Merge remote-tracking branch 'origin/3.0' into fix/tsim

Jenkinsfile2

@@ -130,6 +130,9 @@ def pre_test(){
 def pre_test_win(){
     bat '''
         hostname
+        taskkill /f /t /im python.exe
+        taskkill /f /t /im bash.exe
+        taskkill /f /t /im taosd.exe
         ipconfig
         set
         date /t

include/common/tdatablock.h

@@ -222,6 +222,7 @@ void blockDataCleanup(SSDataBlock* pDataBlock);
 size_t blockDataGetCapacityInRow(const SSDataBlock* pBlock, size_t pageSize);
 
 int32_t blockDataTrimFirstNRows(SSDataBlock* pBlock, size_t n);
+int32_t blockDataKeepFirstNRows(SSDataBlock* pBlock, size_t n);
 
 int32_t      assignOneDataBlock(SSDataBlock* dst, const SSDataBlock* src);
 int32_t      copyDataBlock(SSDataBlock* dst, const SSDataBlock* src);
@@ -235,6 +236,7 @@ SColumnInfoData* bdGetColumnInfoData(SSDataBlock* pBlock, int32_t index);
 void blockEncode(const SSDataBlock* pBlock, char* data, int32_t* dataLen, int32_t numOfCols, int8_t needCompress);
 const char* blockDecode(SSDataBlock* pBlock, int32_t numOfCols, int32_t numOfRows, const char* pData);
 
+void blockDebugShowDataBlock(SSDataBlock* pBlock, const char* flag);
 void blockDebugShowDataBlocks(const SArray* dataBlocks, const char* flag);
 // for debug
 char* dumpBlockData(SSDataBlock* pDataBlock, const char* flag, char** dumpBuf);

include/common/tdataformat.h

@@ -78,6 +78,7 @@ int32_t tEncodeTag(SEncoder *pEncoder, const STag *pTag);
 int32_t tDecodeTag(SDecoder *pDecoder, STag **ppTag);
 int32_t tTagToValArray(const STag *pTag, SArray **ppArray);
 void    debugPrintSTag(STag *pTag, const char *tag, int32_t ln);  // TODO: remove
+int32_t parseJsontoTagData(const char* json, SArray* pTagVals, STag** ppTag, void* pMsgBuf);
 
 // STRUCT =================
 struct STColumn {

include/common/tmsgdef.h

@@ -233,10 +233,12 @@ enum {
   TD_DEF_MSG_TYPE(TDMT_SYNC_PING, "sync-ping", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_SYNC_PING_REPLY, "sync-ping-reply", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_SYNC_CLIENT_REQUEST, "sync-client-request", NULL, NULL)
+  TD_DEF_MSG_TYPE(TDMT_SYNC_CLIENT_REQUEST_BATCH, "sync-client-request-batch", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_SYNC_CLIENT_REQUEST_REPLY, "sync-client-request-reply", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_SYNC_REQUEST_VOTE, "sync-request-vote", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_SYNC_REQUEST_VOTE_REPLY, "sync-request-vote-reply", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_SYNC_APPEND_ENTRIES, "sync-append-entries", NULL, NULL)
+  TD_DEF_MSG_TYPE(TDMT_SYNC_APPEND_ENTRIES_BATCH, "sync-append-entries-batch", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_SYNC_APPEND_ENTRIES_REPLY, "sync-append-entries-reply", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_SYNC_NOOP, "sync-noop", NULL, NULL)
   TD_DEF_MSG_TYPE(TDMT_SYNC_UNKNOWN, "sync-unknown", NULL, NULL)

include/libs/sync/sync.h

@@ -26,6 +26,7 @@ extern "C" {
 
 extern bool gRaftDetailLog;
 
+#define SYNC_MAX_BATCH_SIZE 100
 #define SYNC_INDEX_BEGIN 0
 #define SYNC_INDEX_INVALID -1
 #define SYNC_TERM_INVALID 0xFFFFFFFFFFFFFFFF
@@ -120,7 +121,7 @@ typedef struct SSyncFSM {
   int32_t (*FpGetSnapshot)(struct SSyncFSM* pFsm, SSnapshot* pSnapshot, void* pReaderParam, void** ppReader);
   int32_t (*FpGetSnapshotInfo)(struct SSyncFSM* pFsm, SSnapshot* pSnapshot);
 
-  int32_t (*FpSnapshotStartRead)(struct SSyncFSM* pFsm, void** ppReader);
+  int32_t (*FpSnapshotStartRead)(struct SSyncFSM* pFsm, void* pReaderParam, void** ppReader);
   int32_t (*FpSnapshotStopRead)(struct SSyncFSM* pFsm, void* pReader);
   int32_t (*FpSnapshotDoRead)(struct SSyncFSM* pFsm, void* pReader, void** ppBuf, int32_t* len);
 
@@ -164,6 +165,7 @@ typedef struct SSyncLogStore {
   bool (*syncLogIsEmpty)(struct SSyncLogStore* pLogStore);
   int32_t (*syncLogEntryCount)(struct SSyncLogStore* pLogStore);
   int32_t (*syncLogRestoreFromSnapshot)(struct SSyncLogStore* pLogStore, SyncIndex index);
+  bool (*syncLogExist)(struct SSyncLogStore* pLogStore, SyncIndex index);
 
   SyncIndex (*syncLogWriteIndex)(struct SSyncLogStore* pLogStore);
   SyncIndex (*syncLogLastIndex)(struct SSyncLogStore* pLogStore);
@@ -179,6 +181,7 @@ typedef struct SSyncInfo {
   bool        isStandBy;
   bool        snapshotEnable;
   SyncGroupId vgId;
+  int32_t     batchSize;
   SSyncCfg    syncCfg;
   char        path[TSDB_FILENAME_LEN];
   SWal*       pWal;
@@ -202,6 +205,7 @@ SyncGroupId syncGetVgId(int64_t rid);
 void        syncGetEpSet(int64_t rid, SEpSet* pEpSet);
 void        syncGetRetryEpSet(int64_t rid, SEpSet* pEpSet);
 int32_t     syncPropose(int64_t rid, SRpcMsg* pMsg, bool isWeak);
+// int32_t     syncProposeBatch(int64_t rid, SRpcMsg* pMsgArr, bool* pIsWeakArr, int32_t arrSize);
 bool        syncEnvIsStart();
 const char* syncStr(ESyncState state);
 bool        syncIsRestoreFinish(int64_t rid);

include/libs/sync/syncTools.h

@@ -219,6 +219,34 @@ void syncClientRequestPrint2(char* s, const SyncClientRequest* pMsg);
 void syncClientRequestLog(const SyncClientRequest* pMsg);
 void syncClientRequestLog2(char* s, const SyncClientRequest* pMsg);
 
+// ---------------------------------------------
+typedef struct SOffsetAndContLen {
+  int32_t offset;
+  int32_t contLen;
+} SOffsetAndContLen;
+
+typedef struct SRaftMeta {
+  uint64_t seqNum;
+  bool     isWeak;
+} SRaftMeta;
+
+// block1:
+// block2: SRaftMeta array
+// block3: rpc msg array (with pCont)
+
+typedef struct SyncClientRequestBatch {
+  uint32_t bytes;
+  int32_t  vgId;
+  uint32_t msgType;  // SyncClientRequestBatch msgType
+  uint32_t dataCount;
+  uint32_t dataLen;  // user RpcMsg.contLen
+  char     data[];   // user RpcMsg.pCont
+} SyncClientRequestBatch;
+
+SyncClientRequestBatch* syncClientRequestBatchBuild(SRpcMsg* rpcMsgArr, SRaftMeta* raftArr, int32_t arrSize,
+                                                    int32_t vgId);
+void                    syncClientRequestBatch2RpcMsg(const SyncClientRequestBatch* pSyncMsg, SRpcMsg* pRpcMsg);
+
 // ---------------------------------------------
 typedef struct SyncClientRequestReply {
   uint32_t bytes;
@@ -325,22 +353,53 @@ void syncAppendEntriesLog(const SyncAppendEntries* pMsg);
 void syncAppendEntriesLog2(char* s, const SyncAppendEntries* pMsg);
 
 // ---------------------------------------------
+
+// define ahead
+/*
+typedef struct SOffsetAndContLen {
+  int32_t offset;
+  int32_t contLen;
+} SOffsetAndContLen;
+*/
+
 typedef struct SyncAppendEntriesBatch {
   uint32_t bytes;
   int32_t  vgId;
   uint32_t msgType;
   SRaftId  srcId;
   SRaftId  destId;
+
   // private data
   SyncTerm  term;
   SyncIndex prevLogIndex;
   SyncTerm  prevLogTerm;
   SyncIndex commitIndex;
   SyncTerm  privateTerm;
+  int32_t   dataCount;
   uint32_t  dataLen;
   char      data[];
 } SyncAppendEntriesBatch;
 
+SyncAppendEntriesBatch* syncAppendEntriesBatchBuild(SRpcMsg* rpcMsgArr, int32_t arrSize, int32_t vgId);
+void                    syncAppendEntriesBatchDestroy(SyncAppendEntriesBatch* pMsg);
+void                    syncAppendEntriesBatchSerialize(const SyncAppendEntriesBatch* pMsg, char* buf, uint32_t bufLen);
+void                    syncAppendEntriesBatchDeserialize(const char* buf, uint32_t len, SyncAppendEntriesBatch* pMsg);
+char*                   syncAppendEntriesBatchSerialize2(const SyncAppendEntriesBatch* pMsg, uint32_t* len);
+SyncAppendEntriesBatch* syncAppendEntriesBatchDeserialize2(const char* buf, uint32_t len);
+void                    syncAppendEntriesBatch2RpcMsg(const SyncAppendEntriesBatch* pMsg, SRpcMsg* pRpcMsg);
+void                    syncAppendEntriesBatchFromRpcMsg(const SRpcMsg* pRpcMsg, SyncAppendEntriesBatch* pMsg);
+SyncAppendEntriesBatch* syncAppendEntriesBatchFromRpcMsg2(const SRpcMsg* pRpcMsg);
+cJSON*                  syncAppendEntriesBatch2Json(const SyncAppendEntriesBatch* pMsg);
+char*                   syncAppendEntriesBatch2Str(const SyncAppendEntriesBatch* pMsg);
+void syncAppendEntriesBatch2RpcMsgArray(SyncAppendEntriesBatch* pSyncMsg, SRpcMsg* rpcMsgArr, int32_t maxArrSize,
+                                        int32_t* pRetArrSize);
+
+// for debug ----------------------
+void syncAppendEntriesBatchPrint(const SyncAppendEntriesBatch* pMsg);
+void syncAppendEntriesBatchPrint2(char* s, const SyncAppendEntriesBatch* pMsg);
+void syncAppendEntriesBatchLog(const SyncAppendEntriesBatch* pMsg);
+void syncAppendEntriesBatchLog2(char* s, const SyncAppendEntriesBatch* pMsg);
+
 // ---------------------------------------------
 typedef struct SyncAppendEntriesReply {
   uint32_t bytes;
@@ -542,6 +601,7 @@ int32_t syncNodeOnPingCb(SSyncNode* ths, SyncPing* pMsg);
 int32_t syncNodeOnPingReplyCb(SSyncNode* ths, SyncPingReply* pMsg);
 int32_t syncNodeOnTimeoutCb(SSyncNode* ths, SyncTimeout* pMsg);
 int32_t syncNodeOnClientRequestCb(SSyncNode* ths, SyncClientRequest* pMsg, SyncIndex* pRetIndex);
+int32_t syncNodeOnClientRequestBatchCb(SSyncNode* ths, SyncClientRequestBatch* pMsg);
 int32_t syncNodeOnRequestVoteCb(SSyncNode* ths, SyncRequestVote* pMsg);
 int32_t syncNodeOnRequestVoteReplyCb(SSyncNode* ths, SyncRequestVoteReply* pMsg);
 int32_t syncNodeOnAppendEntriesCb(SSyncNode* ths, SyncAppendEntries* pMsg);

include/libs/wal/wal.h

@@ -61,45 +61,23 @@ extern "C" {
     }                                                             \
   }
 
-#define WAL_HEAD_VER     0
+#define WAL_PROTO_VER    0
 #define WAL_NOSUFFIX_LEN 20
 #define WAL_SUFFIX_AT    (WAL_NOSUFFIX_LEN + 1)
 #define WAL_LOG_SUFFIX   "log"
 #define WAL_INDEX_SUFFIX "idx"
 #define WAL_REFRESH_MS   1000
-#define WAL_MAX_SIZE     (TSDB_MAX_WAL_SIZE + sizeof(SWalHead))
+#define WAL_MAX_SIZE     (TSDB_MAX_WAL_SIZE + sizeof(SWalCkHead))
 #define WAL_PATH_LEN     (TSDB_FILENAME_LEN + 12)
 #define WAL_FILE_LEN     (WAL_PATH_LEN + 32)
 #define WAL_MAGIC        0xFAFBFCFDULL
 
-#pragma pack(push, 1)
 typedef enum {
   TAOS_WAL_NOLOG = 0,
   TAOS_WAL_WRITE = 1,
   TAOS_WAL_FSYNC = 2,
 } EWalType;
 
-// used by sync module
-typedef struct {
-  int8_t   isWeek;
-  uint64_t seqNum;
-  uint64_t term;
-} SSyncLogMeta;
-
-typedef struct SWalReadHead {
-  int8_t  headVer;
-  int8_t  reserved;
-  int16_t msgType;
-  int32_t bodyLen;
-  int64_t ingestTs;  // not implemented
-  int64_t version;
-
-  // sync meta
-  SSyncLogMeta syncMeta;
-
-  char body[];
-} SWalReadHead;
-
 typedef struct {
   int32_t  vgId;
   int32_t  fsyncPeriod;      // millisecond
@@ -110,13 +88,6 @@ typedef struct {
   EWalType level;  // wal level
 } SWalCfg;
 
-typedef struct {
-  uint64_t     magic;
-  uint32_t     cksumHead;
-  uint32_t     cksumBody;
-  SWalReadHead head;
-} SWalHead;
-
 typedef struct SWalVer {
   int64_t firstVer;
   int64_t verInSnapshotting;
@@ -125,6 +96,35 @@ typedef struct SWalVer {
   int64_t lastVer;
 } SWalVer;
 
+#pragma pack(push, 1)
+// used by sync module
+typedef struct {
+  int8_t   isWeek;
+  uint64_t seqNum;
+  uint64_t term;
+} SSyncLogMeta;
+
+typedef struct {
+  int8_t  protoVer;
+  int64_t version;
+  int16_t msgType;
+  int32_t bodyLen;
+  int64_t ingestTs;  // not implemented
+
+  // sync meta
+  SSyncLogMeta syncMeta;
+
+  char body[];
+} SWalCont;
+
+typedef struct {
+  uint64_t magic;
+  uint32_t cksumHead;
+  uint32_t cksumBody;
+  SWalCont head;
+} SWalCkHead;
+#pragma pack(pop)
+
 typedef struct SWal {
   // cfg
   SWalCfg cfg;
@@ -134,7 +134,7 @@ typedef struct SWal {
   TdFilePtr pWriteLogTFile;
   TdFilePtr pWriteIdxTFile;
   int32_t   writeCur;
-  SArray   *fileInfoSet;
+  SArray   *fileInfoSet;  // SArray<SWalFileInfo>
   // status
   int64_t totSize;
   int64_t lastRollSeq;
@@ -146,7 +146,7 @@ typedef struct SWal {
   // path
   char path[WAL_PATH_LEN];
   // reusable write head
-  SWalHead writeHead;
+  SWalCkHead writeHead;
 } SWal;  // WAL HANDLE
 
 typedef struct SWalReadHandle {
@@ -158,11 +158,8 @@ typedef struct SWalReadHandle {
   int64_t       capacity;
   int64_t       status;  // if cursor valid
   TdThreadMutex mutex;
-  SWalHead     *pHead;
+  SWalCkHead   *pHead;
 } SWalReadHandle;
-#pragma pack(pop)
-
-// typedef int32_t (*FWalWrite)(void *ahandle, void *pHead);
 
 // module initialization
 int32_t walInit();
@@ -174,9 +171,9 @@ int32_t walAlter(SWal *, SWalCfg *pCfg);
 void    walClose(SWal *);
 
 // write
-int64_t walWriteWithSyncInfo(SWal *, int64_t index, tmsg_t msgType, SSyncLogMeta syncMeta, const void *body,
+int32_t walWriteWithSyncInfo(SWal *, int64_t index, tmsg_t msgType, SSyncLogMeta syncMeta, const void *body,
                              int32_t bodyLen);
-int64_t walWrite(SWal *, int64_t index, tmsg_t msgType, const void *body, int32_t bodyLen);
+int32_t walWrite(SWal *, int64_t index, tmsg_t msgType, const void *body, int32_t bodyLen);
 void    walFsync(SWal *, bool force);
 
 // apis for lifecycle management
@@ -196,9 +193,9 @@ int32_t         walReadWithHandle(SWalReadHandle *pRead, int64_t ver);
 
 // only for tq usage
 void    walSetReaderCapacity(SWalReadHandle *pRead, int32_t capacity);
-int32_t walFetchHead(SWalReadHandle *pRead, int64_t ver, SWalHead *pHead);
-int32_t walFetchBody(SWalReadHandle *pRead, SWalHead **ppHead);
-int32_t walSkipFetchBody(SWalReadHandle *pRead, const SWalHead *pHead);
+int32_t walFetchHead(SWalReadHandle *pRead, int64_t ver, SWalCkHead *pHead);
+int32_t walFetchBody(SWalReadHandle *pRead, SWalCkHead **ppHead);
+int32_t walSkipFetchBody(SWalReadHandle *pRead, const SWalCkHead *pHead);
 
 typedef struct {
   int64_t refId;

include/os/osFile.h

@@ -25,17 +25,17 @@ extern "C" {
 // If the error is in a third-party library, place this header file under the third-party library header file.
 // When you want to use this feature, you should find or add the same function in the following sectio
 #ifndef ALLOW_FORBID_FUNC
-    #define open OPEN_FUNC_TAOS_FORBID
-    #define fopen FOPEN_FUNC_TAOS_FORBID
-    #define access ACCESS_FUNC_TAOS_FORBID
-    #define stat STAT_FUNC_TAOS_FORBID
-    #define lstat LSTAT_FUNC_TAOS_FORBID
-    #define fstat FSTAT_FUNC_TAOS_FORBID
-    #define close CLOSE_FUNC_TAOS_FORBID
-    #define fclose FCLOSE_FUNC_TAOS_FORBID
-    #define fsync FSYNC_FUNC_TAOS_FORBID
-    #define getline GETLINE_FUNC_TAOS_FORBID
-    // #define fflush FFLUSH_FUNC_TAOS_FORBID
+#define open    OPEN_FUNC_TAOS_FORBID
+#define fopen   FOPEN_FUNC_TAOS_FORBID
+#define access  ACCESS_FUNC_TAOS_FORBID
+#define stat    STAT_FUNC_TAOS_FORBID
+#define lstat   LSTAT_FUNC_TAOS_FORBID
+#define fstat   FSTAT_FUNC_TAOS_FORBID
+#define close   CLOSE_FUNC_TAOS_FORBID
+#define fclose  FCLOSE_FUNC_TAOS_FORBID
+#define fsync   FSYNC_FUNC_TAOS_FORBID
+#define getline GETLINE_FUNC_TAOS_FORBID
+// #define fflush FFLUSH_FUNC_TAOS_FORBID
 #endif
 
 #ifndef PATH_MAX
@@ -43,54 +43,54 @@ extern "C" {
 #endif
 
 typedef struct TdFile *TdFilePtr;
- 
-#define TD_FILE_CREATE    0x0001
-#define TD_FILE_WRITE     0x0002
-#define TD_FILE_READ      0x0004
-#define TD_FILE_TRUNC     0x0008
-#define TD_FILE_APPEND    0x0010
-#define TD_FILE_TEXT      0x0020
-#define TD_FILE_AUTO_DEL  0x0040
-#define TD_FILE_EXCL      0x0080
-#define TD_FILE_STREAM    0x0100   // Only support taosFprintfFile, taosGetLineFile, taosEOFFile
-TdFilePtr taosOpenFile(const char *path,int32_t tdFileOptions);
+
+#define TD_FILE_CREATE   0x0001
+#define TD_FILE_WRITE    0x0002
+#define TD_FILE_READ     0x0004
+#define TD_FILE_TRUNC    0x0008
+#define TD_FILE_APPEND   0x0010
+#define TD_FILE_TEXT     0x0020
+#define TD_FILE_AUTO_DEL 0x0040
+#define TD_FILE_EXCL     0x0080
+#define TD_FILE_STREAM   0x0100  // Only support taosFprintfFile, taosGetLineFile, taosEOFFile
+TdFilePtr taosOpenFile(const char *path, int32_t tdFileOptions);
 
 #define TD_FILE_ACCESS_EXIST_OK 0x1
 #define TD_FILE_ACCESS_READ_OK  0x2
 #define TD_FILE_ACCESS_WRITE_OK 0x4
-bool    taosCheckAccessFile(const char *pathname, int mode);
- 
+bool taosCheckAccessFile(const char *pathname, int mode);
+
 int32_t taosLockFile(TdFilePtr pFile);
 int32_t taosUnLockFile(TdFilePtr pFile);
- 
+
 int32_t taosUmaskFile(int32_t maskVal);
- 
+
 int32_t taosStatFile(const char *path, int64_t *size, int32_t *mtime);
 int32_t taosDevInoFile(TdFilePtr pFile, int64_t *stDev, int64_t *stIno);
 int32_t taosFStatFile(TdFilePtr pFile, int64_t *size, int32_t *mtime);
 bool    taosCheckExistFile(const char *pathname);
- 
+
 int64_t taosLSeekFile(TdFilePtr pFile, int64_t offset, int32_t whence);
 int32_t taosFtruncateFile(TdFilePtr pFile, int64_t length);
 int32_t taosFsyncFile(TdFilePtr pFile);
- 
+
 int64_t taosReadFile(TdFilePtr pFile, void *buf, int64_t count);
 int64_t taosPReadFile(TdFilePtr pFile, void *buf, int64_t count, int64_t offset);
 int64_t taosWriteFile(TdFilePtr pFile, const void *buf, int64_t count);
 void    taosFprintfFile(TdFilePtr pFile, const char *format, ...);
 
-int64_t taosGetLineFile(TdFilePtr pFile, char ** __restrict ptrBuf);
+int64_t taosGetLineFile(TdFilePtr pFile, char **__restrict ptrBuf);
 int64_t taosGetsFile(TdFilePtr pFile, int32_t maxSize, char *__restrict buf);
 
 int32_t taosEOFFile(TdFilePtr pFile);
- 
-int64_t taosCloseFile(TdFilePtr *ppFile);
- 
+
+int32_t taosCloseFile(TdFilePtr *ppFile);
+
 int32_t taosRenameFile(const char *oldName, const char *newName);
 int64_t taosCopyFile(const char *from, const char *to);
 int32_t taosRemoveFile(const char *path);
- 
-void    taosGetTmpfilePath(const char *inputTmpDir, const char *fileNamePrefix, char *dstPath);
+
+void taosGetTmpfilePath(const char *inputTmpDir, const char *fileNamePrefix, char *dstPath);
 
 int64_t taosFSendFile(TdFilePtr pFileOut, TdFilePtr pFileIn, int64_t *offset, int64_t size);
 

include/util/taoserror.h

@@ -388,6 +388,10 @@ int32_t* taosGetErrno();
 #define TSDB_CODE_QRY_TASK_MSG_ERROR            TAOS_DEF_ERROR_CODE(0, 0x0719)
 #define TSDB_CODE_QRY_JOB_FREED                 TAOS_DEF_ERROR_CODE(0, 0x071A)
 #define TSDB_CODE_QRY_TASK_STATUS_ERROR         TAOS_DEF_ERROR_CODE(0, 0x071B)
+//json
+#define TSDB_CODE_QRY_JSON_IN_ERROR             TAOS_DEF_ERROR_CODE(0, 0x071C)
+#define TSDB_CODE_QRY_JSON_NOT_SUPPORT_ERROR    TAOS_DEF_ERROR_CODE(0, 0x071D)
+#define TSDB_CODE_QRY_JSON_IN_GROUP_ERROR       TAOS_DEF_ERROR_CODE(0, 0x071E)
 
 // grant
 #define TSDB_CODE_GRANT_EXPIRED                 TAOS_DEF_ERROR_CODE(0, 0x0800)
@@ -423,6 +427,7 @@ int32_t* taosGetErrno();
 #define TSDB_CODE_SYN_RECONFIG_NOT_READY        TAOS_DEF_ERROR_CODE(0, 0x0910)
 #define TSDB_CODE_SYN_PROPOSE_NOT_READY         TAOS_DEF_ERROR_CODE(0, 0x0911)
 #define TSDB_CODE_SYN_STANDBY_NOT_READY         TAOS_DEF_ERROR_CODE(0, 0x0912)
+#define TSDB_CODE_SYN_BATCH_ERROR               TAOS_DEF_ERROR_CODE(0, 0x0913)
 #define TSDB_CODE_SYN_INTERNAL_ERROR            TAOS_DEF_ERROR_CODE(0, 0x09FF)
 
 // tq

include/util/tutil.h

@@ -44,6 +44,8 @@ uint32_t ip2uint(const char *const ip_addr);
 void     taosIp2String(uint32_t ip, char *str);
 void     taosIpPort2String(uint32_t ip, uint16_t port, char *str);
 
+void *tmemmem(const char *haystack, int hlen, const char *needle, int nlen);
+
 static FORCE_INLINE void taosEncryptPass(uint8_t *inBuf, size_t inLen, char *target) {
   T_MD5_CTX context;
   tMD5Init(&context);
@@ -59,10 +61,10 @@ static FORCE_INLINE void taosEncryptPass_c(uint8_t *inBuf, size_t len, char *tar
   tMD5Final(&context);
   char buf[TSDB_PASSWORD_LEN + 1];
 
-  sprintf(buf, "%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x", context.digest[0],
-          context.digest[1], context.digest[2], context.digest[3], context.digest[4], context.digest[5],
-          context.digest[6], context.digest[7], context.digest[8], context.digest[9], context.digest[10],
-          context.digest[11], context.digest[12], context.digest[13], context.digest[14], context.digest[15]);
+  sprintf(buf, "%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x", context.digest[0], context.digest[1],
+          context.digest[2], context.digest[3], context.digest[4], context.digest[5], context.digest[6],
+          context.digest[7], context.digest[8], context.digest[9], context.digest[10], context.digest[11],
+          context.digest[12], context.digest[13], context.digest[14], context.digest[15]);
   memcpy(target, buf, TSDB_PASSWORD_LEN);
 }
 

source/common/src/tdatablock.c

@@ -716,7 +716,12 @@ int32_t dataBlockCompar(const void* p1, const void* p2, const void* param) {
 
     void* left1 = colDataGetData(pColInfoData, left);
     void* right1 = colDataGetData(pColInfoData, right);
-
+    if (pColInfoData->info.type == TSDB_DATA_TYPE_JSON) {
+      if (tTagIsJson(left1) || tTagIsJson(right1)) {
+        terrno = TSDB_CODE_QRY_JSON_NOT_SUPPORT_ERROR;
+        return 0;
+      }
+    }
     __compar_fn_t fn = getKeyComparFunc(pColInfoData->info.type, pOrder->order);
 
     int ret = fn(left1, right1);
@@ -890,7 +895,7 @@ int32_t blockDataSort(SSDataBlock* pDataBlock, SArray* pOrderInfo) {
         SBlockOrderInfo* pOrder = taosArrayGet(pOrderInfo, 0);
 
         int64_t p0 = taosGetTimestampUs();
-
+        
         __compar_fn_t fn = getKeyComparFunc(pColInfoData->info.type, pOrder->order);
         taosSort(pColInfoData->pData, pDataBlock->info.rows, pColInfoData->info.bytes, fn);
 
@@ -919,6 +924,7 @@ int32_t blockDataSort(SSDataBlock* pDataBlock, SArray* pOrderInfo) {
   }
 
   taosqsort(index, rows, sizeof(int32_t), &helper, dataBlockCompar);
+  if(terrno) return terrno;
 
   int64_t p1 = taosGetTimestampUs();
 
@@ -1431,9 +1437,39 @@ static void doShiftBitmap(char* nullBitmap, size_t n, size_t total) {
   }
 }
 
+static int32_t colDataMoveVarData(SColumnInfoData* pColInfoData, size_t start, size_t end){
+  int32_t dataOffset = -1;
+  int32_t dataLen = 0;
+  int32_t beigin = start;
+  while(beigin < end){
+    int32_t offset = pColInfoData->varmeta.offset[beigin];
+    if(offset == -1) {
+      beigin++;
+      continue;
+    }
+    if(start != 0) {
+      pColInfoData->varmeta.offset[beigin] = dataLen;
+    }
+    char *data = pColInfoData->pData + offset;
+    if(dataOffset == -1) dataOffset = offset;   // mark the begin of data
+    int32_t type = pColInfoData->info.type;
+    if (type == TSDB_DATA_TYPE_JSON) {
+      dataLen += getJsonValueLen(data);
+    } else {
+      dataLen += varDataTLen(data);
+    }
+    beigin++;
+  }
+  if(dataOffset > 0){
+    memmove(pColInfoData->pData, pColInfoData->pData + dataOffset, dataLen);
+    memmove(pColInfoData->varmeta.offset, &pColInfoData->varmeta.offset[start], (end - start) * sizeof(int32_t));
+  }
+  return dataLen;
+}
+
 static void colDataTrimFirstNRows(SColumnInfoData* pColInfoData, size_t n, size_t total) {
   if (IS_VAR_DATA_TYPE(pColInfoData->info.type)) {
-    memmove(pColInfoData->varmeta.offset, &pColInfoData->varmeta.offset[n], (total - n) * sizeof(int32_t));
+    pColInfoData->varmeta.length = colDataMoveVarData(pColInfoData, n, total);
     memset(&pColInfoData->varmeta.offset[total - n], 0, n);
   } else {
     int32_t bytes = pColInfoData->info.bytes;
@@ -1461,6 +1497,33 @@ int32_t blockDataTrimFirstNRows(SSDataBlock* pBlock, size_t n) {
   return TSDB_CODE_SUCCESS;
 }
 
+static void colDataKeepFirstNRows(SColumnInfoData* pColInfoData, size_t n, size_t total) {
+  if (IS_VAR_DATA_TYPE(pColInfoData->info.type)) {
+    pColInfoData->varmeta.length = colDataMoveVarData(pColInfoData, 0, n);
+    memset(&pColInfoData->varmeta.offset[n], 0, total - n);
+  }
+}
+
+int32_t blockDataKeepFirstNRows(SSDataBlock* pBlock, size_t n) {
+  if (n == 0) {
+    blockDataCleanup(pBlock);
+    return TSDB_CODE_SUCCESS;
+  }
+
+  if (pBlock->info.rows <= n) {
+    return TSDB_CODE_SUCCESS;
+  } else {
+    size_t numOfCols = taosArrayGetSize(pBlock->pDataBlock);
+    for (int32_t i = 0; i < numOfCols; ++i) {
+      SColumnInfoData* pColInfoData = taosArrayGet(pBlock->pDataBlock, i);
+      colDataKeepFirstNRows(pColInfoData, n, pBlock->info.rows);
+    }
+
+    pBlock->info.rows = n;
+  }
+  return TSDB_CODE_SUCCESS;
+}
+
 int32_t tEncodeDataBlock(void** buf, const SSDataBlock* pBlock) {
   int64_t tbUid = pBlock->info.uid;
   int16_t numOfCols = taosArrayGetSize(pBlock->pDataBlock);
@@ -1605,6 +1668,13 @@ static char* formatTimestamp(char* buf, int64_t val, int precision) {
   return buf;
 }
 
+void blockDebugShowDataBlock(SSDataBlock* pBlock, const char* flag) {
+  SArray* dataBlocks = taosArrayInit(1, sizeof(SSDataBlock));
+  taosArrayPush(dataBlocks, pBlock);
+  blockDebugShowDataBlocks(dataBlocks, flag);
+  taosArrayDestroy(dataBlocks);
+}
+
 void blockDebugShowDataBlocks(const SArray* dataBlocks, const char* flag) {
   char    pBuf[128] = {0};
   int32_t sz = taosArrayGetSize(dataBlocks);

source/common/src/tvariant.c

@@ -155,7 +155,8 @@ void taosVariantCreateFromBinary(SVariant *pVar, const char *pz, size_t len, uin
 void taosVariantDestroy(SVariant *pVar) {
   if (pVar == NULL) return;
 
-  if (pVar->nType == TSDB_DATA_TYPE_BINARY || pVar->nType == TSDB_DATA_TYPE_NCHAR) {
+  if (pVar->nType == TSDB_DATA_TYPE_BINARY || pVar->nType == TSDB_DATA_TYPE_NCHAR
+     || pVar->nType == TSDB_DATA_TYPE_JSON) {
     taosMemoryFreeClear(pVar->pz);
     pVar->nLen = 0;
   }
@@ -184,7 +185,8 @@ void taosVariantAssign(SVariant *pDst, const SVariant *pSrc) {
   if (pSrc == NULL || pDst == NULL) return;
 
   pDst->nType = pSrc->nType;
-  if (pSrc->nType == TSDB_DATA_TYPE_BINARY || pSrc->nType == TSDB_DATA_TYPE_NCHAR) {
+  if (pSrc->nType == TSDB_DATA_TYPE_BINARY || pSrc->nType == TSDB_DATA_TYPE_NCHAR
+      || pSrc->nType == TSDB_DATA_TYPE_JSON) {
     int32_t len = pSrc->nLen + TSDB_NCHAR_SIZE;
     char   *p = taosMemoryRealloc(pDst->pz, len);
     assert(p);
@@ -976,6 +978,7 @@ char *taosVariantGet(SVariant *pVar, int32_t type) {
     case TSDB_DATA_TYPE_FLOAT:
       return (char *)&pVar->d;
     case TSDB_DATA_TYPE_BINARY:
+    case TSDB_DATA_TYPE_JSON:
       return (char *)pVar->pz;
     case TSDB_DATA_TYPE_NCHAR:
       return (char *)pVar->ucs4;

source/dnode/mnode/impl/src/mndSync.c

@@ -117,7 +117,7 @@ void mndReConfig(struct SSyncFSM *pFsm, const SRpcMsg *pMsg, SReConfigCbMeta cbM
   }
 }
 
-int32_t mndSnapshotStartRead(struct SSyncFSM *pFsm, void **ppReader) {
+int32_t mndSnapshotStartRead(struct SSyncFSM *pFsm, void *pParam, void **ppReader) {
   mDebug("start to read snapshot from sdb");
   SMnode *pMnode = pFsm->data;
   return sdbStartRead(pMnode->pSdb, (SSdbIter **)ppReader, NULL, NULL, NULL);

source/dnode/mnode/impl/src/mndVgroup.c

@@ -426,7 +426,15 @@ static int32_t mndCompareDnodeId(int32_t *dnode1Id, int32_t *dnode2Id) { return
 static int32_t mndCompareDnodeVnodes(SDnodeObj *pDnode1, SDnodeObj *pDnode2) {
   float d1Score = (float)pDnode1->numOfVnodes / pDnode1->numOfSupportVnodes;
   float d2Score = (float)pDnode2->numOfVnodes / pDnode2->numOfSupportVnodes;
+#if 0  
+  if (d1Score == d2Score) {
+    return pDnode2->id - pDnode1->id;
+  } else {
+    return d1Score >= d2Score ? 1 : 0;
+  }
+#else
   return d1Score >= d2Score ? 1 : 0;
+#endif
 }
 
 void mndSortVnodeGid(SVgObj *pVgroup) {

source/dnode/vnode/src/inc/tq.h

@@ -157,7 +157,7 @@ typedef struct {
 static STqMgmt tqMgmt = {0};
 
 // tqRead
-int64_t tqFetchLog(STQ* pTq, STqHandle* pHandle, int64_t* fetchOffset, SWalHead** pHeadWithCkSum);
+int64_t tqFetchLog(STQ* pTq, STqHandle* pHandle, int64_t* fetchOffset, SWalCkHead** pHeadWithCkSum);
 
 // tqExec
 int32_t tqLogScanExec(STQ* pTq, STqExecHandle* pExec, SSubmitReq* pReq, SMqDataRsp* pRsp, int32_t workerId);
@@ -178,6 +178,7 @@ STqOffsetStore* tqOffsetOpen();
 void            tqOffsetClose(STqOffsetStore*);
 STqOffset*      tqOffsetRead(STqOffsetStore* pStore, const char* subscribeKey);
 int32_t         tqOffsetWrite(STqOffsetStore* pStore, const STqOffset* pOffset);
+int32_t         tqOffsetDelete(STqOffsetStore* pStore, const char* subscribeKey);
 int32_t         tqOffsetSnapshot(STqOffsetStore* pStore);
 
 // tqSink

source/dnode/vnode/src/tq/tq.c

@@ -271,8 +271,8 @@ int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg, int32_t workerId) {
     STqOffset* pOffset = tqOffsetRead(pTq->pOffsetStore, pReq->subKey);
     if (pOffset != NULL) {
       fetchOffsetNew = pOffset->val;
-      char formatBuf[50];
-      tFormatOffset(formatBuf, 50, &fetchOffsetNew);
+      char formatBuf[80];
+      tFormatOffset(formatBuf, 80, &fetchOffsetNew);
       tqDebug("tmq poll: consumer %ld, offset reset to %s", consumerId, formatBuf);
     } else {
       if (reqOffset.type == TMQ_OFFSET__RESET_EARLIEAST) {
@@ -302,9 +302,9 @@ int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg, int32_t workerId) {
   tqInitDataRsp(&dataRsp, pReq, pHandle->execHandle.subType);
 
   if (fetchOffsetNew.type == TMQ_OFFSET__LOG) {
-    int64_t   fetchVer = fetchOffsetNew.version + 1;
-    SWalHead* pHeadWithCkSum = taosMemoryMalloc(sizeof(SWalHead) + 2048);
-    if (pHeadWithCkSum == NULL) {
+    int64_t     fetchVer = fetchOffsetNew.version + 1;
+    SWalCkHead* pCkHead = taosMemoryMalloc(sizeof(SWalCkHead) + 2048);
+    if (pCkHead == NULL) {
       return -1;
     }
 
@@ -318,7 +318,7 @@ int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg, int32_t workerId) {
         break;
       }
 
-      if (tqFetchLog(pTq, pHandle, &fetchVer, &pHeadWithCkSum) < 0) {
+      if (tqFetchLog(pTq, pHandle, &fetchVer, &pCkHead) < 0) {
         // TODO add push mgr
 
         tqOffsetResetToLog(&dataRsp.rspOffset, fetchVer);
@@ -329,7 +329,7 @@ int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg, int32_t workerId) {
         goto OVER;
       }
 
-      SWalReadHead* pHead = &pHeadWithCkSum->head;
+      SWalCont* pHead = &pCkHead->head;
 
       tqDebug("tmq poll: consumer %ld (epoch %d) iter log, vg %d offset %ld msgType %d", consumerId, pReq->epoch,
               TD_VID(pTq->pVnode), fetchVer, pHead->msgType);
@@ -373,9 +373,10 @@ int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg, int32_t workerId) {
       }
     }
 
-    taosMemoryFree(pHeadWithCkSum);
+    taosMemoryFree(pCkHead);
   } else if (fetchOffsetNew.type == TMQ_OFFSET__SNAPSHOT_DATA) {
-    tqInfo("retrieve using snapshot req offset: uid %ld ts %ld", dataRsp.reqOffset.uid, dataRsp.reqOffset.ts);
+    tqInfo("retrieve using snapshot req offset: uid %ld ts %ld, actual offset: uid %ld ts %ld", dataRsp.reqOffset.uid,
+           dataRsp.reqOffset.ts, fetchOffsetNew.uid, fetchOffsetNew.ts);
     if (tqScanSnapshot(pTq, &pHandle->execHandle, &dataRsp, fetchOffsetNew, workerId) < 0) {
       ASSERT(0);
     }
@@ -522,7 +523,7 @@ int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg, int32_t workerId) {
       break;
     }
 
-    SWalReadHead* pHead = &pHeadWithCkSum->head;
+    SWalCont* pHead = &pHeadWithCkSum->head;
 
     tqDebug("tmq poll: consumer %ld (epoch %d) iter log, vg %d offset %ld msgType %d", consumerId, pReq->epoch,
             TD_VID(pTq->pVnode), fetchOffset, pHead->msgType);
@@ -597,6 +598,8 @@ int32_t tqProcessVgDeleteReq(STQ* pTq, char* msg, int32_t msgLen) {
   int32_t code = taosHashRemove(pTq->handles, pReq->subKey, strlen(pReq->subKey));
   ASSERT(code == 0);
 
+  tqOffsetDelete(pTq->pOffsetStore, pReq->subKey);
+
   if (tqMetaDeleteHandle(pTq, pReq->subKey) < 0) {
     ASSERT(0);
   }

source/dnode/vnode/src/tq/tqOffset.c

@@ -97,6 +97,10 @@ int32_t tqOffsetWrite(STqOffsetStore* pStore, const STqOffset* pOffset) {
   return taosHashPut(pStore->pHash, pOffset->subKey, strlen(pOffset->subKey), pOffset, sizeof(STqOffset));
 }
 
+int32_t tqOffsetDelete(STqOffsetStore* pStore, const char* subscribeKey) {
+  return taosHashRemove(pStore->pHash, subscribeKey, strlen(subscribeKey));
+}
+
 int32_t tqOffsetSnapshot(STqOffsetStore* pStore) {
   // open file
   // TODO file name should be with a version

source/dnode/vnode/src/tq/tqRead.c

@@ -15,13 +15,13 @@
 
 #include "tq.h"
 
-int64_t tqFetchLog(STQ* pTq, STqHandle* pHandle, int64_t* fetchOffset, SWalHead** ppHeadWithCkSum) {
+int64_t tqFetchLog(STQ* pTq, STqHandle* pHandle, int64_t* fetchOffset, SWalCkHead** ppCkHead) {
   int32_t code = 0;
   taosThreadMutexLock(&pHandle->pWalReader->mutex);
   int64_t offset = *fetchOffset;
 
   while (1) {
-    if (walFetchHead(pHandle->pWalReader, offset, *ppHeadWithCkSum) < 0) {
+    if (walFetchHead(pHandle->pWalReader, offset, *ppCkHead) < 0) {
       tqDebug("tmq poll: consumer %ld (epoch %d) vg %d offset %ld, no more log to return", pHandle->consumerId,
               pHandle->epoch, TD_VID(pTq->pVnode), offset);
       *fetchOffset = offset - 1;
@@ -29,8 +29,8 @@ int64_t tqFetchLog(STQ* pTq, STqHandle* pHandle, int64_t* fetchOffset, SWalHead*
       goto END;
     }
 
-    if ((*ppHeadWithCkSum)->head.msgType == TDMT_VND_SUBMIT) {
-      code = walFetchBody(pHandle->pWalReader, ppHeadWithCkSum);
+    if ((*ppCkHead)->head.msgType == TDMT_VND_SUBMIT) {
+      code = walFetchBody(pHandle->pWalReader, ppCkHead);
 
       if (code < 0) {
         ASSERT(0);
@@ -43,9 +43,9 @@ int64_t tqFetchLog(STQ* pTq, STqHandle* pHandle, int64_t* fetchOffset, SWalHead*
       goto END;
     } else {
       if (pHandle->fetchMeta) {
-        SWalReadHead* pHead = &((*ppHeadWithCkSum)->head);
+        SWalCont* pHead = &((*ppCkHead)->head);
         if (IS_META_MSG(pHead->msgType)) {
-          code = walFetchBody(pHandle->pWalReader, ppHeadWithCkSum);
+          code = walFetchBody(pHandle->pWalReader, ppCkHead);
 
           if (code < 0) {
             ASSERT(0);
@@ -58,7 +58,7 @@ int64_t tqFetchLog(STQ* pTq, STqHandle* pHandle, int64_t* fetchOffset, SWalHead*
           goto END;
         }
       }
-      code = walSkipFetchBody(pHandle->pWalReader, *ppHeadWithCkSum);
+      code = walSkipFetchBody(pHandle->pWalReader, *ppCkHead);
       if (code < 0) {
         ASSERT(0);
         *fetchOffset = offset;

source/dnode/vnode/src/vnd/vnodeSync.c

@@ -119,7 +119,7 @@ static int32_t vnodeProcessAlterReplicaReq(SVnode *pVnode, SRpcMsg *pMsg) {
 }
 
 void vnodeProposeMsg(SQueueInfo *pInfo, STaosQall *qall, int32_t numOfMsgs) {
-  SVnode * pVnode = pInfo->ahandle;
+  SVnode  *pVnode = pInfo->ahandle;
   int32_t  vgId = pVnode->config.vgId;
   int32_t  code = 0;
   SRpcMsg *pMsg = NULL;
@@ -199,7 +199,7 @@ void vnodeProposeMsg(SQueueInfo *pInfo, STaosQall *qall, int32_t numOfMsgs) {
 }
 
 void vnodeApplyMsg(SQueueInfo *pInfo, STaosQall *qall, int32_t numOfMsgs) {
-  SVnode * pVnode = pInfo->ahandle;
+  SVnode  *pVnode = pInfo->ahandle;
   int32_t  vgId = pVnode->config.vgId;
   int32_t  code = 0;
   SRpcMsg *pMsg = NULL;
@@ -240,7 +240,7 @@ int32_t vnodeProcessSyncReq(SVnode *pVnode, SRpcMsg *pMsg, SRpcMsg **pRsp) {
     STraceId *trace = &pMsg->info.traceId;
 
     do {
-      char *         syncNodeStr = sync2SimpleStr(pVnode->sync);
+      char          *syncNodeStr = sync2SimpleStr(pVnode->sync);
       static int64_t vndTick = 0;
       if (++vndTick % 10 == 1) {
         vGTrace("vgId:%d, sync trace msg:%s, %s", syncGetVgId(pVnode->sync), TMSG_INFO(pMsg->msgType), syncNodeStr);
@@ -375,7 +375,7 @@ static void vnodeSyncReconfig(struct SSyncFSM *pFsm, const SRpcMsg *pMsg, SReCon
 }
 
 static void vnodeSyncCommitMsg(SSyncFSM *pFsm, const SRpcMsg *pMsg, SFsmCbMeta cbMeta) {
-  SVnode *  pVnode = pFsm->data;
+  SVnode   *pVnode = pFsm->data;
   SSnapshot snapshot = {0};
   SyncIndex beginIndex = SYNC_INDEX_INVALID;
   char      logBuf[256] = {0};
@@ -409,7 +409,7 @@ static void vnodeSyncRollBackMsg(SSyncFSM *pFsm, const SRpcMsg *pMsg, SFsmCbMeta
   syncRpcMsgLog2(logBuf, (SRpcMsg *)pMsg);
 }
 
-static int32_t vnodeSnapshotStartRead(struct SSyncFSM *pFsm, void **ppReader) { return 0; }
+static int32_t vnodeSnapshotStartRead(struct SSyncFSM *pFsm, void *pParam, void **ppReader) { return 0; }
 
 static int32_t vnodeSnapshotStopRead(struct SSyncFSM *pFsm, void *pReader) { return 0; }
 

source/libs/command/src/command.c

@@ -297,7 +297,7 @@ int32_t appendTagValues(char* buf, int32_t* len, STableCfg* pCfg) {
   SArray* pTagVals = NULL;
   STag*   pTag = (STag*)pCfg->pTags;
 
-  if (pCfg->pTags && pTag->flags & TD_TAG_JSON) {
+  if (pCfg->pTags && tTagIsJson(pTag)) {
     char* pJson = parseTagDatatoJson(pTag);
     if (pJson) {
       *len += sprintf(buf + VARSTR_HEADER_SIZE + *len, "%s", pJson);

source/libs/executor/inc/executorimpl.h

@@ -287,6 +287,7 @@ typedef struct STableScanInfo {
   } lastStatus;
 
   int8_t scanMode;
+  int8_t noTable;
 } STableScanInfo;
 
 typedef struct STagScanInfo {

source/libs/executor/src/executil.c

@@ -63,6 +63,7 @@ size_t getResultRowSize(SqlFunctionCtx* pCtx, int32_t numOfOutput) {
     rowSize += pCtx[i].resDataInfo.interBufSize;
   }
 
+  rowSize += (numOfOutput * sizeof(bool));    // expand rowSize to mark if col is null for top/bottom result(saveTupleData)
   return rowSize;
 }
 
@@ -112,7 +113,9 @@ void initGroupedResultInfo(SGroupResInfo* pGroupResInfo, SHashObj* pHashmap, int
     p->groupId = *(uint64_t*)key;
     p->pos = *(SResultRowPosition*)pData;
     memcpy(p->key, (char*)key + sizeof(uint64_t), keyLen - sizeof(uint64_t));
-
+#ifdef BUF_PAGE_DEBUG
+    qDebug("page_groupRes, groupId:%"PRIu64",pageId:%d,offset:%d\n", p->groupId, p->pos.pageId, p->pos.offset);
+#endif
     taosArrayPush(pGroupResInfo->pRows, &p);
   }
 
@@ -271,6 +274,7 @@ static bool isTableOk(STableKeyInfo* info, SNode* pTagCond, SMeta* metaHandle) {
   SNode*  pNew = NULL;
   int32_t code = scalarCalculateConstants(pTagCondTmp, &pNew);
   if (TSDB_CODE_SUCCESS != code) {
+    terrno = code;
     nodesDestroyNode(pTagCondTmp);
     return false;
   }
@@ -323,12 +327,19 @@ int32_t getTableList(void* metaHandle, SScanPhysiNode* pScanNode, STableListInfo
       code = tsdbGetAllTableList(metaHandle, tableUid, pListInfo->pTableList);
     }
 
-    if (pTagCond) {
+    if (code != TSDB_CODE_SUCCESS) {
+      qError("failed  to  get tableIds, reason: %s, suid: %" PRIu64 "", tstrerror(code), tableUid);
+      terrno = code;
+      return code;
+    }
+
+    if(pTagCond){
       int32_t i = 0;
       while (i < taosArrayGetSize(pListInfo->pTableList)) {
         STableKeyInfo* info = taosArrayGet(pListInfo->pTableList, i);
-        bool           isOk = isTableOk(info, pTagCond, metaHandle);
-        if (!isOk) {
+        bool isOk = isTableOk(info, pTagCond, metaHandle);
+        if(terrno) return terrno;
+        if(!isOk){
           taosArrayRemove(pListInfo->pTableList, i);
           continue;
         }
@@ -586,13 +597,16 @@ static int32_t setSelectValueColumnInfo(SqlFunctionCtx* pCtx, int32_t numOfOutpu
   }
 
   for (int32_t i = 0; i < numOfOutput; ++i) {
-    if (strcmp(pCtx[i].pExpr->pExpr->_function.functionName, "_select_value") == 0) {
+    if (strcmp(pCtx[i].pExpr->pExpr->_function.functionName, "_select_value") == 0 ||
+        strcmp(pCtx[i].pExpr->pExpr->_function.functionName, "_group_key") == 0) {
       pValCtx[num++] = &pCtx[i];
     } else if (fmIsSelectFunc(pCtx[i].functionId)) {
       p = &pCtx[i];
     }
   }
-
+#ifdef BUF_PAGE_DEBUG
+  qDebug("page_setSelect num:%d", num);
+#endif
   if (p != NULL) {
     p->subsidiaries.pCtx = pValCtx;
     p->subsidiaries.num = num;

source/libs/executor/src/executorMain.c

@@ -235,9 +235,11 @@ int32_t qStreamPrepareScan(qTaskInfo_t tinfo, uint64_t uid, int64_t ts) {
   SExecTaskInfo* pTaskInfo = (SExecTaskInfo*)tinfo;
 
   if (uid == 0) {
-    STableKeyInfo* pTableInfo = taosArrayGet(pTaskInfo->tableqinfoList.pTableList, 0);
-    uid = pTableInfo->uid;
-    ts = INT64_MIN;
+    if (taosArrayGetSize(pTaskInfo->tableqinfoList.pTableList) != 0) {
+      STableKeyInfo* pTableInfo = taosArrayGet(pTaskInfo->tableqinfoList.pTableList, 0);
+      uid = pTableInfo->uid;
+      ts = INT64_MIN;
+    }
   }
 
   return doPrepareScan(pTaskInfo->pRoot, uid, ts);

source/libs/executor/src/executorimpl.c

@@ -274,6 +274,9 @@ SResultRow* doSetResultOutBufByKey(SDiskbasedBuf* pResultBuf, SResultRowInfo* pR
 
   // 1. close current opened time window
   if (pResultRowInfo->cur.pageId != -1 && ((pResult == NULL) || (pResult->pageId != pResultRowInfo->cur.pageId))) {
+#ifdef BUF_PAGE_DEBUG
+    qDebug("page_1");
+#endif
     SResultRowPosition pos = pResultRowInfo->cur;
     SFilePage*         pPage = getBufPage(pResultBuf, pos.pageId);
     releaseBufPage(pResultBuf, pPage);
@@ -281,6 +284,9 @@ SResultRow* doSetResultOutBufByKey(SDiskbasedBuf* pResultBuf, SResultRowInfo* pR
 
   // allocate a new buffer page
   if (pResult == NULL) {
+#ifdef BUF_PAGE_DEBUG
+    qDebug("page_2");
+#endif
     ASSERT(pSup->resultRowSize > 0);
     pResult = getNewResultRow(pResultBuf, groupId, pSup->resultRowSize);
 
@@ -538,7 +544,9 @@ static int32_t doAggregateImpl(SOperatorInfo* pOperator, TSKEY startTs, SqlFunct
       if (pCtx[k].fpSet.process == NULL) {
         continue;
       }
-
+#ifdef BUF_PAGE_DEBUG
+      qDebug("page_process");
+#endif
       int32_t code = pCtx[k].fpSet.process(&pCtx[k]);
       if (code != TSDB_CODE_SUCCESS) {
         qError("%s aggregate function error happens, code: %s", GET_TASKID(pOperator->pTaskInfo), tstrerror(code));
@@ -1413,7 +1421,9 @@ void setExecutionContext(SOperatorInfo* pOperator, int32_t numOfOutput, uint64_t
   if (pAggInfo->groupId != INT32_MIN && pAggInfo->groupId == groupId) {
     return;
   }
-
+#ifdef BUF_PAGE_DEBUG
+  qDebug("page_setbuf, groupId:%"PRIu64, groupId);
+#endif
   doSetTableGroupOutputBuf(pOperator, pAggInfo, numOfOutput, groupId);
 
   // record the current active group id
@@ -1489,11 +1499,15 @@ int32_t doCopyToSDataBlock(SExecTaskInfo* pTaskInfo, SSDataBlock* pBlock, SExprI
                            int32_t numOfExprs) {
   int32_t numOfRows = getNumOfTotalRes(pGroupResInfo);
   int32_t start = pGroupResInfo->index;
-
+#ifdef BUF_PAGE_DEBUG
+  qDebug("\npage_copytoblock rows:%d", numOfRows);
+#endif
   for (int32_t i = start; i < numOfRows; i += 1) {
     SResKeyPos* pPos = taosArrayGetP(pGroupResInfo->pRows, i);
     SFilePage*  page = getBufPage(pBuf, pPos->pos.pageId);
-
+#ifdef BUF_PAGE_DEBUG
+    qDebug("page_copytoblock pos pageId:%d, offset:%d", pPos->pos.pageId, pPos->pos.offset);
+#endif
     SResultRow* pRow = (SResultRow*)((char*)page + pPos->pos.offset);
 
     doUpdateNumOfRows(pRow, numOfExprs, rowCellOffset);
@@ -1525,6 +1539,9 @@ int32_t doCopyToSDataBlock(SExecTaskInfo* pTaskInfo, SSDataBlock* pBlock, SExprI
 
       pCtx[j].resultInfo = getResultEntryInfo(pRow, j, rowCellOffset);
       if (pCtx[j].fpSet.finalize) {
+#ifdef BUF_PAGE_DEBUG
+        qDebug("\npage_finalize %d", numOfExprs);
+#endif
         int32_t code = pCtx[j].fpSet.finalize(&pCtx[j], pBlock);
         if (TAOS_FAILED(code)) {
           qError("%s build result data block error, code %s", GET_TASKID(pTaskInfo), tstrerror(code));
@@ -1553,9 +1570,9 @@ int32_t doCopyToSDataBlock(SExecTaskInfo* pTaskInfo, SSDataBlock* pBlock, SExprI
 
     releaseBufPage(pBuf, page);
     pBlock->info.rows += pRow->numOfRows;
-    if (pBlock->info.rows >= pBlock->info.capacity) {  // output buffer is full
-      break;
-    }
+//    if (pBlock->info.rows >= pBlock->info.capacity) {  // output buffer is full
+//      break;
+//    }
   }
 
   qDebug("%s result generated, rows:%d, groupId:%" PRIu64, GET_TASKID(pTaskInfo), pBlock->info.rows,
@@ -2373,8 +2390,7 @@ static int32_t initExchangeOperator(SExchangePhysiNode* pExNode, SExchangeInfo*
   }
 
   pInfo->pSources = taosArrayInit(numOfSources, sizeof(SDownstreamSourceNode));
-  pInfo->pSourceDataInfo = taosArrayInit(numOfSources, sizeof(SSourceDataInfo));
-  if (pInfo->pSourceDataInfo == NULL || pInfo->pSources == NULL) {
+  if (pInfo->pSources == NULL) {
     return TSDB_CODE_OUT_OF_MEMORY;
   }
 
@@ -2829,21 +2845,28 @@ int32_t doPrepareScan(SOperatorInfo* pOperator, uint64_t uid, int64_t ts) {
     pScanInfo->blockType = STREAM_INPUT__DATA_SCAN;
 
     STableScanInfo* pInfo = pScanInfo->pSnapshotReadOp->info;
+    if (uid == 0) {
+      pInfo->noTable = 1;
+      return TSDB_CODE_SUCCESS;
+    }
 
     /*if (pSnapShotScanInfo->dataReader == NULL) {*/
     /*pSnapShotScanInfo->dataReader = tsdbReaderOpen(pHandle->vnode, &pSTInfo->cond, tableList, 0, 0);*/
     /*pSnapShotScanInfo->scanMode = TABLE_SCAN__TABLE_ORDER;*/
     /*}*/
 
+    pInfo->noTable = 0;
+
     if (pInfo->lastStatus.uid != uid || pInfo->lastStatus.ts != ts) {
       tsdbSetTableId(pInfo->dataReader, uid);
       int64_t oldSkey = pInfo->cond.twindows[0].skey;
-      pInfo->cond.twindows[0].skey = ts;
+      pInfo->cond.twindows[0].skey = ts + 1;
       tsdbResetReadHandle(pInfo->dataReader, &pInfo->cond, 0);
       pInfo->cond.twindows[0].skey = oldSkey;
       pInfo->scanTimes = 0;
       pInfo->curTWinIdx = 0;
     }
+    return TSDB_CODE_SUCCESS;
 
   } else {
     if (pOperator->numOfDownstream == 1) {
@@ -2856,8 +2879,6 @@ int32_t doPrepareScan(SOperatorInfo* pOperator, uint64_t uid, int64_t ts) {
       return TSDB_CODE_QRY_APP_ERROR;
     }
   }
-
-  return TSDB_CODE_SUCCESS;
 }
 
 int32_t doGetScanStatus(SOperatorInfo* pOperator, uint64_t* uid, int64_t* ts) {
@@ -3160,8 +3181,9 @@ static int32_t handleLimitOffset(SOperatorInfo* pOperator, SSDataBlock* pBlock)
   }
 
   // check for the limitation in each group
-  if (pProjectInfo->limit.limit > 0 && pProjectInfo->curOutput + pRes->info.rows >= pProjectInfo->limit.limit) {
-    pRes->info.rows = (int32_t)(pProjectInfo->limit.limit - pProjectInfo->curOutput);
+  if (pProjectInfo->limit.limit >= 0 && pProjectInfo->curOutput + pRes->info.rows >= pProjectInfo->limit.limit) {
+    int32_t keepRows = (int32_t)(pProjectInfo->limit.limit - pProjectInfo->curOutput);
+    blockDataKeepFirstNRows(pRes, keepRows);
     if (pProjectInfo->slimit.limit > 0 && pProjectInfo->slimit.limit <= pProjectInfo->curGroupOutput) {
       pOperator->status = OP_EXEC_DONE;
     }
@@ -3235,8 +3257,6 @@ static SSDataBlock* doProjectOperation(SOperatorInfo* pOperator) {
       longjmp(pTaskInfo->env, code);
     }
 
-    doFilter(pProjectInfo->pFilterNode, pBlock);
-
     setInputDataBlock(pOperator, pSup->pCtx, pBlock, order, scanFlag, false);
     blockDataEnsureCapacity(pInfo->pRes, pInfo->pRes->info.rows + pBlock->info.rows);
 
@@ -3247,6 +3267,10 @@ static SSDataBlock* doProjectOperation(SOperatorInfo* pOperator) {
     }
 
     int32_t status = handleLimitOffset(pOperator, pBlock);
+
+    // filter shall be applied after apply functions and limit/offset on the result
+    doFilter(pProjectInfo->pFilterNode, pInfo->pRes);
+
     if (status == PROJECT_RETRIEVE_CONTINUE) {
       continue;
     } else if (status == PROJECT_RETRIEVE_DONE) {
@@ -3473,11 +3497,12 @@ int32_t initAggInfo(SExprSupp* pSup, SAggSupporter* pAggSup, SExprInfo* pExprInf
 }
 
 void initResultSizeInfo(SOperatorInfo* pOperator, int32_t numOfRows) {
+  ASSERT(numOfRows != 0);
   pOperator->resultInfo.capacity = numOfRows;
   pOperator->resultInfo.threshold = numOfRows * 0.75;
 
   if (pOperator->resultInfo.threshold == 0) {
-    pOperator->resultInfo.capacity = numOfRows;
+    pOperator->resultInfo.threshold = numOfRows;
   }
 }
 
@@ -3977,12 +4002,12 @@ static int32_t sortTableGroup(STableListInfo* pTableListInfo, int32_t groupNum)
         return TSDB_CODE_QRY_APP_ERROR;
       }
       if (p == NULL) {
-        if (taosArrayPush(sortSupport, groupId) != NULL) {
+        if (taosArrayPush(sortSupport, groupId) == NULL) {
           qError("taos push support array error");
           taosArrayDestroy(sortSupport);
           return TSDB_CODE_QRY_APP_ERROR;
         }
-        if (taosArrayPush(pTableListInfo->pGroupList, &tGroup) != NULL) {
+        if (taosArrayPush(pTableListInfo->pGroupList, &tGroup) == NULL) {
           qError("taos push group array error");
           taosArrayDestroy(sortSupport);
           return TSDB_CODE_QRY_APP_ERROR;
@@ -4062,6 +4087,7 @@ int32_t generateGroupIdMap(STableListInfo* pTableListInfo, SReadHandle* pHandle,
       } else {
         taosMemoryFree(keyBuf);
         nodesClearList(groupNew);
+        metaReaderClear(&mr);
         return code;
       }
 
@@ -4074,7 +4100,14 @@ int32_t generateGroupIdMap(STableListInfo* pTableListInfo, SReadHandle* pHandle,
       } else {
         isNull[index++] = 0;
         char* data = nodesGetValueFromNode(pValue);
-        if (pValue->node.resType.type == TSDB_DATA_TYPE_JSON) {
+        if (pValue->node.resType.type == TSDB_DATA_TYPE_JSON){
+          if(tTagIsJson(data)){
+            terrno = TSDB_CODE_QRY_JSON_IN_GROUP_ERROR;
+            taosMemoryFree(keyBuf);
+            nodesClearList(groupNew);
+            metaReaderClear(&mr);
+            return terrno;
+          }
           int32_t len = getJsonValueLen(data);
           memcpy(pStart, data, len);
           pStart += len;
@@ -4133,7 +4166,8 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
     } else if (QUERY_NODE_PHYSICAL_PLAN_TABLE_MERGE_SCAN == type) {
       STableMergeScanPhysiNode* pTableScanNode = (STableMergeScanPhysiNode*)pPhyNode;
       int32_t code = createScanTableListInfo(pTableScanNode, pHandle, pTableListInfo, queryId, taskId);
-      if (code) {
+      if(code){
+        pTaskInfo->code = code;
         return NULL;
       }
       code = extractTableSchemaVersion(pHandle, pTableScanNode->scan.uid, pTaskInfo);
@@ -4160,7 +4194,11 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
             .maxTs = INT64_MIN,
       };
       if (pHandle) {
-        createScanTableListInfo(pTableScanNode, pHandle, pTableListInfo, queryId, taskId);
+        int32_t code = createScanTableListInfo(pTableScanNode, pHandle, pTableListInfo, queryId, taskId);
+        if(code){
+          pTaskInfo->code = code;
+          return NULL;
+        }
       }
 
       SOperatorInfo* pOperator =

source/libs/executor/src/groupoperator.c

@@ -141,6 +141,10 @@ static void recordNewGroupKeys(SArray* pGroupCols, SArray* pGroupColVals, SSData
       pkey->isNull = false;
       char* val = colDataGetData(pColInfoData, rowIndex);
       if (pkey->type == TSDB_DATA_TYPE_JSON) {
+        if(tTagIsJson(val)){
+          terrno = TSDB_CODE_QRY_JSON_IN_GROUP_ERROR;
+          return;
+        }
         int32_t dataLen = getJsonValueLen(val);
         memcpy(pkey->pData, val, dataLen);
       } else if (IS_VAR_DATA_TYPE(pkey->type)) {
@@ -227,11 +231,15 @@ static void doHashGroupbyAgg(SOperatorInfo* pOperator, SSDataBlock* pBlock) {
   int32_t     len = 0;
   STimeWindow w = TSWINDOW_INITIALIZER;
 
+  terrno = TSDB_CODE_SUCCESS;
   int32_t num = 0;
   for (int32_t j = 0; j < pBlock->info.rows; ++j) {
     // Compare with the previous row of this column, and do not set the output buffer again if they are identical.
     if (!pInfo->isInit) {
       recordNewGroupKeys(pInfo->pGroupCols, pInfo->pGroupColVals, pBlock, j);
+      if (terrno != TSDB_CODE_SUCCESS) {  // group by json error
+        longjmp(pTaskInfo->env, terrno);
+      }
       pInfo->isInit = true;
       num++;
       continue;
@@ -247,6 +255,9 @@ static void doHashGroupbyAgg(SOperatorInfo* pOperator, SSDataBlock* pBlock) {
     if (j == 0) {
       num++;
       recordNewGroupKeys(pInfo->pGroupCols, pInfo->pGroupColVals, pBlock, j);
+      if (terrno != TSDB_CODE_SUCCESS) {  // group by json error
+        longjmp(pTaskInfo->env, terrno);
+      }
       continue;
     }
 
@@ -661,7 +672,11 @@ static SSDataBlock* hashPartition(SOperatorInfo* pOperator) {
       }
     }
 
+    terrno = TSDB_CODE_SUCCESS;
     doHashPartition(pOperator, pBlock);
+    if (terrno != TSDB_CODE_SUCCESS) {  // group by json error
+      longjmp(pTaskInfo->env, terrno);
+    }
   }
 
   SArray* groupArray = taosArrayInit(taosHashGetSize(pInfo->pGroupSet), sizeof(SDataGroupInfo));

source/libs/executor/src/scanoperator.c

@@ -518,6 +518,7 @@ static SSDataBlock* doTableScan(SOperatorInfo* pOperator) {
 
   // if scan table by table
   if (pInfo->scanMode == TABLE_SCAN__TABLE_ORDER) {
+    if (pInfo->noTable) return NULL;
     while (1) {
       SSDataBlock* result = doTableScanGroup(pOperator);
       if (result) {
@@ -2089,6 +2090,7 @@ int32_t createScanTableListInfo(STableScanPhysiNode* pTableScanNode, SReadHandle
     qDebug("no table qualified for query, TID:0x%" PRIx64 ", QID:0x%" PRIx64, taskId, queryId);
     return TSDB_CODE_SUCCESS;
   }
+  pTableListInfo->needSortTableByGroupId = pTableScanNode->groupSort;
   code = generateGroupIdMap(pTableListInfo, pHandle, pTableScanNode->pGroupTags);
   if (code != TSDB_CODE_SUCCESS) {
     return code;

source/libs/executor/src/tsort.c

@@ -593,7 +593,10 @@ static int32_t createInitialSources(SSortHandle* pHandle) {
       if (size > sortBufSize) {
         // Perform the in-memory sort and then flush data in the buffer into disk.
         int64_t p = taosGetTimestampUs();
-        blockDataSort(pHandle->pDataBlock, pHandle->pSortInfo);
+        code = blockDataSort(pHandle->pDataBlock, pHandle->pSortInfo);
+        if (code != 0) {
+          return code;
+        }
 
         int64_t el = taosGetTimestampUs() - p;
         pHandle->sortElapsed += el;
@@ -608,7 +611,10 @@ static int32_t createInitialSources(SSortHandle* pHandle) {
       // Perform the in-memory sort and then flush data in the buffer into disk.
       int64_t p = taosGetTimestampUs();
 
-      blockDataSort(pHandle->pDataBlock, pHandle->pSortInfo);
+      int32_t code = blockDataSort(pHandle->pDataBlock, pHandle->pSortInfo);
+      if (code != 0) {
+        return code;
+      }
 
       int64_t el = taosGetTimestampUs() - p;
       pHandle->sortElapsed += el;

source/libs/function/src/builtins.c

@@ -1543,7 +1543,7 @@ static int32_t translateToJson(SFunctionNode* pFunc, char* pErrBuf, int32_t len)
     return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
   }
 
-  pFunc->node.resType = (SDataType){.bytes = tDataTypes[TSDB_DATA_TYPE_BINARY].bytes, .type = TSDB_DATA_TYPE_BINARY};
+  pFunc->node.resType = (SDataType){.bytes = tDataTypes[TSDB_DATA_TYPE_JSON].bytes, .type = TSDB_DATA_TYPE_JSON};
   return TSDB_CODE_SUCCESS;
 }
 
@@ -2620,7 +2620,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
   {
     .name = "_group_key",
     .type = FUNCTION_TYPE_GROUP_KEY,
-    .classification = FUNC_MGT_AGG_FUNC,
+    .classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_SELECT_FUNC,
     .translateFunc = translateGroupKey,
     .getEnvFunc   = getGroupKeyFuncEnv,
     .initFunc     = functionSetup,

source/libs/function/src/builtinsimpl.c

@@ -24,6 +24,7 @@
 #include "tglobal.h"
 #include "thistogram.h"
 #include "tpercentile.h"
+#include "query.h"
 
 #define HISTOGRAM_MAX_BINS_NUM 1000
 #define MAVG_MAX_POINTS_NUM    1000
@@ -1472,8 +1473,8 @@ void setSelectivityValue(SqlFunctionCtx* pCtx, SSDataBlock* pBlock, const STuple
   int32_t pageId = pTuplePos->pageId;
   int32_t offset = pTuplePos->offset;
 
-  if (pTuplePos->pageId != -1) {
-    int32_t numOfCols = taosArrayGetSize(pCtx->pSrcBlock->pDataBlock);
+  if (pTuplePos->pageId != -1 && pCtx->subsidiaries.num > 0) {
+    int32_t numOfCols = pCtx->subsidiaries.num;
     SFilePage* pPage = getBufPage(pCtx->pBuf, pageId);
 
     bool* nullList = (bool*)((char*)pPage + offset);
@@ -1484,22 +1485,21 @@ void setSelectivityValue(SqlFunctionCtx* pCtx, SSDataBlock* pBlock, const STuple
       SqlFunctionCtx* pc = pCtx->subsidiaries.pCtx[j];
 
       SFunctParam* pFuncParam = &pc->pExpr->base.pParam[0];
-      int32_t      srcSlotId = pFuncParam->pCol->slotId;
       int32_t      dstSlotId = pc->pExpr->base.resSchema.slotId;
 
       int32_t ps = 0;
-      for (int32_t k = 0; k < srcSlotId; ++k) {
-        SColumnInfoData* pSrcCol = taosArrayGet(pCtx->pSrcBlock->pDataBlock, k);
-        ps += pSrcCol->info.bytes;
-      }
 
       SColumnInfoData* pDstCol = taosArrayGet(pBlock->pDataBlock, dstSlotId);
-      if (nullList[srcSlotId]) {
+      ASSERT(pc->pExpr->base.resSchema.bytes == pDstCol->info.bytes);
+      if (nullList[j]) {
         colDataAppendNULL(pDstCol, rowIndex);
       } else {
-        colDataAppend(pDstCol, rowIndex, (pStart + ps), false);
+        colDataAppend(pDstCol, rowIndex, pStart, false);
       }
+      pStart += pDstCol->info.bytes;
     }
+
+    releaseBufPage(pCtx->pBuf, pPage);
   }
 }
 
@@ -3208,7 +3208,9 @@ void doAddIntoResult(SqlFunctionCtx* pCtx, void* pData, int32_t rowIndex, SSData
     if (pCtx->subsidiaries.num > 0) {
       saveTupleData(pCtx, rowIndex, pSrcBlock, &pItem->tuplePos);
     }
-
+#ifdef BUF_PAGE_DEBUG
+    qDebug("page_saveTuple i:%d, item:%p,pageId:%d, offset:%d\n", pEntryInfo->numOfRes, pItem, pItem->tuplePos.pageId, pItem->tuplePos.offset);
+#endif
     // allocate the buffer and keep the data of this row into the new allocated buffer
     pEntryInfo->numOfRes++;
     taosheapsort((void*)pItems, sizeof(STopBotResItem), pEntryInfo->numOfRes, (const void*)&type, topBotResComparFn,
@@ -3229,7 +3231,9 @@ void doAddIntoResult(SqlFunctionCtx* pCtx, void* pData, int32_t rowIndex, SSData
       if (pCtx->subsidiaries.num > 0) {
         copyTupleData(pCtx, rowIndex, pSrcBlock, &pItem->tuplePos);
       }
-
+#ifdef BUF_PAGE_DEBUG
+      qDebug("page_copyTuple pageId:%d, offset:%d", pItem->tuplePos.pageId, pItem->tuplePos.offset);
+#endif
       taosheapadjust((void*)pItems, sizeof(STopBotResItem), 0, pEntryInfo->numOfRes - 1, (const void*)&type,
                      topBotResComparFn, NULL, !isTopQuery);
     }
@@ -3239,7 +3243,11 @@ void doAddIntoResult(SqlFunctionCtx* pCtx, void* pData, int32_t rowIndex, SSData
 void saveTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pSrcBlock, STuplePos* pPos) {
   SFilePage* pPage = NULL;
 
-  int32_t completeRowSize = pSrcBlock->info.rowSize + (int32_t) taosArrayGetSize(pSrcBlock->pDataBlock) * sizeof(bool);
+  int32_t completeRowSize = pCtx->subsidiaries.num * sizeof(bool);
+  for (int32_t j = 0; j < pCtx->subsidiaries.num; ++j) {
+    SqlFunctionCtx* pc = pCtx->subsidiaries.pCtx[j];
+    completeRowSize += pc->pExpr->base.resSchema.bytes;
+  }
 
   if (pCtx->curBufPage == -1) {
     pPage = getNewBufPage(pCtx->pBuf, 0, &pCtx->curBufPage);
@@ -3257,19 +3265,22 @@ void saveTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pS
   // keep the current row data, extract method
   int32_t offset = 0;
   bool*   nullList = (bool*)((char*)pPage + pPage->num);
-  char*   pStart = (char*)(nullList + sizeof(bool) * (int32_t) taosArrayGetSize(pSrcBlock->pDataBlock));
-  for (int32_t i = 0; i < (int32_t) taosArrayGetSize(pSrcBlock->pDataBlock); ++i) {
-    SColumnInfoData* pCol = taosArrayGet(pSrcBlock->pDataBlock, i);
-    bool             isNull = colDataIsNull_s(pCol, rowIndex);
-    if (isNull) {
-      nullList[i] = true;
+  char*   pStart = (char*)(nullList + sizeof(bool) * pCtx->subsidiaries.num);
+  for (int32_t i = 0; i < pCtx->subsidiaries.num; ++i) {
+    SqlFunctionCtx* pc = pCtx->subsidiaries.pCtx[i];
+
+    SFunctParam* pFuncParam = &pc->pExpr->base.pParam[0];
+    int32_t      srcSlotId = pFuncParam->pCol->slotId;
+
+    SColumnInfoData* pCol = taosArrayGet(pSrcBlock->pDataBlock, srcSlotId);
+    if ((nullList[i] = colDataIsNull_s(pCol, rowIndex)) == true) {
       offset += pCol->info.bytes;
       continue;
     }
 
     char* p = colDataGetData(pCol, rowIndex);
     if (IS_VAR_DATA_TYPE(pCol->info.type)) {
-      memcpy(pStart + offset, p, varDataTLen(p));
+      memcpy(pStart + offset, p, (pCol->info.type == TSDB_DATA_TYPE_JSON) ? getJsonValueLen(p): varDataTLen(p));
     } else {
       memcpy(pStart + offset, p, pCol->info.bytes);
     }
@@ -3287,14 +3298,18 @@ void saveTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pS
 void copyTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pSrcBlock, STuplePos* pPos) {
   SFilePage* pPage = getBufPage(pCtx->pBuf, pPos->pageId);
 
-  int32_t numOfCols = taosArrayGetSize(pSrcBlock->pDataBlock);
+  int32_t numOfCols = pCtx->subsidiaries.num;
 
   bool* nullList = (bool*)((char*)pPage + pPos->offset);
   char* pStart = (char*)(nullList + numOfCols * sizeof(bool));
 
   int32_t offset = 0;
   for (int32_t i = 0; i < numOfCols; ++i) {
-    SColumnInfoData* pCol = taosArrayGet(pSrcBlock->pDataBlock, i);
+    SqlFunctionCtx* pc = pCtx->subsidiaries.pCtx[i];
+    SFunctParam* pFuncParam = &pc->pExpr->base.pParam[0];
+    int32_t      srcSlotId = pFuncParam->pCol->slotId;
+
+    SColumnInfoData* pCol = taosArrayGet(pSrcBlock->pDataBlock, srcSlotId);
     if ((nullList[i] = colDataIsNull_s(pCol, rowIndex)) == true) {
       offset += pCol->info.bytes;
       continue;
@@ -3302,7 +3317,7 @@ void copyTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pS
 
     char* p = colDataGetData(pCol, rowIndex);
     if (IS_VAR_DATA_TYPE(pCol->info.type)) {
-      memcpy(pStart + offset, p, varDataTLen(p));
+      memcpy(pStart + offset, p, (pCol->info.type == TSDB_DATA_TYPE_JSON) ? getJsonValueLen(p): varDataTLen(p));
     } else {
       memcpy(pStart + offset, p, pCol->info.bytes);
     }
@@ -3316,7 +3331,7 @@ void copyTupleData(SqlFunctionCtx* pCtx, int32_t rowIndex, const SSDataBlock* pS
 
 int32_t topBotFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) {
   SResultRowEntryInfo* pEntryInfo = GET_RES_INFO(pCtx);
-  STopBotRes*          pRes = GET_ROWCELL_INTERBUF(pEntryInfo);
+  STopBotRes*          pRes = getTopBotOutputInfo(pCtx);
 
   int16_t type = pCtx->input.pData[0]->info.type;
   int32_t slotId = pCtx->pExpr->base.resSchema.slotId;
@@ -3333,7 +3348,9 @@ int32_t topBotFinalize(SqlFunctionCtx* pCtx, SSDataBlock* pBlock) {
     } else {
       colDataAppend(pCol, currentRow, (const char*)&pItem->v.i, false);
     }
-
+#ifdef BUF_PAGE_DEBUG
+    qDebug("page_finalize i:%d,item:%p,pageId:%d, offset:%d\n", i, pItem, pItem->tuplePos.pageId, pItem->tuplePos.offset);
+#endif
     setSelectivityValue(pCtx, pBlock, &pRes->pItems[i].tuplePos, currentRow);
     currentRow += 1;
   }
@@ -5629,8 +5646,6 @@ int32_t groupKeyFunction(SqlFunctionCtx* pCtx) {
   SInputColumnInfoData* pInput = &pCtx->input;
   SColumnInfoData*   pInputCol = pInput->pData[0];
 
-  int32_t bytes = pInputCol->info.bytes;
-
   int32_t startIndex = pInput->startRowIndex;
 
   //escape rest of data blocks to avoid first entry to be overwritten.
@@ -5645,7 +5660,11 @@ int32_t groupKeyFunction(SqlFunctionCtx* pCtx) {
   }
 
   char* data = colDataGetData(pInputCol, startIndex);
-  memcpy(pInfo->data, data, bytes);
+  if (IS_VAR_DATA_TYPE(pInputCol->info.type)) {
+    memcpy(pInfo->data, data, (pInputCol->info.type == TSDB_DATA_TYPE_JSON) ? getJsonValueLen(data): varDataTLen(data));
+  } else {
+    memcpy(pInfo->data, data, pInputCol->info.bytes);
+  }
   pInfo->hasResult = true;
 
 _group_key_over:

source/libs/nodes/src/nodesCloneFuncs.c

@@ -18,6 +18,7 @@
 #include "querynodes.h"
 #include "taos.h"
 #include "taoserror.h"
+#include "tdatablock.h"
 
 #define COPY_SCALAR_FIELD(fldname)     \
   do {                                 \
@@ -164,7 +165,15 @@ static int32_t valueNodeCopy(const SValueNode* pSrc, SValueNode* pDst) {
       memcpy(pDst->datum.p, pSrc->datum.p, len);
       break;
     }
-    case TSDB_DATA_TYPE_JSON:
+    case TSDB_DATA_TYPE_JSON:{
+      int32_t len = getJsonValueLen(pSrc->datum.p);
+      pDst->datum.p = taosMemoryCalloc(1, len);
+      if (NULL == pDst->datum.p) {
+        return TSDB_CODE_OUT_OF_MEMORY;
+      }
+      memcpy(pDst->datum.p, pSrc->datum.p, len);
+      break;
+    }
     case TSDB_DATA_TYPE_DECIMAL:
     case TSDB_DATA_TYPE_BLOB:
     case TSDB_DATA_TYPE_MEDIUMBLOB:

source/libs/nodes/src/nodesCodeFuncs.c

@@ -20,6 +20,7 @@
 #include "querynodes.h"
 #include "taoserror.h"
 #include "tjson.h"
+#include "tdatablock.h"
 
 static int32_t nodeToJson(const void* pObj, SJson* pJson);
 static int32_t jsonToNode(const SJson* pJson, void* pObj);
@@ -2629,7 +2630,18 @@ static int32_t datumToJson(const void* pObj, SJson* pJson) {
     case TSDB_DATA_TYPE_VARBINARY:
       code = tjsonAddStringToObject(pJson, jkValueDatum, varDataVal(pNode->datum.p));
       break;
-    case TSDB_DATA_TYPE_JSON:
+    case TSDB_DATA_TYPE_JSON:{
+      int32_t len = getJsonValueLen(pNode->datum.p);
+      char* buf = taosMemoryCalloc( len * 2 + 1, sizeof(char));
+      code = taosHexEncode(pNode->datum.p, buf, len);
+      if (code != TSDB_CODE_SUCCESS) {
+        taosMemoryFree(buf);
+        return TSDB_CODE_TSC_INVALID_VALUE;
+      }
+      code = tjsonAddStringToObject(pJson, jkValueDatum, buf);
+      taosMemoryFree(buf);
+      break;
+    }
     case TSDB_DATA_TYPE_DECIMAL:
     case TSDB_DATA_TYPE_BLOB:
       // todo
@@ -2752,7 +2764,30 @@ static int32_t jsonToDatum(const SJson* pJson, void* pObj) {
       }
       break;
     }
-    case TSDB_DATA_TYPE_JSON:
+    case TSDB_DATA_TYPE_JSON:{
+      pNode->datum.p = taosMemoryCalloc(1, pNode->node.resType.bytes);
+      if (NULL == pNode->datum.p) {
+        code = TSDB_CODE_OUT_OF_MEMORY;
+        break;
+      }
+      char* buf = taosMemoryCalloc(1, pNode->node.resType.bytes * 2 + 1);
+      if (NULL == buf) {
+        code = TSDB_CODE_OUT_OF_MEMORY;
+        break;
+      }
+      code = tjsonGetStringValue(pJson, jkValueDatum, buf);
+      if (code != TSDB_CODE_SUCCESS) {
+        taosMemoryFree(buf);
+        break;
+      }
+      code = taosHexDecode(buf, pNode->datum.p, pNode->node.resType.bytes);
+      if (code != TSDB_CODE_SUCCESS) {
+        taosMemoryFree(buf);
+        break;
+      }
+      taosMemoryFree(buf);
+      break;
+    }
     case TSDB_DATA_TYPE_DECIMAL:
     case TSDB_DATA_TYPE_BLOB:
       // todo

source/libs/nodes/src/nodesUtilFuncs.c

@@ -20,6 +20,7 @@
 #include "taos.h"
 #include "taoserror.h"
 #include "thash.h"
+#include "tdatablock.h"
 
 static SNode* makeNode(ENodeType type, size_t size) {
   SNode* p = taosMemoryCalloc(1, size);
@@ -1675,6 +1676,10 @@ void nodesValueNodeToVariant(const SValueNode* pNode, SVariant* pVal) {
       pVal->pz[pVal->nLen + VARSTR_HEADER_SIZE] = 0;
       break;
     case TSDB_DATA_TYPE_JSON:
+      pVal->nLen = getJsonValueLen(pNode->datum.p);
+      pVal->pz = taosMemoryMalloc(pVal->nLen);
+      memcpy(pVal->pz, pNode->datum.p, pVal->nLen);
+      break;
     case TSDB_DATA_TYPE_DECIMAL:
     case TSDB_DATA_TYPE_BLOB:
       // todo

source/libs/parser/inc/parUtil.h

@@ -62,7 +62,6 @@ int32_t       getNumOfColumns(const STableMeta* pTableMeta);
 int32_t       getNumOfTags(const STableMeta* pTableMeta);
 STableComInfo getTableInfo(const STableMeta* pTableMeta);
 STableMeta*   tableMetaDup(const STableMeta* pTableMeta);
-int32_t       parseJsontoTagData(const char* json, SArray* pTagVals, STag** ppTag, SMsgBuf* pMsgBuf);
 
 int32_t trimString(const char* src, int32_t len, char* dst, int32_t dlen);
 

source/libs/parser/src/parUtil.c

@@ -219,6 +219,7 @@ int32_t buildInvalidOperationMsg(SMsgBuf* pBuf, const char* msg) {
 }
 
 int32_t buildSyntaxErrMsg(SMsgBuf* pBuf, const char* additionalInfo, const char* sourceStr) {
+  if(pBuf == NULL) return TSDB_CODE_TSC_SQL_SYNTAX_ERROR;
   const char* msgFormat1 = "syntax error near \'%s\'";
   const char* msgFormat2 = "syntax error near \'%s\' (%s)";
   const char* msgFormat3 = "%s";
@@ -346,7 +347,7 @@ static bool isValidateTag(char* input) {
   return true;
 }
 
-int32_t parseJsontoTagData(const char* json, SArray* pTagVals, STag** ppTag, SMsgBuf* pMsgBuf) {
+int32_t parseJsontoTagData(const char* json, SArray* pTagVals, STag** ppTag, void* pMsgBuf) {
   int32_t   retCode = TSDB_CODE_SUCCESS;
   cJSON*    root = NULL;
   SHashObj* keyHash = NULL;

source/libs/planner/src/planOptimizer.c

@@ -476,10 +476,16 @@ static int32_t pushDownCondOptPushCondToScan(SOptimizeContext* pCxt, SScanLogicN
   return pushDownCondOptAppendCond(&pScan->node.pConditions, pCond);
 }
 
+static int32_t pushDownCondOptPushCondToProject(SOptimizeContext* pCxt, SProjectLogicNode* pProject, SNode** pCond) {
+  return pushDownCondOptAppendCond(&pProject->node.pConditions, pCond);
+}
+
 static int32_t pushDownCondOptPushCondToChild(SOptimizeContext* pCxt, SLogicNode* pChild, SNode** pCond) {
   switch (nodeType(pChild)) {
     case QUERY_NODE_LOGIC_PLAN_SCAN:
       return pushDownCondOptPushCondToScan(pCxt, (SScanLogicNode*)pChild, pCond);
+    case QUERY_NODE_LOGIC_PLAN_PROJECT:
+      return pushDownCondOptPushCondToProject(pCxt, (SProjectLogicNode*)pChild, pCond);
     default:
       break;
   }
@@ -713,7 +719,8 @@ static int32_t pushDownCondOptDealAgg(SOptimizeContext* pCxt, SAggLogicNode* pAg
   }
   //TODO: remove it after full implementation of pushing down to child
   if (1 != LIST_LENGTH(pAgg->node.pChildren) ||
-      QUERY_NODE_LOGIC_PLAN_SCAN != nodeType(nodesListGetNode(pAgg->node.pChildren, 0))) {
+      QUERY_NODE_LOGIC_PLAN_SCAN != nodeType(nodesListGetNode(pAgg->node.pChildren, 0)) &&
+      QUERY_NODE_LOGIC_PLAN_PROJECT != nodeType(nodesListGetNode(pAgg->node.pChildren, 0))) {
     return TSDB_CODE_SUCCESS;
   }
   

source/libs/qcom/src/queryUtil.c

@@ -305,18 +305,21 @@ int32_t dataConverToStr(char* str, int type, void* buf, int32_t bufSize, int32_t
 
 char* parseTagDatatoJson(void* p) {
   char*  string = NULL;
-  cJSON* json = cJSON_CreateObject();
-  if (json == NULL) {
-    goto end;
-  }
-
   SArray* pTagVals = NULL;
+  cJSON* json = NULL;
   if (tTagToValArray((const STag*)p, &pTagVals) != 0) {
     goto end;
   }
 
   int16_t nCols = taosArrayGetSize(pTagVals);
+  if (nCols == 0) {
+    goto end;
+  }
   char    tagJsonKey[256] = {0};
+  json = cJSON_CreateObject();
+  if (json == NULL) {
+    goto end;
+  }
   for (int j = 0; j < nCols; ++j) {
     STagVal* pTagVal = (STagVal*)taosArrayGet(pTagVals, j);
     // json key  encode by binary
@@ -374,6 +377,10 @@ char* parseTagDatatoJson(void* p) {
   string = cJSON_PrintUnformatted(json);
 end:
   cJSON_Delete(json);
+  taosArrayDestroy(pTagVals);
+  if(string == NULL){
+    string = strdup(TSDB_DATA_NULL_STR_L);
+  }
   return string;
 }
 

source/libs/scalar/src/filter.c

@@ -192,6 +192,9 @@ int8_t filterGetCompFuncIdx(int32_t type, int32_t optr) {
       case TSDB_DATA_TYPE_DOUBLE:        
       case TSDB_DATA_TYPE_TIMESTAMP:        
         return 18;
+      case TSDB_DATA_TYPE_JSON:
+        terrno = TSDB_CODE_QRY_JSON_IN_ERROR;
+        return 0;
       default:
         assert(0);
     }
@@ -215,6 +218,9 @@ int8_t filterGetCompFuncIdx(int32_t type, int32_t optr) {
       case TSDB_DATA_TYPE_DOUBLE:        
       case TSDB_DATA_TYPE_TIMESTAMP:        
         return 24;
+      case TSDB_DATA_TYPE_JSON:
+        terrno = TSDB_CODE_QRY_JSON_IN_ERROR;
+        return 0;
       default:
         assert(0);
     }

source/libs/scalar/src/scalar.c

@@ -551,7 +551,9 @@ int32_t sclExecOperator(SOperatorNode *node, SScalarCtx *ctx, SScalarParam *outp
   SScalarParam* pLeft = &params[0];
   SScalarParam* pRight = paramNum > 1 ? &params[1] : NULL;
 
+  terrno = TSDB_CODE_SUCCESS;
   OperatorFn(pLeft, pRight, output, TSDB_ORDER_ASC);
+  code = terrno;
 
 _return:
   for (int32_t i = 0; i < paramNum; ++i) {
@@ -693,7 +695,11 @@ EDealRes sclRewriteFunction(SNode** pNode, SScalarCtx *ctx) {
     res->node.resType.scale = output.columnData->info.scale;
     res->node.resType.precision = output.columnData->info.precision;
     int32_t type = output.columnData->info.type;
-    if (IS_VAR_DATA_TYPE(type)) {
+    if (type == TSDB_DATA_TYPE_JSON){
+      int32_t len = getJsonValueLen(output.columnData->pData);
+      res->datum.p = taosMemoryCalloc(len, 1);
+      memcpy(res->datum.p, output.columnData->pData, len);
+    } else if (IS_VAR_DATA_TYPE(type)) {
       res->datum.p = taosMemoryCalloc(res->node.resType.bytes + VARSTR_HEADER_SIZE + 1, 1);
       memcpy(res->datum.p, output.columnData->pData, varDataTLen(output.columnData->pData));
     } else {

source/libs/scalar/src/sclfunc.c

@@ -1152,42 +1152,30 @@ int32_t toJsonFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOu
 
   char tmp[TSDB_MAX_JSON_TAG_LEN] = {0};
   for (int32_t i = 0; i < pInput[0].numOfRows; ++i) {
-    if (colDataIsNull_s(pInput[0].columnData, i)) {
-      colDataAppendNULL(pOutput->columnData, i);
-      continue;
-    }
-    char *input = pInput[0].columnData->pData + pInput[0].columnData->varmeta.offset[i];
+    SArray* pTagVals = taosArrayInit(8, sizeof(STagVal));
+    STag*   pTag = NULL;
 
-    if(type == TSDB_DATA_TYPE_NCHAR){
-      if (varDataTLen(input) > TSDB_MAX_JSON_TAG_LEN){
-        colDataAppendNULL(pOutput->columnData, i);
-        continue;
-      }
-      int32_t len  = taosUcs4ToMbs((TdUcs4 *)varDataVal(input), varDataLen(input), tmp);
-      if (len < 0) {
-        colDataAppendNULL(pOutput->columnData, i);
-        continue;
-      }
-      tmp[len] = 0;
+    if (colDataIsNull_s(pInput[0].columnData, i)) {
+      tTagNew(pTagVals, 1, true, &pTag);
     }else{
+      char *input = pInput[0].columnData->pData + pInput[0].columnData->varmeta.offset[i];
       if (varDataLen(input) > (TSDB_MAX_JSON_TAG_LEN - VARSTR_HEADER_SIZE) / TSDB_NCHAR_SIZE){
-        colDataAppendNULL(pOutput->columnData, i);
-        continue;
+        taosArrayDestroy(pTagVals);
+        return TSDB_CODE_FAILED;
       }
       memcpy(tmp, varDataVal(input), varDataLen(input));
       tmp[varDataLen(input)] = 0;
+      if(parseJsontoTagData(tmp, pTagVals, &pTag, NULL)){
+        tTagNew(pTagVals, 1, true, &pTag);
+      }
     }
 
-    if(!tjsonValidateJson(tmp)){
-      colDataAppendNULL(pOutput->columnData, i);
-      continue;
-    }
-
-    colDataAppend(pOutput->columnData, i, input, false);
+    colDataAppend(pOutput->columnData, i, (const char*)pTag, false);
+    tTagFree(pTag);
+    taosArrayDestroy(pTagVals);
   }
 
   pOutput->numOfRows = pInput->numOfRows;
-
   return TSDB_CODE_SUCCESS;
 }
 

source/libs/scalar/src/sclvector.c

@@ -160,6 +160,9 @@ int64_t getVectorBigintValue_JSON(void *src, int32_t index){
     return 0;
   } else if(*data == TSDB_DATA_TYPE_NCHAR) {   // json inner type can not be BINARY
     convertNcharToDouble(data+CHAR_BYTES, &out);
+  } else if(tTagIsJson(data)){
+    terrno = TSDB_CODE_QRY_JSON_NOT_SUPPORT_ERROR;
+    return 0;
   } else {
     convertNumberToNumber(data+CHAR_BYTES, &out, *data, TSDB_DATA_TYPE_DOUBLE);
   }
@@ -416,6 +419,9 @@ int32_t vectorConvertFromVarData(const SScalarParam* pIn, SScalarParam* pOut, in
       else if(*data == TSDB_DATA_TYPE_NCHAR) {
         data += CHAR_BYTES;
         convertType = TSDB_DATA_TYPE_NCHAR;
+      } else if(tTagIsJson(data)){
+        terrno = TSDB_CODE_QRY_JSON_NOT_SUPPORT_ERROR;
+        return terrno;
       } else {
         convertNumberToNumber(data+CHAR_BYTES, colDataGetNumData(pOut->columnData, i), *data, outType);
         continue;
@@ -461,7 +467,10 @@ double getVectorDoubleValue_JSON(void *src, int32_t index){
     return out;
   } else if(*data == TSDB_DATA_TYPE_NCHAR) {   // json inner type can not be BINARY
     convertNcharToDouble(data+CHAR_BYTES, &out);
-  } else {
+  } else if(tTagIsJson(data)){
+    terrno = TSDB_CODE_QRY_JSON_NOT_SUPPORT_ERROR;
+    return 0;
+  } else{
     convertNumberToNumber(data+CHAR_BYTES, &out, *data, TSDB_DATA_TYPE_DOUBLE);
   }
   return out;
@@ -493,10 +502,18 @@ bool convertJsonValue(__compar_fn_t *fp, int32_t optr, int8_t typeLeft, int8_t t
   }
 
   if(typeLeft == TSDB_DATA_TYPE_JSON){
+    if(tTagIsJson(*pLeftData)){
+      terrno = TSDB_CODE_QRY_JSON_NOT_SUPPORT_ERROR;
+      return false;
+    }
     typeLeft = **pLeftData;
     (*pLeftData) ++;
   }
   if(typeRight == TSDB_DATA_TYPE_JSON){
+    if(tTagIsJson(*pLeftData)){
+      terrno = TSDB_CODE_QRY_JSON_NOT_SUPPORT_ERROR;
+      return false;
+    }
     typeRight = **pRightData;
     (*pRightData) ++;
   }
@@ -1576,7 +1593,11 @@ void vectorBitOr(SScalarParam* pLeft, SScalarParam* pRight, SScalarParam *pOut,
 void vectorCompareImpl(SScalarParam* pLeft, SScalarParam* pRight, SScalarParam *pOut, int32_t _ord, int32_t optr) {
   int32_t       i = ((_ord) == TSDB_ORDER_ASC) ? 0 : TMAX(pLeft->numOfRows, pRight->numOfRows) - 1;
   int32_t       step = ((_ord) == TSDB_ORDER_ASC) ? 1 : -1;
+
   __compar_fn_t fp = filterGetCompFunc(GET_PARAM_TYPE(pLeft), optr);
+  if(terrno != TSDB_CODE_SUCCESS){
+    return;
+  }
 
   pOut->numOfRows = TMAX(pLeft->numOfRows, pRight->numOfRows);
 
@@ -1709,6 +1730,7 @@ void vectorIsTrue(SScalarParam* pLeft, SScalarParam* pRight, SScalarParam *pOut,
 STagVal getJsonValue(char *json, char *key, bool *isExist) {
   STagVal val = {.pKey = key};
   if (tTagIsJson((const STag *)json) == false){
+    terrno = TSDB_CODE_QRY_JSON_NOT_SUPPORT_ERROR;
     if(isExist){
       *isExist = false;
     }

source/libs/stream/src/streamData.c

@@ -88,3 +88,12 @@ SStreamDataSubmit* streamSubmitRefClone(SStreamDataSubmit* pSubmit) {
   memcpy(pSubmitClone, pSubmit, sizeof(SStreamDataSubmit));
   return pSubmitClone;
 }
+
+void streamDataSubmitRefDec(SStreamDataSubmit* pDataSubmit) {
+  int32_t ref = atomic_sub_fetch_32(pDataSubmit->dataRef, 1);
+  ASSERT(ref >= 0);
+  if (ref == 0) {
+    taosMemoryFree(pDataSubmit->data);
+    taosMemoryFree(pDataSubmit->dataRef);
+  }
+}

source/libs/stream/src/streamExec.c

@@ -63,7 +63,6 @@ static int32_t streamTaskExecImpl(SStreamTask* pTask, void* data, SArray* pRes)
       continue;
     }
 
-    // TODO: do we need free memory?
     SSDataBlock block = {0};
     assignOneDataBlock(&block, output);
     block.info.childId = pTask->selfChildId;

source/libs/stream/src/streamQueue.c

@@ -41,12 +41,3 @@ void streamQueueClose(SStreamQueue* queue) {
       return;
   }
 }
-
-void streamDataSubmitRefDec(SStreamDataSubmit* pDataSubmit) {
-  int32_t ref = atomic_sub_fetch_32(pDataSubmit->dataRef, 1);
-  ASSERT(ref >= 0);
-  if (ref == 0) {
-    taosMemoryFree(pDataSubmit->data);
-    taosMemoryFree(pDataSubmit->dataRef);
-  }
-}

source/libs/sync/inc/syncAppendEntries.h

@@ -94,6 +94,7 @@ extern "C" {
 //
 int32_t syncNodeOnAppendEntriesCb(SSyncNode* ths, SyncAppendEntries* pMsg);
 int32_t syncNodeOnAppendEntriesSnapshotCb(SSyncNode* ths, SyncAppendEntries* pMsg);
+int32_t syncNodeOnAppendEntriesSnapshot2Cb(SSyncNode* ths, SyncAppendEntriesBatch* pMsg);
 
 #ifdef __cplusplus
 }

source/libs/sync/inc/syncAppendEntriesReply.h

@@ -42,6 +42,12 @@ extern "C" {
 //
 int32_t syncNodeOnAppendEntriesReplyCb(SSyncNode* ths, SyncAppendEntriesReply* pMsg);
 int32_t syncNodeOnAppendEntriesReplySnapshotCb(SSyncNode* ths, SyncAppendEntriesReply* pMsg);
+int32_t syncNodeOnAppendEntriesReplySnapshot2Cb(SSyncNode* ths, SyncAppendEntriesReply* pMsg);
+
+typedef struct SReaderParam {
+  SyncIndex start;
+  SyncIndex end;
+} SReaderParam;
 
 #ifdef __cplusplus
 }

source/libs/sync/inc/syncInt.h

@@ -67,6 +67,7 @@ typedef struct SSyncNode {
   char        path[TSDB_FILENAME_LEN];
   char        raftStorePath[TSDB_FILENAME_LEN * 2];
   char        configPath[TSDB_FILENAME_LEN * 2];
+  int32_t     batchSize;
 
   // sync io
   SWal*         pWal;
@@ -170,6 +171,7 @@ void       syncNodeStart(SSyncNode* pSyncNode);
 void       syncNodeStartStandBy(SSyncNode* pSyncNode);
 void       syncNodeClose(SSyncNode* pSyncNode);
 int32_t    syncNodePropose(SSyncNode* pSyncNode, SRpcMsg* pMsg, bool isWeak);
+int32_t    syncNodeProposeBatch(SSyncNode* pSyncNode, SRpcMsg* pMsgArr, bool* pIsWeakArr, int32_t arrSize);
 
 // option
 bool      syncNodeSnapshotEnable(SSyncNode* pSyncNode);

source/libs/sync/inc/syncReplication.h

@@ -53,8 +53,10 @@ extern "C" {
 //
 int32_t syncNodeAppendEntriesPeers(SSyncNode* pSyncNode);
 int32_t syncNodeAppendEntriesPeersSnapshot(SSyncNode* pSyncNode);
+int32_t syncNodeAppendEntriesPeersSnapshot2(SSyncNode* pSyncNode);
 int32_t syncNodeReplicate(SSyncNode* pSyncNode);
 int32_t syncNodeAppendEntries(SSyncNode* pSyncNode, const SRaftId* destRaftId, const SyncAppendEntries* pMsg);
+int32_t syncNodeAppendEntriesBatch(SSyncNode* pSyncNode, const SRaftId* destRaftId, const SyncAppendEntriesBatch* pMsg);
 
 #ifdef __cplusplus
 }

source/libs/sync/inc/syncSnapshot.h

@@ -40,8 +40,8 @@ typedef struct SSyncSnapshotSender {
   bool       start;
   int32_t    seq;
   int32_t    ack;
-  void      *pReader;
-  void      *pCurrentBlock;
+  void *     pReader;
+  void *     pCurrentBlock;
   int32_t    blockLen;
   SSnapshot  snapshot;
   SSyncCfg   lastConfig;
@@ -62,14 +62,14 @@ int32_t              snapshotSend(SSyncSnapshotSender *pSender);
 int32_t              snapshotReSend(SSyncSnapshotSender *pSender);
 
 cJSON *snapshotSender2Json(SSyncSnapshotSender *pSender);
-char  *snapshotSender2Str(SSyncSnapshotSender *pSender);
-char  *snapshotSender2SimpleStr(SSyncSnapshotSender *pSender, char *event);
+char * snapshotSender2Str(SSyncSnapshotSender *pSender);
+char * snapshotSender2SimpleStr(SSyncSnapshotSender *pSender, char *event);
 
 //---------------------------------------------------
 typedef struct SSyncSnapshotReceiver {
   bool       start;
   int32_t    ack;
-  void      *pWriter;
+  void *     pWriter;
   SyncTerm   term;
   SyncTerm   privateTerm;
   SSnapshot  snapshot;
@@ -85,8 +85,8 @@ bool snapshotReceiverIsStart(SSyncSnapshotReceiver *pReceiver);
 void snapshotReceiverStop(SSyncSnapshotReceiver *pReceiver);
 
 cJSON *snapshotReceiver2Json(SSyncSnapshotReceiver *pReceiver);
-char  *snapshotReceiver2Str(SSyncSnapshotReceiver *pReceiver);
-char  *snapshotReceiver2SimpleStr(SSyncSnapshotReceiver *pReceiver, char *event);
+char * snapshotReceiver2Str(SSyncSnapshotReceiver *pReceiver);
+char * snapshotReceiver2SimpleStr(SSyncSnapshotReceiver *pReceiver, char *event);
 
 //---------------------------------------------------
 // on message

source/libs/sync/src/syncAppendEntries.c

@@ -719,6 +719,8 @@ static bool syncNodeOnAppendEntriesLogOK(SSyncNode* pSyncNode, SyncAppendEntries
   return false;
 }
 
+int32_t syncNodeOnAppendEntriesSnapshot2Cb(SSyncNode* ths, SyncAppendEntriesBatch* pMsg) { return 0; }
+
 int32_t syncNodeOnAppendEntriesSnapshotCb(SSyncNode* ths, SyncAppendEntries* pMsg) {
   int32_t ret = 0;
   int32_t code = 0;

source/libs/sync/src/syncAppendEntriesReply.c

@@ -108,48 +108,82 @@ int32_t syncNodeOnAppendEntriesReplyCb(SSyncNode* ths, SyncAppendEntriesReply* p
   return ret;
 }
 
-#if 0
-int32_t syncNodeOnAppendEntriesReplyCb(SSyncNode* ths, SyncAppendEntriesReply* pMsg) {
+// only start once
+static void syncNodeStartSnapshot(SSyncNode* ths, SyncIndex beginIndex, SyncIndex endIndex, SyncTerm lastApplyTerm,
+                                  SyncAppendEntriesReply* pMsg) {
+  // get sender
+  SSyncSnapshotSender* pSender = syncNodeGetSnapshotSender(ths, &(pMsg->srcId));
+  ASSERT(pSender != NULL);
+
+  SSnapshot snapshot = {
+      .data = NULL, .lastApplyIndex = endIndex, .lastApplyTerm = lastApplyTerm, .lastConfigIndex = SYNC_INDEX_INVALID};
+
+  void*        pReader = NULL;
+  SReaderParam readerParam = {.start = beginIndex, .end = endIndex};
+  ths->pFsm->FpSnapshotStartRead(ths->pFsm, &readerParam, &pReader);
+  if (!snapshotSenderIsStart(pSender) && pMsg->privateTerm < pSender->privateTerm) {
+    ASSERT(pReader != NULL);
+    snapshotSenderStart(pSender, snapshot, pReader);
+
+  } else {
+    if (pReader != NULL) {
+      ths->pFsm->FpSnapshotStopRead(ths->pFsm, pReader);
+    }
+  }
+}
+
+int32_t syncNodeOnAppendEntriesReplySnapshot2Cb(SSyncNode* ths, SyncAppendEntriesReply* pMsg) {
   int32_t ret = 0;
 
-  char logBuf[128] = {0};
-  snprintf(logBuf, sizeof(logBuf), "==syncNodeOnAppendEntriesReplyCb== term:%lu", ths->pRaftStore->currentTerm);
-  syncAppendEntriesReplyLog2(logBuf, pMsg);
+  // if already drop replica, do not process
+  if (!syncNodeInRaftGroup(ths, &(pMsg->srcId)) && !ths->pRaftCfg->isStandBy) {
+    syncNodeEventLog(ths, "recv sync-append-entries-reply,  maybe replica already dropped");
+    return -1;
+  }
 
+  // drop stale response
   if (pMsg->term < ths->pRaftStore->currentTerm) {
-    sTrace("DropStaleResponse, receive term:%" PRIu64 ", current term:%" PRIu64 "", pMsg->term,
-           ths->pRaftStore->currentTerm);
-    return ret;
+    char logBuf[128];
+    snprintf(logBuf, sizeof(logBuf), "recv sync-append-entries-reply, recv-term:%lu, drop stale response", pMsg->term);
+    syncNodeEventLog(ths, logBuf);
+    return -1;
   }
 
-  syncIndexMgrLog2("==syncNodeOnAppendEntriesReplyCb== before pNextIndex", ths->pNextIndex);
-  syncIndexMgrLog2("==syncNodeOnAppendEntriesReplyCb== before pMatchIndex", ths->pMatchIndex);
-
-  // no need this code, because if I receive reply.term, then I must have sent for that term.
-  //  if (pMsg->term > ths->pRaftStore->currentTerm) {
-  //    syncNodeUpdateTerm(ths, pMsg->term);
-  //  }
-
+  // error term
   if (pMsg->term > ths->pRaftStore->currentTerm) {
-    char logBuf[128] = {0};
-    snprintf(logBuf, sizeof(logBuf), "syncNodeOnAppendEntriesReplyCb error term, receive:%lu current:%lu", pMsg->term,
-             ths->pRaftStore->currentTerm);
-    syncNodeLog2(logBuf, ths);
-    sError("%s", logBuf);
-    return ret;
+    char logBuf[128];
+    snprintf(logBuf, sizeof(logBuf), "recv sync-append-entries-reply, error term, recv-term:%lu", pMsg->term);
+    syncNodeErrorLog(ths, logBuf);
+    return -1;
   }
 
   ASSERT(pMsg->term == ths->pRaftStore->currentTerm);
 
   if (pMsg->success) {
-    // nextIndex'  = [nextIndex  EXCEPT ![i][j] = m.mmatchIndex + 1]
-    syncIndexMgrSetIndex(ths->pNextIndex, &(pMsg->srcId), pMsg->matchIndex + 1);
+    SyncIndex newNextIndex = pMsg->matchIndex + 1;
+    SyncIndex newMatchIndex = pMsg->matchIndex;
 
-    // matchIndex' = [matchIndex EXCEPT ![i][j] = m.mmatchIndex]
-    syncIndexMgrSetIndex(ths->pMatchIndex, &(pMsg->srcId), pMsg->matchIndex);
+    if (ths->pLogStore->syncLogExist(ths->pLogStore, newNextIndex) &&
+        ths->pLogStore->syncLogExist(ths->pLogStore, newNextIndex - 1)) {
+      // nextIndex'  = [nextIndex  EXCEPT ![i][j] = m.mmatchIndex + 1]
+      syncIndexMgrSetIndex(ths->pNextIndex, &(pMsg->srcId), newNextIndex);
 
-    // maybe commit
-    syncMaybeAdvanceCommitIndex(ths);
+      // matchIndex' = [matchIndex EXCEPT ![i][j] = m.mmatchIndex]
+      syncIndexMgrSetIndex(ths->pMatchIndex, &(pMsg->srcId), newMatchIndex);
+
+      // maybe commit
+      if (ths->state == TAOS_SYNC_STATE_LEADER) {
+        syncMaybeAdvanceCommitIndex(ths);
+      }
+    } else {
+      // start snapshot <match+1, old snapshot.end>
+      SSnapshot snapshot;
+      ths->pFsm->FpGetSnapshotInfo(ths->pFsm, &snapshot);
+      syncNodeStartSnapshot(ths, newMatchIndex + 1, snapshot.lastApplyIndex, snapshot.lastApplyTerm, pMsg);
+
+      syncIndexMgrSetIndex(ths->pNextIndex, &(pMsg->srcId), snapshot.lastApplyIndex + 1);
+      syncIndexMgrSetIndex(ths->pMatchIndex, &(pMsg->srcId), newMatchIndex);
+    }
 
   } else {
     SyncIndex nextIndex = syncIndexMgrGetIndex(ths->pNextIndex, &(pMsg->srcId));
@@ -157,18 +191,35 @@ int32_t syncNodeOnAppendEntriesReplyCb(SSyncNode* ths, SyncAppendEntriesReply* p
     // notice! int64, uint64
     if (nextIndex > SYNC_INDEX_BEGIN) {
       --nextIndex;
+
+      if (ths->pLogStore->syncLogExist(ths->pLogStore, nextIndex) &&
+          ths->pLogStore->syncLogExist(ths->pLogStore, nextIndex - 1)) {
+        // do nothing
+      } else {
+        SSyncRaftEntry* pEntry;
+        int32_t         code = ths->pLogStore->syncLogGetEntry(ths->pLogStore, nextIndex, &pEntry);
+        ASSERT(code == 0);
+        syncNodeStartSnapshot(ths, SYNC_INDEX_BEGIN, nextIndex, pEntry->term, pMsg);
+
+        // get sender
+        SSyncSnapshotSender* pSender = syncNodeGetSnapshotSender(ths, &(pMsg->srcId));
+        ASSERT(pSender != NULL);
+        SyncIndex sentryIndex = pSender->snapshot.lastApplyIndex + 1;
+
+        // update nextIndex to sentryIndex
+        if (nextIndex <= sentryIndex) {
+          nextIndex = sentryIndex;
+        }
+      }
+
     } else {
       nextIndex = SYNC_INDEX_BEGIN;
     }
     syncIndexMgrSetIndex(ths->pNextIndex, &(pMsg->srcId), nextIndex);
   }
 
-  syncIndexMgrLog2("==syncNodeOnAppendEntriesReplyCb== after pNextIndex", ths->pNextIndex);
-  syncIndexMgrLog2("==syncNodeOnAppendEntriesReplyCb== after pMatchIndex", ths->pMatchIndex);
-
-  return ret;
+  return 0;
 }
-#endif
 
 int32_t syncNodeOnAppendEntriesReplySnapshotCb(SSyncNode* ths, SyncAppendEntriesReply* pMsg) {
   int32_t ret = 0;

source/libs/sync/src/syncMain.c

@@ -50,7 +50,6 @@ static int32_t syncNodeAppendNoop(SSyncNode* ths);
 // process message ----
 int32_t syncNodeOnPingCb(SSyncNode* ths, SyncPing* pMsg);
 int32_t syncNodeOnPingReplyCb(SSyncNode* ths, SyncPingReply* pMsg);
-int32_t syncNodeOnClientRequestCb(SSyncNode* ths, SyncClientRequest* pMsg, SyncIndex* pRetIndex);
 
 // life cycle
 static void syncFreeNode(void* param);
@@ -627,6 +626,94 @@ int32_t syncPropose(int64_t rid, SRpcMsg* pMsg, bool isWeak) {
   return ret;
 }
 
+int32_t syncProposeBatch(int64_t rid, SRpcMsg* pMsgArr, bool* pIsWeakArr, int32_t arrSize) {
+  if (arrSize < 0) {
+    terrno = TSDB_CODE_SYN_INTERNAL_ERROR;
+    return -1;
+  }
+
+  int32_t    ret = 0;
+  SSyncNode* pSyncNode = taosAcquireRef(tsNodeRefId, rid);
+  if (pSyncNode == NULL) {
+    terrno = TSDB_CODE_SYN_INTERNAL_ERROR;
+    return -1;
+  }
+  ASSERT(rid == pSyncNode->rid);
+  ret = syncNodeProposeBatch(pSyncNode, pMsgArr, pIsWeakArr, arrSize);
+
+  taosReleaseRef(tsNodeRefId, pSyncNode->rid);
+  return ret;
+}
+
+static bool syncNodeBatchOK(SRpcMsg* pMsgArr, int32_t arrSize) {
+  for (int32_t i = 0; i < arrSize; ++i) {
+    if (pMsgArr[i].msgType == TDMT_SYNC_CONFIG_CHANGE) {
+      return false;
+    }
+
+    if (pMsgArr[i].msgType == TDMT_SYNC_CONFIG_CHANGE_FINISH) {
+      return false;
+    }
+  }
+
+  return true;
+}
+
+int32_t syncNodeProposeBatch(SSyncNode* pSyncNode, SRpcMsg* pMsgArr, bool* pIsWeakArr, int32_t arrSize) {
+  if (!syncNodeBatchOK(pMsgArr, arrSize)) {
+    syncNodeErrorLog(pSyncNode, "sync propose batch error");
+    terrno = TSDB_CODE_SYN_BATCH_ERROR;
+    return -1;
+  }
+
+  if (arrSize > SYNC_MAX_BATCH_SIZE) {
+    syncNodeErrorLog(pSyncNode, "sync propose match batch error");
+    terrno = TSDB_CODE_SYN_BATCH_ERROR;
+    return -1;
+  }
+
+  if (pSyncNode->state == TAOS_SYNC_STATE_LEADER) {
+    syncNodeErrorLog(pSyncNode, "sync propose not leader");
+    terrno = TSDB_CODE_SYN_NOT_LEADER;
+    return -1;
+  }
+
+  if (pSyncNode->changing) {
+    syncNodeErrorLog(pSyncNode, "sync propose not ready");
+    terrno = TSDB_CODE_SYN_PROPOSE_NOT_READY;
+    return -1;
+  }
+
+  SRaftMeta raftArr[SYNC_MAX_BATCH_SIZE];
+  for (int i = 0; i < arrSize; ++i) {
+    SRespStub stub;
+    stub.createTime = taosGetTimestampMs();
+    stub.rpcMsg = pMsgArr[i];
+    uint64_t seqNum = syncRespMgrAdd(pSyncNode->pSyncRespMgr, &stub);
+
+    raftArr[i].isWeak = pIsWeakArr[i];
+    raftArr[i].seqNum = seqNum;
+  }
+
+  SyncClientRequestBatch* pSyncMsg = syncClientRequestBatchBuild(pMsgArr, raftArr, arrSize, pSyncNode->vgId);
+  ASSERT(pSyncMsg != NULL);
+
+  SRpcMsg rpcMsg;
+  syncClientRequestBatch2RpcMsg(pSyncMsg, &rpcMsg);
+  taosMemoryFree(pSyncMsg);  // only free msg body, do not free rpc msg content
+
+  if (pSyncNode->FpEqMsg != NULL && (*pSyncNode->FpEqMsg)(pSyncNode->msgcb, &rpcMsg) == 0) {
+    // enqueue msg ok
+
+  } else {
+    sError("enqueue msg error, FpEqMsg is NULL");
+    terrno = TSDB_CODE_SYN_INTERNAL_ERROR;
+    return -1;
+  }
+
+  return 0;
+}
+
 int32_t syncNodePropose(SSyncNode* pSyncNode, SRpcMsg* pMsg, bool isWeak) {
   int32_t ret = 0;
 
@@ -2362,6 +2449,49 @@ int32_t syncNodeOnClientRequestCb(SSyncNode* ths, SyncClientRequest* pMsg, SyncI
   return ret;
 }
 
+int32_t syncNodeOnClientRequestBatchCb(SSyncNode* ths, SyncClientRequestBatch* pMsg) {
+  int32_t code = 0;
+
+  if (ths->state != TAOS_SYNC_STATE_LEADER) {
+    // call FpCommitCb, delete resp mgr
+    return -1;
+  }
+
+  SyncIndex index = ths->pLogStore->syncLogWriteIndex(ths->pLogStore);
+  SyncTerm  term = ths->pRaftStore->currentTerm;
+
+  int32_t    raftMetaArrayLen = sizeof(SRaftMeta) * pMsg->dataCount;
+  int32_t    rpcArrayLen = sizeof(SRpcMsg) * pMsg->dataCount;
+  SRaftMeta* raftMetaArr = (SRaftMeta*)(pMsg->data);
+  SRpcMsg*   msgArr = (SRpcMsg*)((char*)(pMsg->data) + raftMetaArrayLen);
+  for (int32_t i = 0; i < pMsg->dataCount; ++i) {
+    SSyncRaftEntry* pEntry = syncEntryBuild(msgArr[i].contLen);
+    ASSERT(pEntry != NULL);
+
+    pEntry->originalRpcType = msgArr[i].msgType;
+    pEntry->seqNum = raftMetaArr[i].seqNum;
+    pEntry->isWeak = raftMetaArr[i].isWeak;
+    pEntry->term = term;
+    pEntry->index = index;
+    memcpy(pEntry->data, msgArr[i].pCont, msgArr[i].contLen);
+    ASSERT(msgArr[i].contLen == pEntry->dataLen);
+
+    code = ths->pLogStore->syncLogAppendEntry(ths->pLogStore, pEntry);
+    if (code != 0) {
+      // del resp mgr, call FpCommitCb
+      ASSERT(0);
+      return -1;
+    }
+  }
+
+  // fsync once
+  SSyncLogStoreData* pData = ths->pLogStore->data;
+  SWal*              pWal = pData->pWal;
+  walFsync(pWal, true);
+
+  return 0;
+}
+
 static void syncFreeNode(void* param) {
   SSyncNode* pNode = param;
   // inner object already free

source/libs/sync/src/syncMessage.c

@@ -591,7 +591,7 @@ void syncPingReplySerialize(const SyncPingReply* pMsg, char* buf, uint32_t bufLe
 void syncPingReplyDeserialize(const char* buf, uint32_t len, SyncPingReply* pMsg) {
   memcpy(pMsg, buf, len);
   ASSERT(len == pMsg->bytes);
-  ASSERT(pMsg->bytes == sizeof(SyncPing) + pMsg->dataLen);
+  ASSERT(pMsg->bytes == sizeof(SyncPingReply) + pMsg->dataLen);
 }
 
 char* syncPingReplySerialize2(const SyncPingReply* pMsg, uint32_t* len) {
@@ -956,6 +956,48 @@ void syncClientRequestLog2(char* s, const SyncClientRequest* pMsg) {
   }
 }
 
+// ---- message process SyncClientRequestBatch----
+
+// block1:
+// block2: SRaftMeta array
+// block3: rpc msg array (with pCont)
+
+SyncClientRequestBatch* syncClientRequestBatchBuild(SRpcMsg* rpcMsgArr, SRaftMeta* raftArr, int32_t arrSize,
+                                                    int32_t vgId) {
+  ASSERT(rpcMsgArr != NULL);
+  ASSERT(arrSize > 0);
+
+  int32_t dataLen = 0;
+  int32_t raftMetaArrayLen = sizeof(SRaftMeta) * arrSize;
+  int32_t rpcArrayLen = sizeof(SRpcMsg) * arrSize;
+  dataLen += (raftMetaArrayLen + rpcArrayLen);
+
+  uint32_t                bytes = sizeof(SyncClientRequestBatch) + dataLen;
+  SyncClientRequestBatch* pMsg = taosMemoryMalloc(bytes);
+  memset(pMsg, 0, bytes);
+  pMsg->bytes = bytes;
+  pMsg->vgId = vgId;
+  pMsg->msgType = TDMT_SYNC_CLIENT_REQUEST_BATCH;
+  pMsg->dataCount = arrSize;
+  pMsg->dataLen = dataLen;
+
+  SRaftMeta* raftMetaArr = (SRaftMeta*)(pMsg->data);
+  SRpcMsg*   msgArr = (SRpcMsg*)((char*)(pMsg->data) + raftMetaArrayLen);
+
+  for (int i = 0; i < arrSize; ++i) {
+    // init raftMetaArr
+    raftMetaArr[i].isWeak = raftArr[i].isWeak;
+    raftMetaArr[i].seqNum = raftArr[i].seqNum;
+
+    // init msgArr
+    msgArr[i] = rpcMsgArr[i];
+  }
+
+  return pMsg;
+}
+
+void syncClientRequestBatch2RpcMsg(const SyncClientRequestBatch* pSyncMsg, SRpcMsg* pRpcMsg) {}
+
 // ---- message process SyncRequestVote----
 SyncRequestVote* syncRequestVoteBuild(int32_t vgId) {
   uint32_t         bytes = sizeof(SyncRequestVote);
@@ -1426,6 +1468,279 @@ void syncAppendEntriesLog2(char* s, const SyncAppendEntries* pMsg) {
   }
 }
 
+// ---- message process SyncAppendEntriesBatch----
+
+// block1: SOffsetAndContLen
+// block2: SOffsetAndContLen Array
+// block3: SRpcMsg Array
+// block4: SRpcMsg pCont Array
+
+SyncAppendEntriesBatch* syncAppendEntriesBatchBuild(SRpcMsg* rpcMsgArr, int32_t arrSize, int32_t vgId) {
+  ASSERT(rpcMsgArr != NULL);
+  ASSERT(arrSize > 0);
+
+  int32_t dataLen = 0;
+  int32_t metaArrayLen = sizeof(SOffsetAndContLen) * arrSize;  // <offset, contLen>
+  int32_t rpcArrayLen = sizeof(SRpcMsg) * arrSize;             // SRpcMsg
+  int32_t contArrayLen = 0;
+  for (int i = 0; i < arrSize; ++i) {  // SRpcMsg pCont
+    contArrayLen += rpcMsgArr[i].contLen;
+  }
+  dataLen += (metaArrayLen + rpcArrayLen + contArrayLen);
+
+  uint32_t                bytes = sizeof(SyncAppendEntriesBatch) + dataLen;
+  SyncAppendEntriesBatch* pMsg = taosMemoryMalloc(bytes);
+  memset(pMsg, 0, bytes);
+  pMsg->bytes = bytes;
+  pMsg->vgId = vgId;
+  pMsg->msgType = TDMT_SYNC_APPEND_ENTRIES_BATCH;
+  pMsg->dataCount = arrSize;
+  pMsg->dataLen = dataLen;
+
+  SOffsetAndContLen* metaArr = (SOffsetAndContLen*)(pMsg->data);
+  SRpcMsg*           msgArr = (SRpcMsg*)((char*)(pMsg->data) + metaArrayLen);
+  char*              pData = pMsg->data;
+
+  for (int i = 0; i < arrSize; ++i) {
+    // init <offset, contLen>
+    if (i == 0) {
+      metaArr[i].offset = metaArrayLen + rpcArrayLen;
+      metaArr[i].contLen = rpcMsgArr[i].contLen;
+    } else {
+      metaArr[i].offset = metaArr[i - 1].offset + metaArr[i - 1].contLen;
+      metaArr[i].contLen = rpcMsgArr[i].contLen;
+    }
+
+    // init msgArr
+    msgArr[i] = rpcMsgArr[i];
+
+    // init data
+    ASSERT(rpcMsgArr[i].contLen == metaArr[i].contLen);
+    memcpy(pData + metaArr[i].offset, rpcMsgArr[i].pCont, rpcMsgArr[i].contLen);
+  }
+
+  return pMsg;
+}
+
+void syncAppendEntriesBatchDestroy(SyncAppendEntriesBatch* pMsg) {
+  if (pMsg != NULL) {
+    taosMemoryFree(pMsg);
+  }
+}
+
+void syncAppendEntriesBatchSerialize(const SyncAppendEntriesBatch* pMsg, char* buf, uint32_t bufLen) {
+  ASSERT(pMsg->bytes <= bufLen);
+  memcpy(buf, pMsg, pMsg->bytes);
+}
+
+void syncAppendEntriesBatchDeserialize(const char* buf, uint32_t len, SyncAppendEntriesBatch* pMsg) {
+  memcpy(pMsg, buf, len);
+  ASSERT(len == pMsg->bytes);
+  ASSERT(pMsg->bytes == sizeof(SyncAppendEntriesBatch) + pMsg->dataLen);
+}
+
+char* syncAppendEntriesBatchSerialize2(const SyncAppendEntriesBatch* pMsg, uint32_t* len) {
+  char* buf = taosMemoryMalloc(pMsg->bytes);
+  ASSERT(buf != NULL);
+  syncAppendEntriesBatchSerialize(pMsg, buf, pMsg->bytes);
+  if (len != NULL) {
+    *len = pMsg->bytes;
+  }
+  return buf;
+}
+
+SyncAppendEntriesBatch* syncAppendEntriesBatchDeserialize2(const char* buf, uint32_t len) {
+  uint32_t                bytes = *((uint32_t*)buf);
+  SyncAppendEntriesBatch* pMsg = taosMemoryMalloc(bytes);
+  ASSERT(pMsg != NULL);
+  syncAppendEntriesBatchDeserialize(buf, len, pMsg);
+  ASSERT(len == pMsg->bytes);
+  return pMsg;
+}
+
+void syncAppendEntriesBatch2RpcMsg(const SyncAppendEntriesBatch* pMsg, SRpcMsg* pRpcMsg) {
+  memset(pRpcMsg, 0, sizeof(*pRpcMsg));
+  pRpcMsg->msgType = pMsg->msgType;
+  pRpcMsg->contLen = pMsg->bytes;
+  pRpcMsg->pCont = rpcMallocCont(pRpcMsg->contLen);
+  syncAppendEntriesBatchSerialize(pMsg, pRpcMsg->pCont, pRpcMsg->contLen);
+}
+
+void syncAppendEntriesBatchFromRpcMsg(const SRpcMsg* pRpcMsg, SyncAppendEntriesBatch* pMsg) {
+  syncAppendEntriesBatchDeserialize(pRpcMsg->pCont, pRpcMsg->contLen, pMsg);
+}
+
+SyncAppendEntriesBatch* syncAppendEntriesBatchFromRpcMsg2(const SRpcMsg* pRpcMsg) {
+  SyncAppendEntriesBatch* pMsg = syncAppendEntriesBatchDeserialize2(pRpcMsg->pCont, pRpcMsg->contLen);
+  ASSERT(pMsg != NULL);
+  return pMsg;
+}
+
+cJSON* syncAppendEntriesBatch2Json(const SyncAppendEntriesBatch* pMsg) {
+  char   u64buf[128] = {0};
+  cJSON* pRoot = cJSON_CreateObject();
+
+  if (pMsg != NULL) {
+    cJSON_AddNumberToObject(pRoot, "bytes", pMsg->bytes);
+    cJSON_AddNumberToObject(pRoot, "vgId", pMsg->vgId);
+    cJSON_AddNumberToObject(pRoot, "msgType", pMsg->msgType);
+
+    cJSON* pSrcId = cJSON_CreateObject();
+    snprintf(u64buf, sizeof(u64buf), "%lu", pMsg->srcId.addr);
+    cJSON_AddStringToObject(pSrcId, "addr", u64buf);
+    {
+      uint64_t u64 = pMsg->srcId.addr;
+      cJSON*   pTmp = pSrcId;
+      char     host[128] = {0};
+      uint16_t port;
+      syncUtilU642Addr(u64, host, sizeof(host), &port);
+      cJSON_AddStringToObject(pTmp, "addr_host", host);
+      cJSON_AddNumberToObject(pTmp, "addr_port", port);
+    }
+    cJSON_AddNumberToObject(pSrcId, "vgId", pMsg->srcId.vgId);
+    cJSON_AddItemToObject(pRoot, "srcId", pSrcId);
+
+    cJSON* pDestId = cJSON_CreateObject();
+    snprintf(u64buf, sizeof(u64buf), "%lu", pMsg->destId.addr);
+    cJSON_AddStringToObject(pDestId, "addr", u64buf);
+    {
+      uint64_t u64 = pMsg->destId.addr;
+      cJSON*   pTmp = pDestId;
+      char     host[128] = {0};
+      uint16_t port;
+      syncUtilU642Addr(u64, host, sizeof(host), &port);
+      cJSON_AddStringToObject(pTmp, "addr_host", host);
+      cJSON_AddNumberToObject(pTmp, "addr_port", port);
+    }
+    cJSON_AddNumberToObject(pDestId, "vgId", pMsg->destId.vgId);
+    cJSON_AddItemToObject(pRoot, "destId", pDestId);
+
+    snprintf(u64buf, sizeof(u64buf), "%lu", pMsg->term);
+    cJSON_AddStringToObject(pRoot, "term", u64buf);
+
+    snprintf(u64buf, sizeof(u64buf), "%ld", pMsg->prevLogIndex);
+    cJSON_AddStringToObject(pRoot, "prevLogIndex", u64buf);
+
+    snprintf(u64buf, sizeof(u64buf), "%lu", pMsg->prevLogTerm);
+    cJSON_AddStringToObject(pRoot, "prevLogTerm", u64buf);
+
+    snprintf(u64buf, sizeof(u64buf), "%ld", pMsg->commitIndex);
+    cJSON_AddStringToObject(pRoot, "commitIndex", u64buf);
+
+    snprintf(u64buf, sizeof(u64buf), "%lu", pMsg->privateTerm);
+    cJSON_AddStringToObject(pRoot, "privateTerm", u64buf);
+
+    cJSON_AddNumberToObject(pRoot, "dataCount", pMsg->dataCount);
+    cJSON_AddNumberToObject(pRoot, "dataLen", pMsg->dataLen);
+
+    int32_t metaArrayLen = sizeof(SOffsetAndContLen) * pMsg->dataCount;  // <offset, contLen>
+    int32_t rpcArrayLen = sizeof(SRpcMsg) * pMsg->dataCount;             // SRpcMsg
+    int32_t contArrayLen = pMsg->dataLen - metaArrayLen - rpcArrayLen;
+
+    cJSON_AddNumberToObject(pRoot, "metaArrayLen", metaArrayLen);
+    cJSON_AddNumberToObject(pRoot, "rpcArrayLen", rpcArrayLen);
+    cJSON_AddNumberToObject(pRoot, "contArrayLen", contArrayLen);
+
+    SOffsetAndContLen* metaArr = (SOffsetAndContLen*)(pMsg->data);
+    SRpcMsg*           msgArr = (SRpcMsg*)(pMsg->data + metaArrayLen);
+    void*              pData = (void*)(pMsg->data + metaArrayLen + rpcArrayLen);
+
+    cJSON* pMetaArr = cJSON_CreateArray();
+    cJSON_AddItemToObject(pRoot, "metaArr", pMetaArr);
+    for (int i = 0; i < pMsg->dataCount; ++i) {
+      cJSON* pMeta = cJSON_CreateObject();
+      cJSON_AddNumberToObject(pMeta, "offset", metaArr[i].offset);
+      cJSON_AddNumberToObject(pMeta, "contLen", metaArr[i].contLen);
+      cJSON_AddItemToArray(pMetaArr, pMeta);
+    }
+
+    cJSON* pMsgArr = cJSON_CreateArray();
+    cJSON_AddItemToObject(pRoot, "msgArr", pMsgArr);
+    for (int i = 0; i < pMsg->dataCount; ++i) {
+      cJSON* pRpcMsgJson = cJSON_CreateObject();
+      cJSON_AddNumberToObject(pRpcMsgJson, "code", msgArr[i].code);
+      cJSON_AddNumberToObject(pRpcMsgJson, "contLen", msgArr[i].contLen);
+      cJSON_AddNumberToObject(pRpcMsgJson, "msgType", msgArr[i].msgType);
+      cJSON_AddItemToArray(pMsgArr, pRpcMsgJson);
+    }
+
+    char* s;
+    s = syncUtilprintBin((char*)(pMsg->data), pMsg->dataLen);
+    cJSON_AddStringToObject(pRoot, "data", s);
+    taosMemoryFree(s);
+    s = syncUtilprintBin2((char*)(pMsg->data), pMsg->dataLen);
+    cJSON_AddStringToObject(pRoot, "data2", s);
+    taosMemoryFree(s);
+  }
+
+  cJSON* pJson = cJSON_CreateObject();
+  cJSON_AddItemToObject(pJson, "SyncAppendEntriesBatch", pRoot);
+  return pJson;
+}
+
+char* syncAppendEntriesBatch2Str(const SyncAppendEntriesBatch* pMsg) {
+  cJSON* pJson = syncAppendEntriesBatch2Json(pMsg);
+  char*  serialized = cJSON_Print(pJson);
+  cJSON_Delete(pJson);
+  return serialized;
+}
+
+void syncAppendEntriesBatch2RpcMsgArray(SyncAppendEntriesBatch* pSyncMsg, SRpcMsg* rpcMsgArr, int32_t maxArrSize,
+                                        int32_t* pRetArrSize) {
+  if (pRetArrSize != NULL) {
+    *pRetArrSize = pSyncMsg->dataCount;
+  }
+
+  int32_t arrSize = pSyncMsg->dataCount;
+  if (arrSize > maxArrSize) {
+    arrSize = maxArrSize;
+  }
+
+  int32_t metaArrayLen = sizeof(SOffsetAndContLen) * pSyncMsg->dataCount;  // <offset, contLen>
+  int32_t rpcArrayLen = sizeof(SRpcMsg) * pSyncMsg->dataCount;             // SRpcMsg
+  int32_t contArrayLen = pSyncMsg->dataLen - metaArrayLen - rpcArrayLen;
+
+  SOffsetAndContLen* metaArr = (SOffsetAndContLen*)(pSyncMsg->data);
+  SRpcMsg*           msgArr = (SRpcMsg*)(pSyncMsg->data + metaArrayLen);
+  void*              pData = pSyncMsg->data + metaArrayLen + rpcArrayLen;
+
+  for (int i = 0; i < arrSize; ++i) {
+    rpcMsgArr[i] = msgArr[i];
+    rpcMsgArr[i].pCont = rpcMallocCont(msgArr[i].contLen);
+    void* pRpcCont = pSyncMsg->data + metaArr[i].offset;
+    memcpy(rpcMsgArr[i].pCont, pRpcCont, rpcMsgArr[i].contLen);
+  }
+}
+
+// for debug ----------------------
+void syncAppendEntriesBatchPrint(const SyncAppendEntriesBatch* pMsg) {
+  char* serialized = syncAppendEntriesBatch2Str(pMsg);
+  printf("syncAppendEntriesBatchPrint | len:%lu | %s \n", strlen(serialized), serialized);
+  fflush(NULL);
+  taosMemoryFree(serialized);
+}
+
+void syncAppendEntriesBatchPrint2(char* s, const SyncAppendEntriesBatch* pMsg) {
+  char* serialized = syncAppendEntriesBatch2Str(pMsg);
+  printf("syncAppendEntriesBatchPrint2 | len:%lu | %s | %s \n", strlen(serialized), s, serialized);
+  fflush(NULL);
+  taosMemoryFree(serialized);
+}
+
+void syncAppendEntriesBatchLog(const SyncAppendEntriesBatch* pMsg) {
+  char* serialized = syncAppendEntriesBatch2Str(pMsg);
+  sTrace("syncAppendEntriesBatchLog | len:%lu | %s", strlen(serialized), serialized);
+  taosMemoryFree(serialized);
+}
+
+void syncAppendEntriesBatchLog2(char* s, const SyncAppendEntriesBatch* pMsg) {
+  if (gRaftDetailLog) {
+    char* serialized = syncAppendEntriesBatch2Str(pMsg);
+    sTraceLong("syncAppendEntriesBatchLog2 | len:%lu | %s | %s", strlen(serialized), s, serialized);
+    taosMemoryFree(serialized);
+  }
+}
+
 // ---- message process SyncAppendEntriesReply----
 SyncAppendEntriesReply* syncAppendEntriesReplyBuild(int32_t vgId) {
   uint32_t                bytes = sizeof(SyncAppendEntriesReply);

source/libs/sync/src/syncRaftCfg.c

@@ -101,7 +101,7 @@ cJSON *syncCfg2Json(SSyncCfg *pSyncCfg) {
 
 char *syncCfg2Str(SSyncCfg *pSyncCfg) {
   cJSON *pJson = syncCfg2Json(pSyncCfg);
-  char  *serialized = cJSON_Print(pJson);
+  char * serialized = cJSON_Print(pJson);
   cJSON_Delete(pJson);
   return serialized;
 }
@@ -109,7 +109,7 @@ char *syncCfg2Str(SSyncCfg *pSyncCfg) {
 char *syncCfg2SimpleStr(SSyncCfg *pSyncCfg) {
   if (pSyncCfg != NULL) {
     int32_t len = 512;
-    char   *s = taosMemoryMalloc(len);
+    char *  s = taosMemoryMalloc(len);
     memset(s, 0, len);
 
     snprintf(s, len, "{replica-num:%d, my-index:%d, ", pSyncCfg->replicaNum, pSyncCfg->myIndex);
@@ -205,7 +205,7 @@ cJSON *raftCfg2Json(SRaftCfg *pRaftCfg) {
 
 char *raftCfg2Str(SRaftCfg *pRaftCfg) {
   cJSON *pJson = raftCfg2Json(pRaftCfg);
-  char  *serialized = cJSON_Print(pJson);
+  char * serialized = cJSON_Print(pJson);
   cJSON_Delete(pJson);
   return serialized;
 }
@@ -280,7 +280,7 @@ int32_t raftCfgFromJson(const cJSON *pRoot, SRaftCfg *pRaftCfg) {
     (pRaftCfg->configIndexArr)[i] = atoll(pIndex->valuestring);
   }
 
-  cJSON  *pJsonSyncCfg = cJSON_GetObjectItem(pJson, "SSyncCfg");
+  cJSON * pJsonSyncCfg = cJSON_GetObjectItem(pJson, "SSyncCfg");
   int32_t code = syncCfgFromJson(pJsonSyncCfg, &(pRaftCfg->cfg));
   ASSERT(code == 0);
 

source/libs/sync/src/syncReplication.c

@@ -116,6 +116,73 @@ int32_t syncNodeAppendEntriesPeers(SSyncNode* pSyncNode) {
   return ret;
 }
 
+int32_t syncNodeAppendEntriesPeersSnapshot2(SSyncNode* pSyncNode) {
+  if (pSyncNode->state != TAOS_SYNC_STATE_LEADER) {
+    return -1;
+  }
+
+  int32_t ret = 0;
+  for (int i = 0; i < pSyncNode->peersNum; ++i) {
+    SRaftId* pDestId = &(pSyncNode->peersId[i]);
+
+    // next index
+    SyncIndex nextIndex = syncIndexMgrGetIndex(pSyncNode->pNextIndex, pDestId);
+
+    // pre index, pre term
+    SyncIndex preLogIndex = syncNodeGetPreIndex(pSyncNode, nextIndex);
+    SyncTerm  preLogTerm = syncNodeGetPreTerm(pSyncNode, nextIndex);
+    if (preLogTerm == SYNC_TERM_INVALID) {
+      SSyncSnapshotSender* pSender = syncNodeGetSnapshotSender(pSyncNode, pDestId);
+      ASSERT(pSender != NULL);
+      ASSERT(!snapshotSenderIsStart(pSender));
+
+      SyncIndex newNextIndex = syncNodeGetLastIndex(pSyncNode) + 1;
+      syncIndexMgrSetIndex(pSyncNode->pNextIndex, pDestId, newNextIndex);
+      syncIndexMgrSetIndex(pSyncNode->pMatchIndex, pDestId, SYNC_INDEX_INVALID);
+      sError("vgId:%d sync get pre term error, nextIndex:%ld, update next-index:%ld, match-index:%d, raftid:%ld",
+             pSyncNode->vgId, nextIndex, newNextIndex, SYNC_INDEX_INVALID, pDestId->addr);
+
+      return -1;
+    }
+
+    SRpcMsg rpcMsgArr[SYNC_MAX_BATCH_SIZE];
+    memset(rpcMsgArr, 0, sizeof(rpcMsgArr));
+
+    int32_t getCount = 0;
+    for (int32_t i = 0; i < pSyncNode->batchSize; ++i) {
+      SSyncRaftEntry* pEntry;
+      int32_t         code = pSyncNode->pLogStore->syncLogGetEntry(pSyncNode->pLogStore, nextIndex, &pEntry);
+      if (code == 0) {
+        ASSERT(pEntry != NULL);
+        // get rpc msg [i] from entry
+        syncEntryDestory(pEntry);
+        getCount++;
+      } else {
+        break;
+      }
+    }
+
+    SyncAppendEntriesBatch* pMsg = syncAppendEntriesBatchBuild(rpcMsgArr, getCount, pSyncNode->vgId);
+    ASSERT(pMsg != NULL);
+
+    // prepare msg
+    pMsg->srcId = pSyncNode->myRaftId;
+    pMsg->destId = *pDestId;
+    pMsg->term = pSyncNode->pRaftStore->currentTerm;
+    pMsg->prevLogIndex = preLogIndex;
+    pMsg->prevLogTerm = preLogTerm;
+    pMsg->commitIndex = pSyncNode->commitIndex;
+    pMsg->privateTerm = 0;
+    pMsg->dataCount = getCount;
+
+    // send msg
+    syncNodeAppendEntriesBatch(pSyncNode, pDestId, pMsg);
+    syncAppendEntriesBatchDestroy(pMsg);
+  }
+
+  return 0;
+}
+
 int32_t syncNodeAppendEntriesPeersSnapshot(SSyncNode* pSyncNode) {
   ASSERT(pSyncNode->state == TAOS_SYNC_STATE_LEADER);
 
@@ -234,4 +301,24 @@ int32_t syncNodeAppendEntries(SSyncNode* pSyncNode, const SRaftId* destRaftId, c
   syncAppendEntries2RpcMsg(pMsg, &rpcMsg);
   syncNodeSendMsgById(destRaftId, pSyncNode, &rpcMsg);
   return ret;
+}
+
+int32_t syncNodeAppendEntriesBatch(SSyncNode* pSyncNode, const SRaftId* destRaftId,
+                                   const SyncAppendEntriesBatch* pMsg) {
+  do {
+    char     host[128];
+    uint16_t port;
+    syncUtilU642Addr(destRaftId->addr, host, sizeof(host), &port);
+    sDebug(
+        "vgId:%d, send sync-append-entries-batch to %s:%d, {term:%lu, pre-index:%ld, pre-term:%lu, pterm:%lu, "
+        "commit:%ld, "
+        "datalen:%d, dataCount:%d}",
+        pSyncNode->vgId, host, port, pMsg->term, pMsg->prevLogIndex, pMsg->prevLogTerm, pMsg->privateTerm,
+        pMsg->commitIndex, pMsg->dataLen, pMsg->dataCount);
+  } while (0);
+
+  SRpcMsg rpcMsg;
+  syncAppendEntriesBatch2RpcMsg(pMsg, &rpcMsg);
+  syncNodeSendMsgById(destRaftId, pSyncNode, &rpcMsg);
+  return 0;
 }
\ No newline at end of file

source/libs/sync/src/syncSnapshot.c

@@ -349,14 +349,14 @@ cJSON *snapshotSender2Json(SSyncSnapshotSender *pSender) {
 
 char *snapshotSender2Str(SSyncSnapshotSender *pSender) {
   cJSON *pJson = snapshotSender2Json(pSender);
-  char  *serialized = cJSON_Print(pJson);
+  char * serialized = cJSON_Print(pJson);
   cJSON_Delete(pJson);
   return serialized;
 }
 
 char *snapshotSender2SimpleStr(SSyncSnapshotSender *pSender, char *event) {
   int32_t len = 256;
-  char   *s = taosMemoryMalloc(len);
+  char *  s = taosMemoryMalloc(len);
 
   SRaftId  destId = pSender->pSyncNode->replicasId[pSender->replicaIndex];
   char     host[64];
@@ -604,7 +604,7 @@ cJSON *snapshotReceiver2Json(SSyncSnapshotReceiver *pReceiver) {
     cJSON_AddStringToObject(pFromId, "addr", u64buf);
     {
       uint64_t u64 = pReceiver->fromId.addr;
-      cJSON   *pTmp = pFromId;
+      cJSON *  pTmp = pFromId;
       char     host[128] = {0};
       uint16_t port;
       syncUtilU642Addr(u64, host, sizeof(host), &port);
@@ -637,14 +637,14 @@ cJSON *snapshotReceiver2Json(SSyncSnapshotReceiver *pReceiver) {
 
 char *snapshotReceiver2Str(SSyncSnapshotReceiver *pReceiver) {
   cJSON *pJson = snapshotReceiver2Json(pReceiver);
-  char  *serialized = cJSON_Print(pJson);
+  char * serialized = cJSON_Print(pJson);
   cJSON_Delete(pJson);
   return serialized;
 }
 
 char *snapshotReceiver2SimpleStr(SSyncSnapshotReceiver *pReceiver, char *event) {
   int32_t len = 256;
-  char   *s = taosMemoryMalloc(len);
+  char *  s = taosMemoryMalloc(len);
 
   SRaftId  fromId = pReceiver->fromId;
   char     host[128];

source/libs/sync/test/CMakeLists.txt

@@ -21,6 +21,7 @@ add_executable(syncEntryCacheTest "")
 add_executable(syncRequestVoteTest "")
 add_executable(syncRequestVoteReplyTest "")
 add_executable(syncAppendEntriesTest "")
+add_executable(syncAppendEntriesBatchTest "")
 add_executable(syncAppendEntriesReplyTest "")
 add_executable(syncClientRequestTest "")
 add_executable(syncTimeoutTest "")
@@ -146,6 +147,10 @@ target_sources(syncAppendEntriesTest
     PRIVATE
     "syncAppendEntriesTest.cpp"
 )
+target_sources(syncAppendEntriesBatchTest
+    PRIVATE
+    "syncAppendEntriesBatchTest.cpp"
+)
 target_sources(syncAppendEntriesReplyTest
     PRIVATE
     "syncAppendEntriesReplyTest.cpp"
@@ -387,6 +392,11 @@ target_include_directories(syncAppendEntriesTest
     "${TD_SOURCE_DIR}/include/libs/sync"
     "${CMAKE_CURRENT_SOURCE_DIR}/../inc"
 )
+target_include_directories(syncAppendEntriesBatchTest
+    PUBLIC
+    "${TD_SOURCE_DIR}/include/libs/sync"
+    "${CMAKE_CURRENT_SOURCE_DIR}/../inc"
+)
 target_include_directories(syncAppendEntriesReplyTest
     PUBLIC
     "${TD_SOURCE_DIR}/include/libs/sync"
@@ -636,6 +646,10 @@ target_link_libraries(syncAppendEntriesTest
     sync
     gtest_main
 )
+target_link_libraries(syncAppendEntriesBatchTest
+    sync
+    gtest_main
+)
 target_link_libraries(syncAppendEntriesReplyTest
     sync
     gtest_main

source/libs/sync/test/syncAppendEntriesBatchTest.cpp

+//#include <gtest/gtest.h>
+#include <stdio.h>
+#include "syncIO.h"
+#include "syncInt.h"
+#include "syncMessage.h"
+#include "syncUtil.h"
+#include "trpc.h"
+
+void logTest() {
+  sTrace("--- sync log test: trace");
+  sDebug("--- sync log test: debug");
+  sInfo("--- sync log test: info");
+  sWarn("--- sync log test: warn");
+  sError("--- sync log test: error");
+  sFatal("--- sync log test: fatal");
+}
+
+SRpcMsg *createRpcMsg(int32_t i, int32_t dataLen) {
+  SRpcMsg *pRpcMsg = (SRpcMsg *)taosMemoryMalloc(sizeof(SRpcMsg));
+  memset(pRpcMsg, 0, sizeof(SRpcMsg));
+
+  pRpcMsg->msgType = TDMT_SYNC_PING;
+  pRpcMsg->contLen = dataLen;
+  pRpcMsg->pCont = rpcMallocCont(pRpcMsg->contLen);
+  pRpcMsg->code = 10 * i;
+  snprintf((char *)pRpcMsg->pCont, pRpcMsg->contLen, "value_%d", i);
+
+  return pRpcMsg;
+}
+
+SyncAppendEntriesBatch *createMsg() {
+  SRpcMsg rpcMsgArr[5];
+  memset(rpcMsgArr, 0, sizeof(rpcMsgArr));
+
+  for (int32_t i = 0; i < 5; ++i) {
+    SRpcMsg *pRpcMsg = createRpcMsg(i, 20);
+    rpcMsgArr[i] = *pRpcMsg;
+    taosMemoryFree(pRpcMsg);
+  }
+
+  SyncAppendEntriesBatch *pMsg = syncAppendEntriesBatchBuild(rpcMsgArr, 5, 1234);
+  pMsg->srcId.addr = syncUtilAddr2U64("127.0.0.1", 1234);
+  pMsg->srcId.vgId = 100;
+  pMsg->destId.addr = syncUtilAddr2U64("127.0.0.1", 5678);
+  pMsg->destId.vgId = 100;
+  pMsg->prevLogIndex = 11;
+  pMsg->prevLogTerm = 22;
+  pMsg->commitIndex = 33;
+  pMsg->privateTerm = 44;
+  return pMsg;
+}
+
+void test1() {
+  SyncAppendEntriesBatch *pMsg = createMsg();
+  syncAppendEntriesBatchLog2((char *)"test1:", pMsg);
+
+  SRpcMsg rpcMsgArr[5];
+  int32_t retArrSize;
+  syncAppendEntriesBatch2RpcMsgArray(pMsg, rpcMsgArr, 5, &retArrSize);
+  for (int i = 0; i < retArrSize; ++i) {
+    char logBuf[128];
+    snprintf(logBuf, sizeof(logBuf), "==test1 decode rpc msg %d: msgType:%d, code:%d, contLen:%d, pCont:%s \n", i,
+             rpcMsgArr[i].msgType, rpcMsgArr[i].code, rpcMsgArr[i].contLen, (char *)rpcMsgArr[i].pCont);
+    sTrace("%s", logBuf);
+  }
+
+  syncAppendEntriesBatchDestroy(pMsg);
+}
+
+/*
+void test2() {
+  SyncAppendEntries *pMsg = createMsg();
+  uint32_t           len = pMsg->bytes;
+  char *             serialized = (char *)taosMemoryMalloc(len);
+  syncAppendEntriesSerialize(pMsg, serialized, len);
+  SyncAppendEntries *pMsg2 = syncAppendEntriesBuild(pMsg->dataLen, 1000);
+  syncAppendEntriesDeserialize(serialized, len, pMsg2);
+  syncAppendEntriesLog2((char *)"test2: syncAppendEntriesSerialize -> syncAppendEntriesDeserialize ", pMsg2);
+
+  taosMemoryFree(serialized);
+  syncAppendEntriesDestroy(pMsg);
+  syncAppendEntriesDestroy(pMsg2);
+}
+
+void test3() {
+  SyncAppendEntries *pMsg = createMsg();
+  uint32_t           len;
+  char *             serialized = syncAppendEntriesSerialize2(pMsg, &len);
+  SyncAppendEntries *pMsg2 = syncAppendEntriesDeserialize2(serialized, len);
+  syncAppendEntriesLog2((char *)"test3: syncAppendEntriesSerialize3 -> syncAppendEntriesDeserialize2 ", pMsg2);
+
+  taosMemoryFree(serialized);
+  syncAppendEntriesDestroy(pMsg);
+  syncAppendEntriesDestroy(pMsg2);
+}
+
+void test4() {
+  SyncAppendEntries *pMsg = createMsg();
+  SRpcMsg            rpcMsg;
+  syncAppendEntries2RpcMsg(pMsg, &rpcMsg);
+  SyncAppendEntries *pMsg2 = (SyncAppendEntries *)taosMemoryMalloc(rpcMsg.contLen);
+  syncAppendEntriesFromRpcMsg(&rpcMsg, pMsg2);
+  syncAppendEntriesLog2((char *)"test4: syncAppendEntries2RpcMsg -> syncAppendEntriesFromRpcMsg ", pMsg2);
+
+  rpcFreeCont(rpcMsg.pCont);
+  syncAppendEntriesDestroy(pMsg);
+  syncAppendEntriesDestroy(pMsg2);
+}
+
+void test5() {
+  SyncAppendEntries *pMsg = createMsg();
+  SRpcMsg            rpcMsg;
+  syncAppendEntries2RpcMsg(pMsg, &rpcMsg);
+  SyncAppendEntries *pMsg2 = syncAppendEntriesFromRpcMsg2(&rpcMsg);
+  syncAppendEntriesLog2((char *)"test5: syncAppendEntries2RpcMsg -> syncAppendEntriesFromRpcMsg2 ", pMsg2);
+
+  rpcFreeCont(rpcMsg.pCont);
+  syncAppendEntriesDestroy(pMsg);
+  syncAppendEntriesDestroy(pMsg2);
+}
+*/
+
+int main() {
+  gRaftDetailLog = true;
+  tsAsyncLog = 0;
+  sDebugFlag = DEBUG_DEBUG + DEBUG_TRACE + DEBUG_SCREEN + DEBUG_FILE;
+  logTest();
+
+  test1();
+
+  /*
+   test2();
+   test3();
+   test4();
+   test5();
+ */
+
+  return 0;
+}

source/libs/sync/test/syncConfigChangeSnapshotTest.cpp

@@ -77,7 +77,7 @@ int32_t GetSnapshotCb(struct SSyncFSM* pFsm, SSnapshot* pSnapshot) {
   return 0;
 }
 
-int32_t SnapshotStartRead(struct SSyncFSM* pFsm, void** ppReader) {
+int32_t SnapshotStartRead(struct SSyncFSM* pFsm, void *pParam, void** ppReader) {
   *ppReader = (void*)0xABCD;
   char logBuf[256] = {0};
   snprintf(logBuf, sizeof(logBuf), "==callback== ==SnapshotStartRead== pFsm:%p, *ppReader:%p", pFsm, *ppReader);

source/libs/sync/test/syncSnapshotSenderTest.cpp

@@ -25,7 +25,7 @@ void ReConfigCb(struct SSyncFSM* pFsm, SSyncCfg newCfg, SReConfigCbMeta cbMeta)
 
 int32_t GetSnapshot(struct SSyncFSM* pFsm, SSnapshot* pSnapshot) { return 0; }
 
-int32_t SnapshotStartRead(struct SSyncFSM* pFsm, void** ppReader) { return 0; }
+int32_t SnapshotStartRead(struct SSyncFSM* pFsm, void *pParam, void** ppReader) { return 0; }
 int32_t SnapshotStopRead(struct SSyncFSM* pFsm, void* pReader) { return 0; }
 int32_t SnapshotDoRead(struct SSyncFSM* pFsm, void* pReader, void** ppBuf, int32_t* len) { return 0; }
 

source/libs/sync/test/syncTestTool.cpp

@@ -74,7 +74,7 @@ int32_t GetSnapshotCb(struct SSyncFSM* pFsm, SSnapshot* pSnapshot) {
   return 0;
 }
 
-int32_t SnapshotStartRead(struct SSyncFSM* pFsm, void** ppReader) {
+int32_t SnapshotStartRead(struct SSyncFSM* pFsm, void *pParam, void** ppReader) {
   *ppReader = (void*)0xABCD;
   char logBuf[256] = {0};
   snprintf(logBuf, sizeof(logBuf), "==callback== ==SnapshotStartRead== pFsm:%p, *ppReader:%p", pFsm, *ppReader);

source/libs/wal/inc/walInt.h

@@ -27,7 +27,7 @@ extern "C" {
 #endif
 
 // meta section begin
-typedef struct WalFileInfo {
+typedef struct {
   int64_t firstVer;
   int64_t lastVer;
   int64_t createTs;
@@ -98,20 +98,20 @@ static inline int walBuildIdxName(SWal* pWal, int64_t fileFirstVer, char* buf) {
   return sprintf(buf, "%s/%020" PRId64 "." WAL_INDEX_SUFFIX, pWal->path, fileFirstVer);
 }
 
-static inline int walValidHeadCksum(SWalHead* pHead) {
-  return taosCheckChecksum((uint8_t*)&pHead->head, sizeof(SWalReadHead), pHead->cksumHead);
+static inline int walValidHeadCksum(SWalCkHead* pHead) {
+  return taosCheckChecksum((uint8_t*)&pHead->head, sizeof(SWalCont), pHead->cksumHead);
 }
 
-static inline int walValidBodyCksum(SWalHead* pHead) {
+static inline int walValidBodyCksum(SWalCkHead* pHead) {
   return taosCheckChecksum((uint8_t*)pHead->head.body, pHead->head.bodyLen, pHead->cksumBody);
 }
 
-static inline int walValidCksum(SWalHead* pHead, void* body, int64_t bodyLen) {
+static inline int walValidCksum(SWalCkHead* pHead, void* body, int64_t bodyLen) {
   return walValidHeadCksum(pHead) && walValidBodyCksum(pHead);
 }
 
-static inline uint32_t walCalcHeadCksum(SWalHead* pHead) {
-  return taosCalcChecksum(0, (uint8_t*)&pHead->head, sizeof(SWalReadHead));
+static inline uint32_t walCalcHeadCksum(SWalCkHead* pHead) {
+  return taosCalcChecksum(0, (uint8_t*)&pHead->head, sizeof(SWalCont));
 }
 
 static inline uint32_t walCalcBodyCksum(const void* body, uint32_t len) {

source/libs/wal/src/walMeta.c

@@ -16,7 +16,7 @@
 #include "cJSON.h"
 #include "os.h"
 #include "taoserror.h"
-#include "tref.h"
+#include "tutil.h"
 #include "walInt.h"
 
 bool FORCE_INLINE walLogExist(SWal* pWal, int64_t ver) {
@@ -37,26 +37,9 @@ static FORCE_INLINE int walBuildMetaName(SWal* pWal, int metaVer, char* buf) {
   return sprintf(buf, "%s/meta-ver%d", pWal->path, metaVer);
 }
 
-void* tmemmem(char* haystack, int hlen, char* needle, int nlen) {
-  char* limit;
-
-  if (nlen == 0 || hlen < nlen) {
-    return NULL;
-  }
-
-  limit = haystack + hlen - nlen + 1;
-  while ((haystack = (char*)memchr(haystack, needle[0], limit - haystack)) != NULL) {
-    if (memcmp(haystack, needle, nlen) == 0) {
-      return haystack;
-    }
-    haystack++;
-  }
-  return NULL;
-}
-
 static FORCE_INLINE int64_t walScanLogGetLastVer(SWal* pWal) {
   ASSERT(pWal->fileInfoSet != NULL);
-  int sz = taosArrayGetSize(pWal->fileInfoSet);
+  int32_t sz = taosArrayGetSize(pWal->fileInfoSet);
   ASSERT(sz > 0);
 #if 0
   for (int i = 0; i < sz; i++) {
@@ -101,14 +84,14 @@ static FORCE_INLINE int64_t walScanLogGetLastVer(SWal* pWal) {
   char* candidate;
   while ((candidate = tmemmem(haystack, readSize - (haystack - buf), (char*)&magic, sizeof(uint64_t))) != NULL) {
     // read and validate
-    SWalHead* logContent = (SWalHead*)candidate;
+    SWalCkHead* logContent = (SWalCkHead*)candidate;
     if (walValidHeadCksum(logContent) == 0 && walValidBodyCksum(logContent) == 0) {
       found = candidate;
     }
     haystack = candidate + 1;
   }
   if (found == buf) {
-    SWalHead* logContent = (SWalHead*)found;
+    SWalCkHead* logContent = (SWalCkHead*)found;
     if (walValidHeadCksum(logContent) != 0 || walValidBodyCksum(logContent) != 0) {
       // file has to be deleted
       taosMemoryFree(buf);
@@ -118,7 +101,7 @@ static FORCE_INLINE int64_t walScanLogGetLastVer(SWal* pWal) {
     }
   }
   taosCloseFile(&pFile);
-  SWalHead* lastEntry = (SWalHead*)found;
+  SWalCkHead* lastEntry = (SWalCkHead*)found;
 
   return lastEntry->head.version;
 }

source/libs/wal/src/walMgmt.c

@@ -117,8 +117,8 @@ SWal *walOpen(const char *path, SWalCfg *pCfg) {
   pWal->lastRollSeq = -1;
 
   // init write buffer
-  memset(&pWal->writeHead, 0, sizeof(SWalHead));
-  pWal->writeHead.head.headVer = WAL_HEAD_VER;
+  memset(&pWal->writeHead, 0, sizeof(SWalCkHead));
+  pWal->writeHead.head.protoVer = WAL_PROTO_VER;
   pWal->writeHead.magic = WAL_MAGIC;
 
   if (taosThreadMutexInit(&pWal->mutex, NULL) < 0) {

source/libs/wal/src/walRead.c

@@ -33,7 +33,7 @@ SWalReadHandle *walOpenReadHandle(SWal *pWal) {
 
   taosThreadMutexInit(&pRead->mutex, NULL);
 
-  pRead->pHead = taosMemoryMalloc(sizeof(SWalHead));
+  pRead->pHead = taosMemoryMalloc(sizeof(SWalCkHead));
   if (pRead->pHead == NULL) {
     terrno = TSDB_CODE_WAL_OUT_OF_MEMORY;
     taosMemoryFree(pRead);
@@ -155,7 +155,7 @@ static int32_t walReadSeekVer(SWalReadHandle *pRead, int64_t ver) {
 
 void walSetReaderCapacity(SWalReadHandle *pRead, int32_t capacity) { pRead->capacity = capacity; }
 
-int32_t walFetchHead(SWalReadHandle *pRead, int64_t ver, SWalHead *pHead) {
+int32_t walFetchHead(SWalReadHandle *pRead, int64_t ver, SWalCkHead *pHead) {
   int64_t code;
 
   // TODO: valid ver
@@ -170,8 +170,8 @@ int32_t walFetchHead(SWalReadHandle *pRead, int64_t ver, SWalHead *pHead) {
 
   ASSERT(taosValidFile(pRead->pReadLogTFile) == true);
 
-  code = taosReadFile(pRead->pReadLogTFile, pHead, sizeof(SWalHead));
-  if (code != sizeof(SWalHead)) {
+  code = taosReadFile(pRead->pReadLogTFile, pHead, sizeof(SWalCkHead));
+  if (code != sizeof(SWalCkHead)) {
     return -1;
   }
 
@@ -186,7 +186,7 @@ int32_t walFetchHead(SWalReadHandle *pRead, int64_t ver, SWalHead *pHead) {
   return 0;
 }
 
-int32_t walSkipFetchBody(SWalReadHandle *pRead, const SWalHead *pHead) {
+int32_t walSkipFetchBody(SWalReadHandle *pRead, const SWalCkHead *pHead) {
   int64_t code;
 
   ASSERT(pRead->curVersion == pHead->head.version);
@@ -203,12 +203,12 @@ int32_t walSkipFetchBody(SWalReadHandle *pRead, const SWalHead *pHead) {
   return 0;
 }
 
-int32_t walFetchBody(SWalReadHandle *pRead, SWalHead **ppHead) {
-  SWalReadHead *pReadHead = &((*ppHead)->head);
-  int64_t       ver = pReadHead->version;
+int32_t walFetchBody(SWalReadHandle *pRead, SWalCkHead **ppHead) {
+  SWalCont *pReadHead = &((*ppHead)->head);
+  int64_t   ver = pReadHead->version;
 
   if (pRead->capacity < pReadHead->bodyLen) {
-    void *ptr = taosMemoryRealloc(*ppHead, sizeof(SWalHead) + pReadHead->bodyLen);
+    void *ptr = taosMemoryRealloc(*ppHead, sizeof(SWalCkHead) + pReadHead->bodyLen);
     if (ptr == NULL) {
       terrno = TSDB_CODE_WAL_OUT_OF_MEMORY;
       return -1;
@@ -241,18 +241,18 @@ int32_t walFetchBody(SWalReadHandle *pRead, SWalHead **ppHead) {
   return 0;
 }
 
-int32_t walReadWithHandle_s(SWalReadHandle *pRead, int64_t ver, SWalReadHead **ppHead) {
+int32_t walReadWithHandle_s(SWalReadHandle *pRead, int64_t ver, SWalCont **ppHead) {
   taosThreadMutexLock(&pRead->mutex);
   if (walReadWithHandle(pRead, ver) < 0) {
     taosThreadMutexUnlock(&pRead->mutex);
     return -1;
   }
-  *ppHead = taosMemoryMalloc(sizeof(SWalReadHead) + pRead->pHead->head.bodyLen);
+  *ppHead = taosMemoryMalloc(sizeof(SWalCont) + pRead->pHead->head.bodyLen);
   if (*ppHead == NULL) {
     taosThreadMutexUnlock(&pRead->mutex);
     return -1;
   }
-  memcpy(*ppHead, &pRead->pHead->head, sizeof(SWalReadHead) + pRead->pHead->head.bodyLen);
+  memcpy(*ppHead, &pRead->pHead->head, sizeof(SWalCont) + pRead->pHead->head.bodyLen);
   taosThreadMutexUnlock(&pRead->mutex);
   return 0;
 }
@@ -282,8 +282,8 @@ int32_t walReadWithHandle(SWalReadHandle *pRead, int64_t ver) {
 
   ASSERT(taosValidFile(pRead->pReadLogTFile) == true);
 
-  code = taosReadFile(pRead->pReadLogTFile, pRead->pHead, sizeof(SWalHead));
-  if (code != sizeof(SWalHead)) {
+  code = taosReadFile(pRead->pReadLogTFile, pRead->pHead, sizeof(SWalCkHead));
+  if (code != sizeof(SWalCkHead)) {
     if (code < 0)
       terrno = TAOS_SYSTEM_ERROR(errno);
     else {
@@ -301,7 +301,7 @@ int32_t walReadWithHandle(SWalReadHandle *pRead, int64_t ver) {
   }
 
   if (pRead->capacity < pRead->pHead->head.bodyLen) {
-    void *ptr = taosMemoryRealloc(pRead->pHead, sizeof(SWalHead) + pRead->pHead->head.bodyLen);
+    void *ptr = taosMemoryRealloc(pRead->pHead, sizeof(SWalCkHead) + pRead->pHead->head.bodyLen);
     if (ptr == NULL) {
       terrno = TSDB_CODE_WAL_OUT_OF_MEMORY;
       return -1;

source/libs/wal/src/walWrite.c

@@ -142,10 +142,10 @@ int32_t walRollback(SWal *pWal, int64_t ver) {
     return -1;
   }
   // validate offset
-  SWalHead head;
+  SWalCkHead head;
   ASSERT(taosValidFile(pLogTFile));
-  int64_t size = taosReadFile(pLogTFile, &head, sizeof(SWalHead));
-  if (size != sizeof(SWalHead)) {
+  int64_t size = taosReadFile(pLogTFile, &head, sizeof(SWalCkHead));
+  if (size != sizeof(SWalCkHead)) {
     return -1;
   }
   code = walValidHeadCksum(&head);
@@ -261,7 +261,7 @@ int32_t walEndSnapshot(SWal *pWal) {
 }
 
 int walRoll(SWal *pWal) {
-  int code = 0;
+  int32_t code = 0;
   if (pWal->pWriteIdxTFile != NULL) {
     code = taosCloseFile(&pWal->pWriteIdxTFile);
     if (code != 0) {
@@ -321,12 +321,13 @@ static int walWriteIndex(SWal *pWal, int64_t ver, int64_t offset) {
   return 0;
 }
 
-int64_t walWriteWithSyncInfo(SWal *pWal, int64_t index, tmsg_t msgType, SSyncLogMeta syncMeta, const void *body,
+int32_t walWriteWithSyncInfo(SWal *pWal, int64_t index, tmsg_t msgType, SSyncLogMeta syncMeta, const void *body,
                              int32_t bodyLen) {
-  int code = 0;
+  int32_t code = 0;
 
   // no wal
   if (pWal->cfg.level == TAOS_WAL_NOLOG) return 0;
+
   if (bodyLen > TSDB_MAX_WAL_SIZE) {
     terrno = TSDB_CODE_WAL_SIZE_LIMIT;
     return -1;
@@ -356,6 +357,7 @@ int64_t walWriteWithSyncInfo(SWal *pWal, int64_t index, tmsg_t msgType, SSyncLog
     terrno = TSDB_CODE_WAL_INVALID_VER;
     return -1;
   }
+
   /*if (!tfValid(pWal->pWriteLogTFile)) return -1;*/
 
   ASSERT(pWal->writeCur >= 0);
@@ -380,7 +382,7 @@ int64_t walWriteWithSyncInfo(SWal *pWal, int64_t index, tmsg_t msgType, SSyncLog
   pWal->writeHead.cksumHead = walCalcHeadCksum(&pWal->writeHead);
   pWal->writeHead.cksumBody = walCalcBodyCksum(body, bodyLen);
 
-  if (taosWriteFile(pWal->pWriteLogTFile, &pWal->writeHead, sizeof(SWalHead)) != sizeof(SWalHead)) {
+  if (taosWriteFile(pWal->pWriteLogTFile, &pWal->writeHead, sizeof(SWalCkHead)) != sizeof(SWalCkHead)) {
     // TODO ftruncate
     terrno = TAOS_SYSTEM_ERROR(errno);
     wError("vgId:%d, file:%" PRId64 ".log, failed to write since %s", pWal->cfg.vgId, walGetLastFileFirstVer(pWal),
@@ -405,19 +407,19 @@ int64_t walWriteWithSyncInfo(SWal *pWal, int64_t index, tmsg_t msgType, SSyncLog
   // set status
   if (pWal->vers.firstVer == -1) pWal->vers.firstVer = index;
   pWal->vers.lastVer = index;
-  pWal->totSize += sizeof(SWalHead) + bodyLen;
+  pWal->totSize += sizeof(SWalCkHead) + bodyLen;
   if (walGetCurFileInfo(pWal)->firstVer == -1) {
     walGetCurFileInfo(pWal)->firstVer = index;
   }
   walGetCurFileInfo(pWal)->lastVer = index;
-  walGetCurFileInfo(pWal)->fileSize += sizeof(SWalHead) + bodyLen;
+  walGetCurFileInfo(pWal)->fileSize += sizeof(SWalCkHead) + bodyLen;
 
   taosThreadMutexUnlock(&pWal->mutex);
 
   return 0;
 }
 
-int64_t walWrite(SWal *pWal, int64_t index, tmsg_t msgType, const void *body, int32_t bodyLen) {
+int32_t walWrite(SWal *pWal, int64_t index, tmsg_t msgType, const void *body, int32_t bodyLen) {
   SSyncLogMeta syncMeta = {
       .isWeek = -1,
       .seqNum = UINT64_MAX,
@@ -435,27 +437,3 @@ void walFsync(SWal *pWal, bool forceFsync) {
     }
   }
 }
-
-/*static int walValidateOffset(SWal* pWal, int64_t ver) {*/
-/*int code = 0;*/
-/*SWalHead *pHead = NULL;*/
-/*code = (int)walRead(pWal, &pHead, ver);*/
-/*if(pHead->head.version != ver) {*/
-/*return -1;*/
-/*}*/
-/*return 0;*/
-/*}*/
-
-/*static int64_t walGetOffset(SWal* pWal, int64_t ver) {*/
-/*int code = walSeekVer(pWal, ver);*/
-/*if(code != 0) {*/
-/*return -1;*/
-/*}*/
-
-/*code = walValidateOffset(pWal, ver);*/
-/*if(code != 0) {*/
-/*return -1;*/
-/*}*/
-
-/*return 0;*/
-/*}*/

source/libs/wal/test/walMetaTest.cpp

@@ -148,7 +148,7 @@ TEST_F(WalCleanEnv, createNew) {
   walRollFileInfo(pWal);
   ASSERT(pWal->fileInfoSet != NULL);
   ASSERT_EQ(pWal->fileInfoSet->size, 1);
-  WalFileInfo* pInfo = (WalFileInfo*)taosArrayGetLast(pWal->fileInfoSet);
+  SWalFileInfo* pInfo = (SWalFileInfo*)taosArrayGetLast(pWal->fileInfoSet);
   ASSERT_EQ(pInfo->firstVer, 0);
   ASSERT_EQ(pInfo->lastVer, -1);
   ASSERT_EQ(pInfo->closeTs, -1);

source/os/src/osFile.c

@@ -300,16 +300,14 @@ TdFilePtr taosOpenFile(const char *path, int32_t tdFileOptions) {
   return pFile;
 }
 
-int64_t taosCloseFile(TdFilePtr *ppFile) {
+int32_t taosCloseFile(TdFilePtr *ppFile) {
+  int32_t code = 0;
   if (ppFile == NULL || *ppFile == NULL) {
     return 0;
   }
 #if FILE_WITH_LOCK
   taosThreadRwlockWrlock(&((*ppFile)->rwlock));
 #endif
-  if (ppFile == NULL || *ppFile == NULL) {
-    return 0;
-  }
   if ((*ppFile)->fp != NULL) {
     fflush((*ppFile)->fp);
     fclose((*ppFile)->fp);
@@ -320,9 +318,10 @@ int64_t taosCloseFile(TdFilePtr *ppFile) {
     HANDLE h = (HANDLE)_get_osfhandle((*ppFile)->fd);
     !FlushFileBuffers(h);
 #else
-    fsync((*ppFile)->fd);
+    // warning: never fsync silently in base lib
+    /*fsync((*ppFile)->fd);*/
 #endif
-    close((*ppFile)->fd);
+    code = close((*ppFile)->fd);
     (*ppFile)->fd = -1;
   }
   (*ppFile)->refId = 0;
@@ -332,7 +331,7 @@ int64_t taosCloseFile(TdFilePtr *ppFile) {
 #endif
   taosMemoryFree(*ppFile);
   *ppFile = NULL;
-  return 0;
+  return code;
 }
 
 int64_t taosReadFile(TdFilePtr pFile, void *buf, int64_t count) {
@@ -560,6 +559,8 @@ int32_t taosFsyncFile(TdFilePtr pFile) {
     return 0;
   }
 
+  // this implementation is WRONG
+  // fflush is not a replacement of fsync
   if (pFile->fp != NULL) return fflush(pFile->fp);
   if (pFile->fd >= 0) {
 #ifdef WINDOWS

source/os/src/osString.c

@@ -202,7 +202,7 @@ int32_t taosHexEncode(const char *src, char *dst, int32_t len) {
   }
 
   for (int32_t i = 0; i < len; ++i) {
-    sprintf(dst + i * 2, "%02x", src[i] & 0xff);
+    sprintf(dst + i * 2, "%02x", src[i]);
   }
 
   return 0;
@@ -213,10 +213,10 @@ int32_t taosHexDecode(const char *src, char *dst, int32_t len) {
     return -1;
   }
 
-  uint16_t hn, ln, out;
+  uint8_t hn, ln, out;
   for (int i = 0, j = 0; i < len * 2; i += 2, ++j ) {
-    hn = src[i] > '9' ? src[i] - 'A' + 10 : src[i] - '0';
-    ln = src[i + 1] > '9' ? src[i + 1] - 'A' + 10 : src[i + 1] - '0';
+    hn = src[i] > '9' ? src[i] - 'a' + 10 : src[i] - '0';
+    ln = src[i + 1] > '9' ? src[i + 1] - 'a' + 10 : src[i + 1] - '0';
 
     out = (hn << 4) | ln;
     memcpy(dst + j, &out, 1);

source/util/src/terror.c

@@ -394,6 +394,9 @@ TAOS_DEFINE_ERROR(TSDB_CODE_QRY_DUPLICATTED_OPERATION,    "Duplicatted operation
 TAOS_DEFINE_ERROR(TSDB_CODE_QRY_TASK_MSG_ERROR,           "Task message error")
 TAOS_DEFINE_ERROR(TSDB_CODE_QRY_JOB_FREED,                "Job already freed")
 TAOS_DEFINE_ERROR(TSDB_CODE_QRY_TASK_STATUS_ERROR,        "Task status error")
+TAOS_DEFINE_ERROR(TSDB_CODE_QRY_JSON_IN_ERROR,            "Json not support in in/notin operator")
+TAOS_DEFINE_ERROR(TSDB_CODE_QRY_JSON_NOT_SUPPORT_ERROR,   "Json not support in this place")
+TAOS_DEFINE_ERROR(TSDB_CODE_QRY_JSON_IN_GROUP_ERROR,      "Json not support in group/partition by")
 
 // grant
 TAOS_DEFINE_ERROR(TSDB_CODE_GRANT_EXPIRED,                "License expired")
@@ -429,6 +432,7 @@ TAOS_DEFINE_ERROR(TSDB_CODE_SYN_NEW_CONFIG_ERROR,         "Sync new config error
 TAOS_DEFINE_ERROR(TSDB_CODE_SYN_RECONFIG_NOT_READY,       "Sync not ready for reconfig")
 TAOS_DEFINE_ERROR(TSDB_CODE_SYN_PROPOSE_NOT_READY,        "Sync not ready for propose")
 TAOS_DEFINE_ERROR(TSDB_CODE_SYN_STANDBY_NOT_READY,        "Sync not ready for standby")
+TAOS_DEFINE_ERROR(TSDB_CODE_SYN_BATCH_ERROR,              "Sync batch error")
 TAOS_DEFINE_ERROR(TSDB_CODE_SYN_INTERNAL_ERROR,           "Sync internal error")
 
 // wal

source/util/src/tpagedbuf.c

@@ -193,7 +193,9 @@ static char* doFlushPageToDisk(SDiskbasedBuf* pBuf, SPageInfo* pg) {
 
   char* pDataBuf = pg->pData;
   memset(pDataBuf, 0, getAllocPageSize(pBuf->pageSize));
-
+#ifdef BUF_PAGE_DEBUG
+  uDebug("page_flush %p, pageId:%d, offset:%d", pDataBuf, pg->pageId, pg->offset);
+#endif
   pg->length = size;  // on disk size
   return pDataBuf;
 }
@@ -440,6 +442,9 @@ void* getNewBufPage(SDiskbasedBuf* pBuf, int32_t groupId, int32_t* pageId) {
   }
 
   ((void**)pi->pData)[0] = pi;
+#ifdef BUF_PAGE_DEBUG
+  uDebug("page_getNewBufPage , pi->pData:%p, pageId:%d, offset:%"PRId64, pi->pData, pi->pageId, pi->offset);
+#endif
   return (void*)(GET_DATA_PAYLOAD(pi));
 }
 
@@ -462,7 +467,9 @@ void* getBufPage(SDiskbasedBuf* pBuf, int32_t id) {
 
     lruListMoveToFront(pBuf->lruList, (*pi));
     (*pi)->used = true;
-
+#ifdef BUF_PAGE_DEBUG
+    uDebug("page_getBufPage1 pageId:%d, offset:%"PRId64, (*pi)->pageId, (*pi)->offset);
+#endif
     return (void*)(GET_DATA_PAYLOAD(*pi));
   } else {  // not in memory
     assert((*pi)->pData == NULL && (*pi)->pn == NULL && (((*pi)->length >= 0 && (*pi)->offset >= 0) || ((*pi)->length == -1 && (*pi)->offset == -1)));
@@ -494,7 +501,9 @@ void* getBufPage(SDiskbasedBuf* pBuf, int32_t id) {
         return NULL;
       }
     }
-
+#ifdef BUF_PAGE_DEBUG
+    uDebug("page_getBufPage2 pageId:%d, offset:%"PRId64, (*pi)->pageId, (*pi)->offset);
+#endif
     return (void*)(GET_DATA_PAYLOAD(*pi));
   }
 }
@@ -506,8 +515,11 @@ void releaseBufPage(SDiskbasedBuf* pBuf, void* page) {
 }
 
 void releaseBufPageInfo(SDiskbasedBuf* pBuf, SPageInfo* pi) {
+#ifdef BUF_PAGE_DEBUG
+  uDebug("page_releaseBufPageInfo pageId:%d, used:%d, offset:%"PRId64, pi->pageId, pi->used, pi->offset);
+#endif
   assert(pi->pData != NULL && pi->used == true);
-
+//  assert(pi->pData != NULL);
   pi->used = false;
   pBuf->statis.releasePages += 1;
 }

source/util/src/tutil.c

@@ -16,6 +16,23 @@
 #define _DEFAULT_SOURCE
 #include "tutil.h"
 
+void *tmemmem(const char *haystack, int32_t hlen, const char *needle, int32_t nlen) {
+  const char *limit;
+
+  if (nlen == 0 || hlen < nlen) {
+    return NULL;
+  }
+
+  limit = haystack + hlen - nlen + 1;
+  while ((haystack = (char *)memchr(haystack, needle[0], limit - haystack)) != NULL) {
+    if (memcmp(haystack, needle, nlen) == 0) {
+      return (void *)haystack;
+    }
+    haystack++;
+  }
+  return NULL;
+}
+
 int32_t strdequote(char *z) {
   if (z == NULL) {
     return 0;

tests/script/jenkins/basic.txt

@@ -30,13 +30,6 @@
 ./test.sh -f tsim/dnode/balance2.sim
 ./test.sh -f tsim/dnode/balance3.sim
 ./test.sh -f tsim/dnode/balancex.sim
-#./test.sh -f tsim/dnode/cluster_alter.sim
-#./test.sh -f tsim/dnode/cluster_balance1.sim
-#./test.sh -f tsim/dnode/cluster_balance2.sim
-#./test.sh -f tsim/dnode/cluster_balance3.sim
-#./test.sh -f tsim/dnode/cluster_cache.sim
-#./test.sh -f tsim/dnode/cluster_flowctrl.sim
-#./test.sh -f tsim/dnode/cluster_vgroup100.sim
 ./test.sh -f tsim/dnode/create_dnode.sim
 ./test.sh -f tsim/dnode/drop_dnode_has_mnode.sim
 ./test.sh -f tsim/dnode/drop_dnode_has_qnode_snode.sim

tests/script/tsim/dnode/cluster_alter.sim

-system sh/stop_dnodes.sh
-
-system sh/deploy.sh -n dnode1 -i 1
-system sh/deploy.sh -n dnode2 -i 2
-system sh/deploy.sh -n dnode3 -i 3
-system sh/deploy.sh -n dnode4 -i 4
-
-system sh/cfg.sh -n dnode1 -c role -v 1
-system sh/cfg.sh -n dnode2 -c role -v 2
-system sh/cfg.sh -n dnode3 -c role -v 2
-system sh/cfg.sh -n dnode4 -c role -v 2
-
-system sh/cfg.sh -n dnode1 -c balanceInterval -v 10
-system sh/cfg.sh -n dnode2 -c balanceInterval -v 10
-system sh/cfg.sh -n dnode3 -c balanceInterval -v 10
-system sh/cfg.sh -n dnode4 -c balanceInterval -v 10
-
-system sh/cfg.sh -n dnode1 -c wallevel -v 2
-system sh/cfg.sh -n dnode2 -c wallevel -v 2
-system sh/cfg.sh -n dnode3 -c wallevel -v 2
-system sh/cfg.sh -n dnode4 -c wallevel -v 2
-
-system sh/cfg.sh -n dnode1 -c balance -v 0
-system sh/cfg.sh -n dnode2 -c balance -v 0
-system sh/cfg.sh -n dnode3 -c balance -v 0
-system sh/cfg.sh -n dnode4 -c balance -v 0
-
-print ========== step1
-system sh/exec.sh -n dnode1 -s start
-sql connect
-sleep 2000
-
-sql create dnode $hostname2
-system sh/exec.sh -n dnode2 -s start
-sleep 2000
-
-print ========== step2
-sql create database d1
-sql create table d1.t1 (t timestamp, i int) 
-sql insert into d1.t1 values(now+1s, 15)
-sql insert into d1.t1 values(now+2s, 14)
-sql insert into d1.t1 values(now+3s, 13)
-sql insert into d1.t1 values(now+4s, 12)
-sql insert into d1.t1 values(now+5s, 11)
-
-print ========== step3
-sleep 2000
-sql create dnode $hostname3
-system sh/exec.sh -n dnode3 -s start
-sql create dnode $hostname4
-system sh/exec.sh -n dnode4 -s start
-
-sql show dnodes
-print dnode1 openVnodes $data2_1
-print dnode2 openVnodes $data2_2
-print dnode3 openVnodes $data2_3
-print dnode4 openVnodes $data2_4
-if $data2_1 != 0 then
-	return -1
-endi
-if $data2_2 != 1 then
-	return -1
-endi
-if $data2_3 != 0 then
-	return -1
-endi
-if $data2_4 != 0 then
-	return -1
-endi
-
-print ========== step4
-sql alter dnode 2 balance "vnode:2-dnode:3"
-
-$x = 0
-show4: 
-	$x = $x + 1
-	sleep 2000
-	if $x == 10 then
-		return -1
-	endi
-	
-sql show dnodes
-print dnode1 openVnodes $data2_1
-print dnode2 openVnodes $data2_2
-print dnode3 openVnodes $data2_3
-print dnode4 openVnodes $data2_4
-if $data2_2 != 0 then
-    goto show4
-endi
-if $data2_3 != 1 then
-    goto show4
-endi
-if $data2_4 != 0 then
-    goto show4
-endi
-
-print ========== step5
-sql alter dnode 3 balance "vnode:2-dnode:4"
-
-$x = 0
-show5: 
-	$x = $x + 1
-	sleep 2000
-	if $x == 10 then
-		return -1
-	endi
-	
-sql show dnodes
-print dnode1 openVnodes $data2_1
-print dnode2 openVnodes $data2_2
-print dnode3 openVnodes $data2_3
-print dnode4 openVnodes $data2_4
-if $data2_2 != 0 then
-    goto show5
-endi
-if $data2_3 != 0 then
-    goto show5
-endi
-if $data2_4 != 1 then
-    goto show5
-endi
-
-print ========== step6
-sql alter dnode 4 balance "vnode:2-dnode:2"
-
-$x = 0
-show6: 
-	$x = $x + 1
-	sleep 2000
-	if $x == 10 then
-		return -1
-	endi
-	
-sql show dnodes
-print dnode1 openVnodes $data2_1
-print dnode2 openVnodes $data2_2
-print dnode3 openVnodes $data2_3
-print dnode4 openVnodes $data2_4
-if $data2_2 != 1 then
-    goto show6
-endi
-if $data2_3 != 0 then
-    goto show6
-endi
-if $data2_4 != 0 then
-    goto show6
-endi
-
-print ========== step7
-sql select * from d1.t1 order by t desc
-print $data01 $data11 $data21 $data31 $data41 
-if $data01 != 11 then
-  return -1
-endi
-if $data11 != 12 then
-  return -1
-endi
-if $data21 != 13 then
-  return -1
-endi
-if $data31 != 14 then
-  return -1
-endi
-if $data41 != 15 then
-  return -1
-endi
-
-print ========== step8
-sql_error sql alter dnode 4 balance "vnode:2-dnode:5"
-
-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
\ No newline at end of file

tests/script/tsim/dnode/cluster_balance1.sim

-system sh/stop_dnodes.sh
-system sh/deploy.sh -n dnode1 -i 1
-system sh/deploy.sh -n dnode2 -i 2
-system sh/deploy.sh -n dnode3 -i 3
-system sh/deploy.sh -n dnode4 -i 4
-system sh/deploy.sh -n dnode5 -i 5
-system sh/deploy.sh -n dnode6 -i 6
-system sh/deploy.sh -n dnode7 -i 7
-system sh/deploy.sh -n dnode8 -i 8
-
-system sh/cfg.sh -n dnode1 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode2 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode3 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode4 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode5 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode6 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode7 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode8 -c numOfMnodes -v 3
-
-system sh/cfg.sh -n dnode1 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode2 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode3 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode4 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode5 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode6 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode7 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode8 -c mnodeEqualVnodeNum -v 0
-
-system sh/cfg.sh -n dnode1 -c wallevel -v 1
-system sh/cfg.sh -n dnode2 -c wallevel -v 1
-system sh/cfg.sh -n dnode3 -c wallevel -v 1
-system sh/cfg.sh -n dnode4 -c wallevel -v 1
-system sh/cfg.sh -n dnode5 -c wallevel -v 1
-system sh/cfg.sh -n dnode6 -c wallevel -v 1
-system sh/cfg.sh -n dnode7 -c wallevel -v 1
-system sh/cfg.sh -n dnode8 -c wallevel -v 1
-
-system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode2 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode3 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode4 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode5 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode6 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode7 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode8 -c maxTablesPerVnode -v 4
-
-
-print ============== step1
-print ========= start dnode1
-system sh/exec.sh -n dnode1 -s start
-sql connect
-
-sql create database c_b1_d1 
-sql use c_b1_d1
-
-sql create table c_b1_t1 (t timestamp, i int)
-sql insert into c_b1_t1 values(1520000020015, 15)
-sql insert into c_b1_t1 values(1520000021014, 14)
-sql insert into c_b1_t1 values(1520000022013, 13)
-sql insert into c_b1_t1 values(1520000023012, 12)
-sql insert into c_b1_t1 values(1520000024011, 11)
-
-sql create database c_b1_d2 
-sql use c_b1_d2
-sql create table c_b1_t2 (t timestamp, i int)
-sql insert into c_b1_t2 values(1520000020025, 25)
-sql insert into c_b1_t2 values(1520000021024, 24)
-sql insert into c_b1_t2 values(1520000022023, 23)
-sql insert into c_b1_t2 values(1520000023022, 22)
-sql insert into c_b1_t2 values(1520000024021, 21)
-
-sql show dnodes
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode2Vnodes = $data2_2
-print dnode2 $dnode2Vnodes
-
-if $dnode1Vnodes != 2 then
-  return -1
-endi
-if $dnode2Vnodes != null then
-  return -1
-endi
-
-print ============================== step2
-print ========= start dnode2
-sql create dnode $hostname2
-system sh/exec.sh -n dnode2 -s start
-
-$x = 0
-show2: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 40 then
-	  return -1
-	endi
-sql show dnodes -x show2
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode2Vnodes = $data2_2
-print dnode2 $dnode2Vnodes
-
-if $dnode1Vnodes != 1 then
-  goto show2
-endi
-if $dnode2Vnodes != 1 then
-  goto show2
-endi
-
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode2Role = $data2_2
-print dnode1 ==> $dnode1Role
-print dnode2 ==> $dnode2Role
-
-print ============================== step3
-print ========= add db3
-sql create database c_b1_d3 
-sql use c_b1_d3
-sql create table c_b1_t3 (t timestamp, i int)
-sql insert into c_b1_t3 values(1520000020035, 35)
-sql insert into c_b1_t3 values(1520000021034, 34)
-sql insert into c_b1_t3 values(1520000022033, 33)
-sql insert into c_b1_t3 values(1520000023032, 32)
-sql insert into c_b1_t3 values(1520000024031, 31)
-
-print ============================== step4
-print ========= drop dnode2
-sql drop dnode $hostname2
-
-$x = 0
-show4: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 40 then
-	  return -1
-	endi
-sql show dnodes -x show4
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode2Vnodes = $data2_2
-print dnode2 $dnode2Vnodes
-
-if $dnode1Vnodes != 3 then
-  goto show4
-endi
-if $dnode2Vnodes != null then
-  goto show4
-endi
-
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode2Role = $data2_2
-$dnode3Role = $data2_3
-$dnode4Role = $data2_4
-print dnode1 ==> $dnode1Role
-print dnode2 ==> $dnode2Role
-print dnode3 ==> $dnode3Role
-print dnode4 ==> $dnode4Role
-
-print ============================== step5
-print ========= add dnode2
-system sh/exec.sh -n dnode2 -s stop -x SIGINT
-system sh/exec.sh -n dnode3 -s start
-sql create dnode $hostname3
-
-$x = 0
-show5: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 40 then 
-	  return -1
-	endi
-sql show dnodes -x show5
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode3Vnodes = $data2_3
-print dnode2 $dnode3Vnodes
-
-if $dnode1Vnodes != 2 then
-  goto show5
-endi
-if $dnode3Vnodes != 1 then
-  goto show5
-endi
-
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode3Role = $data2_3
-$dnode4Role = $data2_4
-print dnode1 ==> $dnode1Role
-print dnode3 ==> $dnode3Role
-print dnode4 ==> $dnode4Role
-
-if $dnode1Role != master then
-  return -1
-endi
-if $dnode3Role != slave then
-  return -1
-endi
-
-print ============================== step6
-sql_error drop dnode $hostname1
-
-print ============================== step7
-sql_error create dnode $hostname1
-
-print ============================== step8
-sql drop dnode $hostname3
-
-$x = 0
-show8: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 40 then
-	  return -1
-	endi
-sql show dnodes -x show8
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode3Vnodes = $data2_3
-print dnode3 $dnode3Vnodes
-
-if $dnode1Vnodes != 3 then
-  goto show8
-endi
-if $dnode3Vnodes != null then
-  goto show8
-endi
-
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode3Role = $data2_3
-$dnode4Role = $data2_4
-print dnode1 ==> $dnode1Role
-print dnode3 ==> $dnode3Role
-print dnode4 ==> $dnode4Role
-
-if $dnode1Role != master then
-  return -1
-endi
-
-if $dnode3Role != null then
-  return -1
-endi
-
-print ============================== step9
-sql create dnode $hostname4
-system sh/exec.sh -n dnode4 -s start
-
-$x = 0
-show9: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 40 then
-	  return -1
-	endi
-sql show dnodes -x show9
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode4Vnodes = $data2_4
-print dnode4 $dnode4Vnodes
-
-if $dnode1Vnodes != 2 then
-  goto show9
-endi
-if $dnode4Vnodes != 1 then
-  goto show9
-endi
-
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode4Role = $data2_4
-print dnode1 ==> $dnode1Role
-print dnode4 ==> $dnode4Role
-
-if $dnode1Role != master then
-  return -1
-endi
-
-if $dnode4Role != slave then
-  return -1
-endi
-
-print ============================== step10
-sql create database c_b1_d4 
-sql use c_b1_d4
-sql create table c_b1_t4 (t timestamp, i int)
-sql insert into c_b1_t4 values(1520000020045, 45)
-sql insert into c_b1_t4 values(1520000021044, 44)
-sql insert into c_b1_t4 values(1520000022043, 43)
-sql insert into c_b1_t4 values(1520000023042, 42)
-sql insert into c_b1_t4 values(1520000024041, 41)
-
-$x = 0
-show10: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 40 then
-	  return -1
-	endi
-sql show dnodes -x show10
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode4Vnodes = $data2_4
-print dnode4 $dnode4Vnodes
-
-if $dnode1Vnodes != 2 then
-   goto show10
-endi
-if $dnode4Vnodes != 2 then
-   goto show10
-endi
-
-sql reset query cache
-
-sql use c_b1_d3
-sql insert into c_b1_t3 values(1520000025036, 36)
-
-sql use c_b1_d2
-sql insert into c_b1_t2 values(1520000025026, 26)
-
-print ============================== step12
-sql create database c_b1_d5 
-sql use c_b1_d5
-sql_error create table c_b1_t5 (t timestamp, i int) -x error3
-
-print ============================== step13
-sql create dnode $hostname5
-system sh/exec.sh -n dnode5 -s start
-$x = 0
-step13: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 40 then
-	  return -1
-	endi
-sql show dnodes
-print dnode1 $data4_1
-print dnode2 $data4_2
-print dnode3 $data4_3
-print dnode4 $data4_4
-print dnode5 $data4_5
-
-if $data4_5 != ready then
-  goto step13
-endi  
-
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode4Role = $data2_4
-$dnode5Role = $data2_5
-print dnode1 ==> $dnode1Role
-print dnode4 ==> $dnode4Role
-print dnode5 ==> $dnode5Role
-
-sql use c_b1_d5;
-sql create table c_b1_t5 (t timestamp, i int)
-sql insert into c_b1_t5 values(1520000020055, 55)
-sql insert into c_b1_t5 values(1520000021054, 54)
-sql insert into c_b1_t5 values(1520000022053, 53)
-sql insert into c_b1_t5 values(1520000023052, 52)
-sql insert into c_b1_t5 values(1520000024051, 51)
-
-sql create database c_b1_d6 
-sql use c_b1_d6
-sql create table c_b1_t6 (t timestamp, i int)
-sql insert into c_b1_t6 values(1520000020065, 65)
-sql insert into c_b1_t6 values(1520000021064, 64)
-sql insert into c_b1_t6 values(1520000022063, 63)
-sql insert into c_b1_t6 values(1520000023062, 62)
-sql insert into c_b1_t6 values(1520000024061, 61)
-
-sql show dnodes
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode4Vnodes = $data2_4
-print dnode4 $dnode2Vnodes
-$dnode5Vnodes = $data2_5
-print dnode5 $dnode2Vnodes
-
-if $dnode1Vnodes != 2 then
-  goto step13
-endi
-if $dnode4Vnodes != 2 then
-  goto step13
-endi
-if $dnode5Vnodes != 2 then
-  goto step13
-endi
-
-print ============================== step14
-sql create dnode $hostname6
-system sh/exec.sh -n dnode6 -s start
-$x = 0
-step14: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 40 then
-	  return -1
-	endi
-sql show dnodes
-print dnode1 $data4_1
-print dnode2 $data4_2
-print dnode3 $data4_3
-print dnode4 $data4_4
-print dnode4 $data4_5
-
-if $data4_6 != ready then
-  goto step14
-endi  
-
-sql create database c_b1_d7 
-sql use c_b1_d7
-sql create table c_b1_t7 (t timestamp, i int)
-sql insert into c_b1_t7 values(1520000020075, 75)
-sql insert into c_b1_t7 values(1520000021074, 74)
-sql insert into c_b1_t7 values(1520000022073, 73)
-sql insert into c_b1_t7 values(1520000023072, 72)
-sql insert into c_b1_t7 values(1520000024071, 71)
-
-sql create database c_b1_d8 
-sql use c_b1_d8
-sql create table c_b1_t8 (t timestamp, i int)
-sql insert into c_b1_t8 values(1520000020085, 85)
-sql insert into c_b1_t8 values(1520000021084, 84)
-sql insert into c_b1_t8 values(1520000022083, 83)
-sql insert into c_b1_t8 values(1520000023082, 82)
-sql insert into c_b1_t8 values(1520000024081, 81)
-
-$x = 0
-show14: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 30 then
-	  return -1
-	endi
-sql show dnodes -x show14
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode4Vnodes = $data2_4
-print dnode4 $dnode4Vnodes
-$dnode5Vnodes = $data2_5
-print dnode5 $dnode5Vnodes
-$dnode6Vnodes = $data2_6
-print dnode6 $dnode6Vnodes
-
-if $dnode1Vnodes != 2 then
-  goto show14
-endi
-if $dnode4Vnodes != 2 then
-  goto show14
-endi
-if $dnode5Vnodes != 2 then
-  goto show14
-endi
-if $dnode6Vnodes != 2 then
-  goto show14
-endi
-
-sql reset query cache
-sleep 1000
-
-print ============================== step17
-print ========= check data
-
-sql reset query cache
-sleep 100
-
-sql use c_b1_d1
-sql select * from c_b1_d1.c_b1_t1
-print $rows 
-print $data01 $data11 $data21 $data31 $data41
-if $rows != 5 then
-  return -1
-endi
-
-sql use c_b1_d2
-sql select * from c_b1_d2.c_b1_t2
-print $rows 
-print $data01 $data11 $data21 $data31 $data41
-if $rows != 6 then
-  return -1
-endi
-
-sql use c_b1_d3
-sql select * from c_b1_d3.c_b1_t3 order by t desc
-print $rows 
-print $data01 $data11 $data21 $data31 $data41
-if $rows != 6 then
-  return -1
-endi
-if $data01 != 36 then
-  return -1
-endi
-
-sql use c_b1_d4
-sql select * from c_b1_d4.c_b1_t4 order by t desc
-print $rows 
-print $data01 $data11 $data21 $data31 $data41
-
-sql use c_b1_d5
-sql select * from c_b1_d5.c_b1_t5 order by t desc
-print $rows 
-print $data01 $data11 $data21 $data31 $data41
-if $data01 != 51 then
-  return -1
-endi
-if $data11 != 52 then
-  return -1
-endi
-if $data21 != 53 then
-  return -1
-endi
-if $data31 != 54 then
-  return -1
-endi
-if $data41 != 55 then
-  return -1
-endi
-
-sql use c_b1_d6
-sql select * from c_b1_d6.c_b1_t6 order by t desc
-print $rows 
-print $data01 $data11 $data21 $data31 $data41
-if $data01 != 61 then
-  return -1
-endi
-if $data11 != 62 then
-  return -1
-endi
-if $data21 != 63 then
-  return -1
-endi
-if $data31 != 64 then
-  return -1
-endi
-if $data41 != 65 then
-  return -1
-endi
-
-sql use c_b1_d7
-sql select * from c_b1_d7.c_b1_t7 order by t desc
-print $rows 
-print $data01 $data11 $data21 $data31 $data41
-if $data01 != 71 then
-  return -1
-endi
-if $data11 != 72 then
-  return -1
-endi
-if $data21 != 73 then
-  return -1
-endi
-if $data31 != 74 then
-  return -1
-endi
-if $data41 != 75 then
-  return -1
-endi
-
-sql use c_b1_d8
-sql select * from c_b1_d8.c_b1_t8 order by t desc
-print $rows 
-print $data01 $data11 $data21 $data31 $data41
-if $data01 != 81 then
-  return -1
-endi
-if $data11 != 82 then
-  return -1
-endi
-if $data21 != 83 then
-  return -1
-endi
-if $data31 != 84 then
-  return -1
-endi
-if $data41 != 85 then
-  return -1
-endi
-
-print ============================================ over=
-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
\ No newline at end of file

tests/script/tsim/dnode/cluster_balance2.sim

-system sh/stop_dnodes.sh
-system sh/deploy.sh -n dnode1 -i 1
-system sh/deploy.sh -n dnode2 -i 2
-system sh/deploy.sh -n dnode3 -i 3
-system sh/deploy.sh -n dnode4 -i 4
-system sh/deploy.sh -n dnode5 -i 5
-system sh/deploy.sh -n dnode6 -i 6
-system sh/deploy.sh -n dnode7 -i 7
-system sh/deploy.sh -n dnode8 -i 8
-
-system sh/cfg.sh -n dnode1 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode2 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode3 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode4 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode5 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode6 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode7 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode8 -c numOfMnodes -v 3
-
-system sh/cfg.sh -n dnode1 -c wallevel -v 1
-system sh/cfg.sh -n dnode2 -c wallevel -v 1
-system sh/cfg.sh -n dnode3 -c wallevel -v 1
-system sh/cfg.sh -n dnode4 -c wallevel -v 1
-system sh/cfg.sh -n dnode5 -c wallevel -v 1
-system sh/cfg.sh -n dnode6 -c wallevel -v 1
-system sh/cfg.sh -n dnode7 -c wallevel -v 1
-system sh/cfg.sh -n dnode8 -c wallevel -v 1
-
-system sh/cfg.sh -n dnode1 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode2 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode3 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode4 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode5 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode6 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode7 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode8 -c mnodeEqualVnodeNum -v 0
-
-system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode2 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode3 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode4 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode5 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode6 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode7 -c maxTablesPerVnode -v 4
-system sh/cfg.sh -n dnode8 -c maxTablesPerVnode -v 4
-
-
-print ============== step1
-print ========= start dnode1
-system sh/exec.sh -n dnode1 -s start
-
-sql connect
-sql create dnode $hostname2
-sql create dnode $hostname3
-system sh/exec.sh -n dnode2 -s start
-system sh/exec.sh -n dnode3 -s start
-$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
-print dnode3 $data4_3
-print dnode4 $data4_4
-
-if $data4_1 != ready then
-  goto step1
-endi
-if $data4_2 != ready then
-  goto step1
-endi
-if $data4_3 != ready then
-  goto step1
-endi
-
-sql show mnodes
-print mnode1 $data2_1
-print mnode1 $data2_2
-print mnode1 $data2_3
-if $data2_1 != master then
-  goto step1
-endi
-if $data2_2 != slave then
-  goto step1
-endi
-if $data2_3 != slave then
-  goto step1
-endi
-
-
-sql create database c_b2_d1 replica 2 
-sql use c_b2_d1
-sql create table c_b2_t1 (t timestamp, i int)
-sql insert into c_b2_t1 values(1520000020015, 15)
-sql insert into c_b2_t1 values(1520000021014, 14)
-sql insert into c_b2_t1 values(1520000022013, 13)
-sql insert into c_b2_t1 values(1520000023012, 12)
-sql insert into c_b2_t1 values(1520000024011, 11)
-
-sql create database c_b2_d2 replica 2 
-sql use c_b2_d2
-sql create table c_b2_t2 (t timestamp, i int)
-sql insert into c_b2_t2 values(1520000020025, 25)
-sql insert into c_b2_t2 values(1520000021024, 24)
-sql insert into c_b2_t2 values(1520000022023, 23)
-sql insert into c_b2_t2 values(1520000023022, 22)
-sql insert into c_b2_t2 values(1520000024021, 21)
-
-sql create database c_b2_d3 replica 2 
-sql use c_b2_d3
-sql create table c_b2_t3 (t timestamp, i int)
-sql insert into c_b2_t3 values(1520000020035, 35)
-sql insert into c_b2_t3 values(1520000021034, 34)
-sql insert into c_b2_t3 values(1520000022033, 33)
-sql insert into c_b2_t3 values(1520000023032, 32)
-sql insert into c_b2_t3 values(1520000024031, 31)
-
-$x = 0
-show1: 
-	$x = $x + 1
-	sleep 3000
-	if $x == 20 then 
-	  return -1
-	endi
-sql show dnodes -x show1
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode2Vnodes = $data2_2
-print dnode2 $dnode2Vnodes
-$dnode3Vnodes = $data2_3
-print dnode3 $dnode3Vnodes
-
-if $dnode1Vnodes != 2 then
-  goto show1
-endi
-if $dnode2Vnodes != 2 then
-  goto show1
-endi
-if $dnode3Vnodes != 2 then
-  goto show1
-endi
-
-print ============================== step2
-print ========= drop dnode2
-sql drop dnode $hostname2
-
-$x = 0
-show2: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 30 then 
-	  return -1
-	endi
-sql show dnodes -x show2
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode2Vnodes = $data2_2
-print dnode2 $dnode2Vnodes
-$dnode3Vnodes = $data2_3
-print dnode3 $dnode3Vnodes
-
-if $dnode1Vnodes != 3 then
-  goto show2
-endi
-if $dnode2Vnodes != null then
-  goto show2
-endi
-if $dnode3Vnodes != 3 then
-  goto show2
-endi
-
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode2Role = $data2_2
-$dnode3Role = $data2_3
-$dnode4Role = $data2_4
-print dnode1 ==> $dnode1Role
-print dnode2 ==> $dnode2Role
-print dnode3 ==> $dnode3Role
-print dnode4 ==> $dnode4Role
-
-system sh/exec.sh -n dnode2 -s stop  -x SIGINT
-
-print ============================== step3
-print ========= start dnode4
-sql create dnode $hostname4
-system sh/exec.sh -n dnode4 -s start
-
-$x = 0
-show3: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 30 then 
-	  return -1
-	endi
-sql show dnodes -x show3
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode4Vnodes = $data2_4
-print dnode4 $dnode4Vnodes
-$dnode3Vnodes = $data2_3
-print dnode3 $dnode3Vnodes
-
-if $dnode1Vnodes != 2 then
-  goto show3
-endi
-if $dnode4Vnodes != 2 then
-  goto show3
-endi
-if $dnode3Vnodes != 2 then
-  goto show3
-endi
-
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode2Role = $data2_2
-$dnode3Role = $data2_3
-$dnode4Role = $data2_4
-print dnode1 ==> $dnode1Role
-print dnode2 ==> $dnode2Role
-print dnode3 ==> $dnode3Role
-print dnode4 ==> $dnode4Role
-
-if $dnode1Role != master then
-  return -1
-endi
-if $dnode2Role != null then
-  return -1
-endi
-if $dnode3Role != slave then
-  return -1
-endi
-
-if $dnode4Role != slave then
-  return -1
-endi
-
-print ============================== step4
-print ========= drop dnode3
-sql drop dnode $hostname3
-
-$x = 0
-show4: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 40 then 
-	  return -1
-	endi
-sql show dnodes -x show4
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode4Vnodes = $data2_4
-print dnode4 $dnode4Vnodes
-$dnode3Vnodes = $data2_3
-print dnode3 $dnode3Vnodes
-
-if $dnode1Vnodes != 3 then
-  goto show4
-endi
-if $dnode4Vnodes != 3 then
-  goto show4
-endi
-if $dnode3Vnodes != null then
-  goto show4
-endi
-
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode2Role = $data2_2
-$dnode3Role = $data2_3
-$dnode4Role = $data2_4
-print dnode1 ==> $dnode1Role
-print dnode2 ==> $dnode2Role
-print dnode3 ==> $dnode3Role
-print dnode4 ==> $dnode4Role
-
-if $dnode1Role != master then
-  return -1
-endi
-if $dnode2Role != null then
-  return -1
-endi
-if $dnode3Role != null then
-  return -1
-endi
-
-if $dnode4Role != slave then
-  return -1
-endi
-
-system sh/exec.sh -n dnode3 -s stop  -x SIGINT
-
-print ============================== step5
-print ========= start dnode3
-sql create dnode $hostname5
-system sh/exec.sh -n dnode5 -s start
-
-$x = 0
-show5: 
-	$x = $x + 1
-	sleep 1000
-	if $x == 30 then
-	  return -1
-	endi
-sql show dnodes -x show5
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode4Vnodes = $data2_4
-print dnode4 $dnode4Vnodes
-$dnode5Vnodes = $data2_5
-print dnode5 $dnode5Vnodes
-
-if $dnode1Vnodes != 2 then
-  goto show5
-endi
-if $dnode4Vnodes != 2 then
-  goto show5
-endi
-if $dnode5Vnodes != 2 then
-  goto show5
-endi
-
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode4Role = $data2_4
-$dnode5Role = $data2_5
-print dnode1 ==> $dnode1Role
-print dnode4 ==> $dnode4Role
-print dnode5 ==> $dnode5Role
-
-print ============================== step6
-system sh/exec.sh -n dnode1 -s stop  -x SIGINT
-print stop dnode1 and sleep 3000
-sleep 3000
-
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode4Role = $data2_4
-$dnode5Role = $data2_5
-print dnode1 ==> $dnode1Role
-print dnode4 ==> $dnode4Role
-print dnode5 ==> $dnode5Role
-
-if $dnode1Role != offline then
-  return -1
-endi
-
-print ============================== step6.1
-system sh/exec.sh -n dnode1 -s start
-
-$x = 0
-step6.1:
-	$x = $x + 1
-	sleep 1000
-	if $x == 10 then
-		return -1
-	endi
-
-sql show dnodes
-print dnode1 $data4_1
-
-if $data4_1 != ready then
-  goto step6.1
-endi
-
-sql drop dnode $hostname1
-print drop dnode1 and sleep 9000
-sleep 9000
-
-$x = 0
-show6: 
-	$x = $x + 1
-	sleep 2000
-	if $x == 10 then
-	  return -1
-	endi
-sql show dnodes -x show6
-$dnode1Vnodes = $data2_1
-print dnode1 $dnode1Vnodes
-$dnode4Vnodes = $data2_4
-print dnode4 $dnode4Vnodes
-$dnode5Vnodes = $data2_5
-print dnode5 $dnode5Vnodes
-
-if $dnode1Vnodes != null then
-  goto show6
-endi
-if $dnode4Vnodes != 3 then
-  goto show6
-endi
-if $dnode5Vnodes != 3 then
-  goto show6
-endi
-
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode4Role = $data2_4
-$dnode5Role = $data2_5
-print dnode1 ==> $dnode1Role
-print dnode4 ==> $dnode4Role
-print dnode5 ==> $dnode5Role
-
-#system sh/exec.sh -n dnode1 -s stop  -x SIGINT
-
-print ============================== step12
-print ========= check data
-
-sql reset query cache
-sleep 100
-
-sql select * from c_b2_d1.c_b2_t1 order by t desc
-print $data01 $data11 $data21 $data31 $data41
-if $data01 != 11 then
-  return -1
-endi
-if $data11 != 12 then
-  return -1
-endi
-if $data21 != 13 then
-  return -1
-endi
-if $data31 != 14 then
-  return -1
-endi
-if $data41 != 15 then
-  return -1
-endi
-
-sql select * from c_b2_d2.c_b2_t2 order by t desc
-print $data01 $data11 $data21 $data31 $data41
-
-if $data01 != 21 then
-  return -1
-endi
-if $data11 != 22 then
-  return -1
-endi
-if $data21 != 23 then
-  return -1
-endi
-if $data31 != 24 then
-  return -1
-endi
-if $data41 != 25 then
-  return -1
-endi
-
-sql select * from c_b2_d3.c_b2_t3 order by t desc
-print $data01 $data11 $data21 $data31 $data41
-if $data01 != 31 then
-  return -1
-endi
-if $data11 != 32 then
-  return -1
-endi
-if $data21 != 33 then
-  return -1
-endi
-if $data31 != 34 then
-  return -1
-endi
-if $data41 != 35 then
-  return -1
-endi
-
-print ============================================ over
-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 SIGKILL
-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/tsim/dnode/cluster_cache.sim

-system sh/stop_dnodes.sh
-
-
-system sh/deploy.sh -n dnode1 -i 1
-system sh/deploy.sh -n dnode2 -i 2
-
-system sh/cfg.sh -n dnode1 -c walLevel            -v 1
-system sh/cfg.sh -n dnode2 -c walLevel            -v 1
-system sh/cfg.sh -n dnode1 -c httpMaxThreads      -v 2
-system sh/cfg.sh -n dnode2 -c httpMaxThreads      -v 2
-system sh/cfg.sh -n dnode1 -c monitor             -v 1
-system sh/cfg.sh -n dnode1 -c monitor             -v 2
-system sh/cfg.sh -n dnode2 -c http                -v 1
-system sh/cfg.sh -n dnode1 -c enableHttp          -v 1
-system sh/cfg.sh -n dnode2 -c monitor             -v 1
-system sh/cfg.sh -n dnode1 -c monitorInterval     -v 1
-system sh/cfg.sh -n dnode2 -c monitorInterval     -v 1
-
-system sh/exec.sh -n dnode1 -s start
-sleep 2000
-sql connect
-
-sql create database testdb
-sql use testdb
-sql create table meter1 (ts timestamp, val int)
-
-$x = 0
-$v = -100
-while $x < 30
-    $v = $v + 1
-	$ms = $v . m 
-	sql insert into meter1 values (now  $ms , $x ) 
-	$x = $x + 1
-endw 
-
-sleep 2000
-system sh/exec.sh -n dnode2 -s start
-sql create dnode $hostname2
-
-
-sleep 10000
-
-sql show log.tables;
-if $rows > 6 then
-  return -1
-endi
-
-sql select * from log.dn1
-print ===>rows $rows
-print $data00 $data01 $data02
-print $data10 $data11 $data12
-print $data20 $data21 $data22
-if $rows < 10 then
-  return -1
-endi
-
-#sql create table sys.st as select avg(taosd), avg(system) from sys.cpu interval(30s)
-
-sql show log.vgroups
-if $data05 != master then
-  return -1
-endi
-if $data15 != master then
-  return -1
-endi

tests/script/tsim/dnode/cluster_flowctrl.sim

-system sh/stop_dnodes.sh
-
-system sh/deploy.sh -n dnode1 -i 1
-system sh/deploy.sh -n dnode2 -i 2
-system sh/deploy.sh -n dnode3 -i 3
-
-system sh/cfg.sh -n dnode1 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode2 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode3 -c numOfMnodes -v 3
-
-system sh/cfg.sh -n dnode1 -c mnodeEqualVnodeNum -v 4
-system sh/cfg.sh -n dnode2 -c mnodeEqualVnodeNum -v 4
-system sh/cfg.sh -n dnode3 -c mnodeEqualVnodeNum -v 4
-
-system sh/cfg.sh -n dnode1 -c http -v 0
-system sh/cfg.sh -n dnode2 -c http -v 0
-system sh/cfg.sh -n dnode3 -c http -v 0
-
-system sh/cfg.sh -n dnode1 -c maxTablesPerVnode -v 20000
-system sh/cfg.sh -n dnode2 -c maxTablesPerVnode -v 20000
-system sh/cfg.sh -n dnode3 -c maxTablesPerVnode -v 20000
-
-system sh/cfg.sh -n dnode1 -c maxVgroupsPerDb -v 20
-system sh/cfg.sh -n dnode2 -c maxVgroupsPerDb -v 20
-system sh/cfg.sh -n dnode3 -c maxVgroupsPerDb -v 20
-
-system sh/cfg.sh -n dnode1 -c replica -v 3
-system sh/cfg.sh -n dnode2 -c replica -v 3
-system sh/cfg.sh -n dnode3 -c replica -v 3
-
-print ============== deploy
-
-system sh/exec.sh -n dnode1 -s start 
-sleep 5001
-sql connect
-
-sql create dnode $hostname2
-sql create dnode $hostname3
-system sh/exec.sh -n dnode2 -s start 
-system sh/exec.sh -n dnode3 -s start 
-
-print  =============== step1
-$x = 0
-show1: 
-	$x = $x + 1
-	sleep 2000
-	if $x == 5 then
-	  return -1
-	endi
-sql show mnodes -x show1
-$mnode1Role = $data2_1
-print mnode1Role $mnode1Role
-$mnode2Role = $data2_2
-print mnode2Role $mnode2Role
-$mnode3Role = $data2_3
-print mnode3Role $mnode3Role
-
-if $mnode1Role != master then
-  goto show1
-endi
-if $mnode2Role != slave then
-  goto show1
-endi
-if $mnode3Role != slave then
-  goto show1
-endi
-
-print  =============== step2
-
-sql create database db replica 3
-sql use db
-sql create table tb (ts timestamp, test int)
-
-$x = 0
-while $x < 100
-  $ms = $x . s
-  sql insert into tb values (now + $ms , $x )
-  $x = $x + 1
-endw 
-
-print  =============== step3
-sleep 2000
-
-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
-
-print  =============== step4
-sleep 3000
-system sh/exec.sh -n dnode1 -s start 
-system sh/exec.sh -n dnode2 -s start 
-system sh/exec.sh -n dnode3 -s start 
-
-print  =============== step5
-sleep 8000
-while $x < 200
-  $ms = $x . s
-  sql insert into tb values (now + $ms , $x )
-  $x = $x + 1
-endw 
-
-print  =============== step6
-system sh/exec.sh -n dnode2 -s stop  -x SIGINT
-sleep 2000
-while $x < 300
-  $ms = $x . s
-  sql insert into tb values (now + $ms , $x )
-  $x = $x + 1
-endw 
-
-system sh/exec.sh -n dnode2 -s start 
-
-sleep 6000
-print  =============== step7
-while $x < 400
-  $ms = $x . s
-  sql insert into tb values (now + $ms , $x )
-  $x = $x + 1
-  sleep 1
-endw 
-
-print  =============== step8
-sql select * from tb
-print rows $rows
-if $rows != 400 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
\ No newline at end of file

tests/script/tsim/dnode/cluster_vgroup100.sim

-system sh/stop_dnodes.sh
-system sh/deploy.sh -n dnode1 -i 1
-system sh/deploy.sh -n dnode2 -i 2
-system sh/deploy.sh -n dnode3 -i 3
-
-system sh/cfg.sh -n dnode1 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode2 -c numOfMnodes -v 3
-system sh/cfg.sh -n dnode3 -c numOfMnodes -v 3
-
-system sh/cfg.sh -n dnode1 -c maxTables -v 4
-system sh/cfg.sh -n dnode2 -c maxTables -v 4
-system sh/cfg.sh -n dnode3 -c maxTables -v 4
-
-system sh/cfg.sh -n dnode1 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode2 -c mnodeEqualVnodeNum -v 0
-system sh/cfg.sh -n dnode3 -c mnodeEqualVnodeNum -v 0
-
-print ============================== step1
-system sh/exec.sh -n dnode1 -s start
-sql connect
-
-print ============================== step2
-print ========= start dnode2
-sql create dnode $hostname2
-system sh/exec.sh -n dnode2 -s start
-sql create dnode $hostname3
-system sh/exec.sh -n dnode3 -s start
-
-$maxNum = 102
-$maxNum = 12
-
-$x = 0
-show2: 
-	$x = $x + 1
-	sleep 2000
-	if $x == 10 then
-		return -1
-	endi
-	
-sql show mnodes
-$dnode1Role = $data2_1
-$dnode2Role = $data2_2
-$dnode3Role = $data2_3
-
-print $dnode1Role
-print $dnode2Role 
-print $dnode3Role 
-
-if $dnode1Role != master then
-	goto show2
-endi
-if $dnode2Role != slave then
-	goto show2
-endi
-if $dnode3Role != slave then
-	goto show2
-endi
-
-print ============================== step3
-$count = 2
-while $count < $maxNum
-  $db = d . $count
-  $tb = $db . .t
-  $tb2 = $db . .t2
-  sql create database $db replica 3 cache 1 blocks 3
-  sql create table $tb (ts timestamp, i int)
-  sql insert into $tb values(now, 1)
-  sql create table $tb2 as select count(*) from $tb interval(10s)
-  $count = $count + 1
-  print insert into $tb values(now, 1) ==> finished
-endw
-
-print ============================== step4
-
-$count = 2
-while $count < $maxNum
-  $db = d . $count
-  $tb = $db . .t
-  sql select * from $tb
-  if $rows != 1 then
-    print select * from $tb
-    return -1
-  endi
-  $count = $count + 1
-  print select * from $tb ==> rows: $rows 
-endw
-
-print ============================== step5
-system sh/exec.sh -n dnode1 -s stop  -x SIGINT
-system sh/exec.sh -n dnode2 -s stop  -x SIGINT
-system sh/exec.sh -n dnode3 -s stop  -x SIGINT
-
-print ============================== step6
-system sh/exec.sh -n dnode1 -s start
-system sh/exec.sh -n dnode2 -s start
-system sh/exec.sh -n dnode3 -s start
-
-print ============================== step7
-
-$x = 0
-show7: 
-	$x = $x + 1
-	sleep 2000
-	if $x == 50 then
-		return -1
-	endi
-	
-sql show mnodes -x show7
-$dnode1Role = $data2_1
-$dnode2Role = $data2_2
-$dnode3Role = $data2_3
-if $dnode1Role != master then
-	goto show7
-endi
-if $dnode2Role != slave then
-	goto show7
-endi
-if $dnode2Role != slave then
-	goto show7
-endi
-
-print ============================== step8
-$x = 0
-show8: 
-	$x = $x + 1
-	sleep 2000
-	if $x == 20 then
-		return -1
-	endi
-
-$count = 2
-while $count < $maxNum
-  $db = d . $count
-  $tb = $db . .t
-  sql select * from $tb
-  if $rows != 1 then
-    print select * from $tb
-    goto show8
-  endi
-  $count = $count + 1
-  print select * from $tb ==> rows: $rows 
-endw
-
-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
\ No newline at end of file

tests/system-test/7-tmq/subscribeDb2.py

@@ -5,6 +5,7 @@ import time
 import socket
 import os
 import threading
+import math
 
 from util.log import *
 from util.sql import *
@@ -127,10 +128,14 @@ class TDTestCase:
         #tdLog.debug("doing insert data into stable:%s rows:%d ..."%(stbName, allRows))
         for i in range(ctbNum):
             sql += " %s_%d values "%(stbName,i)
+            batchRows = 0
             for j in range(rowsPerTbl):
                 sql += "(%d, %d, 'tmqrow_%d') "%(startTs + j, j, j)
-                if (j > 0) and ((j%batchNum == 0) or (j == rowsPerTbl - 1)):
+                batchRows += 1
+                # if (j > 0) and ((j%(batchNum-1) == 0) or (j == rowsPerTbl - 1)):
+                if (j > 0) and ((batchRows == batchNum) or (j == rowsPerTbl - 1)):
                     tsql.execute(sql)
+                    batchRows = 0
                     if j < rowsPerTbl - 1:
                         sql = "insert into %s_%d values " %(stbName,i)
                     else:
@@ -171,8 +176,8 @@ class TDTestCase:
                          'dbName':     'db8',    \
                          'vgroups':    4,        \
                          'stbName':    'stb',    \
-                         'ctbNum':     10,       \
-                         'rowsPerTbl': 10000,    \
+                         'ctbNum':     1,       \
+                         'rowsPerTbl': 1000,    \
                          'batchNum':   100,      \
                          'startTs':    1640966400000}  # 2022-01-01 00:00:00.000
         parameterDict['cfg'] = cfgPath
@@ -189,7 +194,7 @@ class TDTestCase:
         
         tdSql.execute("create topic %s as database %s" %(topicName1, parameterDict['dbName']))
         consumerId   = 0
-        expectrowcnt = parameterDict["rowsPerTbl"] * parameterDict["ctbNum"] / 2
+        expectrowcnt = math.ceil(parameterDict["rowsPerTbl"] * parameterDict["ctbNum"] / 2)
         topicList    = topicName1
         ifcheckdata  = 0
         ifManualCommit = 0
@@ -217,7 +222,7 @@ class TDTestCase:
         for i in range(expectRows):
             totalConsumeRows += resultList[i]
         
-        if totalConsumeRows != expectrowcnt:
+        if not (totalConsumeRows >= expectrowcnt):
             tdLog.info("act consume rows: %d, expect consume rows: %d"%(totalConsumeRows, expectrowcnt))
             tdLog.exit("tmq consume rows error!")
 
@@ -267,7 +272,7 @@ class TDTestCase:
         
         tdSql.execute("create topic %s as database %s" %(topicName1, parameterDict['dbName']))
         consumerId   = 0
-        expectrowcnt = parameterDict["rowsPerTbl"] * parameterDict["ctbNum"] / 2
+        expectrowcnt = math.ceil(parameterDict["rowsPerTbl"] * parameterDict["ctbNum"] / 2)
         topicList    = topicName1
         ifcheckdata  = 0
         ifManualCommit = 1

tools/shell/src/shellCommand.c

@@ -482,6 +482,7 @@ int32_t shellReadCommand(char *command) {
           #endif
           break;
         case 4:  // EOF or Ctrl+D
+          taosResetTerminalMode();
           printf("\n");
           return -1;
         case 5:  // ctrl E