_SYSTEM_ERROR(errno);
     dError("failed to write %s since %s", file, terrstr());
     return -1;
   }
@@ -145,7 +145,7 @@ int32_t mmWriteFile(SMnodeMgmt *pMgmt, bool deployed) {
   snprintf(realfile, sizeof(realfile), "%s%smnode.json", pMgmt->path, TD_DIRSEP);
 
   if (taosRenameFile(file, realfile) != 0) {
-    terrno = TAOS_SYSTEM_ERROR(errno);;
+    terrno = TAOS_SYSTEM_ERROR(errno);
     dError("failed to rename %s since %s", file, terrstr());
     return -1;
   }
diff --git a/source/dnode/mgmt/mm/src/mmInt.c b/source/dnode/mgmt/mm/src/mmInt.c
index 3d15c9b9aeeb7768e20139463498885f0372ad3a..f4e00b5921c7c65dfd92037bb0a2a763c15e46f6 100644
--- a/source/dnode/mgmt/mm/src/mmInt.c
+++ b/source/dnode/mgmt/mm/src/mmInt.c
@@ -39,10 +39,6 @@ static int32_t mmRequire(SMgmtWrapper *pWrapper, bool *required) {
 }
 
 static void mmInitOption(SMnodeMgmt *pMgmt, SMnodeOpt *pOption) {
-  SDnode *pDnode = pMgmt->pDnode;
-  pOption->dnodeId = pDnode->dnodeId;
-  pOption->clusterId = pDnode->clusterId;
-
   SMsgCb msgCb = {0};
   msgCb.pWrapper = pMgmt->pWrapper;
   msgCb.queueFps[QUERY_QUEUE] = mmPutMsgToQueryQueue;
@@ -66,6 +62,7 @@ static void mmBuildOptionForDeploy(SMnodeMgmt *pMgmt, SMnodeOpt *pOption) {
   pReplica->id = 1;
   pReplica->port = pDnode->serverPort;
   tstrncpy(pReplica->fqdn, pDnode->localFqdn, TSDB_FQDN_LEN);
+  pOption->deploy = true;
 
   pMgmt->selfIndex = pOption->selfIndex;
   pMgmt->replica = pOption->replica;
@@ -77,6 +74,7 @@ static void mmBuildOptionForOpen(SMnodeMgmt *pMgmt, SMnodeOpt *pOption) {
   pOption->selfIndex = pMgmt->selfIndex;
   pOption->replica = pMgmt->replica;
   memcpy(&pOption->replicas, pMgmt->replicas, sizeof(SReplica) * TSDB_MAX_REPLICA);
+  pOption->deploy = false;
 }
 
 static int32_t mmBuildOptionFromReq(SMnodeMgmt *pMgmt, SMnodeOpt *pOption, SDCreateMnodeReq *pCreate) {
@@ -89,7 +87,7 @@ static int32_t mmBuildOptionFromReq(SMnodeMgmt *pMgmt, SMnodeOpt *pOption, SDCre
     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) {
+    if (pReplica->id == pMgmt->pDnode->dnodeId) {
       pOption->selfIndex = i;
     }
   }
@@ -98,6 +96,7 @@ static int32_t mmBuildOptionFromReq(SMnodeMgmt *pMgmt, SMnodeOpt *pOption, SDCre
     dError("failed to build mnode options since %s", terrstr());
     return -1;
   }
+  pOption->deploy = true;
 
   pMgmt->selfIndex = pOption->selfIndex;
   pMgmt->replica = pOption->replica;
@@ -225,9 +224,7 @@ int32_t mmOpenFromMsg(SMgmtWrapper *pWrapper, SDCreateMnodeReq *pReq) {
   return code;
 }
 
-static int32_t mmOpen(SMgmtWrapper *pWrapper) {
-  return mmOpenFromMsg(pWrapper, NULL);
-}
+static int32_t mmOpen(SMgmtWrapper *pWrapper) { return mmOpenFromMsg(pWrapper, NULL); }
 
 static int32_t mmStart(SMgmtWrapper *pWrapper) {
   dDebug("mnode-mgmt start to run");
@@ -258,7 +255,7 @@ int32_t mmGetUserAuth(SMgmtWrapper *pWrapper, char *user, char *spi, char *encry
 }
 
 int32_t mmMonitorMnodeInfo(SMgmtWrapper *pWrapper, SMonClusterInfo *pClusterInfo, SMonVgroupInfo *pVgroupInfo,
-                         SMonGrantInfo *pGrantInfo) {
+                           SMonGrantInfo *pGrantInfo) {
   SMnodeMgmt *pMgmt = pWrapper->pMgmt;
   return mndGetMonitorInfo(pMgmt->pMnode, pClusterInfo, pVgroupInfo, pGrantInfo);
 }
diff --git a/source/dnode/mnode/impl/inc/mndCluster.h b/source/dnode/mnode/impl/inc/mndCluster.h
index 0206695b8814ee83832aeddeed86b4738649b3b7..5b7bac4486492d4c852a49c4c706526092ff83ec 100644
--- a/source/dnode/mnode/impl/inc/mndCluster.h
+++ b/source/dnode/mnode/impl/inc/mndCluster.h
@@ -25,6 +25,7 @@ extern "C" {
 int32_t mndInitCluster(SMnode *pMnode);
 void    mndCleanupCluster(SMnode *pMnode);
 int32_t mndGetClusterName(SMnode *pMnode, char *clusterName, int32_t len);
+int64_t mndGetClusterId(SMnode *pMnode);
 
 #ifdef __cplusplus
 }
diff --git a/source/dnode/mnode/impl/inc/mndInt.h b/source/dnode/mnode/impl/inc/mndInt.h
index f89e9d8fe0536f4848bd5798ce3b4c4fe806e1f6..1cb0c78aa536a9982940e4d5424a1dc9bc86072b 100644
--- a/source/dnode/mnode/impl/inc/mndInt.h
+++ b/source/dnode/mnode/impl/inc/mndInt.h
@@ -100,7 +100,6 @@ typedef struct {
 } SGrantInfo;
 
 typedef struct SMnode {
-  int32_t           dnodeId;
   int64_t           clusterId;
   int8_t            replica;
   int8_t            selfIndex;
diff --git a/source/dnode/mnode/impl/src/mndCluster.c b/source/dnode/mnode/impl/src/mndCluster.c
index dde7e1fe8fd946c0f9632dcb7273b89cbe246757..94e1efde61b7f9d5f2dab3969b8a87fb12d27a3d 100644
--- a/source/dnode/mnode/impl/src/mndCluster.c
+++ b/source/dnode/mnode/impl/src/mndCluster.c
@@ -17,7 +17,7 @@
 #include "mndCluster.h"
 #include "mndShow.h"
 
-#define TSDB_CLUSTER_VER_NUMBE 1
+#define TSDB_CLUSTER_VER_NUMBE    1
 #define TSDB_CLUSTER_RESERVE_SIZE 64
 
 static SSdbRaw *mndClusterActionEncode(SClusterObj *pCluster);
@@ -61,6 +61,23 @@ int32_t mndGetClusterName(SMnode *pMnode, char *clusterName, int32_t len) {
   return 0;
 }
 
+int64_t mndGetClusterId(SMnode *pMnode) {
+  SSdb   *pSdb = pMnode->pSdb;
+  void   *pIter = NULL;
+  int64_t clusterId = -1;
+
+  while (1) {
+    SClusterObj *pCluster = NULL;
+    pIter = sdbFetch(pSdb, SDB_CLUSTER, pIter, (void **)&pCluster);
+    if (pIter == NULL) break;
+
+    clusterId = pCluster->id;
+    sdbRelease(pSdb, pCluster);
+  }
+
+  return clusterId;
+}
+
 static SSdbRaw *mndClusterActionEncode(SClusterObj *pCluster) {
   terrno = TSDB_CODE_OUT_OF_MEMORY;
 
diff --git a/source/dnode/mnode/impl/src/mndQnode.c b/source/dnode/mnode/impl/src/mndQnode.c
index 4b19a26bc4f2dc7a6c039b551b17835d347951f0..85718e203717a684117bb2a21645652a785a9a3a 100644
--- a/source/dnode/mnode/impl/src/mndQnode.c
+++ b/source/dnode/mnode/impl/src/mndQnode.c
@@ -423,7 +423,7 @@ static int32_t mndProcessDropQnodeReq(SNodeMsg *pReq) {
 
 DROP_QNODE_OVER:
   if (code != 0 && code != TSDB_CODE_MND_ACTION_IN_PROGRESS) {
-    mError("qnode:%d, failed to drop since %s", pMnode->dnodeId, terrstr());
+    mError("qnode:%d, failed to drop since %s", dropReq.dnodeId, terrstr());
   }
 
   mndReleaseQnode(pMnode, pObj);
diff --git a/source/dnode/mnode/impl/src/mndSnode.c b/source/dnode/mnode/impl/src/mndSnode.c
index 5e0d9fae9a4bccbd382355ad2897b99c13ea8929..4f24c6f7eb957992899f848143f87dbdd827e8ef 100644
--- a/source/dnode/mnode/impl/src/mndSnode.c
+++ b/source/dnode/mnode/impl/src/mndSnode.c
@@ -433,7 +433,7 @@ static int32_t mndProcessDropSnodeReq(SNodeMsg *pReq) {
 
 DROP_SNODE_OVER:
   if (code != 0 && code != TSDB_CODE_MND_ACTION_IN_PROGRESS) {
-    mError("snode:%d, failed to drop since %s", pMnode->dnodeId, terrstr());
+    mError("snode:%d, failed to drop since %s", dropReq.dnodeId, terrstr());
   }
 
   mndReleaseSnode(pMnode, pObj);
diff --git a/source/dnode/mnode/impl/src/mnode.c b/source/dnode/mnode/impl/src/mnode.c
index 67b7d6dd45f754a6378c3b25941b5118ce62f61c..5c3dd778e15cef24ebf734d883f5e6def9d805c5 100644
--- a/source/dnode/mnode/impl/src/mnode.c
+++ b/source/dnode/mnode/impl/src/mnode.c
@@ -187,7 +187,7 @@ static int32_t mndAllocStep(SMnode *pMnode, char *name, MndInitFp initFp, MndCle
   return 0;
 }
 
-static int32_t mndInitSteps(SMnode *pMnode) {
+static int32_t mndInitSteps(SMnode *pMnode, bool deploy) {
   if (mndAllocStep(pMnode, "mnode-sdb", mndInitSdb, mndCleanupSdb) != 0) return -1;
   if (mndAllocStep(pMnode, "mnode-trans", mndInitTrans, mndCleanupTrans) != 0) return -1;
   if (mndAllocStep(pMnode, "mnode-cluster", mndInitCluster, mndCleanupCluster) != 0) return -1;
@@ -210,7 +210,7 @@ static int32_t mndInitSteps(SMnode *pMnode) {
   if (mndAllocStep(pMnode, "mnode-infos", mndInitInfos, mndCleanupInfos) != 0) return -1;
   if (mndAllocStep(pMnode, "mnode-db", mndInitDb, mndCleanupDb) != 0) return -1;
   if (mndAllocStep(pMnode, "mnode-func", mndInitFunc, mndCleanupFunc) != 0) return -1;
-  if (pMnode->clusterId <= 0) {
+  if (deploy) {
     if (mndAllocStep(pMnode, "mnode-sdb-deploy", mndDeploySdb, NULL) != 0) return -1;
   } else {
     if (mndAllocStep(pMnode, "mnode-sdb-read", mndReadSdb, NULL) != 0) return -1;
@@ -263,23 +263,15 @@ static int32_t mndExecSteps(SMnode *pMnode) {
     }
   }
 
+  pMnode->clusterId = mndGetClusterId(pMnode);
   return 0;
 }
 
-static int32_t mndSetOptions(SMnode *pMnode, const SMnodeOpt *pOption) {
-  pMnode->dnodeId = pOption->dnodeId;
-  pMnode->clusterId = pOption->clusterId;
+static void mndSetOptions(SMnode *pMnode, const SMnodeOpt *pOption) {
   pMnode->replica = pOption->replica;
   pMnode->selfIndex = pOption->selfIndex;
   memcpy(&pMnode->replicas, pOption->replicas, sizeof(SReplica) * TSDB_MAX_REPLICA);
   pMnode->msgCb = pOption->msgCb;
-
-  if (pMnode->dnodeId < 0 || pMnode->clusterId < 0) {
-    terrno = TSDB_CODE_MND_INVALID_OPTIONS;
-    return -1;
-  }
-
-  return 0;
 }
 
 SMnode *mndOpen(const char *path, const SMnodeOpt *pOption) {
@@ -294,6 +286,7 @@ SMnode *mndOpen(const char *path, const SMnodeOpt *pOption) {
 
   char timestr[24] = "1970-01-01 00:00:00.00";
   (void)taosParseTime(timestr, &pMnode->checkTime, (int32_t)strlen(timestr), TSDB_TIME_PRECISION_MILLI, 0);
+  mndSetOptions(pMnode, pOption);
 
   pMnode->pSteps = taosArrayInit(24, sizeof(SMnodeStep));
   if (pMnode->pSteps == NULL) {
@@ -312,16 +305,7 @@ SMnode *mndOpen(const char *path, const SMnodeOpt *pOption) {
     return NULL;
   }
 
-  code = mndSetOptions(pMnode, pOption);
-  if (code != 0) {
-    code = terrno;
-    mError("failed to open mnode since %s", terrstr());
-    mndClose(pMnode);
-    terrno = code;
-    return NULL;
-  }
-
-  code = mndInitSteps(pMnode);
+  code = mndInitSteps(pMnode, pOption->deploy);
   if (code != 0) {
     code = terrno;
     mError("failed to open mnode since %s", terrstr());