Merge branch '3.0' into feature/stream

cmake/taostools_CMakeLists.txt.in

@@ -2,7 +2,7 @@
 # taos-tools
 ExternalProject_Add(taos-tools
         GIT_REPOSITORY https://github.com/taosdata/taos-tools.git
-        GIT_TAG d58230c
+        GIT_TAG 2849aa4
         SOURCE_DIR "${TD_SOURCE_DIR}/tools/taos-tools"
         BINARY_DIR ""
         #BUILD_IN_SOURCE TRUE

docs/en/14-reference/05-taosbenchmark.md

@@ -231,7 +231,7 @@ The parameters related to database creation are configured in `dbinfo` in the js
 
 - **name**: specify the name of the database.
 
-- **drop**: indicate whether to delete the database before inserting. The default is true.
+- **drop**: indicate whether to delete the database before inserting. The value can be 'yes' or 'no'. No means do not drop. The default is to drop.
 
 #### Stream processing related configuration parameters
 

docs/zh/12-taos-sql/10-function.md

@@ -340,7 +340,7 @@ LTRIM(expr)
 #### RTRIM
 
 ```sql
-LTRIM(expr)
+RTRIM(expr)
 ```
 
 **功能说明**：返回清除右边空格后的字符串。

docs/zh/12-taos-sql/29-changes.md

@@ -94,3 +94,10 @@ description: "TDengine 3.0 版本的语法变更说明"
 | 9 | SAMPLE | 增强 | 可以直接用于超级表了。没有PARTITION BY时，超级表的数据会被合并成一条时间线。
 | 10 | STATECOUNT | 增强 | 可以直接用于超级表了。没有PARTITION BY时，超级表的数据会被合并成一条时间线。
 | 11 | STATEDURATION | 增强 | 可以直接用于超级表了。没有PARTITION BY时，超级表的数据会被合并成一条时间线。
+
+
+## SCHEMALESS 变更
+
+| # | **元素**  | **<div style={{width: 60}}>差异性</div>** | **说明** |
+| - | :------- | :-------- | :------- |
+| 1 | 主键ts 变更为 _ts | 变更 | schemaless自动建的列名用 _ 开头，不同于2.x。

docs/zh/14-reference/05-taosbenchmark.md

@@ -231,7 +231,7 @@ taosBenchmark -A INT,DOUBLE,NCHAR,BINARY\(16\)
 
 - **name** : 数据库名。
 
-- **drop** : 插入前是否删除数据库，默认为 true。
+- **drop** : 插入前是否删除数据库，可选项为 "yes" 或者 "no", 为 "no" 时不创建。默认删除。
 
 #### 流式计算相关配置参数
 

docs/zh/14-reference/13-schemaless/13-schemaless.md

@@ -115,7 +115,7 @@ st,t1=3,t2=4,t3=t3 c1=3i64,c3="passit",c2=false,c4=4f64 1626006833639000000
 
 ## 数据模式映射规则
 
-本节将说明行协议的数据如何映射成为具有模式的数据。每个行协议中数据 measurement 映射为
+本节将说明 InfluxDB 行协议(Line Protocol)的数据如何映射成为具有模式的数据。每个行协议中数据 measurement 映射为
 超级表名称。tag_set 中的 标签名称为 数据模式中的标签名，field_set 中的名称为列名称。以如下数据为例，说明映射规则：
 
 ```json

examples/c/demoapi.c

@@ -1,5 +1,17 @@
-// C api call sequence demo
+/*
-// to compile: gcc -o apidemo apidemo.c -ltaos
+ * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
+ *
+ * This program is free software: you can use, redistribute, and/or modify
+ * it under the terms of the GNU Affero General Public License, version 3
+ * or later ("AGPL"), as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
 
 #include <stdio.h>
 #include <stdlib.h>
@@ -13,23 +25,23 @@
 #include "taos.h"
 
 #define debugPrint(fmt, ...) \
-    do { if (g_args.debug_print || g_args.verbose_print) \
+    do { if (g_args.debug_print || g_args.verbose_print) {\
-      fprintf(stdout, "DEBG: "fmt, __VA_ARGS__); } while(0)
+      fprintf(stdout, "DEBG: "fmt, __VA_ARGS__); }} while (0)
 
 #define warnPrint(fmt, ...) \
     do { fprintf(stderr, "\033[33m"); \
         fprintf(stderr, "WARN: "fmt, __VA_ARGS__); \
-        fprintf(stderr, "\033[0m"); } while(0)
+        fprintf(stderr, "\033[0m"); } while (0)
 
 #define errorPrint(fmt, ...) \
     do { fprintf(stderr, "\033[31m"); \
         fprintf(stderr, "ERROR: "fmt, __VA_ARGS__); \
-        fprintf(stderr, "\033[0m"); } while(0)
+        fprintf(stderr, "\033[0m"); } while (0)
 
 #define okPrint(fmt, ...) \
     do { fprintf(stderr, "\033[32m"); \
         fprintf(stderr, "OK: "fmt, __VA_ARGS__); \
-        fprintf(stderr, "\033[0m"); } while(0)
+        fprintf(stderr, "\033[0m"); } while (0)
 
 int64_t g_num_of_tb = 2;
 int64_t g_num_of_rec = 3;
@@ -74,10 +86,16 @@ static void prepare_data(TAOS* taos) {
     res = taos_query(taos, "create database test;");
     taos_free_result(res);
     taosMsleep(100);
-    taos_select_db(taos, "test");
+    if (taos_select_db(taos, "test")) {
+        errorPrint("%s() LN%d: taos_select_db() failed\n",
+                __func__, __LINE__);
+        return;
+    }
 
     char command[1024] = {0};
-    sprintf(command, "%s", "create table meters(ts timestamp, f float, n int, bin1 binary(20), c nchar(20), bin2 binary(20)) tags(area int, city binary(20), dist nchar(20), street binary(20));");
+    sprintf(command, "%s", "create table meters(ts timestamp, f float, n int, "
+            "bin1 binary(20), c nchar(20), bin2 binary(20)) tags(area int, "
+            "city binary(20), dist nchar(20), street binary(20));");
     res = taos_query(taos, command);
     if ((res) && (0 == taos_errno(res))) {
         okPrint("%s created\n", "meters");
@@ -117,12 +135,11 @@ static void prepare_data(TAOS* taos) {
             sprintf(command, "insert into t%"PRId64" "
                     "values(%" PRId64 ", %f, %"PRId64", "
                     "'%c%d', '%s%c%d', '%c%d')",
-                    i, 1650000000000+j, (float)j, j,
+                    i, 1650000000000+j, j * 1.0, j,
                     'a'+(int)j%25, rand(),
                     // "涛思", 'z' - (int)j%25, rand(),
                     "TAOS", 'z' - (int)j%25, rand(),
-                    'b' - (int)j%25, rand()
+                    'b' - (int)j%25, rand());
-                    );
             res = taos_query(taos,  command);
             if ((res) && (0 == taos_errno(res))) {
                 affected = taos_affected_rows(res);
@@ -150,8 +167,7 @@ static void prepare_data(TAOS* taos) {
                 i, 1650000000000+j+1, (float)j, j,
                 'a'+(int)j%25, rand(),
                 "数据", 'z' - (int)j%25, rand(),
-                'b' - (int)j%25, rand()
+                'b' - (int)j%25, rand());
-               );
         res = taos_query(taos,  command);
         if ((res) && (0 == taos_errno(res))) {
             affected = taos_affected_rows(res);
@@ -179,7 +195,8 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
     }
 
     if (block) {
-        warnPrint("%s", "call taos_fetch_block(), don't call taos_fetch_lengths()\n");
+        warnPrint("%s", "call taos_fetch_block(), "
+                "don't call taos_fetch_lengths()\n");
         int rows = 0;
         while ((rows = taos_fetch_block(res, &row))) {
             int *lengths = taos_fetch_lengths(res);
@@ -190,7 +207,7 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
                     printf("col%d type is %d, no need get offset\n",
                             f, fields[f].type);
                     for (int64_t c = 0; c < rows; c++) {
-                        switch(fields[f].type) {
+                        switch (fields[f].type) {
                             case TSDB_DATA_TYPE_TIMESTAMP:
                                 if (taos_is_null(res, c, f)) {
                                     printf("col%d, row: %"PRId64" "
@@ -199,7 +216,8 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
                                     printf("col%d, row: %"PRId64", "
                                             "value: %"PRId64"\n",
                                             f, c,
-                                            *(int64_t*)((char*)(row[f])+c*sizeof(int64_t)));
+                                            *(int64_t*)((char*)(row[f])
+                                                +c*sizeof(int64_t)));
                                 }
                                 break;
 
@@ -211,7 +229,8 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
                                     printf("col%d, row: %"PRId64", "
                                             "value: %d\n",
                                             f, c,
-                                            *(int32_t*)((char*)(row[f])+c*sizeof(int32_t)));
+                                            *(int32_t*)((char*)(row[f])
+                                                +c*sizeof(int32_t)));
                                 }
                                 break;
 
@@ -223,7 +242,8 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
                                     printf("col%d, row: %"PRId64", "
                                             "value: %f\n",
                                             f, c,
-                                            *(float*)((char*)(row[f])+c*sizeof(float)));
+                                            *(float*)((char*)(row[f])
+                                                +c*sizeof(float)));
                                 }
                                 break;
 
@@ -238,14 +258,18 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
                     if (offsets) {
                         for (int c = 0; c < rows; c++) {
                             if (offsets[c] != -1) {
-                                int length = *(int16_t*)((char*)(row[f]) + offsets[c]);
+                                int length = *(int16_t*)((char*)(row[f])
+                                        + offsets[c]);
                                 char *buf = calloc(1, length + 1);
-                                strncpy(buf, (char *)((char*)(row[f]) + offsets[c] + 2), length);
+                                strncpy(buf, (char *)((char*)(row[f])
-                                printf("row: %d, col: %d, offset: %d, length: %d, content: %s\n",
+                                            + offsets[c] + 2), length);
+                                printf("row: %d, col: %d, offset: %d, "
+                                        "length: %d, content: %s\n",
                                         c, f, offsets[c], length, buf);
                                 free(buf);
                             } else {
-                                printf("row: %d, col: %d, offset: -1, means content is NULL\n",
+                                printf("row: %d, col: %d, offset: -1, "
+                                        "means content is NULL\n",
                                         c, f);
                             }
                         }
@@ -267,7 +291,8 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
             int* lengths = taos_fetch_lengths(res);
             if (lengths) {
                 for (int c = 0; c < num_fields; c++) {
-                    printf("row: %"PRId64", col: %d, is_null: %s, length of column %d is %d\n",
+                    printf("row: %"PRId64", col: %d, is_null: %s, "
+                            "length of column %d is %d\n",
                             num_rows, c,
                             taos_is_null(res, num_rows, c)?"True":"False",
                             c, lengths[c]);
@@ -277,7 +302,7 @@ static int print_result(char *tbname, TAOS_RES* res, int block) {
                         __func__, __LINE__, tbname);
             }
 
-            num_rows ++;
+            num_rows++;
         }
     }
 
@@ -322,7 +347,8 @@ int main(int argc, char *argv[]) {
 #endif
     TAOS* taos = taos_connect(host, user, passwd, "", 0);
     if (taos == NULL) {
-        printf("\033[31mfailed to connect to db, reason:%s\033[0m\n", taos_errstr(taos));
+        printf("\033[31mfailed to connect to db, reason:%s\033[0m\n",
+                taos_errstr(taos));
         exit(1);
     }
 

include/common/tglobal.h

@@ -93,6 +93,7 @@ extern int64_t tsQueryBufferSizeBytes;  // maximum allowed usage buffer size in
 
 // query client
 extern int32_t tsQueryPolicy;
+extern int32_t tsQueryRspPolicy;
 extern int32_t tsQuerySmaOptimize;
 extern int32_t tsQueryRsmaTolerance;
 extern bool    tsQueryPlannerTrace;

include/dnode/mnode/mnode.h

@@ -30,9 +30,10 @@ typedef struct SMnode SMnode;
 
 typedef struct {
   int32_t  dnodeId;
-  bool     standby;
   bool     deploy;
-  SReplica replica;
+  int8_t   selfIndex;
+  int8_t   numOfReplicas;
+  SReplica replicas[TSDB_MAX_REPLICA];
   SMsgCb   msgCb;
 } SMnodeOpt;
 

include/libs/qcom/query.h

@@ -54,6 +54,9 @@ typedef enum {
 #define QUERY_POLICY_QNODE  3
 #define QUERY_POLICY_CLIENT 4
 
+#define QUERY_RSP_POLICY_DELAY 0
+#define QUERY_RSP_POLICY_QUICK 1
+
 typedef struct STableComInfo {
   uint8_t  numOfTags;     // the number of tags in schema
   uint8_t  precision;     // the number of precision

include/libs/sync/sync.h

@@ -211,7 +211,7 @@ typedef struct SSyncInfo {
 
 int32_t     syncInit();
 void        syncCleanUp();
-int64_t     syncOpen(const SSyncInfo* pSyncInfo);
+int64_t     syncOpen(SSyncInfo* pSyncInfo);
 void        syncStart(int64_t rid);
 void        syncStop(int64_t rid);
 int32_t     syncSetStandby(int64_t rid);
@@ -233,7 +233,7 @@ const char* syncStr(ESyncState state);
 bool        syncIsRestoreFinish(int64_t rid);
 int32_t     syncGetSnapshotByIndex(int64_t rid, SyncIndex index, SSnapshot* pSnapshot);
 
-int32_t syncReconfig(int64_t rid, const SSyncCfg* pNewCfg);
+int32_t syncReconfig(int64_t rid, SSyncCfg* pNewCfg);
 
 // build SRpcMsg, need to call syncPropose with SRpcMsg
 int32_t syncReconfigBuild(int64_t rid, const SSyncCfg* pNewCfg, SRpcMsg* pRpcMsg);

include/libs/wal/wal.h

@@ -43,6 +43,7 @@ extern "C" {
 #define WAL_FILE_LEN      (WAL_PATH_LEN + 32)
 #define WAL_MAGIC         0xFAFBFCFDF4F3F2F1ULL
 #define WAL_SCAN_BUF_SIZE (1024 * 1024 * 3)
+#define WAL_RECOV_SIZE_LIMIT (100 * WAL_SCAN_BUF_SIZE)
 
 typedef enum {
   TAOS_WAL_WRITE = 1,

include/os/osEnv.h

@@ -49,9 +49,15 @@ extern SDiskSpace tsTempSpace;
 void osDefaultInit();
 void osUpdate();
 void osCleanup();
+
 bool osLogSpaceAvailable();
 bool osDataSpaceAvailable();
 bool osTempSpaceAvailable();
+
+bool osLogSpaceSufficient();
+bool osDataSpaceSufficient();
+bool osTempSpaceSufficient();
+
 void osSetTimezone(const char *timezone);
 void osSetSystemLocale(const char *inLocale, const char *inCharSet);
 
@@ -59,4 +65,4 @@ void osSetSystemLocale(const char *inLocale, const char *inCharSet);
 }
 #endif
 
-#endif /*_TD_OS_ENV_H_*/
+#endif /*_TD_OS_ENV_H_*/

include/util/taoserror.h

@@ -450,6 +450,7 @@ int32_t* taosGetErrno();
 #define TSDB_CODE_WAL_INVALID_VER               TAOS_DEF_ERROR_CODE(0, 0x1003)
 #define TSDB_CODE_WAL_OUT_OF_MEMORY             TAOS_DEF_ERROR_CODE(0, 0x1004)
 #define TSDB_CODE_WAL_LOG_NOT_EXIST             TAOS_DEF_ERROR_CODE(0, 0x1005)
+#define TSDB_CODE_WAL_CHKSUM_MISMATCH           TAOS_DEF_ERROR_CODE(0, 0x1006)
 
 // tfs
 #define TSDB_CODE_FS_INVLD_CFG                  TAOS_DEF_ERROR_CODE(0, 0x2201)

source/client/src/clientImpl.c

@@ -1036,30 +1036,38 @@ static int32_t asyncExecSchQuery(SRequestObj* pRequest, SQuery* pQuery, SMetaDat
   pRequest->type = pQuery->msgType;
 
   SArray*      pMnodeList = taosArrayInit(4, sizeof(SQueryNodeLoad));
+
   SPlanContext cxt = {.queryId = pRequest->requestId,
-                      .acctId = pRequest->pTscObj->acctId,
+                          .acctId = pRequest->pTscObj->acctId,
-                      .mgmtEpSet = getEpSet_s(&pRequest->pTscObj->pAppInfo->mgmtEp),
+                          .mgmtEpSet = getEpSet_s(&pRequest->pTscObj->pAppInfo->mgmtEp),
-                      .pAstRoot = pQuery->pRoot,
+                          .pAstRoot = pQuery->pRoot,
-                      .showRewrite = pQuery->showRewrite,
+                          .showRewrite = pQuery->showRewrite,
-                      .pMsg = pRequest->msgBuf,
+                          .pMsg = pRequest->msgBuf,
-                      .msgLen = ERROR_MSG_BUF_DEFAULT_SIZE,
+                          .msgLen = ERROR_MSG_BUF_DEFAULT_SIZE,
-                      .pUser = pRequest->pTscObj->user,
+                          .pUser = pRequest->pTscObj->user,
-                      .sysInfo = pRequest->pTscObj->sysInfo,
+                          .sysInfo = pRequest->pTscObj->sysInfo,
-                      .allocatorId = pRequest->allocatorRefId};
+                          .allocatorId = pRequest->allocatorRefId};
-  SQueryPlan*  pDag = NULL;
+
-  int32_t      code = qCreateQueryPlan(&cxt, &pDag, pMnodeList);
+      SAppInstInfo* pAppInfo = getAppInfo(pRequest);
-  if (code) {
+      SQueryPlan*   pDag = NULL;
-    tscError("0x%" PRIx64 " failed to create query plan, code:%s 0x%" PRIx64, pRequest->self, tstrerror(code),
+
-             pRequest->requestId);
+  int64_t st = taosGetTimestampUs();
-  } else {
+  int32_t code = qCreateQueryPlan(&cxt, &pDag, pMnodeList);
-    pRequest->body.subplanNum = pDag->numOfSubplans;
+      if (code) {
+        tscError("0x%" PRIx64 " failed to create query plan, code:%s 0x%" PRIx64, pRequest->self, tstrerror(code),
+                 pRequest->requestId);
+      } else {
+        pRequest->body.subplanNum = pDag->numOfSubplans;
+      }
+
+      pRequest->metric.planEnd = taosGetTimestampUs();
+  if (code == TSDB_CODE_SUCCESS) {
+    tscDebug("0x%" PRIx64 " create query plan success, elapsed time:%.2f ms, 0x%" PRIx64, pRequest->self,
+             (pRequest->metric.planEnd - st)/1000.0, pRequest->requestId);
   }
-
+      if (TSDB_CODE_SUCCESS == code && !pRequest->validateOnly) {
-  pRequest->metric.planEnd = taosGetTimestampUs();
+        SArray* pNodeList = NULL;
-
+        buildAsyncExecNodeList(pRequest, &pNodeList, pMnodeList, pResultMeta);
-  if (TSDB_CODE_SUCCESS == code && !pRequest->validateOnly) {
-    SArray* pNodeList = NULL;
-    buildAsyncExecNodeList(pRequest, &pNodeList, pMnodeList, pResultMeta);
 
     SRequestConnInfo conn = {.pTrans = getAppInfo(pRequest)->pTransporter,
                              .requestId = pRequest->requestId,

source/client/src/clientMain.c

@@ -146,7 +146,7 @@ void taos_close(TAOS *taos) {
 
 int taos_errno(TAOS_RES *res) {
   if (res == NULL || TD_RES_TMQ_META(res)) {
-    if (terrno == TSDB_CODE_RPC_REDIRECT) terrno = TSDB_CODE_RPC_NETWORK_UNAVAIL;
+    if (terrno == TSDB_CODE_RPC_REDIRECT) terrno = TSDB_CODE_QRY_NOT_READY;
     return terrno;
   }
 
@@ -154,13 +154,12 @@ int taos_errno(TAOS_RES *res) {
     return 0;
   }
 
-  return ((SRequestObj *)res)->code == TSDB_CODE_RPC_REDIRECT ? TSDB_CODE_RPC_NETWORK_UNAVAIL
+  return ((SRequestObj *)res)->code == TSDB_CODE_RPC_REDIRECT ? TSDB_CODE_QRY_NOT_READY : ((SRequestObj *)res)->code;
-                                                              : ((SRequestObj *)res)->code;
 }
 
 const char *taos_errstr(TAOS_RES *res) {
   if (res == NULL || TD_RES_TMQ_META(res)) {
-    if (terrno == TSDB_CODE_RPC_REDIRECT) terrno = TSDB_CODE_RPC_NETWORK_UNAVAIL;
+    if (terrno == TSDB_CODE_RPC_REDIRECT) terrno = TSDB_CODE_QRY_NOT_READY;
     return (const char *)tstrerror(terrno);
   }
 
@@ -172,7 +171,7 @@ const char *taos_errstr(TAOS_RES *res) {
   if (NULL != pRequest->msgBuf && (strlen(pRequest->msgBuf) > 0 || pRequest->code == TSDB_CODE_RPC_FQDN_ERROR)) {
     return pRequest->msgBuf;
   } else {
-    return pRequest->code == TSDB_CODE_RPC_REDIRECT ? (const char *)tstrerror(TSDB_CODE_RPC_NETWORK_UNAVAIL)
+    return pRequest->code == TSDB_CODE_RPC_REDIRECT ? (const char *)tstrerror(TSDB_CODE_QRY_NOT_READY)
                                                     : (const char *)tstrerror(pRequest->code);
   }
 }
@@ -222,7 +221,9 @@ void taos_kill_query(TAOS *taos) {
 
   int64_t  rid = *(int64_t *)taos;
   STscObj *pTscObj = acquireTscObj(rid);
-  stopAllRequests(pTscObj->pRequests);
+  if (pTscObj) {
+    stopAllRequests(pTscObj->pRequests);
+  }
   releaseTscObj(rid);
 }
 

source/client/test/clientTests.cpp

@@ -52,15 +52,15 @@ void printResult(TAOS_RES* pRes) {
   int32_t n = 0;
   char    str[512] = {0};
   while ((pRow = taos_fetch_row(pRes)) != NULL) {
-    int32_t* length = taos_fetch_lengths(pRes);
+//    int32_t* length = taos_fetch_lengths(pRes);
-    for (int32_t i = 0; i < numOfFields; ++i) {
+//    for(int32_t i = 0; i < numOfFields; ++i) {
-      printf("(%d):%d ", i, length[i]);
+//      printf("(%d):%d " , i, length[i]);
-    }
+//    }
-    printf("\n");
+//    printf("\n");
-
+//
-    int32_t code = taos_print_row(str, pRow, pFields, numOfFields);
+//    int32_t code = taos_print_row(str, pRow, pFields, numOfFields);
-    printf("%s\n", str);
+//    printf("%s\n", str);
-    memset(str, 0, sizeof(str));
+//    memset(str, 0, sizeof(str));
   }
 }
 
@@ -102,17 +102,6 @@ void queryCallback(void* param, void* res, int32_t code) {
   taos_fetch_raw_block_a(res, fetchCallback, param);
 }
 
-void queryCallback1(void* param, void* res, int32_t code) {
-  if (code != TSDB_CODE_SUCCESS) {
-    printf("failed to execute, reason:%s\n", taos_errstr(res));
-  }
-
-  taos_free_result(res);
-
-  printf("exec query:\n");
-  taos_query_a(param, "select * from tm1", queryCallback, param);
-}
-
 void createNewTable(TAOS* pConn, int32_t index) {
   char str[1024] = {0};
   sprintf(str, "create table tu%d using st2 tags(%d)", index, index);
@@ -123,7 +112,7 @@ void createNewTable(TAOS* pConn, int32_t index) {
   }
   taos_free_result(pRes);
 
-  for (int32_t i = 0; i < 3280; i += 20) {
+  for(int32_t i = 0; i < 10000; i += 20) {
     char sql[1024] = {0};
     sprintf(sql,
             "insert into tu%d values(now+%da, %d)(now+%da, %d)(now+%da, %d)(now+%da, %d)"
@@ -141,10 +130,49 @@ void createNewTable(TAOS* pConn, int32_t index) {
     taos_free_result(p);
   }
 }
+
+void *queryThread(void *arg) {
+  TAOS* pConn = taos_connect("192.168.0.209", "root", "taosdata", NULL, 0);
+  if (pConn == NULL) {
+    printf("failed to connect to db, reason:%s", taos_errstr(pConn));
+    return NULL;
+  }
+
+  int64_t el = 0;
+
+  for (int32_t i = 0; i < 5000000; ++i) {
+    int64_t st = taosGetTimestampUs();
+    TAOS_RES* pRes = taos_query(pConn,
+                      "SELECT _wstart as ts,max(usage_user) FROM benchmarkcpu.host_49 WHERE  ts >= 1451618560000 AND ts < 1451622160000 INTERVAL(1m) ;");
+    if (taos_errno(pRes) != 0) {
+      printf("failed, reason:%s\n", taos_errstr(pRes));
+    } else {
+      printResult(pRes);
+    }
+
+    taos_free_result(pRes);
+    el += (taosGetTimestampUs() - st);
+    if (i % 1000 == 0 && i != 0) {
+      printf("total:%d, avg time:%.2fms\n", i, el/(double)(i*1000));
+    }
+  }
+
+  taos_close(pConn);
+  return NULL;
+}
+
+static int32_t numOfThreads = 1;
 }  // namespace
 
 int main(int argc, char** argv) {
   testing::InitGoogleTest(&argc, argv);
+  if (argc > 1) {
+    numOfThreads = atoi(argv[1]);
+  }
+
+  numOfThreads = TMAX(numOfThreads, 1);
+  printf("the runing threads is:%d", numOfThreads);
+
   return RUN_ALL_TESTS();
 }
 
@@ -664,7 +692,6 @@ TEST(testCase, insert_test) {
   taos_free_result(pRes);
   taos_close(pConn);
 }
-#endif
 
 TEST(testCase, projection_query_tables) {
   TAOS* pConn = taos_connect("localhost", "root", "taosdata", NULL, 0);
@@ -676,13 +703,14 @@ TEST(testCase, projection_query_tables) {
   //  }
   //  taos_free_result(pRes);
 
-  TAOS_RES* pRes = taos_query(pConn, "use abc1");
+  TAOS_RES* pRes = taos_query(pConn, "use benchmarkcpu");
   taos_free_result(pRes);
 
   pRes = taos_query(pConn, "create stable st1 (ts timestamp, k int) tags(a int)");
   if (taos_errno(pRes) != 0) {
     printf("failed to create table tu, reason:%s\n", taos_errstr(pRes));
   }
+
   taos_free_result(pRes);
 
   pRes = taos_query(pConn, "create stable st2 (ts timestamp, k int) tags(a int)");
@@ -722,6 +750,16 @@ TEST(testCase, projection_query_tables) {
   taos_free_result(pRes);
   taos_close(pConn);
 }
+#endif
+
+TEST(testCase, tsbs_perf_test) {
+  TdThread qid[20] = {0};
+
+  for(int32_t i = 0; i < numOfThreads; ++i) {
+    taosThreadCreate(&qid[i], NULL, queryThread, NULL);
+  }
+  getchar();
+}
 
 #if 0
 TEST(testCase, projection_query_stables) {

source/common/src/tglobal.c

@@ -90,6 +90,7 @@ bool tsSmlDataFormat = false;
 
 // query
 int32_t tsQueryPolicy = 1;
+int32_t tsQueryRspPolicy = 0;
 int32_t tsQuerySmaOptimize = 0;
 int32_t tsQueryRsmaTolerance = 1000;  // the tolerance time (ms) to judge from which level to query rsma data.
 bool    tsQueryPlannerTrace = false;
@@ -350,6 +351,7 @@ static int32_t taosAddServerCfg(SConfig *pCfg) {
   if (cfgAddInt32(pCfg, "countAlwaysReturnValue", tsCountAlwaysReturnValue, 0, 1, 0) != 0) return -1;
   if (cfgAddInt32(pCfg, "queryBufferSize", tsQueryBufferSize, -1, 500000000000, 0) != 0) return -1;
   if (cfgAddBool(pCfg, "printAuth", tsPrintAuth, 0) != 0) return -1;
+  if (cfgAddInt32(pCfg, "queryRspPolicy", tsQueryRspPolicy, 0, 1, 0) != 0) return -1;
 
   if (cfgAddInt32(pCfg, "multiProcess", tsMultiProcess, 0, 2, 0) != 0) return -1;
   if (cfgAddInt32(pCfg, "mnodeShmSize", tsMnodeShmSize, TSDB_MAX_MSG_SIZE * 2 + 1024, INT32_MAX, 0) != 0) return -1;
@@ -728,6 +730,7 @@ static int32_t taosSetServerCfg(SConfig *pCfg) {
   tsMonitorPort = (uint16_t)cfgGetItem(pCfg, "monitorPort")->i32;
   tsMonitorMaxLogs = cfgGetItem(pCfg, "monitorMaxLogs")->i32;
   tsMonitorComp = cfgGetItem(pCfg, "monitorComp")->bval;
+  tsQueryRspPolicy = cfgGetItem(pCfg, "queryRspPolicy")->i32;
 
   tsEnableTelem = cfgGetItem(pCfg, "telemetryReporting")->bval;
   tsTelemInterval = cfgGetItem(pCfg, "telemetryInterval")->i32;

source/dnode/mgmt/mgmt_mnode/inc/mmInt.h

@@ -41,8 +41,8 @@ typedef struct SMnodeMgmt {
 } SMnodeMgmt;
 
 // mmFile.c
-int32_t mmReadFile(SMnodeMgmt *pMgmt, SReplica *pReplica, bool *pDeployed);
+int32_t mmReadFile(const char *path, SMnodeOpt *pOption);
-int32_t mmWriteFile(SMnodeMgmt *pMgmt, const SReplica *pReplica, bool deployed);
+int32_t mmWriteFile(const char *path, const SMnodeOpt *pOption);
 
 // mmHandle.c
 SArray *mmGetMsgHandles();

source/dnode/mgmt/mgmt_mnode/src/mmFile.c

@@ -16,7 +16,7 @@
 #define _DEFAULT_SOURCE
 #include "mmInt.h"
 
-int32_t mmReadFile(SMnodeMgmt *pMgmt, SReplica *pReplica, bool *pDeployed) {
+int32_t mmReadFile(const char *path, SMnodeOpt *pOption) {
   int32_t   code = TSDB_CODE_INVALID_JSON_FORMAT;
   int32_t   len = 0;
   int32_t   maxLen = 4096;
@@ -25,7 +25,7 @@ int32_t mmReadFile(SMnodeMgmt *pMgmt, SReplica *pReplica, bool *pDeployed) {
   char      file[PATH_MAX] = {0};
   TdFilePtr pFile = NULL;
 
-  snprintf(file, sizeof(file), "%s%smnode.json", pMgmt->path, TD_DIRSEP);
+  snprintf(file, sizeof(file), "%s%smnode.json", path, TD_DIRSEP);
   pFile = taosOpenFile(file, TD_FILE_READ);
   if (pFile == NULL) {
     code = 0;
@@ -50,38 +50,69 @@ int32_t mmReadFile(SMnodeMgmt *pMgmt, SReplica *pReplica, bool *pDeployed) {
     dError("failed to read %s since deployed not found", file);
     goto _OVER;
   }
-  *pDeployed = deployed->valueint;
+  pOption->deploy = deployed->valueint;
 
-  cJSON *id = cJSON_GetObjectItem(root, "id");
+  cJSON *selfIndex = cJSON_GetObjectItem(root, "selfIndex");
-  if (id) {
+  if (selfIndex) {
-    if (id->type != cJSON_Number) {
+    if (selfIndex->type != cJSON_Number) {
-      dError("failed to read %s since id not found", file);
+      dError("failed to read %s since selfIndex not found", file);
       goto _OVER;
     }
-    if (pReplica) {
+    pOption->selfIndex = selfIndex->valueint;
-      pReplica->id = id->valueint;
-    }
   }
 
-  cJSON *fqdn = cJSON_GetObjectItem(root, "fqdn");
+  cJSON *replicas = cJSON_GetObjectItem(root, "replicas");
-  if (fqdn) {
+  if (replicas) {
-    if (fqdn->type != cJSON_String || fqdn->valuestring == NULL) {
+    if (replicas->type != cJSON_Array) {
-      dError("failed to read %s since fqdn not found", file);
+      dError("failed to read %s since replicas not found", file);
       goto _OVER;
     }
-    if (pReplica) {
-      tstrncpy(pReplica->fqdn, fqdn->valuestring, TSDB_FQDN_LEN);
-    }
-  }
 
-  cJSON *port = cJSON_GetObjectItem(root, "port");
+    int32_t numOfReplicas = cJSON_GetArraySize(replicas);
-  if (port) {
+    if (numOfReplicas <= 0) {
-    if (port->type != cJSON_Number) {
+      dError("failed to read %s since numOfReplicas:%d invalid", file, numOfReplicas);
-      dError("failed to read %s since port not found", file);
       goto _OVER;
     }
-    if (pReplica) {
+    pOption->numOfReplicas = numOfReplicas;
-      pReplica->port = (uint16_t)port->valueint;
+
+    for (int32_t i = 0; i < numOfReplicas; ++i) {
+      SReplica *pReplica = pOption->replicas + i;
+
+      cJSON *replica = cJSON_GetArrayItem(replicas, i);
+      if (replica == NULL) break;
+
+      cJSON *id = cJSON_GetObjectItem(replica, "id");
+      if (id) {
+        if (id->type != cJSON_Number) {
+          dError("failed to read %s since id not found", file);
+          goto _OVER;
+        }
+        if (pReplica) {
+          pReplica->id = id->valueint;
+        }
+      }
+
+      cJSON *fqdn = cJSON_GetObjectItem(replica, "fqdn");
+      if (fqdn) {
+        if (fqdn->type != cJSON_String || fqdn->valuestring == NULL) {
+          dError("failed to read %s since fqdn not found", file);
+          goto _OVER;
+        }
+        if (pReplica) {
+          tstrncpy(pReplica->fqdn, fqdn->valuestring, TSDB_FQDN_LEN);
+        }
+      }
+
+      cJSON *port = cJSON_GetObjectItem(replica, "port");
+      if (port) {
+        if (port->type != cJSON_Number) {
+          dError("failed to read %s since port not found", file);
+          goto _OVER;
+        }
+        if (pReplica) {
+          pReplica->port = (uint16_t)port->valueint;
+        }
+      }
     }
   }
 
@@ -92,18 +123,18 @@ _OVER:
   if (root != NULL) cJSON_Delete(root);
   if (pFile != NULL) taosCloseFile(&pFile);
   if (code == 0) {
-    dDebug("succcessed to read file %s, deployed:%d", file, *pDeployed);
+    dDebug("succcessed to read file %s, deployed:%d", file, pOption->deploy);
   }
 
   terrno = code;
   return code;
 }
 
-int32_t mmWriteFile(SMnodeMgmt *pMgmt, const SReplica *pReplica, bool deployed) {
+int32_t mmWriteFile(const char *path, const SMnodeOpt *pOption) {
   char file[PATH_MAX] = {0};
   char realfile[PATH_MAX] = {0};
-  snprintf(file, sizeof(file), "%s%smnode.json.bak", pMgmt->path, TD_DIRSEP);
+  snprintf(file, sizeof(file), "%s%smnode.json.bak", path, TD_DIRSEP);
-  snprintf(realfile, sizeof(realfile), "%s%smnode.json", pMgmt->path, TD_DIRSEP);
+  snprintf(realfile, sizeof(realfile), "%s%smnode.json", path, TD_DIRSEP);
 
   TdFilePtr pFile = taosOpenFile(file, TD_FILE_CREATE | TD_FILE_WRITE | TD_FILE_TRUNC);
   if (pFile == NULL) {
@@ -117,12 +148,25 @@ int32_t mmWriteFile(SMnodeMgmt *pMgmt, const SReplica *pReplica, bool deployed)
   char   *content = taosMemoryCalloc(1, maxLen + 1);
 
   len += snprintf(content + len, maxLen - len, "{\n");
-  if (pReplica != NULL && pReplica->id > 0) {
+  if (pOption->deploy && pOption->numOfReplicas > 0) {
-    len += snprintf(content + len, maxLen - len, "  \"id\": %d,\n", pReplica->id);
+    len += snprintf(content + len, maxLen - len, "  \"selfIndex\": %d,\n", pOption->selfIndex);
-    len += snprintf(content + len, maxLen - len, "  \"fqdn\": \"%s\",\n", pReplica->fqdn);
+    len += snprintf(content + len, maxLen - len, "  \"replicas\": [{\n");
-    len += snprintf(content + len, maxLen - len, "  \"port\": %u\n,", pReplica->port);
+
+    for (int32_t i = 0; i < pOption->numOfReplicas; ++i) {
+      const SReplica *pReplica = pOption->replicas + i;
+      if (pReplica != NULL && pReplica->id > 0) {
+        len += snprintf(content + len, maxLen - len, "    \"id\": %d,\n", pReplica->id);
+        len += snprintf(content + len, maxLen - len, "    \"fqdn\": \"%s\",\n", pReplica->fqdn);
+        len += snprintf(content + len, maxLen - len, "    \"port\": %u\n", pReplica->port);
+      }
+      if (i < pOption->numOfReplicas - 1) {
+        len += snprintf(content + len, maxLen - len, "  },{\n");
+      } else {
+        len += snprintf(content + len, maxLen - len, "  }],\n");
+      }
+    }
   }
-  len += snprintf(content + len, maxLen - len, "  \"deployed\": %d\n", deployed);
+  len += snprintf(content + len, maxLen - len, "  \"deployed\": %d\n", pOption->deploy);
   len += snprintf(content + len, maxLen - len, "}\n");
 
   taosWriteFile(pFile, content, len);
@@ -136,6 +180,6 @@ int32_t mmWriteFile(SMnodeMgmt *pMgmt, const SReplica *pReplica, bool deployed)
     return -1;
   }
 
-  dDebug("successed to write %s, deployed:%d", realfile, deployed);
+  dDebug("successed to write %s, deployed:%d", realfile, pOption->deploy);
   return 0;
 }

source/dnode/mgmt/mgmt_mnode/src/mmHandle.c

@@ -80,18 +80,21 @@ int32_t mmProcessCreateReq(const SMgmtInputOpt *pInput, SRpcMsg *pMsg) {
     return -1;
   }
 
-  if (createReq.replica != 1) {
+  SMnodeOpt option = {.deploy = true, .numOfReplicas = createReq.replica, .selfIndex = -1};
+  memcpy(option.replicas, createReq.replicas, sizeof(createReq.replicas));
+  for (int32_t i = 0; i < option.numOfReplicas; ++i) {
+    if (createReq.replicas[i].id == pInput->pData->dnodeId) {
+      option.selfIndex = i;
+    }
+  }
+
+  if (option.selfIndex == -1) {
     terrno = TSDB_CODE_INVALID_OPTION;
-    dGError("failed to create mnode since %s", terrstr());
+    dGError("failed to create mnode since %s, selfIndex is -1", terrstr());
     return -1;
   }
 
-  bool deployed = true;
+  if (mmWriteFile(pInput->path, &option) != 0) {
-
-  SMnodeMgmt mgmt = {0};
-  mgmt.path = pInput->path;
-  mgmt.name = pInput->name;
-  if (mmWriteFile(&mgmt, &createReq.replicas[0], deployed) != 0) {
     dGError("failed to write mnode file since %s", terrstr());
     return -1;
   }
@@ -113,12 +116,8 @@ int32_t mmProcessDropReq(const SMgmtInputOpt *pInput, SRpcMsg *pMsg) {
     return -1;
   }
 
-  bool deployed = false;
+  SMnodeOpt option = {.deploy = false};
-
+  if (mmWriteFile(pInput->path, &option) != 0) {
-  SMnodeMgmt mgmt = {0};
-  mgmt.path = pInput->path;
-  mgmt.name = pInput->name;
-  if (mmWriteFile(&mgmt, NULL, deployed) != 0) {
     dGError("failed to write mnode file since %s", terrstr());
     return -1;
   }
@@ -207,7 +206,6 @@ SArray *mmGetMsgHandles() {
   if (dmSetMgmtHandle(pArray, TDMT_MND_HEARTBEAT, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_STATUS, mmPutMsgToReadQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_SYSTABLE_RETRIEVE, mmPutMsgToReadQueue, 0) == NULL) goto _OVER;
-  // if (dmSetMgmtHandle(pArray, TDMT_MND_GRANT, mmPutMsgToWriteQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_AUTH, mmPutMsgToReadQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_SHOW_VARIABLES, mmPutMsgToReadQueue, 0) == NULL) goto _OVER;
   if (dmSetMgmtHandle(pArray, TDMT_MND_SERVER_VERSION, mmPutMsgToReadQueue, 0) == NULL) goto _OVER;

source/dnode/mgmt/mgmt_mnode/src/mmInt.c

@@ -25,38 +25,35 @@ static bool mmDeployRequired(const SMgmtInputOpt *pInput) {
 }
 
 static int32_t mmRequire(const SMgmtInputOpt *pInput, bool *required) {
-  SMnodeMgmt mgmt = {0};
+  SMnodeOpt option = {0};
-  mgmt.path = pInput->path;
+  if (mmReadFile(pInput->path, &option) != 0) {
-  if (mmReadFile(&mgmt, NULL, required) != 0) {
     return -1;
   }
 
-  if (!(*required)) {
+  if (!option.deploy) {
     *required = mmDeployRequired(pInput);
+  } else {
+    *required = true;
   }
 
   return 0;
 }
 
 static void mmBuildOptionForDeploy(SMnodeMgmt *pMgmt, const SMgmtInputOpt *pInput, SMnodeOpt *pOption) {
-  pOption->standby = false;
   pOption->deploy = true;
   pOption->msgCb = pMgmt->msgCb;
   pOption->dnodeId = pMgmt->pData->dnodeId;
-  pOption->replica.id = 1;
+  pOption->selfIndex = 0;
-  pOption->replica.port = tsServerPort;
+  pOption->numOfReplicas = 1;
-  tstrncpy(pOption->replica.fqdn, tsLocalFqdn, TSDB_FQDN_LEN);
+  pOption->replicas[0].id = 1;
+  pOption->replicas[0].port = tsServerPort;
+  tstrncpy(pOption->replicas[0].fqdn, tsLocalFqdn, TSDB_FQDN_LEN);
 }
 
-static void mmBuildOptionForOpen(SMnodeMgmt *pMgmt, const SReplica *pReplica, SMnodeOpt *pOption) {
+static void mmBuildOptionForOpen(SMnodeMgmt *pMgmt, SMnodeOpt *pOption) {
-  pOption->standby = false;
   pOption->deploy = false;
   pOption->msgCb = pMgmt->msgCb;
   pOption->dnodeId = pMgmt->pData->dnodeId;
-  if (pReplica->id > 0) {
-    pOption->standby = true;
-    pOption->replica = *pReplica;
-  }
 }
 
 static void mmClose(SMnodeMgmt *pMgmt) {
@@ -95,22 +92,20 @@ static int32_t mmOpen(SMgmtInputOpt *pInput, SMgmtOutputOpt *pOutput) {
   pMgmt->msgCb.mgmt = pMgmt;
   taosThreadRwlockInit(&pMgmt->lock, NULL);
 
-  bool     deployed = false;
+  SMnodeOpt option = {0};
-  SReplica replica = {0};
+  if (mmReadFile(pMgmt->path, &option) != 0) {
-  if (mmReadFile(pMgmt, &replica, &deployed) != 0) {
     dError("failed to read file since %s", terrstr());
     mmClose(pMgmt);
     return -1;
   }
 
-  SMnodeOpt option = {0};
+  if (!option.deploy) {
-  if (!deployed) {
     dInfo("mnode start to deploy");
     pMgmt->pData->dnodeId = 1;
     mmBuildOptionForDeploy(pMgmt, pInput, &option);
   } else {
     dInfo("mnode start to open");
-    mmBuildOptionForOpen(pMgmt, &replica, &option);
+    mmBuildOptionForOpen(pMgmt, &option);
   }
 
   pMgmt->pMnode = mndOpen(pMgmt->path, &option);
@@ -128,9 +123,10 @@ static int32_t mmOpen(SMgmtInputOpt *pInput, SMgmtOutputOpt *pOutput) {
   }
   tmsgReportStartup("mnode-worker", "initialized");
 
-  if (!deployed || replica.id > 0) {
+  if (option.numOfReplicas > 0) {
-    deployed = true;
+    option.deploy = true;
-    if (mmWriteFile(pMgmt, NULL, deployed) != 0) {
+    option.numOfReplicas = 0;
+    if (mmWriteFile(pMgmt->path, &option) != 0) {
       dError("failed to write mnode file since %s", terrstr());
       return -1;
     }

source/dnode/mgmt/mgmt_mnode/src/mmWorker.c

@@ -72,7 +72,7 @@ static void mmProcessRpcMsg(SQueueInfo *pInfo, SRpcMsg *pMsg) {
     mndPostProcessQueryMsg(pMsg);
   }
 
-  dGTrace("msg:%p, is freed, code:0x%x", pMsg, code);
+  dGTrace("msg:%p is freed, code:%s", pMsg, tstrerror(code));
   rpcFreeCont(pMsg->pCont);
   taosFreeQitem(pMsg);
 }

source/dnode/mgmt/node_mgmt/src/dmEnv.c

@@ -51,26 +51,14 @@ static int32_t dmInitMonitor() {
 
 static bool dmCheckDiskSpace() {
   osUpdate();
-  if (!osDataSpaceAvailable()) {
+  if (!osDataSpaceSufficient()) {
-    dError("free disk size: %f GB, too little, require %f GB at least at least , quit",
+    dWarn("free data disk size: %f GB, not sufficient, expected %f GB at least", (double)tsDataSpace.size.avail / 1024.0 / 1024.0 / 1024.0, (double)tsDataSpace.reserved / 1024.0 / 1024.0 / 1024.0);
-           (double)tsDataSpace.size.avail / 1024.0 / 1024.0 / 1024.0,
-           (double)tsDataSpace.reserved / 1024.0 / 1024.0 / 1024.0);
-    terrno = TSDB_CODE_NO_AVAIL_DISK;
-    return false;
   }
-  if (!osLogSpaceAvailable()) {
+  if (!osLogSpaceSufficient()) {
-    dError("free disk size: %f GB, too little, require %f GB at least at least, quit",
+    dWarn("free log disk size: %f GB, not sufficient, expected %f GB at least", (double)tsLogSpace.size.avail / 1024.0 / 1024.0 / 1024.0, (double)tsLogSpace.reserved / 1024.0 / 1024.0 / 1024.0);
-           (double)tsLogSpace.size.avail / 1024.0 / 1024.0 / 1024.0,
-           (double)tsLogSpace.reserved / 1024.0 / 1024.0 / 1024.0);
-    terrno = TSDB_CODE_NO_AVAIL_DISK;
-    return false;
   }
-  if (!osTempSpaceAvailable()) {
+  if (!osTempSpaceSufficient()) {
-    dError("free disk size: %f GB, too little, require %f GB at least at least, quit",
+    dWarn("free temp disk size: %f GB, not sufficient, expected %f GB at least", (double)tsTempSpace.size.avail / 1024.0 / 1024.0 / 1024.0, (double)tsTempSpace.reserved / 1024.0 / 1024.0 / 1024.0);
-           (double)tsTempSpace.size.avail / 1024.0 / 1024.0 / 1024.0,
-           (double)tsTempSpace.reserved / 1024.0 / 1024.0 / 1024.0);
-    terrno = TSDB_CODE_NO_AVAIL_DISK;
-    return false;
   }
   return true;
 }

source/dnode/mnode/impl/inc/mndInt.h

@@ -87,12 +87,13 @@ typedef struct {
 typedef struct {
   tsem_t   syncSem;
   int64_t  sync;
-  SReplica replica;
   int32_t  errCode;
   int32_t  transId;
   SRWLatch lock;
-  int8_t   standby;
   int8_t   leaderTransferFinish;
+  int8_t   selfIndex;
+  int8_t   numOfReplicas;
+  SReplica replicas[TSDB_MAX_REPLICA];
 } SSyncMgmt;
 
 typedef struct {
@@ -130,11 +131,10 @@ typedef struct SMnode {
 void    mndSetMsgHandle(SMnode *pMnode, tmsg_t msgType, MndMsgFp fp);
 int64_t mndGenerateUid(const char *name, int32_t len);
 
-int32_t mndAcquireRpcRef(SMnode *pMnode);
+void mndSetRestored(SMnode *pMnode, bool restored);
-void    mndReleaseRpcRef(SMnode *pMnode);
+bool mndGetRestored(SMnode *pMnode);
-void    mndSetRestore(SMnode *pMnode, bool restored);
+void mndSetStop(SMnode *pMnode);
-void    mndSetStop(SMnode *pMnode);
+bool mndGetStop(SMnode *pMnode);
-bool    mndGetStop(SMnode *pMnode);
 
 #ifdef __cplusplus
 }

source/dnode/mnode/impl/inc/mndSync.h

@@ -24,7 +24,7 @@ extern "C" {
 
 int32_t mndInitSync(SMnode *pMnode);
 void    mndCleanupSync(SMnode *pMnode);
-bool    mndIsMaster(SMnode *pMnode);
+bool    mndIsLeader(SMnode *pMnode);
 int32_t mndSyncPropose(SMnode *pMnode, SSdbRaw *pRaw, int32_t transId);
 void    mndSyncStart(SMnode *pMnode);
 void    mndSyncStop(SMnode *pMnode);

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

@@ -1809,7 +1809,7 @@ static int32_t mndRetrieveDbs(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pBloc
   }
 
   while (numOfRows < rowsCapacity) {
-    pShow->pIter = sdbFetchAll(pSdb, SDB_DB, pShow->pIter, (void **)&pDb, &objStatus);
+    pShow->pIter = sdbFetchAll(pSdb, SDB_DB, pShow->pIter, (void **)&pDb, &objStatus, true);
     if (pShow->pIter == NULL) break;
 
     if (mndCheckDbPrivilege(pMnode, pReq->info.conn.user, MND_OPER_READ_OR_WRITE_DB, pDb) == 0) {

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

@@ -423,7 +423,7 @@ static int32_t mndProcessStatusReq(SRpcMsg *pReq) {
         goto _OVER;
       } else {
         pDnode->accessTimes++;
-        mTrace("dnode:%d, status received, access times %d", pDnode->id, pDnode->accessTimes);
+        mDebug("dnode:%d, status received, access times %d", pDnode->id, pDnode->accessTimes);
       }
     }
 
@@ -471,6 +471,7 @@ static int32_t mndProcessStatusReq(SRpcMsg *pReq) {
   }
 
   pDnode->lastAccessTime = curMs;
+  pDnode->accessTimes++;
   code = 0;
 
 _OVER:

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

@@ -43,6 +43,37 @@
 #include "mndUser.h"
 #include "mndVgroup.h"
 
+static inline int32_t mndAcquireRpc(SMnode *pMnode) {
+  int32_t code = 0;
+  taosThreadRwlockRdlock(&pMnode->lock);
+  if (pMnode->stopped) {
+    terrno = TSDB_CODE_APP_NOT_READY;
+    code = -1;
+  } else if (!mndIsLeader(pMnode)) {
+    code = -1;
+  } else {
+#if 1
+    atomic_add_fetch_32(&pMnode->rpcRef, 1);
+#else
+    int32_t ref = atomic_add_fetch_32(&pMnode->rpcRef, 1);
+    mTrace("mnode rpc is acquired, ref:%d", ref);
+#endif
+  }
+  taosThreadRwlockUnlock(&pMnode->lock);
+  return code;
+}
+
+static inline void mndReleaseRpc(SMnode *pMnode) {
+  taosThreadRwlockRdlock(&pMnode->lock);
+#if 1
+  atomic_sub_fetch_32(&pMnode->rpcRef, 1);
+#else
+  int32_t ref = atomic_sub_fetch_32(&pMnode->rpcRef, 1);
+  mTrace("mnode rpc is released, ref:%d", ref);
+#endif
+  taosThreadRwlockUnlock(&pMnode->lock);
+}
+
 static void *mndBuildTimerMsg(int32_t *pContLen) {
   SMTimerReq timerReq = {0};
 
@@ -201,7 +232,7 @@ static int32_t mndInitWal(SMnode *pMnode) {
 
   pMnode->pWal = walOpen(path, &cfg);
   if (pMnode->pWal == NULL) {
-    mError("failed to open wal since %s", terrstr());
+    mError("failed to open wal since %s. wal:%s", terrstr(), path);
     return -1;
   }
 
@@ -338,8 +369,9 @@ static int32_t mndExecSteps(SMnode *pMnode) {
 static void mndSetOptions(SMnode *pMnode, const SMnodeOpt *pOption) {
   pMnode->msgCb = pOption->msgCb;
   pMnode->selfDnodeId = pOption->dnodeId;
-  pMnode->syncMgmt.replica = pOption->replica;
+  pMnode->syncMgmt.selfIndex = pOption->selfIndex;
-  pMnode->syncMgmt.standby = pOption->standby;
+  pMnode->syncMgmt.numOfReplicas = pOption->numOfReplicas;
+  memcpy(pMnode->syncMgmt.replicas, pOption->replicas, sizeof(pOption->replicas));
 }
 
 SMnode *mndOpen(const char *path, const SMnodeOpt *pOption) {
@@ -430,7 +462,7 @@ int32_t mndStart(SMnode *pMnode) {
       mError("failed to deploy sdb while start mnode");
       return -1;
     }
-    mndSetRestore(pMnode, true);
+    mndSetRestored(pMnode, true);
   }
 
   grantReset(pMnode, TSDB_GRANT_ALL, 0);
@@ -570,23 +602,27 @@ static int32_t mndCheckMnodeState(SRpcMsg *pMsg) {
       pMsg->msgType == TDMT_SCH_FETCH || pMsg->msgType == TDMT_SCH_MERGE_FETCH || pMsg->msgType == TDMT_SCH_DROP_TASK) {
     return 0;
   }
-  if (mndAcquireRpcRef(pMsg->info.node) == 0) return 0;
+  if (mndAcquireRpc(pMsg->info.node) == 0) return 0;
   if (pMsg->msgType == TDMT_MND_MQ_TIMER || pMsg->msgType == TDMT_MND_TELEM_TIMER ||
       pMsg->msgType == TDMT_MND_TRANS_TIMER || pMsg->msgType == TDMT_MND_TTL_TIMER ||
       pMsg->msgType == TDMT_MND_UPTIME_TIMER) {
     return -1;
   }
 
-  SEpSet epSet = {0};
+  SEpSet  epSet = {0};
-  mndGetMnodeEpSet(pMsg->info.node, &epSet);
+  SMnode *pMnode = pMsg->info.node;
+  mndGetMnodeEpSet(pMnode, &epSet);
 
   const STraceId *trace = &pMsg->info.traceId;
-  mError("msg:%p, failed to check mnode state since %s, type:%s, numOfMnodes:%d inUse:%d", pMsg, terrstr(),
+  mDebug(
-         TMSG_INFO(pMsg->msgType), epSet.numOfEps, epSet.inUse);
+      "msg:%p, failed to check mnode state since %s, mnode restored:%d stopped:%d, sync restored:%d role:%s type:%s "
+      "numOfEps:%d inUse:%d",
+      pMsg, terrstr(), pMnode->restored, pMnode->stopped, syncIsRestoreFinish(pMnode->syncMgmt.sync),
+      syncGetMyRoleStr(pMnode->syncMgmt.sync), TMSG_INFO(pMsg->msgType), epSet.numOfEps, epSet.inUse);
 
   if (epSet.numOfEps > 0) {
     for (int32_t i = 0; i < epSet.numOfEps; ++i) {
-      mInfo("mnode index:%d, ep:%s:%u", i, epSet.eps[i].fqdn, epSet.eps[i].port);
+      mDebug("mnode index:%d, ep:%s:%u", i, epSet.eps[i].fqdn, epSet.eps[i].port);
     }
 
     int32_t contLen = tSerializeSEpSet(NULL, 0, &epSet);
@@ -633,7 +669,7 @@ int32_t mndProcessRpcMsg(SRpcMsg *pMsg) {
 
   mGTrace("msg:%p, start to process in mnode, app:%p type:%s", pMsg, pMsg->info.ahandle, TMSG_INFO(pMsg->msgType));
   int32_t code = (*fp)(pMsg);
-  mndReleaseRpcRef(pMnode);
+  mndReleaseRpc(pMnode);
 
   if (code == TSDB_CODE_ACTION_IN_PROGRESS) {
     mGTrace("msg:%p, won't response immediately since in progress", pMsg);
@@ -669,7 +705,7 @@ int64_t mndGenerateUid(const char *name, int32_t len) {
 
 int32_t mndGetMonitorInfo(SMnode *pMnode, SMonClusterInfo *pClusterInfo, SMonVgroupInfo *pVgroupInfo,
                           SMonStbInfo *pStbInfo, SMonGrantInfo *pGrantInfo) {
-  if (mndAcquireRpcRef(pMnode) != 0) return -1;
+  if (mndAcquireRpc(pMnode) != 0) return -1;
 
   SSdb   *pSdb = pMnode->pSdb;
   int64_t ms = taosGetTimestampMs();
@@ -680,7 +716,7 @@ int32_t mndGetMonitorInfo(SMnode *pMnode, SMonClusterInfo *pClusterInfo, SMonVgr
   pStbInfo->stbs = taosArrayInit(sdbGetSize(pSdb, SDB_STB), sizeof(SMonStbDesc));
   if (pClusterInfo->dnodes == NULL || pClusterInfo->mnodes == NULL || pVgroupInfo->vgroups == NULL ||
       pStbInfo->stbs == NULL) {
-    mndReleaseRpcRef(pMnode);
+    mndReleaseRpc(pMnode);
     return -1;
   }
 
@@ -800,7 +836,7 @@ int32_t mndGetMonitorInfo(SMnode *pMnode, SMonClusterInfo *pClusterInfo, SMonVgr
     pGrantInfo->timeseries_total = INT32_MAX;
   }
 
-  mndReleaseRpcRef(pMnode);
+  mndReleaseRpc(pMnode);
   return 0;
 }
 
@@ -810,32 +846,7 @@ int32_t mndGetLoad(SMnode *pMnode, SMnodeLoad *pLoad) {
   return 0;
 }
 
-int32_t mndAcquireRpcRef(SMnode *pMnode) {
+void mndSetRestored(SMnode *pMnode, bool restored) {
-  int32_t code = 0;
-  taosThreadRwlockRdlock(&pMnode->lock);
-  if (pMnode->stopped) {
-    mTrace("mnode not running");
-    terrno = TSDB_CODE_APP_NOT_READY;
-    code = -1;
-  } else if (!mndIsMaster(pMnode)) {
-    mTrace("mnode not ready, role:%s restored:%d", syncGetMyRoleStr(pMnode->syncMgmt.sync), pMnode->restored);
-    code = -1;
-  } else {
-    int32_t ref = atomic_add_fetch_32(&pMnode->rpcRef, 1);
-    // mTrace("mnode rpc is acquired, ref:%d", ref);
-  }
-  taosThreadRwlockUnlock(&pMnode->lock);
-  return code;
-}
-
-void mndReleaseRpcRef(SMnode *pMnode) {
-  taosThreadRwlockRdlock(&pMnode->lock);
-  int32_t ref = atomic_sub_fetch_32(&pMnode->rpcRef, 1);
-  // mTrace("mnode rpc is released, ref:%d", ref);
-  taosThreadRwlockUnlock(&pMnode->lock);
-}
-
-void mndSetRestore(SMnode *pMnode, bool restored) {
   if (restored) {
     taosThreadRwlockWrlock(&pMnode->lock);
     pMnode->restored = true;

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

@@ -36,6 +36,7 @@ static int32_t  mndProcessAlterMnodeReq(SRpcMsg *pReq);
 static int32_t  mndProcessDropMnodeReq(SRpcMsg *pReq);
 static int32_t  mndRetrieveMnodes(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pBlock, int32_t rows);
 static void     mndCancelGetNextMnode(SMnode *pMnode, void *pIter);
+static void     mndReloadSyncConfig(SMnode *pMnode);
 
 int32_t mndInitMnode(SMnode *pMnode) {
   SSdbTable table = {
@@ -187,6 +188,7 @@ static int32_t mndMnodeActionInsert(SSdb *pSdb, SMnodeObj *pObj) {
   }
 
   pObj->state = TAOS_SYNC_STATE_ERROR;
+  mndReloadSyncConfig(pSdb->pMnode);
   return 0;
 }
 
@@ -203,6 +205,8 @@ static int32_t mndMnodeActionDelete(SSdb *pSdb, SMnodeObj *pObj) {
 static int32_t mndMnodeActionUpdate(SSdb *pSdb, SMnodeObj *pOld, SMnodeObj *pNew) {
   mTrace("mnode:%d, perform update action, old row:%p new row:%p", pOld->id, pOld, pNew);
   pOld->updateTime = pNew->updateTime;
+  mndReloadSyncConfig(pSdb->pMnode);
+
   return 0;
 }
 
@@ -233,7 +237,7 @@ void mndGetMnodeEpSet(SMnode *pMnode, SEpSet *pEpSet) {
     if (pIter == NULL) break;
 
     if (pObj->id == pMnode->selfDnodeId) {
-      if (mndIsMaster(pMnode)) {
+      if (mndIsLeader(pMnode)) {
         pEpSet->inUse = pEpSet->numOfEps;
       } else {
         pEpSet->inUse = (pEpSet->numOfEps + 1) % totalMnodes;
@@ -248,6 +252,10 @@ void mndGetMnodeEpSet(SMnode *pMnode, SEpSet *pEpSet) {
   if (pEpSet->numOfEps == 0) {
     syncGetRetryEpSet(pMnode->syncMgmt.sync, pEpSet);
   }
+
+  if (pEpSet->inUse >= pEpSet->numOfEps) {
+    pEpSet->inUse = 0;
+  }
 }
 
 static int32_t mndSetCreateMnodeRedoLogs(SMnode *pMnode, STrans *pTrans, SMnodeObj *pObj) {
@@ -274,13 +282,72 @@ static int32_t mndSetCreateMnodeCommitLogs(SMnode *pMnode, STrans *pTrans, SMnod
   return 0;
 }
 
+static int32_t mndBuildCreateMnodeRedoAction(STrans *pTrans, SDCreateMnodeReq *pCreateReq, SEpSet *pCreateEpSet) {
+  int32_t contLen = tSerializeSDCreateMnodeReq(NULL, 0, pCreateReq);
+  void   *pReq = taosMemoryMalloc(contLen);
+  tSerializeSDCreateMnodeReq(pReq, contLen, pCreateReq);
+
+  STransAction action = {
+      .epSet = *pCreateEpSet,
+      .pCont = pReq,
+      .contLen = contLen,
+      .msgType = TDMT_DND_CREATE_MNODE,
+      .acceptableCode = TSDB_CODE_NODE_ALREADY_DEPLOYED,
+  };
+
+  if (mndTransAppendRedoAction(pTrans, &action) != 0) {
+    taosMemoryFree(pReq);
+    return -1;
+  }
+  return 0;
+}
+
+static int32_t mndBuildAlterMnodeRedoAction(STrans *pTrans, SDCreateMnodeReq *pAlterReq, SEpSet *pAlterEpSet) {
+  int32_t contLen = tSerializeSDCreateMnodeReq(NULL, 0, pAlterReq);
+  void   *pReq = taosMemoryMalloc(contLen);
+  tSerializeSDCreateMnodeReq(pReq, contLen, pAlterReq);
+
+  STransAction action = {
+      .epSet = *pAlterEpSet,
+      .pCont = pReq,
+      .contLen = contLen,
+      .msgType = TDMT_MND_ALTER_MNODE,
+      .acceptableCode = 0,
+  };
+
+  if (mndTransAppendRedoAction(pTrans, &action) != 0) {
+    taosMemoryFree(pReq);
+    return -1;
+  }
+
+  return 0;
+}
+
+static int32_t mndBuildDropMnodeRedoAction(STrans *pTrans, SDDropMnodeReq *pDropReq, SEpSet *pDroprEpSet) {
+  int32_t contLen = tSerializeSCreateDropMQSBNodeReq(NULL, 0, pDropReq);
+  void   *pReq = taosMemoryMalloc(contLen);
+  tSerializeSCreateDropMQSBNodeReq(pReq, contLen, pDropReq);
+
+  STransAction action = {
+      .epSet = *pDroprEpSet,
+      .pCont = pReq,
+      .contLen = contLen,
+      .msgType = TDMT_DND_DROP_MNODE,
+      .acceptableCode = TSDB_CODE_NODE_NOT_DEPLOYED,
+  };
+
+  if (mndTransAppendRedoAction(pTrans, &action) != 0) {
+    taosMemoryFree(pReq);
+    return -1;
+  }
+  return 0;
+}
+
 static int32_t mndSetCreateMnodeRedoActions(SMnode *pMnode, STrans *pTrans, SDnodeObj *pDnode, SMnodeObj *pObj) {
   SSdb            *pSdb = pMnode->pSdb;
   void            *pIter = NULL;
   int32_t          numOfReplicas = 0;
-  SDAlterMnodeReq  alterReq = {0};
   SDCreateMnodeReq createReq = {0};
-  SEpSet           alterEpset = {0};
   SEpSet           createEpset = {0};
 
   while (1) {
@@ -288,75 +355,25 @@ static int32_t mndSetCreateMnodeRedoActions(SMnode *pMnode, STrans *pTrans, SDno
     pIter = sdbFetch(pSdb, SDB_MNODE, pIter, (void **)&pMObj);
     if (pIter == NULL) break;
 
-    alterReq.replicas[numOfReplicas].id = pMObj->id;
+    createReq.replicas[numOfReplicas].id = pMObj->id;
-    alterReq.replicas[numOfReplicas].port = pMObj->pDnode->port;
+    createReq.replicas[numOfReplicas].port = pMObj->pDnode->port;
-    memcpy(alterReq.replicas[numOfReplicas].fqdn, pMObj->pDnode->fqdn, TSDB_FQDN_LEN);
+    memcpy(createReq.replicas[numOfReplicas].fqdn, pMObj->pDnode->fqdn, TSDB_FQDN_LEN);
-
-    alterEpset.eps[numOfReplicas].port = pMObj->pDnode->port;
-    memcpy(alterEpset.eps[numOfReplicas].fqdn, pMObj->pDnode->fqdn, TSDB_FQDN_LEN);
-    if (pMObj->state == TAOS_SYNC_STATE_LEADER) {
-      alterEpset.inUse = numOfReplicas;
-    }
 
     numOfReplicas++;
     sdbRelease(pSdb, pMObj);
   }
 
-  alterReq.replica = numOfReplicas + 1;
+  createReq.replica = numOfReplicas + 1;
-  alterReq.replicas[numOfReplicas].id = pDnode->id;
+  createReq.replicas[numOfReplicas].id = pDnode->id;
-  alterReq.replicas[numOfReplicas].port = pDnode->port;
+  createReq.replicas[numOfReplicas].port = pDnode->port;
-  memcpy(alterReq.replicas[numOfReplicas].fqdn, pDnode->fqdn, TSDB_FQDN_LEN);
+  memcpy(createReq.replicas[numOfReplicas].fqdn, pDnode->fqdn, TSDB_FQDN_LEN);
-
-  alterEpset.numOfEps = numOfReplicas + 1;
-  alterEpset.eps[numOfReplicas].port = pDnode->port;
-  memcpy(alterEpset.eps[numOfReplicas].fqdn, pDnode->fqdn, TSDB_FQDN_LEN);
-
-  createReq.replica = 1;
-  createReq.replicas[0].id = pDnode->id;
-  createReq.replicas[0].port = pDnode->port;
-  memcpy(createReq.replicas[0].fqdn, pDnode->fqdn, TSDB_FQDN_LEN);
 
+  createEpset.inUse = 0;
   createEpset.numOfEps = 1;
   createEpset.eps[0].port = pDnode->port;
   memcpy(createEpset.eps[0].fqdn, pDnode->fqdn, TSDB_FQDN_LEN);
 
-  {
+  if (mndBuildCreateMnodeRedoAction(pTrans, &createReq, &createEpset) != 0) return -1;
-    int32_t contLen = tSerializeSDCreateMnodeReq(NULL, 0, &createReq);
-    void   *pReq = taosMemoryMalloc(contLen);
-    tSerializeSDCreateMnodeReq(pReq, contLen, &createReq);
-
-    STransAction action = {
-        .epSet = createEpset,
-        .pCont = pReq,
-        .contLen = contLen,
-        .msgType = TDMT_DND_CREATE_MNODE,
-        .acceptableCode = TSDB_CODE_NODE_ALREADY_DEPLOYED,
-    };
-
-    if (mndTransAppendRedoAction(pTrans, &action) != 0) {
-      taosMemoryFree(pReq);
-      return -1;
-    }
-  }
-
-  {
-    int32_t contLen = tSerializeSDCreateMnodeReq(NULL, 0, &alterReq);
-    void   *pReq = taosMemoryMalloc(contLen);
-    tSerializeSDCreateMnodeReq(pReq, contLen, &alterReq);
-
-    STransAction action = {
-        .epSet = alterEpset,
-        .pCont = pReq,
-        .contLen = contLen,
-        .msgType = TDMT_MND_ALTER_MNODE,
-        .acceptableCode = 0,
-    };
-
-    if (mndTransAppendRedoAction(pTrans, &action) != 0) {
-      taosMemoryFree(pReq);
-      return -1;
-    }
-  }
 
   return 0;
 }
@@ -374,9 +391,9 @@ static int32_t mndCreateMnode(SMnode *pMnode, SRpcMsg *pReq, SDnodeObj *pDnode,
   mndTransSetSerial(pTrans);
   mInfo("trans:%d, used to create mnode:%d", pTrans->id, pCreate->dnodeId);
 
+  if (mndSetCreateMnodeRedoActions(pMnode, pTrans, pDnode, &mnodeObj) != 0) goto _OVER;
   if (mndSetCreateMnodeRedoLogs(pMnode, pTrans, &mnodeObj) != 0) goto _OVER;
   if (mndSetCreateMnodeCommitLogs(pMnode, pTrans, &mnodeObj) != 0) goto _OVER;
-  if (mndSetCreateMnodeRedoActions(pMnode, pTrans, pDnode, &mnodeObj) != 0) goto _OVER;
   if (mndTransAppendNullLog(pTrans) != 0) goto _OVER;
   if (mndTransPrepare(pMnode, pTrans) != 0) goto _OVER;
 
@@ -459,107 +476,28 @@ static int32_t mndSetDropMnodeCommitLogs(SMnode *pMnode, STrans *pTrans, SMnodeO
 }
 
 static int32_t mndSetDropMnodeRedoActions(SMnode *pMnode, STrans *pTrans, SDnodeObj *pDnode, SMnodeObj *pObj) {
-  SSdb           *pSdb = pMnode->pSdb;
+  SSdb          *pSdb = pMnode->pSdb;
-  void           *pIter = NULL;
+  void          *pIter = NULL;
-  int32_t         numOfReplicas = 0;
+  int32_t        numOfReplicas = 0;
-  SDAlterMnodeReq alterReq = {0};
+  SDDropMnodeReq dropReq = {0};
-  SDDropMnodeReq  dropReq = {0};
+  SEpSet         dropEpSet = {0};
-  SSetStandbyReq  standbyReq = {0};
-  SEpSet          alterEpset = {0};
-  SEpSet          dropEpSet = {0};
-
-  while (1) {
-    SMnodeObj *pMObj = NULL;
-    pIter = sdbFetch(pSdb, SDB_MNODE, pIter, (void **)&pMObj);
-    if (pIter == NULL) break;
-    if (pMObj->id == pObj->id) {
-      sdbRelease(pSdb, pMObj);
-      continue;
-    }
-
-    alterReq.replicas[numOfReplicas].id = pMObj->id;
-    alterReq.replicas[numOfReplicas].port = pMObj->pDnode->port;
-    memcpy(alterReq.replicas[numOfReplicas].fqdn, pMObj->pDnode->fqdn, TSDB_FQDN_LEN);
-
-    alterEpset.eps[numOfReplicas].port = pMObj->pDnode->port;
-    memcpy(alterEpset.eps[numOfReplicas].fqdn, pMObj->pDnode->fqdn, TSDB_FQDN_LEN);
-    if (pMObj->state == TAOS_SYNC_STATE_LEADER) {
-      alterEpset.inUse = numOfReplicas;
-    }
-
-    numOfReplicas++;
-    sdbRelease(pSdb, pMObj);
-  }
-
-  alterReq.replica = numOfReplicas;
-  alterEpset.numOfEps = numOfReplicas;
 
   dropReq.dnodeId = pDnode->id;
   dropEpSet.numOfEps = 1;
   dropEpSet.eps[0].port = pDnode->port;
   memcpy(dropEpSet.eps[0].fqdn, pDnode->fqdn, TSDB_FQDN_LEN);
 
-  standbyReq.dnodeId = pDnode->id;
+  int32_t totalMnodes = sdbGetSize(pSdb, SDB_MNODE);
-  standbyReq.standby = 1;
+  if (totalMnodes == 2) {
-
+    mInfo("vgId:1, has %d mnodes, exec redo log first", totalMnodes);
-  {
+    if (mndSetDropMnodeRedoLogs(pMnode, pTrans, pObj) != 0) return -1;
-    int32_t contLen = tSerializeSSetStandbyReq(NULL, 0, &standbyReq) + sizeof(SMsgHead);
+    if (mndBuildDropMnodeRedoAction(pTrans, &dropReq, &dropEpSet) != 0) return -1;
-    void   *pReq = taosMemoryMalloc(contLen);
+  } else if (totalMnodes == 3) {
-    tSerializeSSetStandbyReq((char *)pReq + sizeof(SMsgHead), contLen, &standbyReq);
+    mInfo("vgId:1, has %d mnodes, exec redo action first", totalMnodes);
-    SMsgHead *pHead = pReq;
+    if (mndBuildDropMnodeRedoAction(pTrans, &dropReq, &dropEpSet) != 0) return -1;
-    pHead->contLen = htonl(contLen);
+    if (mndSetDropMnodeRedoLogs(pMnode, pTrans, pObj) != 0) return -1;
-    pHead->vgId = htonl(MNODE_HANDLE);
+  } else {
-
+    return -1;
-    STransAction action = {
-        .epSet = dropEpSet,
-        .pCont = pReq,
-        .contLen = contLen,
-        .msgType = TDMT_SYNC_SET_MNODE_STANDBY,
-        .acceptableCode = TSDB_CODE_NODE_NOT_DEPLOYED,
-    };
-
-    if (mndTransAppendRedoAction(pTrans, &action) != 0) {
-      taosMemoryFree(pReq);
-      return -1;
-    }
-  }
-
-  {
-    int32_t contLen = tSerializeSDCreateMnodeReq(NULL, 0, &alterReq);
-    void   *pReq = taosMemoryMalloc(contLen);
-    tSerializeSDCreateMnodeReq(pReq, contLen, &alterReq);
-
-    STransAction action = {
-        .epSet = alterEpset,
-        .pCont = pReq,
-        .contLen = contLen,
-        .msgType = TDMT_MND_ALTER_MNODE,
-        .acceptableCode = 0,
-    };
-
-    if (mndTransAppendRedoAction(pTrans, &action) != 0) {
-      taosMemoryFree(pReq);
-      return -1;
-    }
-  }
-
-  {
-    int32_t contLen = tSerializeSCreateDropMQSBNodeReq(NULL, 0, &dropReq);
-    void   *pReq = taosMemoryMalloc(contLen);
-    tSerializeSCreateDropMQSBNodeReq(pReq, contLen, &dropReq);
-
-    STransAction action = {
-        .epSet = dropEpSet,
-        .pCont = pReq,
-        .contLen = contLen,
-        .msgType = TDMT_DND_DROP_MNODE,
-        .acceptableCode = TSDB_CODE_NODE_NOT_DEPLOYED,
-    };
-
-    if (mndTransAppendRedoAction(pTrans, &action) != 0) {
-      taosMemoryFree(pReq);
-      return -1;
-    }
   }
 
   return 0;
@@ -567,7 +505,6 @@ static int32_t mndSetDropMnodeRedoActions(SMnode *pMnode, STrans *pTrans, SDnode
 
 int32_t mndSetDropMnodeInfoToTrans(SMnode *pMnode, STrans *pTrans, SMnodeObj *pObj) {
   if (pObj == NULL) return 0;
-  if (mndSetDropMnodeRedoLogs(pMnode, pTrans, pObj) != 0) return -1;
   if (mndSetDropMnodeCommitLogs(pMnode, pTrans, pObj) != 0) return -1;
   if (mndSetDropMnodeRedoActions(pMnode, pTrans, pObj->pDnode, pObj) != 0) return -1;
   if (mndTransAppendNullLog(pTrans) != 0) return -1;
@@ -657,7 +594,7 @@ static int32_t mndRetrieveMnodes(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pB
   int64_t    curMs = taosGetTimestampMs();
 
   while (numOfRows < rows) {
-    pShow->pIter = sdbFetchAll(pSdb, SDB_MNODE, pShow->pIter, (void **)&pObj, &objStatus);
+    pShow->pIter = sdbFetchAll(pSdb, SDB_MNODE, pShow->pIter, (void **)&pObj, &objStatus, true);
     if (pShow->pIter == NULL) break;
 
     cols = 0;
@@ -712,6 +649,9 @@ static void mndCancelGetNextMnode(SMnode *pMnode, void *pIter) {
 }
 
 static int32_t mndProcessAlterMnodeReq(SRpcMsg *pReq) {
+#if 1
+  return 0;
+#else
   SMnode         *pMnode = pReq->info.node;
   SDAlterMnodeReq alterReq = {0};
 
@@ -720,41 +660,107 @@ static int32_t mndProcessAlterMnodeReq(SRpcMsg *pReq) {
     return -1;
   }
 
+  SMnodeOpt option = {.deploy = true, .numOfReplicas = alterReq.replica, .selfIndex = -1};
+  memcpy(option.replicas, alterReq.replicas, sizeof(alterReq.replicas));
+  for (int32_t i = 0; i < option.numOfReplicas; ++i) {
+    if (alterReq.replicas[i].id == pMnode->selfDnodeId) {
+      option.selfIndex = i;
+    }
+  }
+
+  if (option.selfIndex == -1) {
+    mInfo("alter mnode not processed since selfIndex is -1", terrstr());
+    return 0;
+  }
+
+  if (mndWriteFile(pMnode->path, &option) != 0) {
+    mError("failed to write mnode file since %s", terrstr());
+    return -1;
+  }
+
   SSyncCfg cfg = {.replicaNum = alterReq.replica, .myIndex = -1};
   for (int32_t i = 0; i < alterReq.replica; ++i) {
     SNodeInfo *pNode = &cfg.nodeInfo[i];
     tstrncpy(pNode->nodeFqdn, alterReq.replicas[i].fqdn, sizeof(pNode->nodeFqdn));
     pNode->nodePort = alterReq.replicas[i].port;
-    if (alterReq.replicas[i].id == pMnode->selfDnodeId) cfg.myIndex = i;
+    if (alterReq.replicas[i].id == pMnode->selfDnodeId) {
+      cfg.myIndex = i;
+    }
   }
 
   if (cfg.myIndex == -1) {
     mError("failed to alter mnode since myindex is -1");
     return -1;
   } else {
-    mInfo("start to alter mnode sync, replica:%d myindex:%d", cfg.replicaNum, cfg.myIndex);
+    mInfo("start to alter mnode sync, replica:%d myIndex:%d", cfg.replicaNum, cfg.myIndex);
     for (int32_t i = 0; i < alterReq.replica; ++i) {
       SNodeInfo *pNode = &cfg.nodeInfo[i];
       mInfo("index:%d, fqdn:%s port:%d", i, pNode->nodeFqdn, pNode->nodePort);
     }
   }
 
-  mInfo("trans:-1, sync reconfig will be proposed");
+  int32_t code = syncReconfig(pMnode->syncMgmt.sync, &cfg);
-
-  SSyncMgmt *pMgmt = &pMnode->syncMgmt;
-  pMgmt->standby = 0;
-  int32_t code = syncReconfig(pMgmt->sync, &cfg);
   if (code != 0) {
-    mError("trans:-1, failed to propose sync reconfig since %s", terrstr());
+    mError("failed to sync reconfig since %s", terrstr());
-    return code;
   } else {
-    pMgmt->errCode = 0;
+    mInfo("alter mnode sync success");
-    taosWLockLatch(&pMgmt->lock);
+  }
-    pMgmt->transId = -1;
+
-    taosWUnLockLatch(&pMgmt->lock);
+  return code;
-    tsem_wait(&pMgmt->syncSem);
+#endif
-    mInfo("alter mnode sync result:0x%x %s", pMgmt->errCode, tstrerror(pMgmt->errCode));
+}
-    terrno = pMgmt->errCode;
+
-    return pMgmt->errCode;
+static void mndReloadSyncConfig(SMnode *pMnode) {
+  SSdb      *pSdb = pMnode->pSdb;
+  SMnodeObj *pObj = NULL;
+  ESdbStatus objStatus = 0;
+  void      *pIter = NULL;
+  bool       hasUpdatingMnode = false;
+  SSyncCfg   cfg = {.myIndex = -1};
+
+  while (1) {
+    pIter = sdbFetchAll(pSdb, SDB_MNODE, pIter, (void **)&pObj, &objStatus, false);
+    if (pIter == NULL) break;
+    if (objStatus == SDB_STATUS_CREATING || objStatus == SDB_STATUS_DROPPING) {
+      mInfo("vgId:1, has updating mnode:%d, status:%s", pObj->id, sdbStatusName(objStatus));
+      hasUpdatingMnode = true;
+    }
+
+    if (objStatus == SDB_STATUS_READY || objStatus == SDB_STATUS_CREATING) {
+      SNodeInfo *pNode = &cfg.nodeInfo[cfg.replicaNum];
+      tstrncpy(pNode->nodeFqdn, pObj->pDnode->fqdn, sizeof(pNode->nodeFqdn));
+      pNode->nodePort = pObj->pDnode->port;
+      if (pObj->pDnode->id == pMnode->selfDnodeId) {
+        cfg.myIndex = cfg.replicaNum;
+      }
+      cfg.replicaNum++;
+    }
+
+    sdbReleaseLock(pSdb, pObj, false);
+  }
+
+  if (cfg.myIndex == -1) {
+    mInfo("vgId:1, mnode not reload since selfIndex is -1");
+    return;
+  }
+
+  if (!mndGetRestored(pMnode)) {
+    mInfo("vgId:1, mnode not reload since restore not finished");
+    return;
+  }
+
+  if (hasUpdatingMnode) {
+    mInfo("vgId:1, start to reload mnode sync, replica:%d myIndex:%d", cfg.replicaNum, cfg.myIndex);
+    for (int32_t i = 0; i < cfg.replicaNum; ++i) {
+      SNodeInfo *pNode = &cfg.nodeInfo[i];
+      mInfo("vgId:1, index:%d, fqdn:%s port:%d", i, pNode->nodeFqdn, pNode->nodePort);
+    }
+
+    int32_t code = syncReconfig(pMnode->syncMgmt.sync, &cfg);
+    if (code != 0) {
+      mError("vgId:1, failed to reconfig mnode sync since %s", terrstr());
+    } else {
+      mInfo("vgId:1, reconfig mnode sync success");
+    }
   }
 }

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

@@ -107,7 +107,7 @@ void mndRestoreFinish(struct SSyncFSM *pFsm) {
   if (!pMnode->deploy) {
     mInfo("vgId:1, sync restore finished, and will handle outstanding transactions");
     mndTransPullup(pMnode);
-    mndSetRestore(pMnode, true);
+    mndSetRestored(pMnode, true);
   } else {
     mInfo("vgId:1, sync restore finished");
   }
@@ -225,18 +225,17 @@ int32_t mndInitSync(SMnode *pMnode) {
   snprintf(syncInfo.path, sizeof(syncInfo.path), "%s%ssync", pMnode->path, TD_DIRSEP);
   syncInfo.pWal = pMnode->pWal;
   syncInfo.pFsm = mndSyncMakeFsm(pMnode);
-  syncInfo.isStandBy = pMgmt->standby;
   syncInfo.snapshotStrategy = SYNC_STRATEGY_STANDARD_SNAPSHOT;
 
-  mInfo("vgId:1, start to open sync, standby:%d", pMgmt->standby);
+  mInfo("vgId:1, start to open sync, selfIndex:%d replica:%d", pMgmt->selfIndex, pMgmt->numOfReplicas);
-  if (pMgmt->standby || pMgmt->replica.id > 0) {
+  SSyncCfg *pCfg = &syncInfo.syncCfg;
-    SSyncCfg *pCfg = &syncInfo.syncCfg;
+  pCfg->replicaNum = pMgmt->numOfReplicas;
-    pCfg->replicaNum = 1;
+  pCfg->myIndex = pMgmt->selfIndex;
-    pCfg->myIndex = 0;
+  for (int32_t i = 0; i < pMgmt->numOfReplicas; ++i) {
-    SNodeInfo *pNode = &pCfg->nodeInfo[0];
+    SNodeInfo *pNode = &pCfg->nodeInfo[i];
-    tstrncpy(pNode->nodeFqdn, pMgmt->replica.fqdn, sizeof(pNode->nodeFqdn));
+    tstrncpy(pNode->nodeFqdn, pMgmt->replicas[i].fqdn, sizeof(pNode->nodeFqdn));
-    pNode->nodePort = pMgmt->replica.port;
+    pNode->nodePort = pMgmt->replicas[i].port;
-    mInfo("vgId:1, ep:%s:%u", pNode->nodeFqdn, pNode->nodePort);
+    mInfo("vgId:1, index:%d ep:%s:%u", i, pNode->nodeFqdn, pNode->nodePort);
   }
 
   tsem_init(&pMgmt->syncSem, 0, 0);
@@ -250,10 +249,6 @@ int32_t mndInitSync(SMnode *pMnode) {
   setPingTimerMS(pMgmt->sync, 5000);
   setElectTimerMS(pMgmt->sync, 3000);
   setHeartbeatTimerMS(pMgmt->sync, 500);
-  /*
-    setElectTimerMS(pMgmt->sync, 600);
-    setHeartbeatTimerMS(pMgmt->sync, 300);
-  */
 
   mInfo("mnode-sync is opened, id:%" PRId64, pMgmt->sync);
   return 0;
@@ -319,7 +314,7 @@ void mndSyncStart(SMnode *pMnode) {
   SSyncMgmt *pMgmt = &pMnode->syncMgmt;
   syncSetMsgCb(pMgmt->sync, &pMnode->msgCb);
   syncStart(pMgmt->sync);
-  mInfo("vgId:1, sync started, id:%" PRId64 " standby:%d", pMgmt->sync, pMgmt->standby);
+  mInfo("vgId:1, sync started, id:%" PRId64, pMgmt->sync);
 }
 
 void mndSyncStop(SMnode *pMnode) {
@@ -331,7 +326,7 @@ void mndSyncStop(SMnode *pMnode) {
   taosWUnLockLatch(&pMnode->syncMgmt.lock);
 }
 
-bool mndIsMaster(SMnode *pMnode) {
+bool mndIsLeader(SMnode *pMnode) {
   SSyncMgmt *pMgmt = &pMnode->syncMgmt;
 
   if (!syncIsReady(pMgmt->sync)) {
@@ -340,7 +335,7 @@ bool mndIsMaster(SMnode *pMnode) {
     return false;
   }
 
-  if (!pMnode->restored) {
+  if (!mndGetRestored(pMnode)) {
     terrno = TSDB_CODE_APP_NOT_READY;
     return false;
   }

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

@@ -53,7 +53,7 @@ static bool    mndTransPerformCommitActionStage(SMnode *pMnode, STrans *pTrans);
 static bool    mndTransPerformCommitStage(SMnode *pMnode, STrans *pTrans);
 static bool    mndTransPerformRollbackStage(SMnode *pMnode, STrans *pTrans);
 static bool    mndTransPerfromFinishedStage(SMnode *pMnode, STrans *pTrans);
-static bool    mndCannotExecuteTransAction(SMnode *pMnode) { return !pMnode->deploy && !mndIsMaster(pMnode); }
+static bool    mndCannotExecuteTransAction(SMnode *pMnode) { return !pMnode->deploy && !mndIsLeader(pMnode); }
 
 static void    mndTransSendRpcRsp(SMnode *pMnode, STrans *pTrans);
 static int32_t mndProcessTransTimer(SRpcMsg *pReq);

source/dnode/mnode/sdb/inc/sdb.h

@@ -298,6 +298,7 @@ void *sdbAcquire(SSdb *pSdb, ESdbType type, const void *pKey);
  * @param pObj The object of the row.
  */
 void sdbRelease(SSdb *pSdb, void *pObj);
+void sdbReleaseLock(SSdb *pSdb, void *pObj, bool lock);
 
 /**
  * @brief Traverse a sdb table
@@ -309,7 +310,7 @@ void sdbRelease(SSdb *pSdb, void *pObj);
  * @return void* The next iterator of the table.
  */
 void *sdbFetch(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj);
-void *sdbFetchAll(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj, ESdbStatus *status);
+void *sdbFetchAll(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj, ESdbStatus *status, bool lock);
 
 /**
  * @brief Cancel a traversal

source/dnode/mnode/sdb/src/sdbHash.c

@@ -327,14 +327,16 @@ static void sdbCheckRow(SSdb *pSdb, SSdbRow *pRow) {
   taosThreadRwlockUnlock(pLock);
 }
 
-void sdbRelease(SSdb *pSdb, void *pObj) {
+void sdbReleaseLock(SSdb *pSdb, void *pObj, bool lock) {
   if (pObj == NULL) return;
 
   SSdbRow *pRow = (SSdbRow *)((char *)pObj - sizeof(SSdbRow));
   if (pRow->type >= SDB_MAX) return;
 
   TdThreadRwlock *pLock = &pSdb->locks[pRow->type];
-  taosThreadRwlockWrlock(pLock);
+  if (lock) {
+    taosThreadRwlockWrlock(pLock);
+  }
 
   int32_t ref = atomic_sub_fetch_32(&pRow->refCount, 1);
   sdbPrintOper(pSdb, pRow, "release");
@@ -342,9 +344,13 @@ void sdbRelease(SSdb *pSdb, void *pObj) {
     sdbFreeRow(pSdb, pRow, true);
   }
 
-  taosThreadRwlockUnlock(pLock);
+  if (lock) {
+    taosThreadRwlockUnlock(pLock);
+  }
 }
 
+void sdbRelease(SSdb *pSdb, void *pObj) { sdbReleaseLock(pSdb, pObj, true); }
+
 void *sdbFetch(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj) {
   *ppObj = NULL;
 
@@ -372,14 +378,16 @@ void *sdbFetch(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj) {
   return ppRow;
 }
 
-void *sdbFetchAll(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj, ESdbStatus *status) {
+void *sdbFetchAll(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj, ESdbStatus *status, bool lock) {
   *ppObj = NULL;
 
   SHashObj *hash = sdbGetHash(pSdb, type);
   if (hash == NULL) return NULL;
 
   TdThreadRwlock *pLock = &pSdb->locks[type];
-  taosThreadRwlockRdlock(pLock);
+  if (lock) {
+    taosThreadRwlockRdlock(pLock);
+  }
 
   SSdbRow **ppRow = taosHashIterate(hash, pIter);
   while (ppRow != NULL) {
@@ -395,7 +403,9 @@ void *sdbFetchAll(SSdb *pSdb, ESdbType type, void *pIter, void **ppObj, ESdbStat
     *status = pRow->status;
     break;
   }
-  taosThreadRwlockUnlock(pLock);
+  if (lock) {
+    taosThreadRwlockUnlock(pLock);
+  }
 
   return ppRow;
 }

source/dnode/vnode/src/meta/metaTable.c

@@ -693,6 +693,7 @@ static int metaAlterTableColumn(SMeta *pMeta, int64_t version, SVAlterTbReq *pAl
     if (iCol >= pSchema->nCols) break;
     pColumn = &pSchema->pSchema[iCol];
 
+    ASSERT(pAlterTbReq->colName);
     if (strcmp(pColumn->name, pAlterTbReq->colName) == 0) break;
     iCol++;
   }

source/dnode/vnode/src/tsdb/tsdbCache.c

@@ -378,10 +378,10 @@ static int32_t getTableDelSkyline(STbData *pMem, STbData *pIMem, SDelFReader *pD
     if (code) goto _err;
   }
 
+_err:
   if (aDelData) {
     taosArrayDestroy(aDelData);
   }
-_err:
   return code;
 }
 
@@ -399,14 +399,13 @@ static int32_t getTableDelIdx(SDelFReader *pDelFReader, tb_uid_t suid, tb_uid_t
 
   // code = tMapDataSearch(&delIdxMap, &idx, tGetDelIdx, tCmprDelIdx, pDelIdx);
   SDelIdx *pIdx = taosArraySearch(pDelIdxArray, &idx, tCmprDelIdx, TD_EQ);
-  if (code) goto _err;
 
   *pDelIdx = *pIdx;
 
+_err:
   if (pDelIdxArray) {
     taosArrayDestroy(pDelIdxArray);
   }
-_err:
   return code;
 }
 
@@ -599,8 +598,6 @@ static int32_t getNextRowFromFS(void *iter, TSDBROW **ppRow) {
        * &state->blockIdx);
        */
       state->pBlockIdx = taosArraySearch(state->aBlockIdx, state->pBlockIdxExp, tCmprBlockIdx, TD_EQ);
-      if (code) goto _err;
-
       if (!state->pBlockIdx) {
         goto _next_fileset;
       }
@@ -898,10 +895,16 @@ static int32_t nextRowIterOpen(CacheNextRowIter *pIter, tb_uid_t uid, STsdb *pTs
     if (code) goto _err;
 
     code = getTableDelIdx(pDelFReader, suid, uid, &delIdx);
-    if (code) goto _err;
+    if (code) {
+      tsdbDelFReaderClose(&pDelFReader);
+      goto _err;
+    }
 
     code = getTableDelSkyline(pMem, pIMem, pDelFReader, &delIdx, pIter->pSkyline);
-    if (code) goto _err;
+    if (code) {
+      tsdbDelFReaderClose(&pDelFReader);
+      goto _err;
+    }
 
     tsdbDelFReaderClose(&pDelFReader);
   } else {
@@ -1231,6 +1234,8 @@ static int32_t mergeLast(tb_uid_t uid, STsdb *pTsdb, SArray **ppLastArray) {
 _err:
   nextRowIterClose(&iter);
   taosMemoryFreeClear(pTSchema);
+  *ppLastArray = NULL;
+  taosArrayDestroy(pColArray);
   return code;
 }
 

source/dnode/vnode/src/tsdb/tsdbCommit.c

@@ -1045,7 +1045,9 @@ static int32_t tsdbEndCommit(SCommitter *pCommitter, int32_t eno) {
   STsdb     *pTsdb = pCommitter->pTsdb;
   SMemTable *pMemTable = pTsdb->imem;
 
-  ASSERT(eno == 0);
+  ASSERT(eno == 0 &&
+    "tsdbCommit failure"
+    "Restart taosd");
 
   code = tsdbFSCommit1(pTsdb, &pCommitter->fs);
   TSDB_CHECK_CODE(code, lino, _exit);

source/dnode/vnode/src/tsdb/tsdbFS.c

@@ -961,7 +961,7 @@ int32_t tsdbFSCommit2(STsdb *pTsdb, STsdbFS *pFSNew) {
     nRef = atomic_sub_fetch_32(&pSetOld->pHeadF->nRef, 1);
     if (nRef == 0) {
       tsdbHeadFileName(pTsdb, pSetOld->diskId, pSetOld->fid, pSetOld->pHeadF, fname);
-      taosRemoveFile(fname);
+      (void)taosRemoveFile(fname);
       taosMemoryFree(pSetOld->pHeadF);
     }
 
@@ -1114,7 +1114,7 @@ void tsdbFSUnref(STsdb *pTsdb, STsdbFS *pFS) {
     ASSERT(nRef >= 0);
     if (nRef == 0) {
       tsdbHeadFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pHeadF, fname);
-      taosRemoveFile(fname);
+      (void)taosRemoveFile(fname);
       taosMemoryFree(pSet->pHeadF);
     }
 

source/dnode/vnode/src/tsdb/tsdbMemTable.c

@@ -160,6 +160,7 @@ int32_t tsdbDeleteTableData(STsdb *pTsdb, int64_t version, tb_uid_t suid, tb_uid
     goto _err;
   }
 
+  ASSERT(pPool != NULL);
   // do delete
   SDelData *pDelData = (SDelData *)vnodeBufPoolMalloc(pPool, sizeof(*pDelData));
   if (pDelData == NULL) {
@@ -353,6 +354,7 @@ static int32_t tsdbGetOrCreateTbData(SMemTable *pMemTable, tb_uid_t suid, tb_uid
   SVBufPool *pPool = pMemTable->pTsdb->pVnode->inUse;
   int8_t     maxLevel = pMemTable->pTsdb->pVnode->config.tsdbCfg.slLevel;
 
+  ASSERT(pPool != NULL);
   pTbData = vnodeBufPoolMalloc(pPool, sizeof(*pTbData) + SL_NODE_SIZE(maxLevel) * 2);
   if (pTbData == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;
@@ -492,6 +494,7 @@ static int32_t tbDataDoPut(SMemTable *pMemTable, STbData *pTbData, SMemSkipListN
 
   // node
   level = tsdbMemSkipListRandLevel(&pTbData->sl);
+  ASSERT(pPool != NULL);
   pNode = (SMemSkipListNode *)vnodeBufPoolMalloc(pPool, SL_NODE_SIZE(level) + tPutTSDBRow(NULL, pRow));
   if (pNode == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;

source/dnode/vnode/src/tsdb/tsdbMergeTree.c

@@ -133,8 +133,15 @@ static SBlockData *loadLastBlock(SLDataIter *pIter, const char *idStr) {
       id.uid = pIter->uid;
     }
 
-    tBlockDataInit(pBlock, &id, pInfo->pSchema, pInfo->colIds, pInfo->numOfCols);
+    code = tBlockDataInit(pBlock, &id, pInfo->pSchema, pInfo->colIds, pInfo->numOfCols);
+    if (code != TSDB_CODE_SUCCESS) {
+      goto _exit;
+    }
+
     code = tsdbReadSttBlock(pIter->pReader, pIter->iStt, pIter->pSttBlk, pBlock);
+    if (code != TSDB_CODE_SUCCESS) {
+      goto _exit;
+    }
 
     double el = (taosGetTimestampUs() - st) / 1000.0;
     pInfo->elapsedTime += el;
@@ -543,6 +550,7 @@ int32_t tMergeTreeOpen(SMergeTree *pMTree, int8_t backward, SDataFReader *pFRead
 
   pMTree->pLoadInfo = pBlockLoadInfo;
   pMTree->destroyLoadInfo = destroyLoadInfo;
+  ASSERT(pMTree->pLoadInfo != NULL);
 
   for (int32_t i = 0; i < pFReader->pSet->nSttF; ++i) {  // open all last file
     struct SLDataIter *pIter = NULL;

source/dnode/vnode/src/tsdb/tsdbRead.c

@@ -2676,9 +2676,9 @@ static STsdb* getTsdbByRetentions(SVnode* pVnode, TSKEY winSKey, SRetention* ret
     int8_t  level = 0;
     int8_t  precision = pVnode->config.tsdbCfg.precision;
     int64_t now = taosGetTimestamp(precision);
-    int64_t offset = tsQueryRsmaTolerance * ((precision == TSDB_TIME_PRECISION_MILLI)   ? 1
+    int64_t offset = tsQueryRsmaTolerance * ((precision == TSDB_TIME_PRECISION_MILLI)   ? 1L
-                                             : (precision == TSDB_TIME_PRECISION_MICRO) ? 1000
+                                             : (precision == TSDB_TIME_PRECISION_MICRO) ? 1000L
-                                                                                        : 1000000);
+                                                                                        : 1000000L);
 
     for (int8_t i = 0; i < TSDB_RETENTION_MAX; ++i) {
       SRetention* pRetention = retentions + level;
@@ -3372,14 +3372,13 @@ int32_t tsdbReaderOpen(SVnode* pVnode, SQueryTableDataCond* pCond, SArray* pTabl
   if (pCond->suid != 0) {
     pReader->pSchema = metaGetTbTSchema(pReader->pTsdb->pVnode->pMeta, pReader->suid, -1, 1);
     if (pReader->pSchema == NULL) {
-      tsdbError("failed to get table schema, suid:%" PRIu64 ", ver:%" PRId64 " , %s", pReader->suid, -1,
+      tsdbError("failed to get table schema, suid:%" PRIu64 ", ver:-1, %s", pReader->suid, pReader->idStr);
-                pReader->idStr);
     }
   } else if (taosArrayGetSize(pTableList) > 0) {
     STableKeyInfo* pKey = taosArrayGet(pTableList, 0);
     pReader->pSchema = metaGetTbTSchema(pReader->pTsdb->pVnode->pMeta, pKey->uid, -1, 1);
     if (pReader->pSchema == NULL) {
-      tsdbError("failed to get table schema, uid:%" PRIu64 ", ver:%" PRId64 " , %s", pKey->uid, -1, pReader->idStr);
+      tsdbError("failed to get table schema, uid:%" PRIu64 ", ver:-1, %s", pKey->uid, pReader->idStr);
     }
   }
 
@@ -3395,19 +3394,20 @@ int32_t tsdbReaderOpen(SVnode* pVnode, SQueryTableDataCond* pCond, SArray* pTabl
     goto _err;
   }
 
-  code = tsdbTakeReadSnap(pReader->pTsdb, &pReader->pReadSnap, pReader->idStr);
+  if (numOfTables > 0) {
-  if (code != TSDB_CODE_SUCCESS) {
+    code = tsdbTakeReadSnap(pReader->pTsdb, &pReader->pReadSnap, pReader->idStr);
-    goto _err;
-  }
-
-  if (pReader->type == TIMEWINDOW_RANGE_CONTAINED) {
-    code = doOpenReaderImpl(pReader);
     if (code != TSDB_CODE_SUCCESS) {
-      return code;
+      goto _err;
     }
-  } else {
+
-    STsdbReader* pPrevReader = pReader->innerReader[0];
+    if (pReader->type == TIMEWINDOW_RANGE_CONTAINED) {
-    STsdbReader* pNextReader = pReader->innerReader[1];
+      code = doOpenReaderImpl(pReader);
+      if (code != TSDB_CODE_SUCCESS) {
+        return code;
+      }
+    } else {
+      STsdbReader* pPrevReader = pReader->innerReader[0];
+      STsdbReader* pNextReader = pReader->innerReader[1];
 
     // we need only one row
     pPrevReader->capacity = 1;
@@ -3422,19 +3422,20 @@ int32_t tsdbReaderOpen(SVnode* pVnode, SQueryTableDataCond* pCond, SArray* pTabl
     pNextReader->pMemSchema = pReader->pMemSchema;
     pNextReader->pReadSnap = pReader->pReadSnap;
 
-    code = doOpenReaderImpl(pPrevReader);
+      code = doOpenReaderImpl(pPrevReader);
-    if (code != TSDB_CODE_SUCCESS) {
+      if (code != TSDB_CODE_SUCCESS) {
-      return code;
+        return code;
-    }
+      }
 
-    code = doOpenReaderImpl(pNextReader);
+      code = doOpenReaderImpl(pNextReader);
-    if (code != TSDB_CODE_SUCCESS) {
+      if (code != TSDB_CODE_SUCCESS) {
-      return code;
+        return code;
-    }
+      }
 
-    code = doOpenReaderImpl(pReader);
+      code = doOpenReaderImpl(pReader);
-    if (code != TSDB_CODE_SUCCESS) {
+      if (code != TSDB_CODE_SUCCESS) {
-      return code;
+        return code;
+      }
     }
   }
 
@@ -3513,6 +3514,10 @@ void tsdbReaderClose(STsdbReader* pReader) {
     taosMemoryFree(pLReader);
   }
 
+  if (pReader->innerReader[0] != 0) {
+    tsdbUntakeReadSnap(pReader->innerReader[0]->pTsdb, pReader->innerReader[0]->pReadSnap, pReader->idStr);
+  }
+
   tsdbDebug(
       "%p :io-cost summary: head-file:%" PRIu64 ", head-file time:%.2f ms, SMA:%" PRId64
       " SMA-time:%.2f ms, fileBlocks:%" PRId64

source/dnode/vnode/src/tsdb/tsdbReaderWriter.c

@@ -124,7 +124,7 @@ static int32_t tsdbReadFilePage(STsdbFD *pFD, int64_t pgno) {
   }
 
   // check
-  if (!taosCheckChecksumWhole(pFD->pBuf, pFD->szPage)) {
+  if (pgno > 1 && !taosCheckChecksumWhole(pFD->pBuf, pFD->szPage)) {
     code = TSDB_CODE_FILE_CORRUPTED;
     goto _exit;
   }
@@ -843,7 +843,8 @@ _err:
 // SDataFReader ====================================================
 int32_t tsdbDataFReaderOpen(SDataFReader **ppReader, STsdb *pTsdb, SDFileSet *pSet) {
   int32_t       code = 0;
-  SDataFReader *pReader;
+  int32_t       lino = 0;
+  SDataFReader *pReader = NULL;
   int32_t       szPage = pTsdb->pVnode->config.tsdbPageSize;
   char          fname[TSDB_FILENAME_LEN];
 
@@ -851,7 +852,7 @@ int32_t tsdbDataFReaderOpen(SDataFReader **ppReader, STsdb *pTsdb, SDFileSet *pS
   pReader = (SDataFReader *)taosMemoryCalloc(1, sizeof(*pReader));
   if (pReader == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;
-    goto _err;
+    TSDB_CHECK_CODE(code, lino, _exit);
   }
   pReader->pTsdb = pTsdb;
   pReader->pSet = pSet;
@@ -859,31 +860,40 @@ int32_t tsdbDataFReaderOpen(SDataFReader **ppReader, STsdb *pTsdb, SDFileSet *pS
   // head
   tsdbHeadFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pHeadF, fname);
   code = tsdbOpenFile(fname, szPage, TD_FILE_READ, &pReader->pHeadFD);
-  if (code) goto _err;
+  TSDB_CHECK_CODE(code, lino, _exit);
 
   // data
   tsdbDataFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pDataF, fname);
   code = tsdbOpenFile(fname, szPage, TD_FILE_READ, &pReader->pDataFD);
-  if (code) goto _err;
+  TSDB_CHECK_CODE(code, lino, _exit);
 
   // sma
   tsdbSmaFileName(pTsdb, pSet->diskId, pSet->fid, pSet->pSmaF, fname);
   code = tsdbOpenFile(fname, szPage, TD_FILE_READ, &pReader->pSmaFD);
-  if (code) goto _err;
+  TSDB_CHECK_CODE(code, lino, _exit);
 
   // stt
   for (int32_t iStt = 0; iStt < pSet->nSttF; iStt++) {
     tsdbSttFileName(pTsdb, pSet->diskId, pSet->fid, pSet->aSttF[iStt], fname);
     code = tsdbOpenFile(fname, szPage, TD_FILE_READ, &pReader->aSttFD[iStt]);
-    if (code) goto _err;
+    TSDB_CHECK_CODE(code, lino, _exit);
   }
 
-  *ppReader = pReader;
+_exit:
-  return code;
+  if (code) {
+    *ppReader = NULL;
+    tsdbError("vgId:%d %s failed at line %d since %s", TD_VID(pTsdb->pVnode), __func__, lino, tstrerror(code));
 
-_err:
+    if (pReader) {
-  tsdbError("vgId:%d, tsdb data file reader open failed since %s", TD_VID(pTsdb->pVnode), tstrerror(code));
+      for (int32_t iStt = 0; iStt < pSet->nSttF; iStt++) tsdbCloseFile(&pReader->aSttFD[iStt]);
-  *ppReader = NULL;
+      tsdbCloseFile(&pReader->pSmaFD);
+      tsdbCloseFile(&pReader->pDataFD);
+      tsdbCloseFile(&pReader->pHeadFD);
+      taosMemoryFree(pReader);
+    }
+  } else {
+    *ppReader = pReader;
+  }
   return code;
 }
 
@@ -1588,4 +1598,4 @@ int32_t tsdbReadDelIdx(SDelFReader *pReader, SArray *aDelIdx) {
 _err:
   tsdbError("vgId:%d, read del idx failed since %s", TD_VID(pReader->pTsdb->pVnode), tstrerror(code));
   return code;
-}
+}

source/dnode/vnode/src/tsdb/tsdbSnapshot.c

@@ -445,13 +445,14 @@ _err:
 
 int32_t tsdbSnapReaderOpen(STsdb* pTsdb, int64_t sver, int64_t ever, int8_t type, STsdbSnapReader** ppReader) {
   int32_t          code = 0;
+  int32_t          lino = 0;
   STsdbSnapReader* pReader = NULL;
 
   // alloc
   pReader = (STsdbSnapReader*)taosMemoryCalloc(1, sizeof(*pReader));
   if (pReader == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;
-    goto _err;
+    TSDB_CHECK_CODE(code, lino, _exit);
   }
   pReader->pTsdb = pTsdb;
   pReader->sver = sver;
@@ -461,19 +462,19 @@ int32_t tsdbSnapReaderOpen(STsdb* pTsdb, int64_t sver, int64_t ever, int8_t type
   code = taosThreadRwlockRdlock(&pTsdb->rwLock);
   if (code) {
     code = TAOS_SYSTEM_ERROR(code);
-    goto _err;
+    TSDB_CHECK_CODE(code, lino, _exit);
   }
 
   code = tsdbFSRef(pTsdb, &pReader->fs);
   if (code) {
     taosThreadRwlockUnlock(&pTsdb->rwLock);
-    goto _err;
+    TSDB_CHECK_CODE(code, lino, _exit);
   }
 
   code = taosThreadRwlockUnlock(&pTsdb->rwLock);
   if (code) {
     code = TAOS_SYSTEM_ERROR(code);
-    goto _err;
+    TSDB_CHECK_CODE(code, lino, _exit);
   }
 
   // data
@@ -485,43 +486,52 @@ int32_t tsdbSnapReaderOpen(STsdb* pTsdb, int64_t sver, int64_t ever, int8_t type
       pIter->aBlockIdx = taosArrayInit(0, sizeof(SBlockIdx));
       if (pIter->aBlockIdx == NULL) {
         code = TSDB_CODE_OUT_OF_MEMORY;
-        goto _err;
+        TSDB_CHECK_CODE(code, lino, _exit);
       }
     } else {
       pIter->aSttBlk = taosArrayInit(0, sizeof(SSttBlk));
       if (pIter->aSttBlk == NULL) {
         code = TSDB_CODE_OUT_OF_MEMORY;
-        goto _err;
+        TSDB_CHECK_CODE(code, lino, _exit);
       }
     }
 
     code = tBlockDataCreate(&pIter->bData);
-    if (code) goto _err;
+    TSDB_CHECK_CODE(code, lino, _exit);
   }
 
   code = tBlockDataCreate(&pReader->bData);
-  if (code) goto _err;
+  TSDB_CHECK_CODE(code, lino, _exit);
 
   // del
   pReader->aDelIdx = taosArrayInit(0, sizeof(SDelIdx));
   if (pReader->aDelIdx == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;
-    goto _err;
+    TSDB_CHECK_CODE(code, lino, _exit);
   }
   pReader->aDelData = taosArrayInit(0, sizeof(SDelData));
   if (pReader->aDelData == NULL) {
     code = TSDB_CODE_OUT_OF_MEMORY;
-    goto _err;
+    TSDB_CHECK_CODE(code, lino, _exit);
   }
 
-  tsdbInfo("vgId:%d, vnode snapshot tsdb reader opened for %s", TD_VID(pTsdb->pVnode), pTsdb->path);
+_exit:
-  *ppReader = pReader;
+  if (code) {
-  return code;
+    tsdbError("vgId:%d %s failed at line %d since %s, TSDB path: %s", TD_VID(pTsdb->pVnode), lino, tstrerror(code),
+              pTsdb->path);
+    *ppReader = NULL;
 
-_err:
+    if (pReader) {
-  tsdbError("vgId:%d, vnode snapshot tsdb reader open for %s failed since %s", TD_VID(pTsdb->pVnode), pTsdb->path,
+      taosArrayDestroy(pReader->aDelData);
-            tstrerror(code));
+      taosArrayDestroy(pReader->aDelIdx);
-  *ppReader = NULL;
+      tBlockDataDestroy(&pReader->bData, 1);
+      tsdbFSDestroy(&pReader->fs);
+      taosMemoryFree(pReader);
+    }
+  } else {
+    *ppReader = pReader;
+    tsdbInfo("vgId:%d, vnode snapshot tsdb reader opened for %s", TD_VID(pTsdb->pVnode), pTsdb->path);
+  }
   return code;
 }
 

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

@@ -112,6 +112,8 @@ void vnodeBufPoolReset(SVBufPool *pPool) {
 void *vnodeBufPoolMalloc(SVBufPool *pPool, int size) {
   SVBufPoolNode *pNode;
   void          *p = NULL;
+  ASSERT(pPool != NULL);
+
   taosThreadSpinLock(&pPool->lock);
   if (pPool->node.size >= pPool->ptr - pPool->node.data + size) {
     // allocate from the anchor node

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

@@ -73,7 +73,7 @@ int vnodeBegin(SVnode *pVnode) {
 
 int vnodeShouldCommit(SVnode *pVnode) {
   if (pVnode->inUse) {
-    return pVnode->inUse->size > pVnode->inUse->node.size;
+    return osDataSpaceAvailable() && (pVnode->inUse->size > pVnode->inUse->node.size);
   }
   return false;
 }
@@ -89,6 +89,7 @@ int vnodeSaveInfo(const char *dir, const SVnodeInfo *pInfo) {
   data = NULL;
 
   if (vnodeEncodeInfo(pInfo, &data) < 0) {
+    vError("failed to encode json info.");
     return -1;
   }
 
@@ -101,7 +102,7 @@ int vnodeSaveInfo(const char *dir, const SVnodeInfo *pInfo) {
   }
 
   if (taosWriteFile(pFile, data, strlen(data)) < 0) {
-    vError("failed to write info file:%s data:%s", fname, terrstr());
+    vError("failed to write info file:%s error:%s", fname, terrstr());
     terrno = TAOS_SYSTEM_ERROR(errno);
     goto _err;
   }
@@ -233,15 +234,15 @@ int vnodeCommit(SVnode *pVnode) {
     snprintf(dir, TSDB_FILENAME_LEN, "%s", pVnode->path);
   }
   if (vnodeSaveInfo(dir, &info) < 0) {
-    ASSERT(0);
+    vError("vgId:%d, failed to save vnode info since %s", TD_VID(pVnode), tstrerror(terrno));
     return -1;
   }
   walBeginSnapshot(pVnode->pWal, pVnode->state.applied);
 
   // preCommit
   // smaSyncPreCommit(pVnode->pSma);
-  if (smaAsyncPreCommit(pVnode->pSma) < 0) {
+  if(smaAsyncPreCommit(pVnode->pSma) < 0){
-    ASSERT(0);
+    vError("vgId:%d, failed to async pre-commit sma since %s", TD_VID(pVnode), tstrerror(terrno));
     return -1;
   }
 
@@ -250,44 +251,44 @@ int vnodeCommit(SVnode *pVnode) {
 
   // commit each sub-system
   if (metaCommit(pVnode->pMeta) < 0) {
-    ASSERT(0);
+    vError("vgId:%d, failed to commit meta since %s", TD_VID(pVnode), tstrerror(terrno));
     return -1;
   }
 
   if (VND_IS_RSMA(pVnode)) {
     if (smaAsyncCommit(pVnode->pSma) < 0) {
-      ASSERT(0);
+      vError("vgId:%d, failed to async commit sma since %s", TD_VID(pVnode), tstrerror(terrno));
       return -1;
     }
 
     if (tsdbCommit(VND_RSMA0(pVnode)) < 0) {
-      ASSERT(0);
+      vError("vgId:%d, failed to commit tsdb rsma0 since %s", TD_VID(pVnode), tstrerror(terrno));
       return -1;
     }
     if (tsdbCommit(VND_RSMA1(pVnode)) < 0) {
-      ASSERT(0);
+      vError("vgId:%d, failed to commit tsdb rsma1 since %s", TD_VID(pVnode), tstrerror(terrno));
       return -1;
     }
     if (tsdbCommit(VND_RSMA2(pVnode)) < 0) {
-      ASSERT(0);
+      vError("vgId:%d, failed to commit tsdb rsma2 since %s", TD_VID(pVnode), tstrerror(terrno));
       return -1;
     }
   } else {
     if (tsdbCommit(pVnode->pTsdb) < 0) {
-      ASSERT(0);
+      vError("vgId:%d, failed to commit tsdb since %s", TD_VID(pVnode), tstrerror(terrno));
       return -1;
     }
   }
 
   if (tqCommit(pVnode->pTq) < 0) {
-    ASSERT(0);
+    vError("vgId:%d, failed to commit tq since %s", TD_VID(pVnode), tstrerror(terrno));
     return -1;
   }
   // walCommit (TODO)
 
   // commit info
   if (vnodeCommitInfo(dir, &info) < 0) {
-    ASSERT(0);
+    vError("vgId:%d, failed to commit vnode info since %s", TD_VID(pVnode), tstrerror(terrno));
     return -1;
   }
 
@@ -296,7 +297,7 @@ int vnodeCommit(SVnode *pVnode) {
   // postCommit
   // smaSyncPostCommit(pVnode->pSma);
   if (smaAsyncPostCommit(pVnode->pSma) < 0) {
-    ASSERT(0);
+    vError("vgId:%d, failed to async post-commit sma since %s", TD_VID(pVnode), tstrerror(terrno));
     return -1;
   }
 

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

@@ -140,7 +140,7 @@ SVnode *vnodeOpen(const char *path, STfs *pTfs, SMsgCb msgCb) {
 
   pVnode->pWal = walOpen(tdir, &(pVnode->config.walCfg));
   if (pVnode->pWal == NULL) {
-    vError("vgId:%d, failed to open vnode wal since %s", TD_VID(pVnode), tstrerror(terrno));
+    vError("vgId:%d, failed to open vnode wal since %s. wal:%s", TD_VID(pVnode), tstrerror(terrno), tdir);
     goto _err;
   }
 

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

@@ -43,14 +43,23 @@ int32_t vnodePreProcessWriteMsg(SVnode *pVnode, SRpcMsg *pMsg) {
       tDecoderInit(&dc, (uint8_t *)pMsg->pCont + sizeof(SMsgHead), pMsg->contLen - sizeof(SMsgHead));
       tStartDecode(&dc);
 
-      tDecodeI32v(&dc, &nReqs);
+      if (tDecodeI32v(&dc, &nReqs) < 0) {
+        code = TSDB_CODE_INVALID_MSG;
+        goto _err;
+      }
       for (int32_t iReq = 0; iReq < nReqs; iReq++) {
         tb_uid_t uid = tGenIdPI64();
         char    *name = NULL;
         tStartDecode(&dc);
 
-        tDecodeI32v(&dc, NULL);
+        if (tDecodeI32v(&dc, NULL) < 0) {
-        tDecodeCStr(&dc, &name);
+          code = TSDB_CODE_INVALID_MSG;
+          return code;
+        }
+        if (tDecodeCStr(&dc, &name) < 0) {
+          code = TSDB_CODE_INVALID_MSG;
+          return code;
+        }
         *(int64_t *)(dc.data + dc.pos) = uid;
         *(int64_t *)(dc.data + dc.pos + 8) = ctime;
 
@@ -81,10 +90,19 @@ int32_t vnodePreProcessWriteMsg(SVnode *pVnode, SRpcMsg *pMsg) {
           char *name = NULL;
 
           tDecoderInit(&dc, pBlock->data, msgIter.schemaLen);
-          tStartDecode(&dc);
+          if (tStartDecode(&dc) < 0) {
+            code = TSDB_CODE_INVALID_MSG;
+            return code;
+          }
 
-          tDecodeI32v(&dc, NULL);
+          if (tDecodeI32v(&dc, NULL) < 0) {
-          tDecodeCStr(&dc, &name);
+            code = TSDB_CODE_INVALID_MSG;
+            return code;
+          }
+          if (tDecodeCStr(&dc, &name) < 0) {
+            code = TSDB_CODE_INVALID_MSG;
+            return code;
+          }
 
           uid = metaGetTableEntryUidByName(pVnode->pMeta, name);
           if (uid == 0) {
@@ -148,6 +166,12 @@ int32_t vnodeProcessWriteMsg(SVnode *pVnode, SRpcMsg *pMsg, int64_t version, SRp
   int32_t len;
   int32_t ret;
 
+  if (!pVnode->inUse) {
+     terrno = TSDB_CODE_VND_NOT_SYNCED;
+     vError("vgId:%d, not ready to write since %s", TD_VID(pVnode), terrstr());
+     return -1;
+  }
+
   vDebug("vgId:%d, start to process write request %s, index:%" PRId64, TD_VID(pVnode), TMSG_INFO(pMsg->msgType),
          version);
 
@@ -268,10 +292,16 @@ int32_t vnodeProcessWriteMsg(SVnode *pVnode, SRpcMsg *pMsg, int64_t version, SRp
   _do_commit:
     vInfo("vgId:%d, commit at version %" PRId64, TD_VID(pVnode), version);
     // commit current change
-    vnodeCommit(pVnode);
+    if (vnodeCommit(pVnode) < 0) {
+        vError("vgId:%d, failed to commit vnode since %s.", TD_VID(pVnode), tstrerror(terrno));
+        goto _err;
+    }
 
     // start a new one
-    vnodeBegin(pVnode);
+    if (vnodeBegin(pVnode) < 0) {
+        vError("vgId:%d, failed to begin vnode since %s.", TD_VID(pVnode), tstrerror(terrno));
+        goto _err;
+    }
   }
 
   return 0;

source/libs/executor/inc/executorimpl.h

@@ -212,6 +212,7 @@ typedef struct SExprSupp {
   int32_t         numOfExprs;  // the number of scalar expression in group operator
   SqlFunctionCtx* pCtx;
   int32_t*        rowEntryInfoOffset;  // offset value for each row result cell info
+  SFilterInfo*    pFilterInfo;
 } SExprSupp;
 
 typedef struct SOperatorInfo {
@@ -926,7 +927,7 @@ int32_t getTableScanInfo(SOperatorInfo* pOperator, int32_t* order, int32_t* scan
 int32_t getBufferPgSize(int32_t rowSize, uint32_t* defaultPgsz, uint32_t* defaultBufsz);
 
 void    doSetOperatorCompleted(SOperatorInfo* pOperator);
-void    doFilter(const SNode* pFilterNode, SSDataBlock* pBlock, const SArray* pColMatchInfo);
+void    doFilter(const SNode* pFilterNode, SSDataBlock* pBlock, const SArray* pColMatchInfo, SFilterInfo* pFilterInfo);
 int32_t addTagPseudoColumnData(SReadHandle* pHandle, SExprInfo* pPseudoExpr, int32_t numOfPseudoExpr,
                                SSDataBlock* pBlock, const char* idStr);
 

source/libs/executor/src/dataInserter.c

@@ -81,7 +81,7 @@ int32_t inserterCallback(void* param, SDataBuf* pMsg, int32_t code) {
     }
 
     pInserter->submitRes.affectedRows += pInserter->submitRes.pRsp->affectedRows;
-    qDebug("submit rsp received, affectedRows:%d, total:%d", pInserter->submitRes.pRsp->affectedRows,
+    qDebug("submit rsp received, affectedRows:%d, total:%"PRId64, pInserter->submitRes.pRsp->affectedRows,
            pInserter->submitRes.affectedRows);
 
     tFreeSSubmitRsp(pInserter->submitRes.pRsp);

source/libs/executor/src/executil.c

@@ -1201,14 +1201,21 @@ void createExprFromOneNode(SExprInfo* pExp, SNode* pNode, int16_t slotId) {
     pExp->base.resSchema =
         createResSchema(pType->type, pType->bytes, slotId, pType->scale, pType->precision, pFuncNode->node.aliasName);
 
-    pExp->pExpr->_function.functionId = pFuncNode->funcId;
+    tExprNode* pExprNode = pExp->pExpr;
-    pExp->pExpr->_function.pFunctNode = pFuncNode;
+
+    pExprNode->_function.functionId = pFuncNode->funcId;
+    pExprNode->_function.pFunctNode = pFuncNode;
+
+    tstrncpy(pExprNode->_function.functionName, pFuncNode->functionName, tListLen(pExprNode->_function.functionName));
 
-    strncpy(pExp->pExpr->_function.functionName, pFuncNode->functionName,
-            tListLen(pExp->pExpr->_function.functionName));
 #if 1
     // todo refactor: add the parameter for tbname function
-    if (!pFuncNode->pParameterList && (strcmp(pExp->pExpr->_function.functionName, "tbname") == 0)) {
+    const char* name = "tbname";
+    int32_t len = strlen(name);
+
+    if (!pFuncNode->pParameterList && (memcmp(pExprNode->_function.functionName, name, len) == 0) &&
+        pExprNode->_function.functionName[len] == 0) {
+
       pFuncNode->pParameterList = nodesMakeList();
       ASSERT(LIST_LENGTH(pFuncNode->pParameterList) == 0);
       SValueNode* res = (SValueNode*)nodesMakeNode(QUERY_NODE_VALUE);

source/libs/executor/src/executor.c

@@ -856,15 +856,15 @@ int32_t qStreamPrepareScan(qTaskInfo_t tinfo, STqOffsetVal* pOffset, int8_t subT
     tDeleteSSchemaWrapper(pTaskInfo->streamInfo.schema);
     pTaskInfo->streamInfo.schema = mtInfo.schema;
 
-    qDebug("tmqsnap qStreamPrepareScan snapshot data uid %ld ts %ld", mtInfo.uid, pOffset->ts);
+    qDebug("tmqsnap qStreamPrepareScan snapshot data uid %ld ts %"PRId64, mtInfo.uid, pOffset->ts);
   } else if (pOffset->type == TMQ_OFFSET__SNAPSHOT_META) {
     SStreamRawScanInfo* pInfo = pOperator->info;
     SSnapContext*       sContext = pInfo->sContext;
     if (setForSnapShot(sContext, pOffset->uid) != 0) {
-      qError("setForSnapShot error. uid:%" PRIi64 " ,version:%" PRIi64, pOffset->uid);
+      qError("setForSnapShot error. uid:%" PRIu64 " ,version:%" PRId64, pOffset->uid, pOffset->version);
       return -1;
     }
-    qDebug("tmqsnap qStreamPrepareScan snapshot meta uid %ld ts %ld", pOffset->uid);
+    qDebug("tmqsnap qStreamPrepareScan snapshot meta uid %ld ts %"PRId64, pOffset->uid, pOffset->ts);
   } else if (pOffset->type == TMQ_OFFSET__LOG) {
     SStreamRawScanInfo* pInfo = pOperator->info;
     tsdbReaderClose(pInfo->dataReader);

source/libs/executor/src/executorimpl.c

@@ -1113,15 +1113,24 @@ void setResultRowInitCtx(SResultRow* pResult, SqlFunctionCtx* pCtx, int32_t numO
 static void extractQualifiedTupleByFilterResult(SSDataBlock* pBlock, const SColumnInfoData* p, bool keep,
                                                 int32_t status);
 
-void doFilter(const SNode* pFilterNode, SSDataBlock* pBlock, const SArray* pColMatchInfo) {
+void doFilter(const SNode* pFilterNode, SSDataBlock* pBlock, const SArray* pColMatchInfo, SFilterInfo* pFilterInfo) {
   if (pFilterNode == NULL || pBlock->info.rows == 0) {
     return;
   }
 
-  SFilterInfo* filter = NULL;
+  SFilterInfo* filter = pFilterInfo;
+  int64_t st = taosGetTimestampUs();
+
+//  pError("start filter");
 
   // todo move to the initialization function
-  int32_t            code = filterInitFromNode((SNode*)pFilterNode, &filter, 0);
+  int32_t code = 0;
+  bool needFree = false;
+  if (filter == NULL) {
+    needFree = true;
+    code = filterInitFromNode((SNode*)pFilterNode, &filter, 0);
+  }
+
   SFilterColumnParam param1 = {.numOfCols = taosArrayGetSize(pBlock->pDataBlock), .pDataBlock = pBlock->pDataBlock};
   code = filterSetDataFromSlotId(filter, &param1);
 
@@ -1130,7 +1139,10 @@ void doFilter(const SNode* pFilterNode, SSDataBlock* pBlock, const SArray* pColM
 
   // todo the keep seems never to be True??
   bool keep = filterExecute(filter, pBlock, &p, NULL, param1.numOfCols, &status);
-  filterFreeInfo(filter);
+
+  if (needFree) {
+    filterFreeInfo(filter);
+  }
 
   extractQualifiedTupleByFilterResult(pBlock, p, keep, status);
 
@@ -1845,6 +1857,7 @@ static int32_t doSendFetchDataRequest(SExchangeInfo* pExchangeInfo, SExecTaskInf
     SResFetchReq* pMsg = taosMemoryCalloc(1, sizeof(SResFetchReq));
     if (NULL == pMsg) {
       pTaskInfo->code = TSDB_CODE_QRY_OUT_OF_MEMORY;
+      taosMemoryFree(pWrapper);
       return pTaskInfo->code;
     }
 
@@ -2479,7 +2492,7 @@ static SSDataBlock* getAggregateResult(SOperatorInfo* pOperator) {
   blockDataEnsureCapacity(pInfo->pRes, pOperator->resultInfo.capacity);
   while (1) {
     doBuildResultDatablock(pOperator, pInfo, &pAggInfo->groupResInfo, pAggInfo->aggSup.pResultBuf);
-    doFilter(pAggInfo->pCondition, pInfo->pRes, NULL);
+    doFilter(pAggInfo->pCondition, pInfo->pRes, NULL, NULL);
 
     if (!hasRemainResults(&pAggInfo->groupResInfo)) {
       doSetOperatorCompleted(pOperator);
@@ -2873,7 +2886,7 @@ static SSDataBlock* doFill(SOperatorInfo* pOperator) {
       break;
     }
 
-    doFilter(pInfo->pCondition, fillResult, pInfo->pColMatchColInfo);
+    doFilter(pInfo->pCondition, fillResult, pInfo->pColMatchColInfo, NULL);
     if (fillResult->info.rows > 0) {
       break;
     }
@@ -3049,6 +3062,12 @@ void cleanupExprSupp(SExprSupp* pSupp) {
     destroyExprInfo(pSupp->pExprInfo, pSupp->numOfExprs);
     taosMemoryFreeClear(pSupp->pExprInfo);
   }
+
+  if (pSupp->pFilterInfo != NULL) {
+    filterFreeInfo(pSupp->pFilterInfo);
+    pSupp->pFilterInfo = NULL;
+  }
+
   taosMemoryFree(pSupp->rowEntryInfoOffset);
 }
 

source/libs/executor/src/groupoperator.c

@@ -311,7 +311,7 @@ static SSDataBlock* buildGroupResultDataBlock(SOperatorInfo* pOperator) {
   SSDataBlock* pRes = pInfo->binfo.pRes;
   while (1) {
     doBuildResultDatablock(pOperator, &pInfo->binfo, &pInfo->groupResInfo, pInfo->aggSup.pResultBuf);
-    doFilter(pInfo->pCondition, pRes, NULL);
+    doFilter(pInfo->pCondition, pRes, NULL, NULL);
 
     if (!hasRemainResults(&pInfo->groupResInfo)) {
       doSetOperatorCompleted(pOperator);

source/libs/executor/src/joinoperator.c

@@ -387,7 +387,7 @@ SSDataBlock* doMergeJoin(struct SOperatorInfo* pOperator) {
       break;
     }
     if (pJoinInfo->pCondAfterMerge != NULL) {
-      doFilter(pJoinInfo->pCondAfterMerge, pRes, NULL);
+      doFilter(pJoinInfo->pCondAfterMerge, pRes, NULL, NULL);
     }
     if (pRes->info.rows >= pOperator->resultInfo.threshold) {
       break;

source/libs/executor/src/projectoperator.c

@@ -315,7 +315,7 @@ SSDataBlock* doProjectOperation(SOperatorInfo* pOperator) {
       }
 
       // do apply filter
-      doFilter(pProjectInfo->pFilterNode, pFinalRes, NULL);
+      doFilter(pProjectInfo->pFilterNode, pFinalRes, NULL, NULL);
 
       // when apply the limit/offset for each group, pRes->info.rows may be 0, due to limit constraint.
       if (pFinalRes->info.rows > 0 || (pOperator->status == OP_EXEC_DONE)) {
@@ -325,7 +325,7 @@ SSDataBlock* doProjectOperation(SOperatorInfo* pOperator) {
     } else {
       // do apply filter
       if (pRes->info.rows > 0) {
-        doFilter(pProjectInfo->pFilterNode, pRes, NULL);
+        doFilter(pProjectInfo->pFilterNode, pRes, NULL, NULL);
         if (pRes->info.rows == 0) {
           continue;
         }
@@ -518,7 +518,7 @@ SSDataBlock* doApplyIndefinitFunction(SOperatorInfo* pOperator) {
       }
     }
 
-    doFilter(pIndefInfo->pCondition, pInfo->pRes, NULL);
+    doFilter(pIndefInfo->pCondition, pInfo->pRes, NULL, NULL);
     size_t rows = pInfo->pRes->info.rows;
     if (rows > 0 || pOperator->status == OP_EXEC_DONE) {
       break;
@@ -620,7 +620,7 @@ SSDataBlock* doGenerateSourceData(SOperatorInfo* pOperator) {
   }
 
   pRes->info.rows = 1;
-  doFilter(pProjectInfo->pFilterNode, pRes, NULL);
+  doFilter(pProjectInfo->pFilterNode, pRes, NULL, NULL);
 
   /*int32_t status = */ doIngroupLimitOffset(&pProjectInfo->limitInfo, 0, pRes, pOperator);
 

source/libs/executor/src/scanoperator.c

@@ -385,7 +385,7 @@ static int32_t loadDataBlock(SOperatorInfo* pOperator, STableScanInfo* pTableSca
 
   if (pTableScanInfo->pFilterNode != NULL) {
     int64_t st = taosGetTimestampUs();
-    doFilter(pTableScanInfo->pFilterNode, pBlock, pTableScanInfo->pColMatchInfo);
+    doFilter(pTableScanInfo->pFilterNode, pBlock, pTableScanInfo->pColMatchInfo, pOperator->exprSupp.pFilterInfo);
 
     double el = (taosGetTimestampUs() - st) / 1000.0;
     pTableScanInfo->readRecorder.filterTime += el;
@@ -754,6 +754,11 @@ SOperatorInfo* createTableScanOperatorInfo(STableScanPhysiNode* pTableScanNode,
   pInfo->dataBlockLoadFlag = pTableScanNode->dataRequired;
   pInfo->pResBlock = createResDataBlock(pDescNode);
   pInfo->pFilterNode = pTableScanNode->scan.node.pConditions;
+
+  if (pInfo->pFilterNode != NULL) {
+    code = filterInitFromNode((SNode*)pInfo->pFilterNode, &pOperator->exprSupp.pFilterInfo, 0);
+  }
+
   pInfo->scanFlag = MAIN_SCAN;
   pInfo->pColMatchInfo = pColList;
   pInfo->currentGroupId = -1;
@@ -1123,7 +1128,7 @@ static SSDataBlock* doRangeScan(SStreamScanInfo* pInfo, SSDataBlock* pSDB, int32
       return NULL;
     }
 
-    doFilter(pInfo->pCondition, pResult, NULL);
+    doFilter(pInfo->pCondition, pResult, NULL, NULL);
     if (pResult->info.rows == 0) {
       continue;
     }
@@ -1474,7 +1479,7 @@ static int32_t setBlockIntoRes(SStreamScanInfo* pInfo, const SSDataBlock* pBlock
   }
 
   if (filter) {
-    doFilter(pInfo->pCondition, pInfo->pRes, NULL);
+    doFilter(pInfo->pCondition, pInfo->pRes, NULL, NULL);
   }
   blockDataUpdateTsWindow(pInfo->pRes, pInfo->primaryTsIndex);
   blockDataFreeRes((SSDataBlock*)pBlock);
@@ -1896,7 +1901,7 @@ FETCH_NEXT_BLOCK:
           }
         }
 
-        doFilter(pInfo->pCondition, pInfo->pRes, NULL);
+        doFilter(pInfo->pCondition, pInfo->pRes, NULL, NULL);
         blockDataUpdateTsWindow(pInfo->pRes, pInfo->primaryTsIndex);
 
         if (pBlockInfo->rows > 0 || pInfo->pUpdateDataRes->info.rows > 0) {
@@ -2382,7 +2387,7 @@ static SSDataBlock* doFilterResult(SSysTableScanInfo* pInfo) {
     return pInfo->pRes->info.rows == 0 ? NULL : pInfo->pRes;
   }
 
-  doFilter(pInfo->pCondition, pInfo->pRes, NULL);
+  doFilter(pInfo->pCondition, pInfo->pRes, NULL, NULL);
   return pInfo->pRes->info.rows == 0 ? NULL : pInfo->pRes;
 }
 
@@ -2563,16 +2568,27 @@ static SSDataBlock* sysTableScanUserTags(SOperatorInfo* pOperator) {
 
     SMetaReader smrChildTable = {0};
     metaReaderInit(&smrChildTable, pInfo->readHandle.meta, 0);
-    metaGetTableEntryByName(&smrChildTable, condTableName);
+    int32_t code = metaGetTableEntryByName(&smrChildTable, condTableName);
+    if (code != TSDB_CODE_SUCCESS) {
+      // terrno has been set by metaGetTableEntryByName, therefore, return directly
+      return NULL;
+    }
+
     if (smrChildTable.me.type != TSDB_CHILD_TABLE) {
       metaReaderClear(&smrChildTable);
       blockDataDestroy(dataBlock);
       pInfo->loadInfo.totalRows = 0;
       return NULL;
     }
+
     SMetaReader smrSuperTable = {0};
     metaReaderInit(&smrSuperTable, pInfo->readHandle.meta, META_READER_NOLOCK);
-    metaGetTableEntryByUid(&smrSuperTable, smrChildTable.me.ctbEntry.suid);
+    code = metaGetTableEntryByUid(&smrSuperTable, smrChildTable.me.ctbEntry.suid);
+    if (code != TSDB_CODE_SUCCESS) {
+      // terrno has been set by metaGetTableEntryByUid
+      return NULL;
+    }
+
     sysTableUserTagsFillOneTableTags(pInfo, &smrSuperTable, &smrChildTable, dbname, tableName, &numOfRows, dataBlock);
     metaReaderClear(&smrSuperTable);
     metaReaderClear(&smrChildTable);
@@ -3455,7 +3471,7 @@ static int32_t loadDataBlockFromOneTable(SOperatorInfo* pOperator, STableMergeSc
 
   if (pTableScanInfo->pFilterNode != NULL) {
     int64_t st = taosGetTimestampMs();
-    doFilter(pTableScanInfo->pFilterNode, pBlock, pTableScanInfo->pColMatchInfo);
+    doFilter(pTableScanInfo->pFilterNode, pBlock, pTableScanInfo->pColMatchInfo, NULL);
 
     double el = (taosGetTimestampUs() - st) / 1000.0;
     pTableScanInfo->readRecorder.filterTime += el;
@@ -3790,6 +3806,7 @@ SOperatorInfo* createTableMergeScanOperatorInfo(STableScanPhysiNode* pTableScanN
 
   int32_t code = initQueryTableDataCond(&pInfo->cond, pTableScanNode);
   if (code != TSDB_CODE_SUCCESS) {
+    taosArrayDestroy(pColList);
     goto _error;
   }
 

source/libs/executor/src/sortoperator.c

@@ -216,7 +216,7 @@ SSDataBlock* doSort(SOperatorInfo* pOperator) {
       return NULL;
     }
 
-    doFilter(pInfo->pCondition, pBlock, pInfo->pColMatchInfo);
+    doFilter(pInfo->pCondition, pBlock, pInfo->pColMatchInfo, NULL);
     if (blockDataGetNumOfRows(pBlock) == 0) {
       continue;
     }
@@ -658,7 +658,7 @@ SSDataBlock* getMultiwaySortedBlockData(SSortHandle* pHandle, SSDataBlock* pData
 
   blockDataDestroy(p);
 
-  qDebug("%s get sorted block, groupId:%0x" PRIx64 " rows:%d", GET_TASKID(pTaskInfo), pDataBlock->info.groupId,
+  qDebug("%s get sorted block, groupId:0x%" PRIx64 " rows:%d", GET_TASKID(pTaskInfo), pDataBlock->info.groupId,
          pDataBlock->info.rows);
   return (pDataBlock->info.rows > 0) ? pDataBlock : NULL;
 }

source/libs/executor/src/tfill.c

@@ -1501,7 +1501,7 @@ static SSDataBlock* doStreamFill(SOperatorInfo* pOperator) {
     }
 
     doStreamFillImpl(pOperator);
-    doFilter(pInfo->pCondition, pInfo->pRes, pInfo->pColMatchColInfo);
+    doFilter(pInfo->pCondition, pInfo->pRes, pInfo->pColMatchColInfo, NULL);
     pOperator->resultInfo.totalRows += pInfo->pRes->info.rows;
     if (pInfo->pRes->info.rows > 0) {
       break;
@@ -1681,7 +1681,11 @@ SOperatorInfo* createStreamFillOperatorInfo(SOperatorInfo* downstream, SStreamFi
                                                  &numOfOutputCols, COL_MATCH_FROM_SLOT_ID);
   pInfo->pCondition = pPhyFillNode->node.pConditions;
   pInfo->pColMatchColInfo = pColMatchColInfo;
-  initExprSupp(&pOperator->exprSupp, pFillExprInfo, numOfFillCols);
+  int32_t code = initExprSupp(&pOperator->exprSupp, pFillExprInfo, numOfFillCols);
+  if (code != TSDB_CODE_SUCCESS) {
+    goto _error;
+  }
+
   pInfo->srcRowIndex = 0;
 
   pOperator->name = "FillOperator";
@@ -1693,7 +1697,7 @@ SOperatorInfo* createStreamFillOperatorInfo(SOperatorInfo* downstream, SStreamFi
   pOperator->fpSet = createOperatorFpSet(operatorDummyOpenFn, doStreamFill, NULL, NULL, destroyStreamFillOperatorInfo,
                                          NULL, NULL, NULL);
 
-  int32_t code = appendDownstream(pOperator, &downstream, 1);
+  code = appendDownstream(pOperator, &downstream, 1);
   if (code != TSDB_CODE_SUCCESS) {
     goto _error;
   }
@@ -1702,5 +1706,6 @@ SOperatorInfo* createStreamFillOperatorInfo(SOperatorInfo* downstream, SStreamFi
 _error:
   destroyStreamFillOperatorInfo(pInfo);
   taosMemoryFreeClear(pOperator);
+  pTaskInfo->code = code;
   return NULL;
 }

source/libs/executor/src/timewindowoperator.c

@@ -1277,7 +1277,7 @@ static SSDataBlock* doStateWindowAgg(SOperatorInfo* pOperator) {
   blockDataEnsureCapacity(pBInfo->pRes, pOperator->resultInfo.capacity);
   while (1) {
     doBuildResultDatablock(pOperator, &pInfo->binfo, &pInfo->groupResInfo, pInfo->aggSup.pResultBuf);
-    doFilter(pInfo->pCondition, pBInfo->pRes, NULL);
+    doFilter(pInfo->pCondition, pBInfo->pRes, NULL, NULL);
 
     bool hasRemain = hasRemainResults(&pInfo->groupResInfo);
     if (!hasRemain) {
@@ -1315,7 +1315,7 @@ static SSDataBlock* doBuildIntervalResult(SOperatorInfo* pOperator) {
     blockDataEnsureCapacity(pBlock, pOperator->resultInfo.capacity);
     while (1) {
       doBuildResultDatablock(pOperator, &pInfo->binfo, &pInfo->groupResInfo, pInfo->aggSup.pResultBuf);
-      doFilter(pInfo->pCondition, pBlock, NULL);
+      doFilter(pInfo->pCondition, pBlock, NULL, NULL);
 
       bool hasRemain = hasRemainResults(&pInfo->groupResInfo);
       if (!hasRemain) {
@@ -1420,8 +1420,9 @@ static void doDeleteWindows(SOperatorInfo* pOperator, SInterval* pInterval, SSDa
   SColumnInfoData*             pGpCol = taosArrayGet(pBlock->pDataBlock, GROUPID_COLUMN_INDEX);
   uint64_t*                    pGpDatas = (uint64_t*)pGpCol->pData;
   for (int32_t i = 0; i < pBlock->info.rows; i++) {
-    SResultRowInfo dumyInfo;
+    SResultRowInfo dumyInfo = {0};
     dumyInfo.cur.pageId = -1;
+
     STimeWindow win = {0};
     if (IS_FINAL_OP(pInfo)) {
       win.skey = startTsCols[i];
@@ -2019,7 +2020,7 @@ static SSDataBlock* doSessionWindowAgg(SOperatorInfo* pOperator) {
   if (pOperator->status == OP_RES_TO_RETURN) {
     while (1) {
       doBuildResultDatablock(pOperator, &pInfo->binfo, &pInfo->groupResInfo, pInfo->aggSup.pResultBuf);
-      doFilter(pInfo->pCondition, pBInfo->pRes, NULL);
+      doFilter(pInfo->pCondition, pBInfo->pRes, NULL, NULL);
 
       bool hasRemain = hasRemainResults(&pInfo->groupResInfo);
       if (!hasRemain) {
@@ -2062,7 +2063,7 @@ static SSDataBlock* doSessionWindowAgg(SOperatorInfo* pOperator) {
   blockDataEnsureCapacity(pBInfo->pRes, pOperator->resultInfo.capacity);
   while (1) {
     doBuildResultDatablock(pOperator, &pInfo->binfo, &pInfo->groupResInfo, pInfo->aggSup.pResultBuf);
-    doFilter(pInfo->pCondition, pBInfo->pRes, NULL);
+    doFilter(pInfo->pCondition, pBInfo->pRes, NULL, NULL);
 
     bool hasRemain = hasRemainResults(&pInfo->groupResInfo);
     if (!hasRemain) {
@@ -5251,7 +5252,7 @@ static void doMergeAlignedIntervalAgg(SOperatorInfo* pOperator) {
     setInputDataBlock(pOperator, pSup->pCtx, pBlock, pIaInfo->inputOrder, scanFlag, true);
     doMergeAlignedIntervalAggImpl(pOperator, &pIaInfo->binfo.resultRowInfo, pBlock, pRes);
 
-    doFilter(pMiaInfo->pCondition, pRes, NULL);
+    doFilter(pMiaInfo->pCondition, pRes, NULL, NULL);
     if (pRes->info.rows >= pOperator->resultInfo.capacity) {
       break;
     }
@@ -5828,27 +5829,30 @@ SOperatorInfo* createStreamIntervalOperatorInfo(SOperatorInfo* downstream, SPhys
   pInfo->ignoreExpiredData = pIntervalPhyNode->window.igExpired;
   pInfo->isFinal = false;
 
-  if (pIntervalPhyNode->window.pExprs != NULL) {
+
-    int32_t    numOfScalar = 0;
-    SExprInfo* pScalarExprInfo = createExprInfo(pIntervalPhyNode->window.pExprs, NULL, &numOfScalar);
-    int32_t    code = initExprSupp(&pInfo->scalarSupp, pScalarExprInfo, numOfScalar);
-    if (code != TSDB_CODE_SUCCESS) {
-      goto _error;
-    }
-  }
 
   pInfo->primaryTsIndex = ((SColumnNode*)pIntervalPhyNode->window.pTspk)->slotId;
   initResultSizeInfo(&pOperator->resultInfo, 4096);
   SExprSupp* pSup = &pOperator->exprSupp;
+
+  initBasicInfo(&pInfo->binfo, pResBlock);
+  initStreamFunciton(pSup->pCtx, pSup->numOfExprs);
+  initExecTimeWindowInfo(&pInfo->twAggSup.timeWindowData, &pTaskInfo->window);
+
   size_t     keyBufSize = sizeof(int64_t) + sizeof(int64_t) + POINTER_BYTES;
   int32_t    code = initAggInfo(pSup, &pInfo->aggSup, pExprInfo, numOfCols, keyBufSize, pTaskInfo->id.str);
   if (code != TSDB_CODE_SUCCESS) {
     goto _error;
   }
 
-  initBasicInfo(&pInfo->binfo, pResBlock);
+  if (pIntervalPhyNode->window.pExprs != NULL) {
-  initStreamFunciton(pSup->pCtx, pSup->numOfExprs);
+    int32_t    numOfScalar = 0;
-  initExecTimeWindowInfo(&pInfo->twAggSup.timeWindowData, &pTaskInfo->window);
+    SExprInfo* pScalarExprInfo = createExprInfo(pIntervalPhyNode->window.pExprs, NULL, &numOfScalar);
+    code = initExprSupp(&pInfo->scalarSupp, pScalarExprInfo, numOfScalar);
+    if (code != TSDB_CODE_SUCCESS) {
+      goto _error;
+    }
+  }
 
   pInfo->invertible = allInvertible(pSup->pCtx, numOfCols);
   pInfo->invertible = false;

source/libs/nodes/src/nodesMsgFuncs.c

@@ -478,7 +478,10 @@ static int32_t tlvDecodeValueEnum(STlvDecoder* pDecoder, void* pValue, int16_t l
   return code;
 }
 
-static int32_t tlvDecodeCStr(STlv* pTlv, char* pValue) {
+static int32_t tlvDecodeCStr(STlv* pTlv, char* pValue, int32_t size) {
+  if (pTlv->len > size - 1) {
+    return TSDB_CODE_FAILED;
+  }
   memcpy(pValue, pTlv->value, pTlv->len);
   return TSDB_CODE_SUCCESS;
 }
@@ -919,9 +922,14 @@ static int32_t msgToDatum(STlv* pTlv, void* pObj) {
       }
       break;
     }
-    case TSDB_DATA_TYPE_JSON:
+    case TSDB_DATA_TYPE_JSON: {
+      if (pTlv->len <= 0 || pTlv->len > TSDB_MAX_JSON_TAG_LEN) {
+        code = TSDB_CODE_FAILED;
+        break;
+      }
       code = tlvDecodeDynBinary(pTlv, (void**)&pNode->datum.p);
       break;
+    }
     case TSDB_DATA_TYPE_DECIMAL:
     case TSDB_DATA_TYPE_BLOB:
       // todo
@@ -1097,7 +1105,7 @@ static int32_t msgToFunctionNode(STlvDecoder* pDecoder, void* pObj) {
         code = tlvDecodeObjFromTlv(pTlv, msgToExprNode, &pNode->node);
         break;
       case FUNCTION_CODE_FUNCTION_NAME:
-        code = tlvDecodeCStr(pTlv, pNode->functionName);
+        code = tlvDecodeCStr(pTlv, pNode->functionName, sizeof(pNode->functionName));
         break;
       case FUNCTION_CODE_FUNCTION_ID:
         code = tlvDecodeI32(pTlv, &pNode->funcId);
@@ -1226,10 +1234,10 @@ static int32_t msgToName(STlvDecoder* pDecoder, void* pObj) {
         code = tlvDecodeI32(pTlv, &pNode->acctId);
         break;
       case NAME_CODE_DB_NAME:
-        code = tlvDecodeCStr(pTlv, pNode->dbname);
+        code = tlvDecodeCStr(pTlv, pNode->dbname, sizeof(pNode->dbname));
         break;
       case NAME_CODE_TABLE_NAME:
-        code = tlvDecodeCStr(pTlv, pNode->tname);
+        code = tlvDecodeCStr(pTlv, pNode->tname, sizeof(pNode->tname));
         break;
       default:
         break;
@@ -1538,7 +1546,7 @@ static int32_t msgToEp(STlvDecoder* pDecoder, void* pObj) {
   tlvForEach(pDecoder, pTlv, code) {
     switch (pTlv->type) {
       case EP_CODE_FQDN:
-        code = tlvDecodeCStr(pTlv, pNode->fqdn);
+        code = tlvDecodeCStr(pTlv, pNode->fqdn, sizeof(pNode->fqdn));
         break;
       case EP_CODE_port:
         code = tlvDecodeU16(pTlv, &pNode->port);
@@ -3207,7 +3215,7 @@ static int32_t msgToPhysiQueryInsertNode(STlvDecoder* pDecoder, void* pObj) {
         code = tlvDecodeI8(pTlv, &pNode->tableType);
         break;
       case PHY_QUERY_INSERT_CODE_TABLE_NAME:
-        code = tlvDecodeCStr(pTlv, pNode->tableName);
+        code = tlvDecodeCStr(pTlv, pNode->tableName, sizeof(pNode->tableName));
         break;
       case PHY_QUERY_INSERT_CODE_VG_ID:
         code = tlvDecodeI32(pTlv, &pNode->vgId);
@@ -3284,10 +3292,10 @@ static int32_t msgToPhysiDeleteNode(STlvDecoder* pDecoder, void* pObj) {
         code = tlvDecodeI8(pTlv, &pNode->tableType);
         break;
       case PHY_DELETER_CODE_TABLE_FNAME:
-        code = tlvDecodeCStr(pTlv, pNode->tableFName);
+        code = tlvDecodeCStr(pTlv, pNode->tableFName, sizeof(pNode->tableFName));
         break;
       case PHY_DELETER_CODE_TS_COL_NAME:
-        code = tlvDecodeCStr(pTlv, pNode->tsColName);
+        code = tlvDecodeCStr(pTlv, pNode->tsColName, sizeof(pNode->tsColName));
         break;
       case PHY_DELETER_CODE_DELETE_TIME_RANGE:
         code = tlvDecodeObjFromTlv(pTlv, msgToTimeWindow, &pNode->deleteTimeRange);

source/libs/nodes/src/nodesUtilFuncs.c

@@ -209,6 +209,10 @@ int64_t nodesMakeAllocatorWeakRef(int64_t allocatorId) {
   }
 
   SNodeAllocator* pAllocator = taosAcquireRef(g_allocatorReqRefPool, allocatorId);
+  if (NULL == pAllocator) {
+    nodesError("allocator id %" PRIx64 " weak reference failed", allocatorId);
+    return -1;
+  }
   return pAllocator->self;
 }
 
@@ -1716,9 +1720,10 @@ static EDealRes doCollect(SCollectColumnsCxt* pCxt, SColumnNode* pCol, SNode* pN
   char    name[TSDB_TABLE_NAME_LEN + TSDB_COL_NAME_LEN];
   int32_t len = 0;
   if ('\0' == pCol->tableAlias[0]) {
-    len = sprintf(name, "%s", pCol->colName);
+    len = snprintf(name, sizeof(name), "%s", pCol->colName);
+  } else {
+    len = snprintf(name, sizeof(name), "%s.%s", pCol->tableAlias, pCol->colName);
   }
-  len = sprintf(name, "%s.%s", pCol->tableAlias, pCol->colName);
   if (NULL == taosHashGet(pCxt->pColHash, name, len)) {
     pCxt->errCode = taosHashPut(pCxt->pColHash, name, len, NULL, 0);
     if (TSDB_CODE_SUCCESS == pCxt->errCode) {

source/libs/parser/inc/parToken.h

@@ -175,8 +175,6 @@ _end:
 
 void taosCleanupKeywordsTable();
 
-SToken tscReplaceStrToken(char **str, SToken *token, const char *newToken);
-
 SToken taosTokenDup(SToken *pToken, char *buf, int32_t len);
 
 #ifdef __cplusplus

source/libs/parser/src/parAstCreater.c

@@ -518,7 +518,7 @@ SNode* createRealTableNode(SAstCreateContext* pCxt, SToken* pDbName, SToken* pTa
   if (NULL != pDbName) {
     COPY_STRING_FORM_ID_TOKEN(realTable->table.dbName, pDbName);
   } else {
-    strcpy(realTable->table.dbName, pCxt->pQueryCxt->db);
+    snprintf(realTable->table.dbName, sizeof(realTable->table.dbName), "%s", pCxt->pQueryCxt->db);
   }
   if (NULL != pTableAlias && TK_NK_NIL != pTableAlias->type) {
     COPY_STRING_FORM_ID_TOKEN(realTable->table.tableAlias, pTableAlias);
@@ -599,7 +599,7 @@ SNode* createStateWindowNode(SAstCreateContext* pCxt, SNode* pExpr) {
   state->pCol = createPrimaryKeyCol(pCxt, NULL);
   if (NULL == state->pCol) {
     nodesDestroyNode((SNode*)state);
-    CHECK_OUT_OF_MEM(state->pCol);
+    CHECK_OUT_OF_MEM(NULL);
   }
   state->pExpr = pExpr;
   return (SNode*)state;
@@ -613,7 +613,7 @@ SNode* createIntervalWindowNode(SAstCreateContext* pCxt, SNode* pInterval, SNode
   interval->pCol = createPrimaryKeyCol(pCxt, NULL);
   if (NULL == interval->pCol) {
     nodesDestroyNode((SNode*)interval);
-    CHECK_OUT_OF_MEM(interval->pCol);
+    CHECK_OUT_OF_MEM(NULL);
   }
   interval->pInterval = pInterval;
   interval->pOffset = pOffset;
@@ -631,7 +631,7 @@ SNode* createFillNode(SAstCreateContext* pCxt, EFillMode mode, SNode* pValues) {
   fill->pWStartTs = nodesMakeNode(QUERY_NODE_FUNCTION);
   if (NULL == fill->pWStartTs) {
     nodesDestroyNode((SNode*)fill);
-    CHECK_OUT_OF_MEM(fill->pWStartTs);
+    CHECK_OUT_OF_MEM(NULL);
   }
   strcpy(((SFunctionNode*)fill->pWStartTs)->functionName, "_wstart");
   return (SNode*)fill;
@@ -1495,10 +1495,10 @@ SNode* createCreateIndexStmt(SAstCreateContext* pCxt, EIndexType type, bool igno
   CHECK_OUT_OF_MEM(pStmt);
   pStmt->indexType = type;
   pStmt->ignoreExists = ignoreExists;
-  strcpy(pStmt->indexDbName, ((SRealTableNode*)pIndexName)->table.dbName);
+  snprintf(pStmt->indexDbName, sizeof(pStmt->indexDbName), "%s", ((SRealTableNode*)pIndexName)->table.dbName);
-  strcpy(pStmt->indexName, ((SRealTableNode*)pIndexName)->table.tableName);
+  snprintf(pStmt->indexName, sizeof(pStmt->indexName), "%s", ((SRealTableNode*)pIndexName)->table.tableName);
-  strcpy(pStmt->dbName, ((SRealTableNode*)pRealTable)->table.dbName);
+  snprintf(pStmt->dbName, sizeof(pStmt->dbName), "%s", ((SRealTableNode*)pRealTable)->table.dbName);
-  strcpy(pStmt->tableName, ((SRealTableNode*)pRealTable)->table.tableName);
+  snprintf(pStmt->tableName, sizeof(pStmt->tableName), "%s", ((SRealTableNode*)pRealTable)->table.tableName);
   nodesDestroyNode(pIndexName);
   nodesDestroyNode(pRealTable);
   pStmt->pCols = pCols;
@@ -1524,8 +1524,8 @@ SNode* createDropIndexStmt(SAstCreateContext* pCxt, bool ignoreNotExists, SNode*
   SDropIndexStmt* pStmt = (SDropIndexStmt*)nodesMakeNode(QUERY_NODE_DROP_INDEX_STMT);
   CHECK_OUT_OF_MEM(pStmt);
   pStmt->ignoreNotExists = ignoreNotExists;
-  strcpy(pStmt->indexDbName, ((SRealTableNode*)pIndexName)->table.dbName);
+  snprintf(pStmt->indexDbName, sizeof(pStmt->indexDbName), "%s", ((SRealTableNode*)pIndexName)->table.dbName);
-  strcpy(pStmt->indexName, ((SRealTableNode*)pIndexName)->table.tableName);
+  snprintf(pStmt->indexName, sizeof(pStmt->indexName), "%s", ((SRealTableNode*)pIndexName)->table.tableName);
   nodesDestroyNode(pIndexName);
   return (SNode*)pStmt;
 }
@@ -1818,7 +1818,7 @@ SNode* createRevokeStmt(SAstCreateContext* pCxt, int64_t privileges, SToken* pDb
 SNode* createFuncForDelete(SAstCreateContext* pCxt, const char* pFuncName) {
   SFunctionNode* pFunc = (SFunctionNode*)nodesMakeNode(QUERY_NODE_FUNCTION);
   CHECK_OUT_OF_MEM(pFunc);
-  strcpy(pFunc->functionName, pFuncName);
+  snprintf(pFunc->functionName, sizeof(pFunc->functionName), "%s", pFuncName);
   if (TSDB_CODE_SUCCESS != nodesListMakeStrictAppend(&pFunc->pParameterList, createPrimaryKeyCol(pCxt, NULL))) {
     nodesDestroyNode((SNode*)pFunc);
     CHECK_OUT_OF_MEM(NULL);

source/libs/parser/src/parInsert.c

@@ -378,7 +378,9 @@ static int parseTime(char** end, SToken* pToken, int16_t timePrec, int64_t* time
   } else if (pToken->type == TK_TODAY) {
     ts = taosGetTimestampToday(timePrec);
   } else if (pToken->type == TK_NK_INTEGER) {
-    toInteger(pToken->z, pToken->n, 10, &ts);
+    if (TSDB_CODE_SUCCESS != toInteger(pToken->z, pToken->n, 10, &ts)) {
+      return buildSyntaxErrMsg(pMsgBuf, "invalid timestamp format", pToken->z);
+    }
   } else {  // parse the RFC-3339/ISO-8601 timestamp format string
     if (taosParseTime(pToken->z, time, pToken->n, timePrec, tsDaylight) != TSDB_CODE_SUCCESS) {
       return buildSyntaxErrMsg(pMsgBuf, "invalid timestamp format", pToken->z);
@@ -591,8 +593,6 @@ static int32_t parseValueToken(char** end, SToken* pToken, SSchema* pSchema, int
     case TSDB_DATA_TYPE_BIGINT: {
       if (TSDB_CODE_SUCCESS != toInteger(pToken->z, pToken->n, 10, &iv)) {
         return buildSyntaxErrMsg(pMsgBuf, "invalid bigint data", pToken->z);
-      } else if (!IS_VALID_BIGINT(iv)) {
-        return buildSyntaxErrMsg(pMsgBuf, "bigint data overflow", pToken->z);
       }
       return func(pMsgBuf, &iv, pSchema->bytes, param);
     }
@@ -894,10 +894,7 @@ static int32_t parseTagToken(char** end, SToken* pToken, SSchema* pSchema, int16
     case TSDB_DATA_TYPE_BIGINT: {
       if (TSDB_CODE_SUCCESS != toInteger(pToken->z, pToken->n, 10, &iv)) {
         return buildSyntaxErrMsg(pMsgBuf, "invalid bigint data", pToken->z);
-      } else if (!IS_VALID_BIGINT(iv)) {
-        return buildSyntaxErrMsg(pMsgBuf, "bigint data overflow", pToken->z);
       }
-
       val->i64 = iv;
       break;
     }
@@ -1173,7 +1170,7 @@ static int32_t parseUsingClause(SInsertParseContext* pCxt, int32_t tbNo, SName*
   NEXT_TOKEN(pCxt->pSql, sToken);
 
   SName sname;
-  createSName(&sname, &sToken, pCxt->pComCxt->acctId, pCxt->pComCxt->db, &pCxt->msg);
+  CHECK_CODE(createSName(&sname, &sToken, pCxt->pComCxt->acctId, pCxt->pComCxt->db, &pCxt->msg));
   char dbFName[TSDB_DB_FNAME_LEN];
   tNameGetFullDbName(&sname, dbFName);
   strcpy(pCxt->sTableName, sname.tname);
@@ -1605,7 +1602,7 @@ static int32_t parseInsertBody(SInsertParseContext* pCxt) {
       if (!pCxt->pComCxt->needMultiParse) {
         continue;
       } else {
-        parserInfo("0x%" PRIx64 " insert from csv. File is too large, do it in batches.", pCxt->pComCxt->requestId);
+        parserDebug("0x%" PRIx64 " insert from csv. File is too large, do it in batches.", pCxt->pComCxt->requestId);
         break;
       }
     }
@@ -1613,7 +1610,7 @@ static int32_t parseInsertBody(SInsertParseContext* pCxt) {
     return buildSyntaxErrMsg(&pCxt->msg, "keyword VALUES or FILE is expected", sToken.z);
   }
 
-  parserInfo("0x%" PRIx64 " insert input rows: %d", pCxt->pComCxt->requestId, pCxt->totalNum);
+  parserDebug("0x%" PRIx64 " insert input rows: %d", pCxt->pComCxt->requestId, pCxt->totalNum);
 
   if (TSDB_QUERY_HAS_TYPE(pCxt->pOutput->insertType, TSDB_QUERY_TYPE_STMT_INSERT)) {
     SParsedDataColInfo* tags = taosMemoryMalloc(sizeof(pCxt->tags));
@@ -1650,7 +1647,7 @@ static int32_t parseInsertBodyAgain(SInsertParseContext* pCxt) {
     pCxt->pComCxt->needMultiParse = false;
     return TSDB_CODE_SUCCESS;
   }
-  parserInfo("0x%" PRIx64 " insert again input rows: %d", pCxt->pComCxt->requestId, pCxt->totalNum);
+  parserDebug("0x%" PRIx64 " insert again input rows: %d", pCxt->pComCxt->requestId, pCxt->totalNum);
   // merge according to vgId
   if (taosHashGetSize(pCxt->pTableBlockHashObj) > 0) {
     CHECK_CODE(mergeTableDataBlocks(pCxt->pTableBlockHashObj, pCxt->pOutput->payloadType, &pCxt->pVgDataBlocks));

source/libs/parser/src/parInsertData.c

@@ -12,7 +12,7 @@
  * You should have received a copy of the GNU Affero General Public License
  * along with this program. If not, see <http://www.gnu.org/licenses/>.
  */
-// clang-format off
+
 #include "parInsertData.h"
 
 #include "catalog.h"
@@ -25,8 +25,8 @@
   (((int)(t)) == PAYLOAD_TYPE_RAW)  // 0: K-V payload for non-prepare insert, 1: rawPayload for prepare insert
 
 typedef struct SBlockKeyTuple {
-  TSKEY skey;
+  TSKEY   skey;
-  void* payloadAddr;
+  void*   payloadAddr;
   int16_t index;
 } SBlockKeyTuple;
 
@@ -194,8 +194,8 @@ static int32_t createDataBlock(size_t defaultSize, int32_t rowSize, int32_t star
 
 int32_t buildCreateTbMsg(STableDataBlocks* pBlocks, SVCreateTbReq* pCreateTbReq) {
   SEncoder coder = {0};
-  char* pBuf;
+  char*    pBuf;
-  int32_t len;
+  int32_t  len;
 
   int32_t ret = 0;
   tEncodeSize(tEncodeSVCreateTbReq, pCreateTbReq, len, ret);
@@ -211,19 +211,19 @@ int32_t buildCreateTbMsg(STableDataBlocks* pBlocks, SVCreateTbReq* pCreateTbReq)
     }
   }
 
-  pBuf= pBlocks->pData + pBlocks->size;
+  pBuf = pBlocks->pData + pBlocks->size;
 
   tEncoderInit(&coder, pBuf, len);
-  tEncodeSVCreateTbReq(&coder, pCreateTbReq);
+  int32_t code = tEncodeSVCreateTbReq(&coder, pCreateTbReq);
   tEncoderClear(&coder);
-
   pBlocks->size += len;
   pBlocks->createTbReqLen = len;
-  return TSDB_CODE_SUCCESS;
+
+  return code;
 }
 
-int32_t getDataBlockFromList(SHashObj* pHashList, void* id, int32_t idLen, int32_t size, int32_t startOffset, int32_t rowSize,
+int32_t getDataBlockFromList(SHashObj* pHashList, void* id, int32_t idLen, int32_t size, int32_t startOffset,
-                             STableMeta* pTableMeta, STableDataBlocks** dataBlocks, SArray* pBlockList,
+                             int32_t rowSize, STableMeta* pTableMeta, STableDataBlocks** dataBlocks, SArray* pBlockList,
                              SVCreateTbReq* pCreateTbReq) {
   *dataBlocks = NULL;
   STableDataBlocks** t1 = (STableDataBlocks**)taosHashGet(pHashList, (const char*)id, idLen);
@@ -272,12 +272,12 @@ static void destroyDataBlock(STableDataBlocks* pDataBlock) {
   }
 
   taosMemoryFreeClear(pDataBlock->pData);
-//  if (!pDataBlock->cloned) {
+  //  if (!pDataBlock->cloned) {
-    // free the refcount for metermeta
+  // free the refcount for metermeta
-    taosMemoryFreeClear(pDataBlock->pTableMeta);
+  taosMemoryFreeClear(pDataBlock->pTableMeta);
 
-    destroyBoundColumnInfo(&pDataBlock->boundColumnInfo);
+  destroyBoundColumnInfo(&pDataBlock->boundColumnInfo);
-//  }
+  //  }
   taosMemoryFreeClear(pDataBlock);
 }
 
@@ -687,16 +687,16 @@ int32_t mergeTableDataBlocks(SHashObj* pHashObj, uint8_t payloadType, SArray** p
   STableDataBlocks** p = taosHashIterate(pHashObj, NULL);
   STableDataBlocks*  pOneTableBlock = *p;
   SBlockKeyInfo      blkKeyInfo = {0};  // share by pOneTableBlock
-  SBlockRowMerger    *pBlkRowMerger = NULL;
+  SBlockRowMerger*   pBlkRowMerger = NULL;
 
   while (pOneTableBlock) {
     SSubmitBlk* pBlocks = (SSubmitBlk*)pOneTableBlock->pData;
     if (pBlocks->numOfRows > 0) {
       STableDataBlocks* dataBuf = NULL;
-      pOneTableBlock->pTableMeta->vgId = pOneTableBlock->vgId;    // for schemaless, restore origin vgId
+      pOneTableBlock->pTableMeta->vgId = pOneTableBlock->vgId;  // for schemaless, restore origin vgId
-      int32_t           ret =
+      int32_t ret = getDataBlockFromList(pVnodeDataBlockHashList, &pOneTableBlock->vgId, sizeof(pOneTableBlock->vgId),
-          getDataBlockFromList(pVnodeDataBlockHashList, &pOneTableBlock->vgId, sizeof(pOneTableBlock->vgId), TSDB_PAYLOAD_SIZE, INSERT_HEAD_SIZE, 0,
+                                         TSDB_PAYLOAD_SIZE, INSERT_HEAD_SIZE, 0, pOneTableBlock->pTableMeta, &dataBuf,
-                               pOneTableBlock->pTableMeta, &dataBuf, pVnodeDataBlockList, NULL);
+                                         pVnodeDataBlockList, NULL);
       if (ret != TSDB_CODE_SUCCESS) {
         tdFreeSBlockRowMerger(pBlkRowMerger);
         taosHashCleanup(pVnodeDataBlockHashList);
@@ -708,7 +708,8 @@ int32_t mergeTableDataBlocks(SHashObj* pHashObj, uint8_t payloadType, SArray** p
       // the maximum expanded size in byte when a row-wise data is converted to SDataRow format
       int32_t expandSize = isRawPayload ? getRowExpandSize(pOneTableBlock->pTableMeta) : 0;
       int64_t destSize = dataBuf->size + pOneTableBlock->size + pBlocks->numOfRows * expandSize +
-                         sizeof(STColumn) * getNumOfColumns(pOneTableBlock->pTableMeta) + pOneTableBlock->createTbReqLen;
+                         sizeof(STColumn) * getNumOfColumns(pOneTableBlock->pTableMeta) +
+                         pOneTableBlock->createTbReqLen;
 
       if (dataBuf->nAllocSize < destSize) {
         dataBuf->nAllocSize = (uint32_t)(destSize * 1.5);
@@ -861,7 +862,7 @@ int32_t qCloneStmtDataBlock(void** pDst, void* pSrc) {
 
   STableDataBlocks* pBlock = (STableDataBlocks*)(*pDst);
   if (pBlock->pTableMeta) {
-    void *pNewMeta = taosMemoryMalloc(TABLE_META_SIZE(pBlock->pTableMeta));
+    void* pNewMeta = taosMemoryMalloc(TABLE_META_SIZE(pBlock->pTableMeta));
     if (NULL == pNewMeta) {
       taosMemoryFreeClear(*pDst);
       return TSDB_CODE_OUT_OF_MEMORY;
@@ -887,20 +888,18 @@ int32_t qRebuildStmtDataBlock(void** pDst, void* pSrc, uint64_t uid, int32_t vgI
   }
 
   pBlock->vgId = vgId;
-  
+
   if (pBlock->pTableMeta) {
     pBlock->pTableMeta->uid = uid;
     pBlock->pTableMeta->vgId = vgId;
   }
-  
+
   memset(pBlock->pData, 0, sizeof(SSubmitBlk));
 
   return TSDB_CODE_SUCCESS;
 }
 
-STableMeta *qGetTableMetaInDataBlock(void* pDataBlock) {
+STableMeta* qGetTableMetaInDataBlock(void* pDataBlock) { return ((STableDataBlocks*)pDataBlock)->pTableMeta; }
-  return ((STableDataBlocks*)pDataBlock)->pTableMeta;
-}
 
 void qFreeStmtDataBlock(void* pDataBlock) {
   if (pDataBlock == NULL) {

source/libs/parser/src/parTokenizer.c

@@ -613,27 +613,6 @@ uint32_t tGetToken(const char* z, uint32_t* tokenId) {
   return 0;
 }
 
-SToken tscReplaceStrToken(char** str, SToken* token, const char* newToken) {
-  char*   src = *str;
-  size_t  nsize = strlen(newToken);
-  int32_t size = (int32_t)strlen(*str) - token->n + (int32_t)nsize + 1;
-  int32_t bsize = (int32_t)((uint64_t)token->z - (uint64_t)src);
-  SToken  ntoken;
-
-  *str = taosMemoryCalloc(1, size);
-
-  strncpy(*str, src, bsize);
-  strcat(*str, newToken);
-  strcat(*str, token->z + token->n);
-
-  ntoken.n = (uint32_t)nsize;
-  ntoken.z = *str + bsize;
-
-  taosMemoryFreeClear(src);
-
-  return ntoken;
-}
-
 SToken tStrGetToken(const char* str, int32_t* i, bool isPrevOptr) {
   SToken t0 = {0};
 

source/libs/parser/src/parTranslater.c

@@ -292,9 +292,13 @@ static int32_t addNamespace(STranslateContext* pCxt, void* pTable) {
   } else {
     do {
       SArray* pTables = taosArrayInit(TARRAY_MIN_SIZE, POINTER_BYTES);
+      if (NULL == pTables) {
+        return TSDB_CODE_OUT_OF_MEMORY;
+      }
       if (pCxt->currLevel == currTotalLevel) {
         taosArrayPush(pTables, &pTable);
         if (hasSameTableAlias(pTables)) {
+          taosArrayDestroy(pTables);
           return generateSyntaxErrMsgExt(&pCxt->msgBuf, TSDB_CODE_PAR_NOT_UNIQUE_TABLE_ALIAS,
                                          "Not unique table/alias: '%s'", ((STableNode*)pTable)->tableAlias);
         }
@@ -308,7 +312,7 @@ static int32_t addNamespace(STranslateContext* pCxt, void* pTable) {
 
 static int32_t collectUseDatabaseImpl(const char* pFullDbName, SHashObj* pDbs) {
   SFullDatabaseName name = {0};
-  strcpy(name.fullDbName, pFullDbName);
+  snprintf(name.fullDbName, sizeof(name.fullDbName), "%s", pFullDbName);
   return taosHashPut(pDbs, pFullDbName, strlen(pFullDbName), &name, sizeof(SFullDatabaseName));
 }
 
@@ -540,15 +544,17 @@ static int32_t getTableIndex(STranslateContext* pCxt, const SName* pName, SArray
   if (TSDB_CODE_SUCCESS == code) {
     code = collectUseTable(pName, pCxt->pTables);
   }
-  if (pParCxt->async) {
+  if (TSDB_CODE_SUCCESS == code) {
-    code = getTableIndexFromCache(pCxt->pMetaCache, pName, pIndexes);
+    if (pParCxt->async) {
-  } else {
+      code = getTableIndexFromCache(pCxt->pMetaCache, pName, pIndexes);
-    SRequestConnInfo conn = {.pTrans = pParCxt->pTransporter,
+    } else {
-                             .requestId = pParCxt->requestId,
+      SRequestConnInfo conn = {.pTrans = pParCxt->pTransporter,
-                             .requestObjRefId = pParCxt->requestRid,
+                               .requestId = pParCxt->requestId,
-                             .mgmtEps = pParCxt->mgmtEpSet};
+                               .requestObjRefId = pParCxt->requestRid,
+                               .mgmtEps = pParCxt->mgmtEpSet};
 
-    code = catalogGetTableIndex(pParCxt->pCatalog, &conn, pName, pIndexes);
+      code = catalogGetTableIndex(pParCxt->pCatalog, &conn, pName, pIndexes);
+    }
   }
   if (TSDB_CODE_SUCCESS != code) {
     parserError("0x%" PRIx64 " getTableIndex error, code:%s, dbName:%s, tbName:%s", pCxt->pParseCxt->requestId,
@@ -990,9 +996,9 @@ static int32_t parseTimeFromValueNode(STranslateContext* pCxt, SValueNode* pVal)
       return pCxt->errCode;
     }
     if (IS_UNSIGNED_NUMERIC_TYPE(pVal->node.resType.type)) {
-      pVal->datum.i = pVal->datum.u;
+      pVal->datum.i = (int64_t)pVal->datum.u;
     } else if (IS_FLOAT_TYPE(pVal->node.resType.type)) {
-      pVal->datum.i = pVal->datum.d;
+      pVal->datum.i = (int64_t)pVal->datum.d;
     } else if (TSDB_DATA_TYPE_BOOL == pVal->node.resType.type) {
       pVal->datum.i = pVal->datum.b;
     }
@@ -1069,7 +1075,7 @@ static EDealRes translateNormalValue(STranslateContext* pCxt, SValueNode* pVal,
     }
     case TSDB_DATA_TYPE_BIGINT: {
       code = toInteger(pVal->literal, strlen(pVal->literal), 10, &pVal->datum.i);
-      if (strict && (TSDB_CODE_SUCCESS != code || !IS_VALID_BIGINT(pVal->datum.i))) {
+      if (strict && TSDB_CODE_SUCCESS != code) {
         return generateDealNodeErrMsg(pCxt, TSDB_CODE_PAR_WRONG_VALUE_TYPE, pVal->literal);
       }
       *(int64_t*)&pVal->typeData = pVal->datum.i;
@@ -3430,10 +3436,10 @@ static SNode* createSetOperProject(const char* pTableAlias, SNode* pNode) {
     return NULL;
   }
   pCol->node.resType = ((SExprNode*)pNode)->resType;
-  strcpy(pCol->tableAlias, pTableAlias);
+  snprintf(pCol->tableAlias, sizeof(pCol->tableAlias), "%s", pTableAlias);
-  strcpy(pCol->colName, ((SExprNode*)pNode)->aliasName);
+  snprintf(pCol->colName, sizeof(pCol->colName), "%s", ((SExprNode*)pNode)->aliasName);
-  strcpy(pCol->node.aliasName, pCol->colName);
+  snprintf(pCol->node.aliasName, sizeof(pCol->node.aliasName), "%s", pCol->colName);
-  strcpy(pCol->node.userAlias, ((SExprNode*)pNode)->userAlias);
+  snprintf(pCol->node.userAlias, sizeof(pCol->node.userAlias), "%s", ((SExprNode*)pNode)->userAlias);
   return (SNode*)pCol;
 }
 
@@ -3740,8 +3746,8 @@ static int32_t buildCreateDbReq(STranslateContext* pCxt, SCreateDatabaseStmt* pS
 static int32_t checkRangeOption(STranslateContext* pCxt, int32_t code, const char* pName, int32_t val, int32_t minVal,
                                 int32_t maxVal) {
   if (val >= 0 && (val < minVal || val > maxVal)) {
-    return generateSyntaxErrMsgExt(&pCxt->msgBuf, code, "Invalid option %s: %" PRId64 " valid range: [%d, %d]", pName,
+    return generateSyntaxErrMsgExt(&pCxt->msgBuf, code, "Invalid option %s: %d valid range: [%d, %d]", pName, val,
-                                   val, minVal, maxVal);
+                                   minVal, maxVal);
   }
   return TSDB_CODE_SUCCESS;
 }
@@ -3878,7 +3884,7 @@ static int32_t checkDbStrictOption(STranslateContext* pCxt, SDatabaseOptions* pO
 static int32_t checkDbEnumOption(STranslateContext* pCxt, const char* pName, int32_t val, int32_t v1, int32_t v2) {
   if (val >= 0 && val != v1 && val != v2) {
     return generateSyntaxErrMsgExt(&pCxt->msgBuf, TSDB_CODE_PAR_INVALID_DB_OPTION,
-                                   "Invalid option %s: %" PRId64 ", only %d, %d allowed", pName, val, v1, v2);
+                                   "Invalid option %s: %d, only %d, %d allowed", pName, val, v1, v2);
   }
   return TSDB_CODE_SUCCESS;
 }
@@ -4463,8 +4469,8 @@ static int32_t buildSampleAst(STranslateContext* pCxt, SSampleAstInfo* pInfo, ch
     nodesDestroyNode((SNode*)pSelect);
     return TSDB_CODE_OUT_OF_MEMORY;
   }
-  strcpy(pTable->table.dbName, pInfo->pDbName);
+  snprintf(pTable->table.dbName, sizeof(pTable->table.dbName), "%s", pInfo->pDbName);
-  strcpy(pTable->table.tableName, pInfo->pTableName);
+  snprintf(pTable->table.tableName, sizeof(pTable->table.tableName), "%s", pInfo->pTableName);
   TSWAP(pTable->pMeta, pInfo->pRollupTableMeta);
   pSelect->pFromTable = (SNode*)pTable;
 
@@ -4940,9 +4946,9 @@ static int32_t translateAlterUser(STranslateContext* pCxt, SAlterUserStmt* pStmt
   alterReq.superUser = 0;
   alterReq.enable = pStmt->enable;
   alterReq.sysInfo = pStmt->sysinfo;
-  strcpy(alterReq.pass, pStmt->password);
+  snprintf(alterReq.pass, sizeof(alterReq.pass), "%s", pStmt->password);
   if (NULL != pCxt->pParseCxt->db) {
-    strcpy(alterReq.dbname, pCxt->pParseCxt->db);
+    snprintf(alterReq.dbname, sizeof(alterReq.dbname), "%s", pCxt->pParseCxt->db);
   }
 
   return buildCmdMsg(pCxt, TDMT_MND_ALTER_USER, (FSerializeFunc)tSerializeSAlterUserReq, &alterReq);
@@ -6049,7 +6055,7 @@ static SNode* createProjectCol(const char* pProjCol) {
   if (NULL == pCol) {
     return NULL;
   }
-  strcpy(pCol->colName, pProjCol);
+  snprintf(pCol->colName, sizeof(pCol->colName), "%s", pProjCol);
   return (SNode*)pCol;
 }
 
@@ -6082,9 +6088,9 @@ static int32_t createSimpleSelectStmt(const char* pDb, const char* pTable, int32
     nodesDestroyNode((SNode*)pSelect);
     return TSDB_CODE_OUT_OF_MEMORY;
   }
-  strcpy(pRealTable->table.dbName, pDb);
+  snprintf(pRealTable->table.dbName, sizeof(pRealTable->table.dbName), "%s", pDb);
-  strcpy(pRealTable->table.tableName, pTable);
+  snprintf(pRealTable->table.tableName, sizeof(pRealTable->table.tableName), "%s", pTable);
-  strcpy(pRealTable->table.tableAlias, pTable);
+  snprintf(pRealTable->table.tableAlias, sizeof(pRealTable->table.tableAlias), "%s", pTable);
   pSelect->pFromTable = (SNode*)pRealTable;
 
   if (numOfProjs >= 0) {
@@ -7125,8 +7131,9 @@ static int32_t buildUpdateOptionsReq(STranslateContext* pCxt, SAlterTableStmt* p
       pReq->newComment = strdup(pStmt->pOptions->comment);
       if (NULL == pReq->newComment) {
         code = TSDB_CODE_OUT_OF_MEMORY;
+      } else {
+        pReq->newCommentLen = strlen(pReq->newComment);
       }
-      pReq->newCommentLen = strlen(pReq->newComment);
     } else {
       pReq->newCommentLen = -1;
     }

source/libs/parser/src/parUtil.c

@@ -381,6 +381,7 @@ int32_t parseJsontoTagData(const char* json, SArray* pTagVals, STag** ppTag, voi
         uError("charset:%s to %s. val:%s, errno:%s, convert failed.", DEFAULT_UNICODE_ENCODEC, tsCharset, jsonValue,
                strerror(errno));
         retCode = buildSyntaxErrMsg(pMsgBuf, "charset convert json error", jsonValue);
+        taosMemoryFree(tmp);
         goto end;
       }
       val.nData = valLen;
@@ -652,8 +653,8 @@ static int32_t buildCatalogReqForInsert(SParseContext* pCxt, const SParseMetaCac
     }
 
     SUserAuthInfo auth = {0};
-    strcpy(auth.user, pCxt->pUser);
+    snprintf(auth.user, sizeof(auth.user), "%s", pCxt->pUser);
-    strcpy(auth.dbFName, p->dbFName);
+    snprintf(auth.dbFName, sizeof(auth.dbFName), "%s", p->dbFName);
     auth.type = AUTH_TYPE_WRITE;
     taosArrayPush(pCatalogReq->pUser, &auth);
 

source/libs/planner/src/planLogicCreater.c

@@ -471,9 +471,9 @@ static SColumnNode* createColumnByExpr(const char* pStmtName, SExprNode* pExpr)
     return NULL;
   }
   pCol->node.resType = pExpr->resType;
-  strcpy(pCol->colName, pExpr->aliasName);
+  snprintf(pCol->colName, sizeof(pCol->colName), "%s", pExpr->aliasName);
   if (NULL != pStmtName) {
-    strcpy(pCol->tableAlias, pStmtName);
+    snprintf(pCol->tableAlias, sizeof(pCol->tableAlias), "%s", pStmtName);
   }
   return pCol;
 }

source/libs/planner/src/planOptimizer.c

@@ -1124,7 +1124,7 @@ static int32_t sortPriKeyOptGetSequencingNodes(SLogicNode* pNode, SNodeList** pS
   bool    notOptimize = false;
   int32_t code = sortPriKeyOptGetSequencingNodesImpl(pNode, &notOptimize, pSequencingNodes);
   if (TSDB_CODE_SUCCESS != code || notOptimize) {
-    nodesClearList(*pSequencingNodes);
+    NODES_CLEAR_LIST(*pSequencingNodes);
   }
   return code;
 }
@@ -1361,74 +1361,6 @@ static int32_t smaIndexOptCouldApplyIndex(SScanLogicNode* pScan, STableIndexInfo
   return code;
 }
 
-static SNode* smaIndexOptCreateWStartTs() {
-  SFunctionNode* pWStart = (SFunctionNode*)nodesMakeNode(QUERY_NODE_FUNCTION);
-  if (NULL == pWStart) {
-    return NULL;
-  }
-  strcpy(pWStart->functionName, "_wstart");
-  snprintf(pWStart->node.aliasName, sizeof(pWStart->node.aliasName), "%s.%p", pWStart->functionName, pWStart);
-  if (TSDB_CODE_SUCCESS != fmGetFuncInfo(pWStart, NULL, 0)) {
-    nodesDestroyNode((SNode*)pWStart);
-    return NULL;
-  }
-  return (SNode*)pWStart;
-}
-
-static int32_t smaIndexOptCreateMergeKey(SNode* pCol, SNodeList** pMergeKeys) {
-  SOrderByExprNode* pMergeKey = (SOrderByExprNode*)nodesMakeNode(QUERY_NODE_ORDER_BY_EXPR);
-  if (NULL == pMergeKey) {
-    return TSDB_CODE_OUT_OF_MEMORY;
-  }
-  pMergeKey->pExpr = nodesCloneNode(pCol);
-  if (NULL == pMergeKey->pExpr) {
-    nodesDestroyNode((SNode*)pMergeKey);
-    return TSDB_CODE_OUT_OF_MEMORY;
-  }
-  pMergeKey->order = ORDER_ASC;
-  pMergeKey->nullOrder = NULL_ORDER_FIRST;
-  return nodesListMakeStrictAppend(pMergeKeys, (SNode*)pMergeKey);
-}
-
-static int32_t smaIndexOptRewriteInterval(SWindowLogicNode* pInterval, int32_t wstrartIndex, SNodeList** pMergeKeys) {
-  if (wstrartIndex < 0) {
-    SNode* pWStart = smaIndexOptCreateWStartTs();
-    if (NULL == pWStart) {
-      return TSDB_CODE_OUT_OF_MEMORY;
-    }
-    int32_t code = createColumnByRewriteExpr(pWStart, &pInterval->node.pTargets);
-    if (TSDB_CODE_SUCCESS != code) {
-      nodesDestroyNode(pWStart);
-      return code;
-    }
-    wstrartIndex = LIST_LENGTH(pInterval->node.pTargets) - 1;
-  }
-  return smaIndexOptCreateMergeKey(nodesListGetNode(pInterval->node.pTargets, wstrartIndex), pMergeKeys);
-}
-
-static int32_t smaIndexOptApplyIndexExt(SLogicSubplan* pLogicSubplan, SScanLogicNode* pScan, STableIndexInfo* pIndex,
-                                        SNodeList* pSmaCols, int32_t wstrartIndex) {
-  SWindowLogicNode* pInterval = (SWindowLogicNode*)pScan->node.pParent;
-  SNodeList*        pMergeTargets = nodesCloneList(pInterval->node.pTargets);
-  if (NULL == pMergeTargets) {
-    return TSDB_CODE_OUT_OF_MEMORY;
-  }
-  SLogicNode* pSmaScan = NULL;
-  SLogicNode* pMerge = NULL;
-  SNodeList*  pMergeKeys = NULL;
-  int32_t     code = smaIndexOptRewriteInterval(pInterval, wstrartIndex, &pMergeKeys);
-  if (TSDB_CODE_SUCCESS == code) {
-    code = smaIndexOptCreateSmaScan(pScan, pIndex, pSmaCols, &pSmaScan);
-  }
-  if (TSDB_CODE_SUCCESS == code) {
-    code = smaIndexOptCreateMerge(pScan->node.pParent, pMergeKeys, pMergeTargets, &pMerge);
-  }
-  if (TSDB_CODE_SUCCESS == code) {
-    code = smaIndexOptRecombinationNode(pLogicSubplan, pScan->node.pParent, pMerge, pSmaScan);
-  }
-  return code;
-}
-
 static int32_t smaIndexOptApplyIndex(SLogicSubplan* pLogicSubplan, SScanLogicNode* pScan, STableIndexInfo* pIndex,
                                      SNodeList* pSmaCols, int32_t wstrartIndex) {
   SLogicNode* pSmaScan = NULL;
@@ -1559,7 +1491,7 @@ static SNode* partTagsCreateWrapperFunc(const char* pFuncName, SNode* pNode) {
     return NULL;
   }
 
-  strcpy(pFunc->functionName, pFuncName);
+  snprintf(pFunc->functionName, sizeof(pFunc->functionName), "%s", pFuncName);
   if (QUERY_NODE_COLUMN == nodeType(pNode) && COLUMN_TYPE_TBNAME != ((SColumnNode*)pNode)->colType) {
     SColumnNode* pCol = (SColumnNode*)pNode;
     partTagsSetAlias(pFunc->node.aliasName, sizeof(pFunc->node.aliasName), pCol->tableAlias, pCol->colName);
@@ -2028,7 +1960,7 @@ static SNode* rewriteUniqueOptCreateFirstFunc(SFunctionNode* pSelectValue, SNode
   if (NULL != pSelectValue) {
     strcpy(pFunc->node.aliasName, pSelectValue->node.aliasName);
   } else {
-    snprintf(pFunc->node.aliasName, sizeof(pFunc->node.aliasName), "%s.%p", pFunc->functionName, pFunc);
+    snprintf(pFunc->node.aliasName, sizeof(pFunc->node.aliasName), "%s.%p", pFunc->functionName, (void*)pFunc);
   }
   int32_t code = nodesListMakeStrictAppend(&pFunc->pParameterList, nodesCloneNode(pCol));
   if (TSDB_CODE_SUCCESS == code) {

source/libs/planner/src/planPhysiCreater.c

@@ -67,7 +67,7 @@ static SNode* createSlotDesc(SPhysiPlanContext* pCxt, const char* pName, const S
   if (NULL == pSlot) {
     return NULL;
   }
-  strcpy(pSlot->name, pName);
+  snprintf(pSlot->name, sizeof(pSlot->name), "%s", pName);
   pSlot->slotId = slotId;
   pSlot->dataType = ((SExprNode*)pNode)->resType;
   pSlot->reserve = reserve;
@@ -663,13 +663,17 @@ static int32_t createJoinPhysiNode(SPhysiPlanContext* pCxt, SNodeList* pChildren
     code = addDataBlockSlots(pCxt, pJoin->pTargets, pJoin->node.pOutputDataBlockDesc);
   }
 
-  SNodeList* condCols = nodesMakeList();
   if (TSDB_CODE_SUCCESS == code && NULL != pJoinLogicNode->pOnConditions) {
-    code = nodesCollectColumnsFromNode(pJoinLogicNode->pOnConditions, NULL, COLLECT_COL_TYPE_ALL, &condCols);
+    SNodeList* pCondCols = nodesMakeList();
-  }
+    if (NULL == pCondCols) {
-  if (TSDB_CODE_SUCCESS == code) {
+      code = TSDB_CODE_OUT_OF_MEMORY;
-    code = addDataBlockSlots(pCxt, condCols, pJoin->node.pOutputDataBlockDesc);
+    } else {
-    nodesDestroyList(condCols);
+      code = nodesCollectColumnsFromNode(pJoinLogicNode->pOnConditions, NULL, COLLECT_COL_TYPE_ALL, &pCondCols);
+    }
+    if (TSDB_CODE_SUCCESS == code) {
+      code = addDataBlockSlots(pCxt, pCondCols, pJoin->node.pOutputDataBlockDesc);
+    }
+    nodesDestroyList(pCondCols);
   }
 
   if (TSDB_CODE_SUCCESS == code && NULL != pJoinLogicNode->pOnConditions) {
@@ -1634,7 +1638,7 @@ static SSubplan* makeSubplan(SPhysiPlanContext* pCxt, SLogicSubplan* pLogicSubpl
   pSubplan->subplanType = pLogicSubplan->subplanType;
   pSubplan->level = pLogicSubplan->level;
   if (NULL != pCxt->pPlanCxt->pUser) {
-    strcpy(pSubplan->user, pCxt->pPlanCxt->pUser);
+    snprintf(pSubplan->user, sizeof(pSubplan->user), "%s", pCxt->pPlanCxt->pUser);
   }
   return pSubplan;
 }
@@ -1825,7 +1829,7 @@ static int32_t pushSubplan(SPhysiPlanContext* pCxt, SNode* pSubplan, int32_t lev
       return TSDB_CODE_OUT_OF_MEMORY;
     }
   }
-  return nodesListStrictAppend(pGroup->pNodeList, (SNode*)pSubplan);
+  return nodesListAppend(pGroup->pNodeList, (SNode*)pSubplan);
 }
 
 static int32_t buildPhysiPlan(SPhysiPlanContext* pCxt, SLogicSubplan* pLogicSubplan, SSubplan* pParent,

source/libs/planner/src/planSpliter.c

@@ -342,7 +342,7 @@ static int32_t stbSplAppendWStart(SNodeList* pFuncs, int32_t* pIndex) {
     return TSDB_CODE_OUT_OF_MEMORY;
   }
   strcpy(pWStart->functionName, "_wstart");
-  snprintf(pWStart->node.aliasName, sizeof(pWStart->node.aliasName), "%s.%p", pWStart->functionName, pWStart);
+  snprintf(pWStart->node.aliasName, sizeof(pWStart->node.aliasName), "%s.%p", pWStart->functionName, (void*)pWStart);
   int32_t code = fmGetFuncInfo(pWStart, NULL, 0);
   if (TSDB_CODE_SUCCESS == code) {
     code = nodesListStrictAppend(pFuncs, (SNode*)pWStart);
@@ -367,7 +367,7 @@ static int32_t stbSplAppendWEnd(SWindowLogicNode* pWin, int32_t* pIndex) {
     return TSDB_CODE_OUT_OF_MEMORY;
   }
   strcpy(pWEnd->functionName, "_wend");
-  snprintf(pWEnd->node.aliasName, sizeof(pWEnd->node.aliasName), "%s.%p", pWEnd->functionName, pWEnd);
+  snprintf(pWEnd->node.aliasName, sizeof(pWEnd->node.aliasName), "%s.%p", pWEnd->functionName, (void*)pWEnd);
   int32_t code = fmGetFuncInfo(pWEnd, NULL, 0);
   if (TSDB_CODE_SUCCESS == code) {
     code = nodesListStrictAppend(pWin->pFuncs, (SNode*)pWEnd);
@@ -389,22 +389,19 @@ static int32_t stbSplCreatePartWindowNode(SWindowLogicNode* pMergeWindow, SLogic
   SNode* pConditions = pMergeWindow->node.pConditions;
   pMergeWindow->node.pConditions = NULL;
 
-  int32_t           code = TSDB_CODE_SUCCESS;
   SWindowLogicNode* pPartWin = (SWindowLogicNode*)nodesCloneNode((SNode*)pMergeWindow);
   if (NULL == pPartWin) {
-    code = TSDB_CODE_OUT_OF_MEMORY;
+    return TSDB_CODE_OUT_OF_MEMORY;
   }
 
   pPartWin->node.groupAction = GROUP_ACTION_KEEP;
+  pMergeWindow->node.pTargets = pTargets;
+  pMergeWindow->node.pConditions = pConditions;
+  pPartWin->node.pChildren = pChildren;
+  splSetParent((SLogicNode*)pPartWin);
 
-  if (TSDB_CODE_SUCCESS == code) {
-    pMergeWindow->node.pTargets = pTargets;
-    pMergeWindow->node.pConditions = pConditions;
-    pPartWin->node.pChildren = pChildren;
-    splSetParent((SLogicNode*)pPartWin);
-    code = stbSplRewriteFuns(pFunc, &pPartWin->pFuncs, &pMergeWindow->pFuncs);
-  }
   int32_t index = 0;
+  int32_t code = stbSplRewriteFuns(pFunc, &pPartWin->pFuncs, &pMergeWindow->pFuncs);
   if (TSDB_CODE_SUCCESS == code) {
     code = stbSplAppendWStart(pPartWin->pFuncs, &index);
   }
@@ -721,15 +718,16 @@ static int32_t stbSplCreatePartAggNode(SAggLogicNode* pMergeAgg, SLogicNode** pO
   SNode* pConditions = pMergeAgg->node.pConditions;
   pMergeAgg->node.pConditions = NULL;
 
-  int32_t        code = TSDB_CODE_SUCCESS;
   SAggLogicNode* pPartAgg = (SAggLogicNode*)nodesCloneNode((SNode*)pMergeAgg);
   if (NULL == pPartAgg) {
-    code = TSDB_CODE_OUT_OF_MEMORY;
+    return TSDB_CODE_OUT_OF_MEMORY;
   }
 
   pPartAgg->node.groupAction = GROUP_ACTION_KEEP;
 
-  if (TSDB_CODE_SUCCESS == code && NULL != pGroupKeys) {
+  int32_t code = TSDB_CODE_SUCCESS;
+
+  if (NULL != pGroupKeys) {
     pPartAgg->pGroupKeys = pGroupKeys;
     code = createColumnByRewriteExprs(pPartAgg->pGroupKeys, &pPartAgg->node.pTargets);
   }

source/libs/qworker/inc/qwInt.h

@@ -123,6 +123,7 @@ typedef struct SQWTaskCtx {
   int32_t  execId;
   int32_t  level;
 
+  bool    queryGotData;
   bool    queryRsped;
   bool    queryEnd;
   bool    queryContinue;

source/libs/qworker/src/qworker.c

@@ -10,6 +10,7 @@
 #include "tdatablock.h"
 #include "tmsg.h"
 #include "tname.h"
+#include "tglobal.h"
 
 SQWorkerMgmt gQwMgmt = {
     .lock = 0,
@@ -92,6 +93,19 @@ int32_t qwHandleTaskComplete(QW_FPARAMS_DEF, SQWTaskCtx *ctx) {
   return TSDB_CODE_SUCCESS;
 }
 
+int32_t qwSendQueryRsp(QW_FPARAMS_DEF, int32_t msgType, SQWTaskCtx *ctx, int32_t rspCode, bool quickRsp) {
+  if ((!quickRsp) || QUERY_RSP_POLICY_QUICK == tsQueryRspPolicy) {
+    if (!ctx->localExec) {
+      qwBuildAndSendQueryRsp(msgType, &ctx->ctrlConnInfo, rspCode, ctx);
+      QW_TASK_DLOG("query msg rsped, handle:%p, code:%x - %s", ctx->ctrlConnInfo.handle, rspCode, tstrerror(rspCode));
+    }
+
+    ctx->queryRsped = true;
+  }
+
+  return TSDB_CODE_SUCCESS;
+}
+
 int32_t qwExecTask(QW_FPARAMS_DEF, SQWTaskCtx *ctx, bool *queryStop) {
   int32_t        code = 0;
   bool           qcontinue = true;
@@ -144,7 +158,7 @@ int32_t qwExecTask(QW_FPARAMS_DEF, SQWTaskCtx *ctx, bool *queryStop) {
       if (numOfResBlock == 0) {
         QW_TASK_DLOG("qExecTask end with empty res, useconds:%" PRIu64, useconds);
       } else {
-        QW_TASK_DLOG("qExecTask done", "");
+        QW_TASK_DLOG("qExecTask done, useconds:%" PRIu64, useconds);
       }
 
       dsEndPut(sinkHandle, useconds);
@@ -234,6 +248,10 @@ int32_t qwGetQueryResFromSink(QW_FPARAMS_DEF, SQWTaskCtx *ctx, int32_t *dataLen,
   int32_t            code = 0;
   SOutputData        output = {0};
 
+  if (NULL == ctx->sinkHandle) {
+    return TSDB_CODE_SUCCESS;
+  }
+
   *dataLen = 0;
 
   while (true) {
@@ -407,6 +425,11 @@ int32_t qwHandlePrePhaseEvents(QW_FPARAMS_DEF, int8_t phase, SQWPhaseInput *inpu
         QW_ERR_JRET(TSDB_CODE_QRY_DUPLICATTED_OPERATION);
       }
 
+      if (ctx->rspCode) {
+        QW_TASK_ELOG("task already failed cause of %s, phase:%s", tstrerror(ctx->rspCode), qwPhaseStr(phase));
+        QW_ERR_JRET(ctx->rspCode);
+      }
+
       if (!ctx->queryRsped) {
         QW_TASK_ELOG("ready msg has not been processed, phase:%s", qwPhaseStr(phase));
         QW_ERR_JRET(TSDB_CODE_QRY_TASK_MSG_ERROR);
@@ -419,6 +442,11 @@ int32_t qwHandlePrePhaseEvents(QW_FPARAMS_DEF, int8_t phase, SQWPhaseInput *inpu
         QW_ERR_JRET(TSDB_CODE_QRY_TASK_DROPPED);
       }
 
+      if (ctx->rspCode) {
+        QW_TASK_ELOG("task already failed cause of %s, phase:%s", tstrerror(ctx->rspCode), qwPhaseStr(phase));
+        QW_ERR_JRET(ctx->rspCode);
+      }
+
       if (QW_EVENT_RECEIVED(ctx, QW_EVENT_DROP)) {
         QW_ERR_JRET(qwDropTask(QW_FPARAMS()));
 
@@ -499,21 +527,17 @@ _return:
 
   if (TSDB_CODE_SUCCESS == code && QW_PHASE_POST_QUERY == phase) {
     qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_PART_SUCC);
+    ctx->queryGotData = true;
   }
 
-  if (QW_PHASE_POST_QUERY == phase && ctx) {
+  if (QW_PHASE_POST_QUERY == phase && ctx && !ctx->queryRsped) {
-    if (!ctx->localExec) {
+    bool   rsped = false;
-      bool   rsped = false;
+    SQWMsg qwMsg = {.msgType = ctx->msgType, .connInfo = ctx->ctrlConnInfo};
-      SQWMsg qwMsg = {.msgType = ctx->msgType, .connInfo = ctx->ctrlConnInfo};
+    qwDbgSimulateRedirect(&qwMsg, ctx, &rsped);
-      qwDbgSimulateRedirect(&qwMsg, ctx, &rsped);
+    qwDbgSimulateDead(QW_FPARAMS(), ctx, &rsped);
-      qwDbgSimulateDead(QW_FPARAMS(), ctx, &rsped);
+    if (!rsped) {      
-      if (!rsped) {
+      qwSendQueryRsp(QW_FPARAMS(), input->msgType + 1, ctx, code, false);
-        qwBuildAndSendQueryRsp(input->msgType + 1, &ctx->ctrlConnInfo, code, ctx);
-        QW_TASK_DLOG("query msg rsped, handle:%p, code:%x - %s", ctx->ctrlConnInfo.handle, code, tstrerror(code));
-      }
     }
-
-    ctx->queryRsped = true;
   }
 
   if (ctx) {
@@ -551,6 +575,7 @@ int32_t qwPreprocessQuery(QW_FPARAMS_DEF, SQWMsg *qwMsg) {
   QW_ERR_JRET(qwAcquireTaskCtx(QW_FPARAMS(), &ctx));
 
   ctx->ctrlConnInfo = qwMsg->connInfo;
+  ctx->phase = -1;
 
   QW_ERR_JRET(qwAddTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_INIT));
 
@@ -604,6 +629,8 @@ int32_t qwProcessQuery(QW_FPARAMS_DEF, SQWMsg *qwMsg, char *sql) {
     QW_ERR_JRET(TSDB_CODE_QRY_APP_ERROR);
   }
 
+  qwSendQueryRsp(QW_FPARAMS(), qwMsg->msgType + 1, ctx, code, true);
+
   ctx->level = plan->level;
   atomic_store_ptr(&ctx->taskHandle, pTaskInfo);
   atomic_store_ptr(&ctx->sinkHandle, sinkHandle);
@@ -619,6 +646,31 @@ _return:
   input.msgType = qwMsg->msgType;
   code = qwHandlePostPhaseEvents(QW_FPARAMS(), QW_PHASE_POST_QUERY, &input, NULL);
 
+  if (ctx != NULL && QW_EVENT_RECEIVED(ctx, QW_EVENT_FETCH)) {
+    void         *rsp = NULL;
+    int32_t       dataLen = 0;
+    SOutputData sOutput = {0};
+    QW_ERR_JRET(qwGetQueryResFromSink(QW_FPARAMS(), ctx, &dataLen, &rsp, &sOutput));
+
+    if (rsp) {
+      bool qComplete = (DS_BUF_EMPTY == sOutput.bufStatus && sOutput.queryEnd);
+
+      qwBuildFetchRsp(rsp, &sOutput, dataLen, qComplete);
+      if (qComplete) {
+        atomic_store_8((int8_t *)&ctx->queryEnd, true);
+      }
+
+      qwMsg->connInfo = ctx->dataConnInfo;
+      QW_SET_EVENT_PROCESSED(ctx, QW_EVENT_FETCH);
+
+      qwBuildAndSendFetchRsp(ctx->fetchType, &qwMsg->connInfo, rsp, dataLen, code);
+      rsp = NULL;
+
+      QW_TASK_DLOG("fetch rsp send, handle:%p, code:%x - %s, dataLen:%d", qwMsg->connInfo.handle, code,
+                   tstrerror(code), dataLen);
+    }
+  }
+
   QW_RET(TSDB_CODE_SUCCESS);
 }
 
@@ -740,7 +792,9 @@ int32_t qwProcessFetch(QW_FPARAMS_DEF, SQWMsg *qwMsg) {
     locked = true;
 
     // RC WARNING
-    if (QW_QUERY_RUNNING(ctx)) {
+    if (-1 == ctx->phase || false == ctx->queryGotData) {
+      QW_TASK_DLOG_E("task query unfinished");
+    } else if (QW_QUERY_RUNNING(ctx)) {
       atomic_store_8((int8_t *)&ctx->queryContinue, 1);
     } else if (0 == atomic_load_8((int8_t *)&ctx->queryInQueue)) {
       qwUpdateTaskStatus(QW_FPARAMS(), JOB_TASK_STATUS_EXEC);

source/libs/scheduler/src/schTask.c

@@ -913,7 +913,7 @@ int32_t schLaunchRemoteTask(SSchJob *pJob, SSchTask *pTask) {
   SCH_ERR_RET(schSetTaskCandidateAddrs(pJob, pTask));
 
   if (SCH_IS_QUERY_JOB(pJob)) {
-    SCH_ERR_RET(schEnsureHbConnection(pJob, pTask));
+//    SCH_ERR_RET(schEnsureHbConnection(pJob, pTask));
   }
 
   SCH_RET(schBuildAndSendMsg(pJob, pTask, NULL, plan->msgType));
@@ -993,6 +993,12 @@ int32_t schLaunchTaskImpl(void *param) {
     SCH_ERR_JRET(schLaunchRemoteTask(pJob, pTask));
   }
 
+#if 0
+  if (SCH_IS_QUERY_JOB(pJob)) {
+    SCH_ERR_JRET(schEnsureHbConnection(pJob, pTask));
+  }
+#endif
+
 _return:
 
   if (pJob->taskNum >= SCH_MIN_AYSNC_EXEC_NUM) {

source/libs/scheduler/src/scheduler.c

@@ -155,8 +155,7 @@ void schedulerFreeJob(int64_t *jobId, int32_t errCode) {
     return;
   }
 
-  SCH_JOB_DLOG("start to free job 0x%" PRIx64 ", errCode:0x%x", *jobId, errCode);
+  SCH_JOB_DLOG("start to free job 0x%" PRIx64 ", code:%s", *jobId, tstrerror(errCode));
-
   schHandleJobDrop(pJob, errCode);
 
   schReleaseJob(*jobId);

source/libs/stream/src/streamUpdate.c

@@ -174,14 +174,9 @@ bool updateInfoIsUpdated(SUpdateInfo *pInfo, uint64_t tableId, TSKEY ts) {
       if (res == TSDB_CODE_SUCCESS) {
         return false;
       } else {
-        qDebug("===stream===Update close window sbf. tableId:%" PRIu64 ", maxTs:%" PRIu64 ", mapMaxTs:%" PRIu64
-               ", ts:%" PRIu64,
-               tableId, maxTs, *pMapMaxTs, ts);
         return true;
       }
     }
-    qDebug("===stream===Update close window. tableId:%" PRIu64 ", maxTs:%" PRIu64 ", mapMaxTs:%" PRIu64 ", ts:%" PRIu64,
-           tableId, maxTs, *pMapMaxTs, ts);
     return true;
   }
 
@@ -203,14 +198,10 @@ bool updateInfoIsUpdated(SUpdateInfo *pInfo, uint64_t tableId, TSKEY ts) {
   }
 
   if (ts < pInfo->minTS) {
-    qDebug("===stream===Update min ts. tableId:%" PRIu64 ", maxTs:%" PRIu64 ", mapMaxTs:%" PRIu64 ", ts:%" PRIu64,
-           tableId, maxTs, *pMapMaxTs, ts);
     return true;
   } else if (res == TSDB_CODE_SUCCESS) {
     return false;
   }
-  qDebug("===stream===Update. tableId:%" PRIu64 ", maxTs:%" PRIu64 ", mapMaxTs:%" PRIu64 ", ts:%" PRIu64, tableId,
-         maxTs, *pMapMaxTs, ts);
   // check from tsdb api
   return true;
 }

source/libs/sync/inc/syncInt.h

@@ -169,7 +169,7 @@ typedef struct SSyncNode {
 } SSyncNode;
 
 // open/close --------------
-SSyncNode* syncNodeOpen(const SSyncInfo* pSyncInfo);
+SSyncNode* syncNodeOpen(SSyncInfo* pSyncInfo);
 void       syncNodeStart(SSyncNode* pSyncNode);
 void       syncNodeStartStandBy(SSyncNode* pSyncNode);
 void       syncNodeClose(SSyncNode* pSyncNode);

source/libs/sync/src/syncAppendEntries.c

@@ -579,6 +579,7 @@ int32_t syncNodeOnAppendEntriesSnapshot2Cb(SSyncNode* ths, SyncAppendEntriesBatc
             SSyncRaftEntry* pAppendEntry = (SSyncRaftEntry*)(pMsg->data + metaTableArr[i].offset);
             code = ths->pLogStore->syncLogAppendEntry(ths->pLogStore, pAppendEntry);
             if (code != 0) {
+              sError("vgId:%d, failed to append log entry since %s",  ths->vgId, tstrerror(terrno));
               return -1;
             }
 
@@ -710,6 +711,7 @@ int32_t syncNodeOnAppendEntriesSnapshot2Cb(SSyncNode* ths, SyncAppendEntriesBatc
             SSyncRaftEntry* pAppendEntry = (SSyncRaftEntry*)(pMsg->data + metaTableArr[i].offset);
             code = ths->pLogStore->syncLogAppendEntry(ths->pLogStore, pAppendEntry);
             if (code != 0) {
+              sError("vgId:%d, failed to append log entry since %s",  ths->vgId, tstrerror(terrno));
               return -1;
             }
 
@@ -859,6 +861,7 @@ int32_t syncNodeOnAppendEntriesSnapshotCb(SSyncNode* ths, SyncAppendEntries* pMs
 
         code = ths->pLogStore->syncLogAppendEntry(ths->pLogStore, pAppendEntry);
         if (code != 0) {
+          sError("vgId:%d, failed to append log entry since %s",  ths->vgId, tstrerror(terrno));
           return -1;
         }
 
@@ -974,6 +977,7 @@ int32_t syncNodeOnAppendEntriesSnapshotCb(SSyncNode* ths, SyncAppendEntries* pMs
 
         code = ths->pLogStore->syncLogAppendEntry(ths->pLogStore, pAppendEntry);
         if (code != 0) {
+          sError("vgId:%d, failed to append log entry since %s",  ths->vgId, tstrerror(terrno));
           return -1;
         }
 

source/libs/sync/src/syncCommit.c

@@ -76,7 +76,10 @@ void syncMaybeAdvanceCommitIndex(SSyncNode* pSyncNode) {
         pEntry = (SSyncRaftEntry*)taosLRUCacheValue(pCache, h);
       } else {
         pEntry = pSyncNode->pLogStore->getEntry(pSyncNode->pLogStore, index);
-        ASSERT(pEntry != NULL);
+        if (pEntry == NULL) {
+           sError("failed to get entry since %s. index:%lld", tstrerror(terrno), index);
+           return;
+        }
       }
       // cannot commit, even if quorum agree. need check term!
       if (pEntry->term <= pSyncNode->pRaftStore->currentTerm) {
@@ -127,7 +130,9 @@ void syncMaybeAdvanceCommitIndex(SSyncNode* pSyncNode) {
     // execute fsm
     if (pSyncNode->pFsm != NULL) {
       int32_t code = syncNodeCommit(pSyncNode, beginIndex, endIndex, pSyncNode->state);
-      ASSERT(code == 0);
+      if (code != 0) {
+         wError("failed to commit sync node since %s", tstrerror(terrno));
+      }
     }
   }
 }

source/libs/sync/src/syncMain.c

@@ -86,10 +86,10 @@ void syncCleanUp() {
   }
 }
 
-int64_t syncOpen(const SSyncInfo* pSyncInfo) {
+int64_t syncOpen(SSyncInfo* pSyncInfo) {
   SSyncNode* pSyncNode = syncNodeOpen(pSyncInfo);
   if (pSyncNode == NULL) {
-    sError("failed to open sync node. vgId:%d", pSyncInfo->vgId);
+    sError("vgId:%d, failed to open sync node since %s", pSyncInfo->vgId, terrstr());
     return -1;
   }
 
@@ -230,7 +230,7 @@ int32_t syncReconfigBuild(int64_t rid, const SSyncCfg* pNewCfg, SRpcMsg* pRpcMsg
   return ret;
 }
 
-int32_t syncReconfig(int64_t rid, const SSyncCfg* pNewCfg) {
+int32_t syncReconfig(int64_t rid, SSyncCfg* pNewCfg) {
   SSyncNode* pSyncNode = (SSyncNode*)taosAcquireRef(tsNodeRefId, rid);
   if (pSyncNode == NULL) {
     terrno = TSDB_CODE_SYN_INTERNAL_ERROR;
@@ -238,6 +238,7 @@ int32_t syncReconfig(int64_t rid, const SSyncCfg* pNewCfg) {
   }
   ASSERT(rid == pSyncNode->rid);
 
+#if 0
   if (!syncNodeCheckNewConfig(pSyncNode, pNewCfg)) {
     taosReleaseRef(tsNodeRefId, pSyncNode->rid);
     terrno = TSDB_CODE_SYN_NEW_CONFIG_ERROR;
@@ -259,6 +260,12 @@ int32_t syncReconfig(int64_t rid, const SSyncCfg* pNewCfg) {
 
   taosReleaseRef(tsNodeRefId, pSyncNode->rid);
   return ret;
+#else
+  syncNodeUpdateNewConfigIndex(pSyncNode, pNewCfg);
+  syncNodeDoConfigChange(pSyncNode, pNewCfg, SYNC_INDEX_INVALID);
+  taosReleaseRef(tsNodeRefId, pSyncNode->rid);
+  return 0;
+#endif
 }
 
 int32_t syncLeaderTransfer(int64_t rid) {
@@ -919,9 +926,7 @@ _END:
 }
 
 // open/close --------------
-SSyncNode* syncNodeOpen(const SSyncInfo* pOldSyncInfo) {
+SSyncNode* syncNodeOpen(SSyncInfo* pSyncInfo) {
-  SSyncInfo* pSyncInfo = (SSyncInfo*)pOldSyncInfo;
-
   SSyncNode* pSyncNode = (SSyncNode*)taosMemoryCalloc(1, sizeof(SSyncNode));
   if (pSyncNode == NULL) {
     terrno = TSDB_CODE_OUT_OF_MEMORY;
@@ -957,7 +962,9 @@ SSyncNode* syncNodeOpen(const SSyncInfo* pOldSyncInfo) {
       sError("failed to open raft cfg file. path:%s", pSyncNode->configPath);
       goto _error;
     }
-    pSyncInfo->syncCfg = pSyncNode->pRaftCfg->cfg;
+    if (pSyncInfo->syncCfg.replicaNum == 0) {
+      pSyncInfo->syncCfg = pSyncNode->pRaftCfg->cfg;
+    }
 
     raftCfgClose(pSyncNode->pRaftCfg);
     pSyncNode->pRaftCfg = NULL;
@@ -2650,7 +2657,10 @@ static int32_t syncNodeAppendNoop(SSyncNode* ths) {
 
   if (ths->state == TAOS_SYNC_STATE_LEADER) {
     int32_t code = ths->pLogStore->syncLogAppendEntry(ths->pLogStore, pEntry);
-    ASSERT(code == 0);
+    if (code != 0) {
+      sError("vgId:%d, failed to append log entry since %s",  ths->vgId, tstrerror(terrno));
+      return -1;
+    }
     syncNodeReplicate(ths, false);
   }
 
@@ -2726,7 +2736,7 @@ int32_t syncNodeOnClientRequestCb(SSyncNode* ths, SyncClientRequest* pMsg, SyncI
     code = ths->pLogStore->syncLogAppendEntry(ths->pLogStore, pEntry);
     if (code != 0) {
       // del resp mgr, call FpCommitCb
-      ASSERT(0);
+      sError("vgId:%d, failed to append log entry since %s",  ths->vgId, tstrerror(terrno));
       return -1;
     }
 
@@ -2790,8 +2800,8 @@ int32_t syncNodeOnClientRequestBatchCb(SSyncNode* ths, SyncClientRequestBatch* p
 
     code = ths->pLogStore->syncLogAppendEntry(ths->pLogStore, pEntry);
     if (code != 0) {
+      sError("vgId:%d, failed to append log entry since %s", ths->vgId, tstrerror(terrno));
       // del resp mgr, call FpCommitCb
-      ASSERT(0);
       return -1;
     }
 
@@ -3043,7 +3053,10 @@ int32_t syncNodeCommit(SSyncNode* ths, SyncIndex beginIndex, SyncIndex endIndex,
           pEntry = (SSyncRaftEntry*)taosLRUCacheValue(pCache, h);
         } else {
           code = ths->pLogStore->syncLogGetEntry(ths->pLogStore, i, &pEntry);
-          ASSERT(code == 0);
+          if (code != 0) {
+              sError("vgId:%d, failed to get log entry since %s. index:%lld", ths->vgId, tstrerror(terrno), i);
+              return -1;
+           }
           ASSERT(pEntry != NULL);
         }
 

source/libs/sync/src/syncRaftLog.c

@@ -234,8 +234,6 @@ static int32_t raftLogAppendEntry(struct SSyncLogStore* pLogStore, SSyncRaftEntr
     snprintf(logBuf, sizeof(logBuf), "wal write error, index:%" PRId64 ", err:%d %X, msg:%s, syserr:%d, sysmsg:%s",
              pEntry->index, err, err, errStr, sysErr, sysErrStr);
     syncNodeErrorLog(pData->pSyncNode, logBuf);
-
-    ASSERT(0);
     return -1;
   }
   pEntry->index = index;
@@ -414,13 +412,11 @@ int32_t logStoreAppendEntry(SSyncLogStore* pLogStore, SSyncRaftEntry* pEntry) {
   return 0;
 }
 
-SSyncRaftEntry* logStoreGetEntry(SSyncLogStore* pLogStore, SyncIndex index) {
+SSyncRaftEntry* logStoreGetEntryWithoutLock(SSyncLogStore* pLogStore, SyncIndex index) {
   SSyncLogStoreData* pData = pLogStore->data;
   SWal*              pWal = pData->pWal;
 
   if (index >= SYNC_INDEX_BEGIN && index <= logStoreLastIndex(pLogStore)) {
-    taosThreadMutexLock(&(pData->mutex));
-
     // SWalReadHandle* pWalHandle = walOpenReadHandle(pWal);
     SWalReader* pWalHandle = pData->pWalHandle;
     ASSERT(pWalHandle != NULL);
@@ -444,7 +440,8 @@ SSyncRaftEntry* logStoreGetEntry(SSyncLogStore* pLogStore, SyncIndex index) {
         }
       } while (0);
 
-      ASSERT(0);
+      sError("failed to read ver since %s. index:%lld", tstrerror(terrno), index);
+      return NULL;
     }
 
     SSyncRaftEntry* pEntry = syncEntryBuild(pWalHandle->pHead->head.bodyLen);
@@ -465,7 +462,6 @@ SSyncRaftEntry* logStoreGetEntry(SSyncLogStore* pLogStore, SyncIndex index) {
         terrno = saveErr;
     */
 
-    taosThreadMutexUnlock(&(pData->mutex));
     return pEntry;
 
   } else {
@@ -473,6 +469,16 @@ SSyncRaftEntry* logStoreGetEntry(SSyncLogStore* pLogStore, SyncIndex index) {
   }
 }
 
+SSyncRaftEntry* logStoreGetEntry(SSyncLogStore* pLogStore, SyncIndex index) {
+    SSyncLogStoreData* pData = pLogStore->data;
+    SSyncRaftEntry *pEntry = NULL;
+
+    taosThreadMutexLock(&pData->mutex);
+    pEntry = logStoreGetEntryWithoutLock(pLogStore, index);
+    taosThreadMutexUnlock(&pData->mutex);
+    return pEntry;
+}
+
 int32_t logStoreTruncate(SSyncLogStore* pLogStore, SyncIndex fromIndex) {
   SSyncLogStoreData* pData = pLogStore->data;
   SWal*              pWal = pData->pWal;

source/libs/tdb/src/db/tdbBtree.c

@@ -121,6 +121,7 @@ int tdbBtreeOpen(int keyLen, int valLen, SPager *pPager, char const *tbname, SPg
 
     ret = tdbPagerWrite(pPager, pPage);
     if (ret < 0) {
+      tdbError("failed to write page since %s", terrstr());
       return -1;
     }
 
@@ -483,9 +484,8 @@ static int tdbBtreeBalanceDeeper(SBTree *pBt, SPage *pRoot, SPage **ppChild, TXN
 
   ret = tdbPagerWrite(pPager, pChild);
   if (ret < 0) {
-    // TODO
+    tdbError("failed to write page since %s", terrstr());
-    ASSERT(0);
+    return -1;
-    return 0;
   }
 
   // Copy the root page content to the child page
@@ -556,8 +556,7 @@ static int tdbBtreeBalanceNonRoot(SBTree *pBt, SPage *pParent, int idx, TXN *pTx
 
       ret = tdbPagerWrite(pBt->pPager, pOlds[i]);
       if (ret < 0) {
-        // TODO
+        tdbError("failed to write page since %s", terrstr());
-        ASSERT(0);
         return -1;
       }
     }
@@ -583,8 +582,7 @@ static int tdbBtreeBalanceNonRoot(SBTree *pBt, SPage *pParent, int idx, TXN *pTx
 
     ret = tdbPagerWrite(pBt->pPager, pParent);
     if (ret < 0) {
-      // TODO
+      tdbError("failed to write page since %s", terrstr());
-      ASSERT(0);
       return -1;
     }
 
@@ -719,8 +717,7 @@ static int tdbBtreeBalanceNonRoot(SBTree *pBt, SPage *pParent, int idx, TXN *pTx
 
         ret = tdbPagerWrite(pBt->pPager, pNews[iNew]);
         if (ret < 0) {
-          // TODO
+          tdbError("failed to write page since %s", terrstr());
-          ASSERT(0);
           return -1;
         }
       }
@@ -937,7 +934,7 @@ static int tdbFetchOvflPage(SPgno *pPgno, SPage **ppOfp, TXN *pTxn, SBTree *pBt)
   // mark dirty
   ret = tdbPagerWrite(pBt->pPager, *ppOfp);
   if (ret < 0) {
-    ASSERT(0);
+    tdbError("failed to write page since %s", terrstr());
     return -1;
   }
 
@@ -987,7 +984,7 @@ static int tdbBtreeEncodePayload(SPage *pPage, SCell *pCell, int nHeader, const
 
     // fetch a new ofp and make it dirty
     SPgno  pgno = 0;
-    SPage *ofp, *nextOfp;
+    SPage *ofp = NULL, *nextOfp = NULL;
 
     ret = tdbFetchOvflPage(&pgno, &ofp, pTxn, pBt);
     if (ret < 0) {
@@ -1942,7 +1939,7 @@ int tdbBtcDelete(SBTC *pBtc) {
   // drop the cell on the leaf
   ret = tdbPagerWrite(pPager, pBtc->pPage);
   if (ret < 0) {
-    ASSERT(0);
+    tdbError("failed to write page since %s", terrstr());
     return -1;
   }
 
@@ -1964,7 +1961,7 @@ int tdbBtcDelete(SBTC *pBtc) {
         if (idx < nCells) {
           ret = tdbPagerWrite(pPager, pPage);
           if (ret < 0) {
-            ASSERT(0);
+            tdbError("failed to write page since %s", terrstr());
             return -1;
           }
 
@@ -2029,7 +2026,7 @@ int tdbBtcUpsert(SBTC *pBtc, const void *pKey, int kLen, const void *pData, int
   // mark dirty
   ret = tdbPagerWrite(pBtc->pBt->pPager, pBtc->pPage);
   if (ret < 0) {
-    ASSERT(0);
+    tdbError("failed to write page since %s", terrstr());
     return -1;
   }
 

source/libs/tdb/src/db/tdbDb.c

@@ -106,7 +106,7 @@ int32_t tdbBegin(TDB *pDb, TXN *pTxn) {
   for (pPager = pDb->pgrList; pPager; pPager = pPager->pNext) {
     ret = tdbPagerBegin(pPager, pTxn);
     if (ret < 0) {
-      ASSERT(0);
+      tdbError("failed to begin pager since %s. dbName:%s, txnId:%d", tstrerror(terrno), pDb->dbName, pTxn->txnId);
       return -1;
     }
   }
@@ -121,7 +121,7 @@ int32_t tdbCommit(TDB *pDb, TXN *pTxn) {
   for (pPager = pDb->pgrList; pPager; pPager = pPager->pNext) {
     ret = tdbPagerCommit(pPager, pTxn);
     if (ret < 0) {
-      ASSERT(0);
+      tdbError("failed to commit pager since %s. dbName:%s, txnId:%d", tstrerror(terrno), pDb->dbName, pTxn->txnId);
       return -1;
     }
   }
@@ -136,7 +136,7 @@ int32_t tdbAbort(TDB *pDb, TXN *pTxn) {
   for (pPager = pDb->pgrList; pPager; pPager = pPager->pNext) {
     ret = tdbPagerAbort(pPager, pTxn);
     if (ret < 0) {
-      ASSERT(0);
+      tdbError("failed to abort pager since %s. dbName:%s, txnId:%d", tstrerror(terrno), pDb->dbName, pTxn->txnId);
       return -1;
     }
   }

source/libs/tdb/src/db/tdbPager.c

@@ -156,6 +156,7 @@ int tdbPagerOpenDB(SPager *pPager, SPgno *ppgno, bool toCreate, SBTree *pBt) {
 
     ret = tdbPagerWrite(pPager, pPage);
     if (ret < 0) {
+      tdbError("failed to write page since %s", terrstr());
       return -1;
     }
 
@@ -210,7 +211,7 @@ int tdbPagerWrite(SPager *pPager, SPage *pPage) {
   if (TDB_PAGE_PGNO(pPage) <= pPager->dbOrigSize) {
     ret = tdbPagerWritePageToJournal(pPager, pPage);
     if (ret < 0) {
-      ASSERT(0);
+      tdbError("failed to write page to journal since %s", tstrerror(terrno));
       return -1;
     }
   }
@@ -226,6 +227,8 @@ int tdbPagerBegin(SPager *pPager, TXN *pTxn) {
   // Open the journal
   pPager->jfd = tdbOsOpen(pPager->jFileName, TDB_O_CREAT | TDB_O_RDWR, 0755);
   if (pPager->jfd < 0) {
+    tdbError("failed to open file due to %s. jFileName:%s", strerror(errno), pPager->jFileName);
+    terrno = TAOS_SYSTEM_ERROR(errno);
     return -1;
   }
 
@@ -243,9 +246,9 @@ int tdbPagerCommit(SPager *pPager, TXN *pTxn) {
   // sync the journal file
   ret = tdbOsFSync(pPager->jfd);
   if (ret < 0) {
-    // TODO
+    tdbError("failed to fsync jfd due to %s. jFileName:%s", strerror(errno), pPager->jFileName);
-    ASSERT(0);
+    terrno = TAOS_SYSTEM_ERROR(errno);
-    return 0;
+    return -1;
   }
 
   // loop to write the dirty pages to file
@@ -255,7 +258,7 @@ int tdbPagerCommit(SPager *pPager, TXN *pTxn) {
     pPage = (SPage *)pNode;
     ret = tdbPagerWritePageToDB(pPager, pPage);
     if (ret < 0) {
-      ASSERT(0);
+      tdbError("failed to write page to db since %s", tstrerror(terrno));
       return -1;
     }
   }
@@ -277,11 +280,25 @@ int tdbPagerCommit(SPager *pPager, TXN *pTxn) {
   tRBTreeCreate(&pPager->rbt, pageCmpFn);
 
   // sync the db file
-  tdbOsFSync(pPager->fd);
+  if (tdbOsFSync(pPager->fd) < 0) {
+    tdbError("failed to fsync fd due to %s. file:%s", strerror(errno), pPager->dbFileName);
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    return -1;
+  }
 
   // remove the journal file
-  tdbOsClose(pPager->jfd);
+  if (tdbOsClose(pPager->jfd) < 0) {
-  tdbOsRemove(pPager->jFileName);
+    tdbError("failed to close jfd due to %s. file:%s", strerror(errno), pPager->jFileName);
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    return -1;
+  }
+
+  if (tdbOsRemove(pPager->jFileName) < 0 && errno != ENOENT) {
+    tdbError("failed to remove file due to %s. file:%s", strerror(errno), pPager->jFileName);
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    return -1;
+  }
+
   pPager->inTran = 0;
 
   return 0;
@@ -297,14 +314,14 @@ int tdbPagerAbort(SPager *pPager, TXN *pTxn) {
   // 0, sync the journal file
   ret = tdbOsFSync(pPager->jfd);
   if (ret < 0) {
-    // TODO
+    tdbError("failed to fsync jfd due to %s. file:%s", strerror(errno), pPager->jFileName);
-    ASSERT(0);
+    terrno = TAOS_SYSTEM_ERROR(errno);
-    return 0;
+    return -1;
   }
 
   tdb_fd_t jfd = tdbOsOpen(pPager->jFileName, TDB_O_RDWR, 0755);
   if (jfd == NULL) {
-    return 0;
+    return -1;
   }
 
   ret = tdbGetFileSize(jfd, pPager->pageSize, &journalSize);
@@ -516,11 +533,15 @@ static int tdbPagerWritePageToJournal(SPager *pPager, SPage *pPage) {
 
   ret = tdbOsWrite(pPager->jfd, &pgno, sizeof(pgno));
   if (ret < 0) {
+    tdbError("failed to write pgno due to %s. file:%s, pgno:%u", strerror(errno), pPager->jFileName, pgno);
+    terrno = TAOS_SYSTEM_ERROR(errno);
     return -1;
   }
 
   ret = tdbOsWrite(pPager->jfd, pPage->pData, pPage->pageSize);
   if (ret < 0) {
+    tdbError("failed to write page data due to %s. file:%s, pageSize:%ld", strerror(errno), pPager->jFileName, pPage->pageSize);
+    terrno = TAOS_SYSTEM_ERROR(errno);
     return -1;
   }
 
@@ -540,13 +561,15 @@ static int tdbPagerWritePageToDB(SPager *pPager, SPage *pPage) {
 
   offset = (i64)pPage->pageSize * (TDB_PAGE_PGNO(pPage) - 1);
   if (tdbOsLSeek(pPager->fd, offset, SEEK_SET) < 0) {
-    ASSERT(0);
+    tdbError("failed to lseek due to %s. file:%s, offset:%ld", strerror(errno), pPager->dbFileName, offset);
+    terrno = TAOS_SYSTEM_ERROR(errno);
     return -1;
   }
 
   ret = tdbOsWrite(pPager->fd, pPage->pData, pPage->pageSize);
   if (ret < 0) {
-    ASSERT(0);
+    tdbError("failed to write page data due to %s. file:%s, pageSize:%ld", strerror(errno), pPager->dbFileName, pPage->pageSize);
+    terrno = TAOS_SYSTEM_ERROR(errno);
     return -1;
   }
 
@@ -590,23 +613,38 @@ int tdbPagerRestore(SPager *pPager, SBTree *pBt) {
 
     i64 offset = pPager->pageSize * (pgno - 1);
     if (tdbOsLSeek(pPager->fd, offset, SEEK_SET) < 0) {
-      ASSERT(0);
+      tdbError("failed to lseek fd due to %s. file:%s, offset:%ld", strerror(errno), pPager->dbFileName, offset);
+      terrno = TAOS_SYSTEM_ERROR(errno);
       return -1;
     }
 
     ret = tdbOsWrite(pPager->fd, pageBuf, pPager->pageSize);
     if (ret < 0) {
-      ASSERT(0);
+      tdbError("failed to write buf due to %s. file: %s, bufsize:%d", strerror(errno), pPager->dbFileName, pPager->pageSize);
+      terrno = TAOS_SYSTEM_ERROR(errno);
       return -1;
     }
   }
 
-  tdbOsFSync(pPager->fd);
+  if (tdbOsFSync(pPager->fd) < 0) {
+    tdbError("failed to fsync fd due to %s. dbfile:%s", strerror(errno), pPager->dbFileName);
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    return -1;
+  }
 
   tdbOsFree(pageBuf);
 
-  tdbOsClose(jfd);
+  if (tdbOsClose(jfd) < 0) {
-  tdbOsRemove(pPager->jFileName);
+    tdbError("failed to close jfd due to %s. jFileName:%s", strerror(errno), pPager->jFileName);
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    return -1;
+  }
+
+  if (tdbOsRemove(pPager->jFileName) < 0 && errno != ENOENT) {
+    tdbError("failed to remove file due to %s. jFileName:%s", strerror(errno), pPager->jFileName);
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    return -1;
+  }
 
   return 0;
 }

source/libs/transport/src/transCli.c

@@ -597,6 +597,7 @@ static int32_t specifyConnRef(SCliConn* conn, bool update, int64_t handle) {
 static void cliAllocRecvBufferCb(uv_handle_t* handle, size_t suggested_size, uv_buf_t* buf) {
   SCliConn*    conn = handle->data;
   SConnBuffer* pBuf = &conn->readBuf;
+  tDebug("%s conn %p alloc read buf", CONN_GET_INST_LABEL(conn), conn);
   transAllocBuffer(pBuf, buf);
 }
 static void cliRecvCb(uv_stream_t* handle, ssize_t nread, const uv_buf_t* buf) {
@@ -609,7 +610,7 @@ static void cliRecvCb(uv_stream_t* handle, ssize_t nread, const uv_buf_t* buf) {
   if (nread > 0) {
     pBuf->len += nread;
     while (transReadComplete(pBuf)) {
-      tTrace("%s conn %p read complete", CONN_GET_INST_LABEL(conn), conn);
+      tDebug("%s conn %p read complete", CONN_GET_INST_LABEL(conn), conn);
       if (pBuf->invalid) {
         cliHandleExcept(conn);
         break;

source/libs/transport/src/transComm.c

@@ -203,7 +203,7 @@ bool transReadComplete(SConnBuffer* connBuf) {
 }
 
 int transSetConnOption(uv_tcp_t* stream) {
-  uv_tcp_nodelay(stream, 1);
+  uv_tcp_nodelay(stream, 0);
   int ret = uv_tcp_keepalive(stream, 5, 60);
   return ret;
 }

source/libs/transport/src/transSvr.c

@@ -332,7 +332,10 @@ void uvOnSendCb(uv_write_t* req, int status) {
   if (status == 0) {
     tTrace("conn %p data already was written on stream", conn);
     if (!transQueueEmpty(&conn->srvMsgs)) {
-      SSvrMsg* msg = transQueuePop(&conn->srvMsgs);
+      SSvrMsg*  msg = transQueuePop(&conn->srvMsgs);
+      STraceId* trace = &msg->msg.info.traceId;
+      tGDebug("conn %p write data out", conn);
+
       destroySmsg(msg);
       // send cached data
       if (!transQueueEmpty(&conn->srvMsgs)) {

source/libs/wal/src/walMeta.c

@@ -35,113 +35,251 @@ int64_t FORCE_INLINE walGetCommittedVer(SWal* pWal) { return pWal->vers.commitVe
 
 int64_t FORCE_INLINE walGetAppliedVer(SWal* pWal) { return pWal->vers.appliedVer; }
 
-static FORCE_INLINE void walBuildMetaName(SWal* pWal, int metaVer, char* buf) {
+static FORCE_INLINE int walBuildMetaName(SWal* pWal, int metaVer, char* buf) {
-  sprintf(buf, "%s/meta-ver%d", pWal->path, metaVer);
+  return sprintf(buf, "%s/meta-ver%d", pWal->path, metaVer);
 }
 
-static FORCE_INLINE int64_t walScanLogGetLastVer(SWal* pWal) {
+static FORCE_INLINE int walBuildTmpMetaName(SWal* pWal, char* buf) {
-  int32_t sz = taosArrayGetSize(pWal->fileInfoSet);
+  return sprintf(buf, "%s/meta-ver.tmp", pWal->path);
-  ASSERT(sz > 0);
+}
-#if 0
-  for (int i = 0; i < sz; i++) {
-    SWalFileInfo* pFileInfo = taosArrayGet(pWal->fileInfoSet, i);
-  }
-#endif
 
-  SWalFileInfo* pLastFileInfo = taosArrayGet(pWal->fileInfoSet, sz - 1);
+static FORCE_INLINE int64_t walScanLogGetLastVer(SWal* pWal, int32_t fileIdx) {
+  int32_t sz = taosArrayGetSize(pWal->fileInfoSet);
+  terrno = TSDB_CODE_SUCCESS;
+  ASSERT(fileIdx >= 0 && fileIdx < sz);
+
+  SWalFileInfo* pFileInfo = taosArrayGet(pWal->fileInfoSet, fileIdx);
   char          fnameStr[WAL_FILE_LEN];
-  walBuildLogName(pWal, pLastFileInfo->firstVer, fnameStr);
+  walBuildLogName(pWal, pFileInfo->firstVer, fnameStr);
 
   int64_t fileSize = 0;
   taosStatFile(fnameStr, &fileSize, NULL);
-  int32_t readSize = TMIN(WAL_SCAN_BUF_SIZE, fileSize);
-  pLastFileInfo->fileSize = fileSize;
 
-  TdFilePtr pFile = taosOpenFile(fnameStr, TD_FILE_READ);
+  TdFilePtr pFile = taosOpenFile(fnameStr, TD_FILE_READ | TD_FILE_WRITE);
   if (pFile == NULL) {
+    wError("vgId:%d, failed to open file due to %s. file:%s", pWal->cfg.vgId, strerror(errno), fnameStr);
     terrno = TAOS_SYSTEM_ERROR(errno);
     return -1;
   }
 
+  // ensure size as non-negative
+  pFileInfo->fileSize = TMAX(0, pFileInfo->fileSize);
+
   uint64_t magic = WAL_MAGIC;
+  int64_t  walCkHeadSz = sizeof(SWalCkHead);
+  int64_t  end = fileSize;
+  int64_t  offset = 0;
+  int32_t  capacity = 0;
+  int32_t  readSize = 0;
+  char*    buf = NULL;
+  char*    found = NULL;
+  bool     firstTrial = pFileInfo->fileSize < fileSize;
 
-  char* buf = taosMemoryMalloc(readSize + 5);
+  // search for the valid last WAL entry, e.g. block by block
-  if (buf == NULL) {
-    taosCloseFile(&pFile);
-    terrno = TSDB_CODE_WAL_OUT_OF_MEMORY;
-    return -1;
-  }
-
-  int64_t offset;
-  offset = taosLSeekFile(pFile, -readSize, SEEK_END);
-  if (readSize != taosReadFile(pFile, buf, readSize)) {
-    taosMemoryFree(buf);
-    taosCloseFile(&pFile);
-    terrno = TAOS_SYSTEM_ERROR(errno);
-    return -1;
-  }
-
-  char* found = NULL;
   while (1) {
+    offset = (firstTrial) ? pFileInfo->fileSize : TMAX(0, end - WAL_SCAN_BUF_SIZE);
+    ASSERT(offset <= end);
+    readSize = end - offset;
+    capacity = readSize + sizeof(magic);
+
+    int64_t limit = WAL_RECOV_SIZE_LIMIT;
+    if (limit < readSize) {
+      wError("vgId:%d, possibly corrupted WAL range exceeds size limit (i.e. %" PRId64 " bytes). offset:%" PRId64
+             ", end:%" PRId64 ", file:%s",
+             pWal->cfg.vgId, limit, offset, end, fnameStr);
+      terrno = TSDB_CODE_WAL_SIZE_LIMIT;
+      goto _err;
+    }
+
+    void* ptr = taosMemoryRealloc(buf, capacity);
+    if (ptr == NULL) {
+      terrno = TSDB_CODE_WAL_OUT_OF_MEMORY;
+      goto _err;
+    }
+    buf = ptr;
+
+    int64_t ret = taosLSeekFile(pFile, offset, SEEK_SET);
+    if (ret < 0) {
+      wError("vgId:%d, failed to lseek file due to %s. offset:%" PRId64 "", pWal->cfg.vgId, strerror(errno), offset);
+      terrno = TAOS_SYSTEM_ERROR(errno);
+      goto _err;
+    }
+
+    if (readSize != taosReadFile(pFile, buf, readSize)) {
+      wError("vgId:%d, failed to read file due to %s. readSize:%" PRId64 ", file:%s", pWal->cfg.vgId, strerror(errno),
+             readSize, fnameStr);
+      terrno = TAOS_SYSTEM_ERROR(errno);
+      goto _err;
+    }
+
+    char* candidate = NULL;
     char* haystack = buf;
-    char* candidate;
+
-    while ((candidate = tmemmem(haystack, readSize - (haystack - buf), (char*)&magic, sizeof(uint64_t))) != NULL) {
+    while ((candidate = tmemmem(haystack, readSize - (haystack - buf), (char*)&magic, sizeof(magic))) != NULL) {
-      // read and validate
+      // validate head
-      SWalCkHead* logContent = (SWalCkHead*)candidate;
+      int64_t len = readSize - (candidate - buf);
-      if (walValidHeadCksum(logContent) == 0 && walValidBodyCksum(logContent) == 0) {
+      if (len < walCkHeadSz) {
-        found = candidate;
+        break;
       }
+      SWalCkHead* logContent = (SWalCkHead*)candidate;
+      if (walValidHeadCksum(logContent) != 0) {
+        wError("vgId:%d, failed to validate checksum of wal entry header. offset:% %" PRId64 ", file:%s",
+               ((char*)(logContent)-buf), fnameStr);
+        haystack = candidate + 1;
+        if (firstTrial) {
+          break;
+        } else {
+          continue;
+        }
+      }
+
+      // validate body
+      int64_t size = walCkHeadSz + logContent->head.bodyLen;
+      if (len < size) {
+        int64_t extraSize = size - len;
+        if (capacity < readSize + extraSize + sizeof(magic)) {
+          capacity += extraSize;
+          void* ptr = taosMemoryRealloc(buf, capacity);
+          if (ptr == NULL) {
+            terrno = TSDB_CODE_OUT_OF_MEMORY;
+            goto _err;
+          }
+          buf = ptr;
+        }
+        int64_t ret = taosLSeekFile(pFile, offset + readSize, SEEK_SET);
+        if (ret < 0) {
+          wError("vgId:%d, failed to lseek file due to %s. offset:%" PRId64 "", pWal->cfg.vgId, strerror(errno),
+                 offset);
+          terrno = TAOS_SYSTEM_ERROR(errno);
+          break;
+        }
+        if (extraSize != taosReadFile(pFile, buf + readSize, extraSize)) {
+          wError("vgId:%d, failed to read file due to %s. offset:%" PRId64 ", extraSize:%" PRId64 ", file:%s",
+                 pWal->cfg.vgId, strerror(errno), offset + readSize, extraSize, fnameStr);
+          terrno = TAOS_SYSTEM_ERROR(errno);
+          break;
+        }
+      }
+      if (walValidBodyCksum(logContent) != 0) {
+        terrno = TSDB_CODE_WAL_CHKSUM_MISMATCH;
+        wError("vgId:%d, failed to validate checksum of wal entry body. offset:% %" PRId64 ", file:%s",
+               ((char*)(logContent)-buf), fnameStr);
+        haystack = candidate + 1;
+        if (firstTrial) {
+          break;
+        } else {
+          continue;
+        }
+      }
+
+      // found one
+      found = candidate;
       haystack = candidate + 1;
     }
+
     if (found || offset == 0) break;
-    offset = TMIN(0, offset - readSize + sizeof(uint64_t));
+
-    int64_t offset2 = taosLSeekFile(pFile, offset, SEEK_SET);
+    // go backwards, e.g. by at most one WAL scan buf size
-    ASSERT(offset == offset2);
+    end = offset + walCkHeadSz - 1;
-    if (readSize != taosReadFile(pFile, buf, readSize)) {
+    firstTrial = false;
-      taosMemoryFree(buf);
-      taosCloseFile(&pFile);
-      terrno = TAOS_SYSTEM_ERROR(errno);
-      return -1;
-    }
-#if 0
-    if (found == buf) {
-      SWalCkHead* logContent = (SWalCkHead*)found;
-      if (walValidHeadCksum(logContent) != 0 || walValidBodyCksum(logContent) != 0) {
-        // file has to be deleted
-        taosMemoryFree(buf);
-        taosCloseFile(&pFile);
-        terrno = TSDB_CODE_WAL_FILE_CORRUPTED;
-        return -1;
-      }
-    }
-#endif
   }
 
-  if (found == NULL) {
+  // determine end of last entry
-    // file corrupted, no complete log
+  SWalCkHead* lastEntry = (SWalCkHead*)found;
-    // TODO delete and search in previous files
+  int64_t     retVer = -1;
-    /*ASSERT(0);*/
+  int64_t     lastEntryBeginOffset = 0;
-    terrno = TSDB_CODE_WAL_FILE_CORRUPTED;
+  int64_t     lastEntryEndOffset = 0;
-    return -1;
+
+  if (lastEntry == NULL) {
+    terrno = TSDB_CODE_WAL_LOG_NOT_EXIST;
+  } else {
+    retVer = lastEntry->head.version;
+    lastEntryBeginOffset = offset + (int64_t)((char*)lastEntry - (char*)buf);
+    lastEntryEndOffset = lastEntryBeginOffset + sizeof(SWalCkHead) + lastEntry->head.bodyLen;
   }
 
   // truncate file
-  SWalCkHead* lastEntry = (SWalCkHead*)found;
-  int64_t     retVer = lastEntry->head.version;
-  int64_t     lastEntryBeginOffset = offset + (int64_t)((char*)found - (char*)buf);
-  int64_t     lastEntryEndOffset = lastEntryBeginOffset + sizeof(SWalCkHead) + lastEntry->head.bodyLen;
   if (lastEntryEndOffset != fileSize) {
-    wWarn("vgId:%d repair meta truncate file %s to %ld, orig size %ld", pWal->cfg.vgId, fnameStr, lastEntryEndOffset,
+    wWarn("vgId:%d, repair meta truncate file %s to %ld, orig size %ld", pWal->cfg.vgId, fnameStr, lastEntryEndOffset,
           fileSize);
-    taosFtruncateFile(pFile, lastEntryEndOffset);
+    if (taosFtruncateFile(pFile, lastEntryEndOffset) < 0) {
-    ((SWalFileInfo*)taosArrayGetLast(pWal->fileInfoSet))->fileSize = lastEntryEndOffset;
+      wError("failed to truncate file due to %s. file:%s", strerror(errno), fnameStr);
-    pWal->totSize -= (fileSize - lastEntryEndOffset);
+      terrno = TAOS_SYSTEM_ERROR(errno);
+      goto _err;
+    }
+    if (taosFsyncFile(pFile) < 0) {
+      wError("failed to fsync file due to %s. file:%s", strerror(errno), fnameStr);
+      terrno = TAOS_SYSTEM_ERROR(errno);
+      goto _err;
+    }
   }
+  pFileInfo->fileSize = lastEntryEndOffset;
 
   taosCloseFile(&pFile);
   taosMemoryFree(buf);
-
   return retVer;
+
+_err:
+  taosCloseFile(&pFile);
+  taosMemoryFree(buf);
+  return -1;
+}
+
+static void walRebuildFileInfoSet(SArray* metaLogList, SArray* actualLogList) {
+  int metaFileNum = taosArrayGetSize(metaLogList);
+  int actualFileNum = taosArrayGetSize(actualLogList);
+  int j = 0;
+
+  // both of the lists in asc order
+  for (int i = 0; i < actualFileNum; i++) {
+    SWalFileInfo* pLogInfo = taosArrayGet(actualLogList, i);
+    while (j < metaFileNum) {
+      SWalFileInfo* pMetaInfo = taosArrayGet(metaLogList, j);
+      ASSERT(pMetaInfo != NULL);
+      if (pMetaInfo->firstVer < pLogInfo->firstVer) {
+        j++;
+      } else if (pMetaInfo->firstVer == pLogInfo->firstVer) {
+        (*pLogInfo) = *pMetaInfo;
+        j++;
+        break;
+      } else {
+        break;
+      }
+    }
+  }
+
+  taosArrayClear(metaLogList);
+
+  for (int i = 0; i < actualFileNum; i++) {
+    SWalFileInfo* pFileInfo = taosArrayGet(actualLogList, i);
+    taosArrayPush(metaLogList, pFileInfo);
+  }
+}
+
+void walAlignVersions(SWal* pWal) {
+  if (pWal->vers.firstVer > pWal->vers.snapshotVer + 1) {
+    wWarn("vgId:%d, firstVer:%" PRId64 " is larger than snapshotVer:%" PRId64 " + 1. align with it.", pWal->cfg.vgId,
+          pWal->vers.firstVer, pWal->vers.snapshotVer);
+    pWal->vers.firstVer = pWal->vers.snapshotVer + 1;
+  }
+  if (pWal->vers.lastVer < pWal->vers.snapshotVer) {
+    wWarn("vgId:%d, lastVer:%" PRId64 " is less than snapshotVer:%" PRId64 ". align with it.", pWal->cfg.vgId,
+          pWal->vers.lastVer, pWal->vers.snapshotVer);
+    pWal->vers.lastVer = pWal->vers.snapshotVer;
+  }
+  if (pWal->vers.commitVer < pWal->vers.snapshotVer) {
+    wWarn("vgId:%d, commitVer:%" PRId64 " is less than snapshotVer:%" PRId64 ". align with it.", pWal->cfg.vgId,
+          pWal->vers.commitVer, pWal->vers.snapshotVer);
+    pWal->vers.commitVer = pWal->vers.snapshotVer;
+  }
+  if (pWal->vers.appliedVer < pWal->vers.snapshotVer) {
+    wWarn("vgId:%d, appliedVer:%" PRId64 " is less than snapshotVer:%" PRId64 ". align with it.", pWal->cfg.vgId,
+          pWal->vers.appliedVer, pWal->vers.snapshotVer);
+    pWal->vers.appliedVer = pWal->vers.snapshotVer;
+  }
+
+  pWal->vers.commitVer = TMIN(pWal->vers.lastVer, pWal->vers.commitVer);
+  pWal->vers.appliedVer = TMIN(pWal->vers.commitVer, pWal->vers.appliedVer);
 }
 
 int walCheckAndRepairMeta(SWal* pWal) {
@@ -150,7 +288,6 @@ int walCheckAndRepairMeta(SWal* pWal) {
   const char* idxPattern = "^[0-9]+.idx$";
   regex_t     logRegPattern;
   regex_t     idxRegPattern;
-  bool        fixed = false;
 
   regcomp(&logRegPattern, logPattern, REG_EXTENDED);
   regcomp(&idxRegPattern, idxPattern, REG_EXTENDED);
@@ -184,225 +321,237 @@ int walCheckAndRepairMeta(SWal* pWal) {
 
   taosArraySort(actualLog, compareWalFileInfo);
 
-  int metaFileNum = taosArrayGetSize(pWal->fileInfoSet);
+  int     metaFileNum = taosArrayGetSize(pWal->fileInfoSet);
-  int actualFileNum = taosArrayGetSize(actualLog);
+  int     actualFileNum = taosArrayGetSize(actualLog);
+  int64_t firstVerPrev = pWal->vers.firstVer;
+  int64_t lastVerPrev = pWal->vers.lastVer;
+  int64_t totSize = 0;
+  bool    updateMeta = (metaFileNum != actualFileNum);
 
-#if 0
+  // rebuild meta of file info
-  for (int32_t fileNo = actualFileNum - 1; fileNo >= 0; fileNo--) {
+  walRebuildFileInfoSet(pWal->fileInfoSet, actualLog);
-    SWalFileInfo* pFileInfo = taosArrayGet(pLogInfoArray, fileNo);
+  taosArrayDestroy(actualLog);
+
+  int32_t sz = taosArrayGetSize(pWal->fileInfoSet);
+  ASSERT(sz == actualFileNum);
+
+  // scan and determine the lastVer
+  int32_t fileIdx = sz;
+
+  while (--fileIdx >= 0) {
     char          fnameStr[WAL_FILE_LEN];
+    int64_t       fileSize = 0;
+    SWalFileInfo* pFileInfo = taosArrayGet(pWal->fileInfoSet, fileIdx);
+
     walBuildLogName(pWal, pFileInfo->firstVer, fnameStr);
-    int64_t fileSize = 0;
+    int32_t code = taosStatFile(fnameStr, &fileSize, NULL);
-    taosStatFile(fnameStr, &fileSize, NULL);
+    if (code < 0) {
-    if (fileSize == 0) {
+      terrno = TAOS_SYSTEM_ERROR(errno);
+      wError("failed to stat file since %s. file:%s", terrstr(), fnameStr);
+      return -1;
+    }
+
+    ASSERT(pFileInfo->firstVer >= 0);
+
+    if (pFileInfo->lastVer >= pFileInfo->firstVer && fileSize == pFileInfo->fileSize) {
+      totSize += pFileInfo->fileSize;
+      continue;
+    }
+    updateMeta = true;
+
+    int64_t lastVer = walScanLogGetLastVer(pWal, fileIdx);
+    if (lastVer < 0) {
+      if (terrno != TSDB_CODE_WAL_LOG_NOT_EXIST) {
+        wError("failed to scan wal last ver since %s", terrstr());
+        return -1;
+      }
+      ASSERT(pFileInfo->fileSize == 0);
+      // remove the empty wal log, and its idx
       taosRemoveFile(fnameStr);
       walBuildIdxName(pWal, pFileInfo->firstVer, fnameStr);
       taosRemoveFile(fnameStr);
-      taosArrayPop(pLogInfoArray);
+      // remove its meta entry
-    } else {
+      taosArrayRemove(pWal->fileInfoSet, fileIdx);
-      break;
+      continue;
     }
+
+    // update lastVer
+    pFileInfo->lastVer = lastVer;
+    totSize += pFileInfo->fileSize;
   }
 
-  actualFileNum = taosArrayGetSize(pLogInfoArray);
+  // reset vers info and so on
-#endif
-
-  {
-    int32_t i = 0, j = 0;
-    while (i < actualFileNum && j < metaFileNum) {
-      SWalFileInfo* pActualFile = taosArrayGet(actualLog, i);
-      SWalFileInfo* pMetaFile = taosArrayGet(pWal->fileInfoSet, j);
-      if (pActualFile->firstVer < pMetaFile->firstVer) {
-        char fNameStr[WAL_FILE_LEN];
-        walBuildLogName(pWal, pActualFile->firstVer, fNameStr);
-        taosRemoveFile(fNameStr);
-        walBuildIdxName(pWal, pActualFile->firstVer, fNameStr);
-        taosRemoveFile(fNameStr);
-        i++;
-      } else if (pActualFile->firstVer > pMetaFile->firstVer) {
-        taosArrayRemove(pWal->fileInfoSet, j);
-        metaFileNum--;
-      } else {
-        i++;
-        j++;
-      }
-    }
-    if (i == actualFileNum && j == metaFileNum) {
-      if (j > 0) {
-        SWalFileInfo* pLastInfo = taosArrayGet(pWal->fileInfoSet, j - 1);
-        int64_t       fsize = 0;
-        char          fNameStr[WAL_FILE_LEN];
-        walBuildLogName(pWal, pLastInfo->firstVer, fNameStr);
-        taosStatFile(fNameStr, &fsize, NULL);
-        if (pLastInfo->fileSize != fsize) {
-          fixed = true;
-          pLastInfo->fileSize = fsize;
-          pLastInfo->lastVer = walScanLogGetLastVer(pWal);
-        }
-      }
-    } else {
-      fixed = true;
-      while (i < actualFileNum) {
-        SWalFileInfo* pActualFile = taosArrayGet(actualLog, i);
-        char          fNameStr[WAL_FILE_LEN];
-        walBuildLogName(pWal, pActualFile->firstVer, fNameStr);
-        taosStatFile(fNameStr, &pActualFile->fileSize, NULL);
-
-        if (pActualFile->fileSize == 0) {
-          ASSERT(i == actualFileNum - 1);
-          taosRemoveFile(fNameStr);
-
-          walBuildIdxName(pWal, pActualFile->firstVer, fNameStr);
-          taosRemoveFile(fNameStr);
-          break;
-        }
-
-        if (i < actualFileNum - 1) {
-          pActualFile->lastVer = ((SWalFileInfo*)taosArrayGet(actualLog, i + 1))->firstVer - 1;
-          taosArrayPush(pWal->fileInfoSet, pActualFile);
-          i++;
-        } else {
-          pActualFile = taosArrayPush(pWal->fileInfoSet, pActualFile);
-          pActualFile->lastVer = walScanLogGetLastVer(pWal);
-          if (pActualFile->lastVer == -1) {
-            taosRemoveFile(fNameStr);
-
-            walBuildIdxName(pWal, pActualFile->firstVer, fNameStr);
-            taosRemoveFile(fNameStr);
-            taosArrayPop(pWal->fileInfoSet);
-          }
-          break;
-        }
-      }
-    }
-  }
-
-#if 0
-  if (metaFileNum > actualFileNum) {
-    taosArrayPopFrontBatch(pWal->fileInfoSet, metaFileNum - actualFileNum);
-  } else if (metaFileNum < actualFileNum) {
-    for (int i = metaFileNum; i < actualFileNum; i++) {
-      SWalFileInfo* pFileInfo = taosArrayGet(actualLog, i);
-      taosArrayPush(pWal->fileInfoSet, pFileInfo);
-    }
-  }
-#endif
-
-  taosArrayDestroy(actualLog);
-
   actualFileNum = taosArrayGetSize(pWal->fileInfoSet);
   pWal->writeCur = actualFileNum - 1;
-
+  pWal->totSize = totSize;
+  pWal->vers.lastVer = -1;
   if (actualFileNum > 0) {
-    int64_t fLastVer = ((SWalFileInfo*)taosArrayGet(pWal->fileInfoSet, pWal->writeCur))->lastVer;
+    pWal->vers.firstVer = ((SWalFileInfo*)taosArrayGet(pWal->fileInfoSet, 0))->firstVer;
-    if (fLastVer != -1 && pWal->vers.lastVer != fLastVer) {
+    pWal->vers.lastVer = ((SWalFileInfo*)taosArrayGetLast(pWal->fileInfoSet))->lastVer;
-      fixed = true;
+  }
-      pWal->vers.lastVer = fLastVer;
+  (void)walAlignVersions(pWal);
-    }
+
-    int64_t fFirstVer = ((SWalFileInfo*)taosArrayGet(pWal->fileInfoSet, 0))->firstVer;
+  // update meta file
-    if (fFirstVer != pWal->vers.firstVer) {
+  if (updateMeta) {
-      fixed = true;
+    (void)walSaveMeta(pWal);
-      pWal->vers.firstVer = fFirstVer;
+  }
-    }
+  return 0;
+}
+
+int walReadLogHead(TdFilePtr pLogFile, int64_t offset, SWalCkHead* pCkHead) {
+  if (taosLSeekFile(pLogFile, offset, SEEK_SET) < 0) {
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    return -1;
   }
 
-  if (fixed) {
+  if (taosReadFile(pLogFile, pCkHead, sizeof(SWalCkHead)) != sizeof(SWalCkHead)) {
-    walSaveMeta(pWal);
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    return -1;
+  }
+
+  if (walValidHeadCksum(pCkHead) != 0) {
+    terrno = TSDB_CODE_WAL_CHKSUM_MISMATCH;
+    return -1;
   }
 
   return 0;
 }
 
-int walCheckAndRepairIdx(SWal* pWal) {
+int walCheckAndRepairIdxFile(SWal* pWal, int32_t fileIdx) {
   int32_t sz = taosArrayGetSize(pWal->fileInfoSet);
-  for (int32_t i = 0; i < sz; i++) {
+  ASSERT(fileIdx >= 0 && fileIdx < sz);
-    SWalFileInfo* pFileInfo = taosArrayGet(pWal->fileInfoSet, i);
+  SWalFileInfo* pFileInfo = taosArrayGet(pWal->fileInfoSet, fileIdx);
+  char          fnameStr[WAL_FILE_LEN];
+  walBuildIdxName(pWal, pFileInfo->firstVer, fnameStr);
+  char fLogNameStr[WAL_FILE_LEN];
+  walBuildLogName(pWal, pFileInfo->firstVer, fLogNameStr);
+  int64_t fileSize = 0;
 
-    char fnameStr[WAL_FILE_LEN];
+  if (taosStatFile(fnameStr, &fileSize, NULL) < 0 && errno != ENOENT) {
-    walBuildIdxName(pWal, pFileInfo->firstVer, fnameStr);
+    wError("vgId:%d, failed to stat file due to %s. file:%s", pWal->cfg.vgId, strerror(errno), fnameStr);
-    int64_t   fsize;
+    terrno = TAOS_SYSTEM_ERROR(errno);
-    TdFilePtr pIdxFile = taosOpenFile(fnameStr, TD_FILE_READ | TD_FILE_WRITE | TD_FILE_CREATE);
+    return -1;
-    if (pIdxFile == NULL) {
+  }
-      ASSERT(0);
+
+  ASSERT(pFileInfo->fileSize > 0 && pFileInfo->firstVer >= 0 && pFileInfo->lastVer >= pFileInfo->firstVer);
+  if (fileSize == (pFileInfo->lastVer - pFileInfo->firstVer + 1) * sizeof(SWalIdxEntry)) {
+    return 0;
+  }
+
+  // start to repair
+  int64_t      offset = fileSize - fileSize % sizeof(SWalIdxEntry);
+  TdFilePtr    pLogFile = NULL;
+  TdFilePtr    pIdxFile = NULL;
+  SWalIdxEntry idxEntry = {.ver = pFileInfo->firstVer - 1, .offset = -sizeof(SWalCkHead)};
+  SWalCkHead   ckHead;
+  memset(&ckHead, 0, sizeof(ckHead));
+  ckHead.head.version = idxEntry.ver;
+
+  pIdxFile = taosOpenFile(fnameStr, TD_FILE_READ | TD_FILE_WRITE | TD_FILE_CREATE);
+  if (pIdxFile == NULL) {
+    wError("vgId:%d, failed to open file due to %s. file:%s", pWal->cfg.vgId, strerror(errno), fnameStr);
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    goto _err;
+  }
+
+  pLogFile = taosOpenFile(fLogNameStr, TD_FILE_READ);
+  if (pLogFile == NULL) {
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    wError("vgId:%d, cannot open file %s, since %s", pWal->cfg.vgId, fLogNameStr, terrstr());
+    goto _err;
+  }
+
+  // determine the last valid entry end, i.e. offset
+  while ((offset -= sizeof(SWalIdxEntry)) >= 0) {
+    if (taosLSeekFile(pIdxFile, offset, SEEK_SET) < 0) {
+      wError("vgId:%d, failed to seek file due to %s. offset:" PRId64 ", file:%s", pWal->cfg.vgId, strerror(errno),
+             offset, fnameStr);
       terrno = TAOS_SYSTEM_ERROR(errno);
-      wError("vgId:%d, cannot open file %s, since %s", pWal->cfg.vgId, fnameStr, terrstr());
+      goto _err;
-      return -1;
     }
 
-    taosFStatFile(pIdxFile, &fsize, NULL);
+    if (taosReadFile(pIdxFile, &idxEntry, sizeof(SWalIdxEntry)) != sizeof(SWalIdxEntry)) {
-    if (fsize == (pFileInfo->lastVer - pFileInfo->firstVer + 1) * sizeof(SWalIdxEntry)) {
+      wError("vgId:%d, failed to read file due to %s. offset:%" PRId64 ", file:%s", pWal->cfg.vgId, strerror(errno),
-      taosCloseFile(&pIdxFile);
+             offset, fnameStr);
+      terrno = TAOS_SYSTEM_ERROR(errno);
+      goto _err;
+    }
+
+    if (idxEntry.ver > pFileInfo->lastVer) {
       continue;
     }
 
-    int32_t left = fsize % sizeof(SWalIdxEntry);
+    if (walReadLogHead(pLogFile, idxEntry.offset, &ckHead) < 0) {
-    int64_t offset = taosLSeekFile(pIdxFile, -left, SEEK_END);
+      wWarn("vgId:%d, failed to read log file since %s. file:%s, offset:%" PRId64 ", idx entry ver:%" PRId64 "",
-    if (left != 0) {
+            pWal->cfg.vgId, terrstr(), fLogNameStr, idxEntry.offset, idxEntry.ver);
-      taosFtruncateFile(pIdxFile, offset);
+      continue;
-      wWarn("vgId:%d wal truncate file %s to offset %ld since size invalid, file size %ld", pWal->cfg.vgId, fnameStr,
-            offset, fsize);
-    }
-    offset -= sizeof(SWalIdxEntry);
-
-    SWalIdxEntry idxEntry = {.ver = pFileInfo->firstVer};
-    while (1) {
-      if (offset < 0) {
-        taosLSeekFile(pIdxFile, 0, SEEK_SET);
-        taosWriteFile(pIdxFile, &idxEntry, sizeof(SWalIdxEntry));
-        break;
-      }
-      if (taosLSeekFile(pIdxFile, offset, SEEK_SET) < 0) {
-        terrno = TAOS_SYSTEM_ERROR(errno);
-        wError("vgId:%d cannot seek offset %ld when repair idx since %s", pWal->cfg.vgId, offset, terrstr());
-      }
-      int64_t contLen = taosReadFile(pIdxFile, &idxEntry, sizeof(SWalIdxEntry));
-      if (contLen < 0 || contLen != sizeof(SWalIdxEntry)) {
-        terrno = TAOS_SYSTEM_ERROR(errno);
-        return -1;
-      }
-      if ((idxEntry.ver - pFileInfo->firstVer) * sizeof(SWalIdxEntry) != offset) {
-        taosFtruncateFile(pIdxFile, offset);
-        wWarn("vgId:%d wal truncate file %s to offset %ld since entry invalid, entry ver %ld, entry offset %ld",
-              pWal->cfg.vgId, fnameStr, offset, idxEntry.ver, idxEntry.offset);
-        offset -= sizeof(SWalIdxEntry);
-      } else {
-        break;
-      }
     }
 
-    if (idxEntry.ver < pFileInfo->lastVer) {
+    if (idxEntry.ver == ckHead.head.version) {
-      char fLogNameStr[WAL_FILE_LEN];
+      break;
-      walBuildLogName(pWal, pFileInfo->firstVer, fLogNameStr);
+    }
-      TdFilePtr pLogFile = taosOpenFile(fLogNameStr, TD_FILE_READ);
+  }
-      if (pLogFile == NULL) {
+  offset += sizeof(SWalIdxEntry);
-        terrno = TAOS_SYSTEM_ERROR(errno);
+
-        wError("vgId:%d, cannot open file %s, since %s", pWal->cfg.vgId, fLogNameStr, terrstr());
+  // ftruncate idx file
-        return -1;
+  if (offset < fileSize) {
-      }
+    if (taosFtruncateFile(pIdxFile, offset) < 0) {
-      while (idxEntry.ver < pFileInfo->lastVer) {
+      wError("vgId:%d, failed to ftruncate file due to %s. offset:%" PRId64 ", file:%s", pWal->cfg.vgId,
-        if (taosLSeekFile(pLogFile, idxEntry.offset, SEEK_SET) == -1) {
+             strerror(errno), offset, fnameStr);
-          terrno = TAOS_SYSTEM_ERROR(errno);
+      terrno = TAOS_SYSTEM_ERROR(errno);
-          wError("vgId:%d, cannot seek file %s at %ld, since %s", pWal->cfg.vgId, fLogNameStr, idxEntry.offset,
+      goto _err;
-                 terrstr());
+    }
-          return -1;
+  }
-        }
+
-        SWalCkHead ckHead;
+  // rebuild idx file
-        taosReadFile(pLogFile, &ckHead, sizeof(SWalCkHead));
+  if (taosLSeekFile(pIdxFile, 0, SEEK_END) < 0) {
-        if (idxEntry.ver != ckHead.head.version) {
+    wError("vgId:%d, failed to seek file due to %s. offset:" PRId64 ", file:%s", pWal->cfg.vgId, strerror(errno),
-          // todo truncate this idx also
+           offset, fnameStr);
-          taosCloseFile(&pLogFile);
+    terrno = TAOS_SYSTEM_ERROR(errno);
-          wError("vgId:%d, invalid repair case, log seek to %ld to find ver %ld, actual ver %ld", pWal->cfg.vgId,
+    goto _err;
-                 idxEntry.offset, idxEntry.ver, ckHead.head.version);
+  }
-          return -1;
+
-        }
+  while (idxEntry.ver < pFileInfo->lastVer) {
-        idxEntry.ver = ckHead.head.version + 1;
+    ASSERT(idxEntry.ver == ckHead.head.version);
-        idxEntry.offset = idxEntry.offset + sizeof(SWalCkHead) + ckHead.head.bodyLen;
+
-        wWarn("vgId:%d wal idx append new entry %ld %ld", pWal->cfg.vgId, idxEntry.ver, idxEntry.offset);
+    idxEntry.ver += 1;
-        taosWriteFile(pIdxFile, &idxEntry, sizeof(SWalIdxEntry));
+    idxEntry.offset += sizeof(SWalCkHead) + ckHead.head.bodyLen;
-      }
+
-      taosCloseFile(&pLogFile);
+    if (walReadLogHead(pLogFile, idxEntry.offset, &ckHead) < 0) {
+      wError("vgId:%d, failed to read wal log head since %s. offset:%" PRId64 ", file:%s", pWal->cfg.vgId, terrstr(),
+             idxEntry.offset, fLogNameStr);
+      goto _err;
+    }
+    wWarn("vgId:%d wal idx append new entry %ld %ld", pWal->cfg.vgId, idxEntry.ver, idxEntry.offset);
+    if (taosWriteFile(pIdxFile, &idxEntry, sizeof(SWalIdxEntry)) < 0) {
+      wError("vgId:%d, failed to append file since %s. file:%s", pWal->cfg.vgId, terrstr(), fnameStr);
+      goto _err;
+    }
+  }
+
+  if (taosFsyncFile(pIdxFile) < 0) {
+    wError("vgId:%d, faild to fsync file since %s. file:%s", pWal->cfg.vgId, terrstr(), fnameStr);
+    goto _err;
+  }
+
+  (void)taosCloseFile(&pLogFile);
+  (void)taosCloseFile(&pIdxFile);
+  return 0;
+
+_err:
+  (void)taosCloseFile(&pLogFile);
+  (void)taosCloseFile(&pIdxFile);
+  return -1;
+}
+
+int walCheckAndRepairIdx(SWal* pWal) {
+  int32_t sz = taosArrayGetSize(pWal->fileInfoSet);
+  int32_t fileIdx = sz;
+  while (--fileIdx >= 0) {
+    if (walCheckAndRepairIdxFile(pWal, fileIdx) < 0) {
+      wError("vgId:%d, failed to repair idx file since %s. fileIdx:%d", pWal->cfg.vgId, terrstr(), fileIdx);
+      return -1;
     }
-    taosCloseFile(&pIdxFile);
   }
   return 0;
 }
@@ -495,14 +644,20 @@ int walMetaDeserialize(SWal* pWal, const char* bytes) {
   ASSERT(taosArrayGetSize(pWal->fileInfoSet) == 0);
   cJSON *pRoot, *pMeta, *pFiles, *pInfoJson, *pField;
   pRoot = cJSON_Parse(bytes);
+  if (!pRoot) goto _err;
   pMeta = cJSON_GetObjectItem(pRoot, "meta");
+  if (!pMeta) goto _err;
   pField = cJSON_GetObjectItem(pMeta, "firstVer");
+  if (!pField) goto _err;
   pWal->vers.firstVer = atoll(cJSON_GetStringValue(pField));
   pField = cJSON_GetObjectItem(pMeta, "snapshotVer");
+  if (!pField) goto _err;
   pWal->vers.snapshotVer = atoll(cJSON_GetStringValue(pField));
   pField = cJSON_GetObjectItem(pMeta, "commitVer");
+  if (!pField) goto _err;
   pWal->vers.commitVer = atoll(cJSON_GetStringValue(pField));
   pField = cJSON_GetObjectItem(pMeta, "lastVer");
+  if (!pField) goto _err;
   pWal->vers.lastVer = atoll(cJSON_GetStringValue(pField));
 
   pFiles = cJSON_GetObjectItem(pRoot, "files");
@@ -512,17 +667,23 @@ int walMetaDeserialize(SWal* pWal, const char* bytes) {
   taosArrayEnsureCap(pArray, sz);
   SWalFileInfo* pData = pArray->pData;
   for (int i = 0; i < sz; i++) {
-    cJSON*        pInfoJson = cJSON_GetArrayItem(pFiles, i);
+    cJSON* pInfoJson = cJSON_GetArrayItem(pFiles, i);
+    if (!pInfoJson) goto _err;
     SWalFileInfo* pInfo = &pData[i];
     pField = cJSON_GetObjectItem(pInfoJson, "firstVer");
+    if (!pField) goto _err;
     pInfo->firstVer = atoll(cJSON_GetStringValue(pField));
     pField = cJSON_GetObjectItem(pInfoJson, "lastVer");
+    if (!pField) goto _err;
     pInfo->lastVer = atoll(cJSON_GetStringValue(pField));
     pField = cJSON_GetObjectItem(pInfoJson, "createTs");
+    if (!pField) goto _err;
     pInfo->createTs = atoll(cJSON_GetStringValue(pField));
     pField = cJSON_GetObjectItem(pInfoJson, "closeTs");
+    if (!pField) goto _err;
     pInfo->closeTs = atoll(cJSON_GetStringValue(pField));
     pField = cJSON_GetObjectItem(pInfoJson, "fileSize");
+    if (!pField) goto _err;
     pInfo->fileSize = atoll(cJSON_GetStringValue(pField));
   }
   taosArraySetSize(pArray, sz);
@@ -530,6 +691,10 @@ int walMetaDeserialize(SWal* pWal, const char* bytes) {
   pWal->writeCur = sz - 1;
   cJSON_Delete(pRoot);
   return 0;
+
+_err:
+  cJSON_Delete(pRoot);
+  return -1;
 }
 
 static int walFindCurMetaVer(SWal* pWal) {
@@ -565,22 +730,63 @@ static int walFindCurMetaVer(SWal* pWal) {
 int walSaveMeta(SWal* pWal) {
   int  metaVer = walFindCurMetaVer(pWal);
   char fnameStr[WAL_FILE_LEN];
-  walBuildMetaName(pWal, metaVer + 1, fnameStr);
+  char tmpFnameStr[WAL_FILE_LEN];
-  TdFilePtr pMetaFile = taosOpenFile(fnameStr, TD_FILE_CREATE | TD_FILE_WRITE);
+  int  n;
-  if (pMetaFile == NULL) {
+
+  // fsync the idx and log file at first to ensure validity of meta
+  if (taosFsyncFile(pWal->pIdxFile) < 0) {
+    wError("vgId:%d, failed to sync idx file due to %s", pWal->cfg.vgId, strerror(errno));
+    terrno = TAOS_SYSTEM_ERROR(errno);
     return -1;
   }
+
+  if (taosFsyncFile(pWal->pLogFile) < 0) {
+    wError("vgId:%d, failed to sync log file due to %s", pWal->cfg.vgId, strerror(errno));
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    return -1;
+  }
+
+  // flush to a tmpfile
+  n = walBuildTmpMetaName(pWal, tmpFnameStr);
+  ASSERT(n < sizeof(tmpFnameStr) && "Buffer overflow of file name");
+
+  TdFilePtr pMetaFile = taosOpenFile(tmpFnameStr, TD_FILE_CREATE | TD_FILE_WRITE | TD_FILE_TRUNC);
+  if (pMetaFile == NULL) {
+    wError("vgId:%d, failed to open file due to %s. file:%s", pWal->cfg.vgId, strerror(errno), tmpFnameStr);
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    return -1;
+  }
+
   char* serialized = walMetaSerialize(pWal);
   int   len = strlen(serialized);
   if (len != taosWriteFile(pMetaFile, serialized, len)) {
-    // TODO:clean file
+    wError("vgId:%d, failed to write file due to %s. file:%s", pWal->cfg.vgId, strerror(errno), tmpFnameStr);
-
+    terrno = TAOS_SYSTEM_ERROR(errno);
-    taosCloseFile(&pMetaFile);
+    goto _err;
-    taosRemoveFile(fnameStr);
+  }
-    return -1;
+
+  if (taosFsyncFile(pMetaFile) < 0) {
+    wError("vgId:%d, failed to sync file due to %s. file:%s", pWal->cfg.vgId, strerror(errno), tmpFnameStr);
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    goto _err;
+  }
+
+  if (taosCloseFile(&pMetaFile) < 0) {
+    wError("vgId:%d, failed to close file due to %s. file:%s", pWal->cfg.vgId, strerror(errno), tmpFnameStr);
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    goto _err;
+  }
+
+  // rename it
+  n = walBuildMetaName(pWal, metaVer + 1, fnameStr);
+  ASSERT(n < sizeof(fnameStr) && "Buffer overflow of file name");
+
+  if (taosRenameFile(tmpFnameStr, fnameStr) < 0) {
+    wError("failed to rename file due to %s. dest:%s", strerror(errno), fnameStr);
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    goto _err;
   }
 
-  taosCloseFile(&pMetaFile);
   // delete old file
   if (metaVer > -1) {
     walBuildMetaName(pWal, metaVer, fnameStr);
@@ -588,6 +794,11 @@ int walSaveMeta(SWal* pWal) {
   }
   taosMemoryFree(serialized);
   return 0;
+
+_err:
+  taosCloseFile(&pMetaFile);
+  taosMemoryFree(serialized);
+  return -1;
 }
 
 int walLoadMeta(SWal* pWal) {
@@ -629,6 +840,10 @@ int walLoadMeta(SWal* pWal) {
   }
   // load into fileInfoSet
   int code = walMetaDeserialize(pWal, buf);
+  if (code < 0) {
+    wError("failed to deserialize wal meta. file:%s", fnameStr);
+    terrno = TSDB_CODE_WAL_FILE_CORRUPTED;
+  }
   taosCloseFile(&pFile);
   taosMemoryFree(buf);
   return code;

source/libs/wal/src/walMgmt.c

@@ -81,6 +81,12 @@ SWal *walOpen(const char *path, SWalCfg *pCfg) {
     return NULL;
   }
 
+  if (taosThreadMutexInit(&pWal->mutex, NULL) < 0) {
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    taosMemoryFree(pWal);
+    return NULL;
+  }
+
   // set config
   memcpy(&pWal->cfg, pCfg, sizeof(SWalCfg));
 
@@ -98,15 +104,14 @@ SWal *walOpen(const char *path, SWalCfg *pCfg) {
   tstrncpy(pWal->path, path, sizeof(pWal->path));
   if (taosMkDir(pWal->path) != 0) {
     wError("vgId:%d, path:%s, failed to create directory since %s", pWal->cfg.vgId, pWal->path, strerror(errno));
-    taosMemoryFree(pWal);
+    goto _err;
-    return NULL;
   }
 
   // init ref
   pWal->pRefHash = taosHashInit(64, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), true, HASH_ENTRY_LOCK);
   if (pWal->pRefHash == NULL) {
-    taosMemoryFree(pWal);
+    wError("failed to init hash since %s", tstrerror(terrno));
-    return NULL;
+    goto _err;
   }
 
   // open meta
@@ -117,9 +122,7 @@ SWal *walOpen(const char *path, SWalCfg *pCfg) {
   pWal->fileInfoSet = taosArrayInit(8, sizeof(SWalFileInfo));
   if (pWal->fileInfoSet == NULL) {
     wError("vgId:%d, path:%s, failed to init taosArray %s", pWal->cfg.vgId, pWal->path, strerror(errno));
-    taosHashCleanup(pWal->pRefHash);
+    goto _err;
-    taosMemoryFree(pWal);
-    return NULL;
   }
 
   // init status
@@ -131,46 +134,37 @@ SWal *walOpen(const char *path, SWalCfg *pCfg) {
   pWal->writeHead.head.protoVer = WAL_PROTO_VER;
   pWal->writeHead.magic = WAL_MAGIC;
 
-  if (taosThreadMutexInit(&pWal->mutex, NULL) < 0) {
+  // load meta
-    taosArrayDestroy(pWal->fileInfoSet);
+  (void)walLoadMeta(pWal);
-    taosHashCleanup(pWal->pRefHash);
-    taosMemoryFree(pWal);
-    return NULL;
-  }
-
-  pWal->refId = taosAddRef(tsWal.refSetId, pWal);
-  if (pWal->refId < 0) {
-    taosHashCleanup(pWal->pRefHash);
-    taosThreadMutexDestroy(&pWal->mutex);
-    taosArrayDestroy(pWal->fileInfoSet);
-    taosMemoryFree(pWal);
-    return NULL;
-  }
-
-  walLoadMeta(pWal);
 
   if (walCheckAndRepairMeta(pWal) < 0) {
     wError("vgId:%d cannot open wal since repair meta file failed", pWal->cfg.vgId);
-    taosHashCleanup(pWal->pRefHash);
+    goto _err;
-    taosRemoveRef(tsWal.refSetId, pWal->refId);
-    taosThreadMutexDestroy(&pWal->mutex);
-    taosArrayDestroy(pWal->fileInfoSet);
-    return NULL;
   }
 
   if (walCheckAndRepairIdx(pWal) < 0) {
     wError("vgId:%d cannot open wal since repair idx file failed", pWal->cfg.vgId);
-    taosHashCleanup(pWal->pRefHash);
+    goto _err;
-    taosRemoveRef(tsWal.refSetId, pWal->refId);
+  }
-    taosThreadMutexDestroy(&pWal->mutex);
+
-    taosArrayDestroy(pWal->fileInfoSet);
+  // add ref
-    return NULL;
+  pWal->refId = taosAddRef(tsWal.refSetId, pWal);
+  if (pWal->refId < 0) {
+    wError("failed to add ref for Wal since %s", tstrerror(terrno));
+    goto _err;
   }
 
   wDebug("vgId:%d, wal:%p is opened, level:%d fsyncPeriod:%d", pWal->cfg.vgId, pWal, pWal->cfg.level,
          pWal->cfg.fsyncPeriod);
-
   return pWal;
+
+_err:
+  taosArrayDestroy(pWal->fileInfoSet);
+  taosHashCleanup(pWal->pRefHash);
+  taosThreadMutexDestroy(&pWal->mutex);
+  taosMemoryFree(pWal);
+  pWal = NULL;
+  return NULL;
 }
 
 int32_t walAlter(SWal *pWal, SWalCfg *pCfg) {
@@ -195,11 +189,11 @@ int32_t walAlter(SWal *pWal, SWalCfg *pCfg) {
 
 void walClose(SWal *pWal) {
   taosThreadMutexLock(&pWal->mutex);
+  (void)walSaveMeta(pWal);
   taosCloseFile(&pWal->pLogFile);
   pWal->pLogFile = NULL;
   taosCloseFile(&pWal->pIdxFile);
   pWal->pIdxFile = NULL;
-  walSaveMeta(pWal);
   taosArrayDestroy(pWal->fileInfoSet);
   pWal->fileInfoSet = NULL;
   taosHashCleanup(pWal->pRefHash);

source/libs/wal/src/walRead.c

@@ -181,7 +181,11 @@ int32_t walReadSeekVerImpl(SWalReader *pReader, int64_t ver) {
   SWalFileInfo tmpInfo;
   tmpInfo.firstVer = ver;
   SWalFileInfo *pRet = taosArraySearch(pWal->fileInfoSet, &tmpInfo, compareWalFileInfo, TD_LE);
-  ASSERT(pRet != NULL);
+  if (pRet == NULL) {
+    wError("failed to find WAL log file with ver:%lld", ver);
+    terrno = TSDB_CODE_WAL_INVALID_VER;
+    return -1;
+  }
   if (pReader->curFileFirstVer != pRet->firstVer) {
     // error code was set inner
     if (walReadChangeFile(pReader, pRet->firstVer) < 0) {
@@ -472,7 +476,8 @@ int32_t walReadVer(SWalReader *pReader, int64_t ver) {
       } else {
         terrno = TSDB_CODE_WAL_FILE_CORRUPTED;
       }
-      ASSERT(0);
+      wError("vgId:%d, failed to read WAL record head, index:%" PRId64 ", from log file since %s",
+             pReader->pWal->cfg.vgId, ver, terrstr());
       taosThreadMutexUnlock(&pReader->mutex);
       return -1;
     }
@@ -505,6 +510,8 @@ int32_t walReadVer(SWalReader *pReader, int64_t ver) {
     else {
       terrno = TSDB_CODE_WAL_FILE_CORRUPTED;
     }
+    wError("vgId:%d, failed to read WAL record body, index:%" PRId64 ", from log file since %s",
+           pReader->pWal->cfg.vgId, ver, terrstr());
     taosThreadMutexUnlock(&pReader->mutex);
     return -1;
   }

source/libs/wal/src/walSeek.c

@@ -79,6 +79,11 @@ int64_t walChangeWrite(SWal* pWal, int64_t ver) {
   TdFilePtr pIdxTFile, pLogTFile;
   char      fnameStr[WAL_FILE_LEN];
   if (pWal->pLogFile != NULL) {
+    code = taosFsyncFile(pWal->pLogFile);
+    if (code != 0) {
+      terrno = TAOS_SYSTEM_ERROR(errno);
+      return -1;
+    }
     code = taosCloseFile(&pWal->pLogFile);
     if (code != 0) {
       terrno = TAOS_SYSTEM_ERROR(errno);
@@ -86,6 +91,11 @@ int64_t walChangeWrite(SWal* pWal, int64_t ver) {
     }
   }
   if (pWal->pIdxFile != NULL) {
+    code = taosFsyncFile(pWal->pIdxFile);
+    if (code != 0) {
+      terrno = TAOS_SYSTEM_ERROR(errno);
+      return -1;
+    }
     code = taosCloseFile(&pWal->pIdxFile);
     if (code != 0) {
       terrno = TAOS_SYSTEM_ERROR(errno);

source/libs/wal/src/walWrite.c

@@ -221,10 +221,12 @@ int32_t walRollback(SWal *pWal, int64_t ver) {
   taosCloseFile(&pIdxFile);
   taosCloseFile(&pLogFile);
 
-  taosFsyncFile(pWal->pLogFile);
+  code = walSaveMeta(pWal);
-  taosFsyncFile(pWal->pIdxFile);
+  if (code < 0) {
-
+    wError("vgId:%d, failed to save meta since %s", pWal->cfg.vgId, terrstr());
-  walSaveMeta(pWal);
+    taosThreadMutexUnlock(&pWal->mutex);
+    return -1;
+  }
 
   // unlock
   taosThreadMutexUnlock(&pWal->mutex);
@@ -396,7 +398,11 @@ int32_t walRollImpl(SWal *pWal) {
 
   pWal->lastRollSeq = walGetSeq();
 
-  walSaveMeta(pWal);
+  code = walSaveMeta(pWal);
+  if (code < 0) {
+    wError("vgId:%d, failed to save meta since %s", pWal->cfg.vgId, terrstr());
+    goto END;
+  }
 
 END:
   return code;
@@ -552,6 +558,11 @@ int32_t walWrite(SWal *pWal, int64_t index, tmsg_t msgType, const void *body, in
 
 void walFsync(SWal *pWal, bool forceFsync) {
   if (forceFsync || (pWal->cfg.level == TAOS_WAL_FSYNC && pWal->cfg.fsyncPeriod == 0)) {
+    wTrace("vgId:%d, fileId:%" PRId64 ".idx, do fsync", pWal->cfg.vgId, walGetCurFileFirstVer(pWal));
+    if (taosFsyncFile(pWal->pIdxFile) < 0) {
+      wError("vgId:%d, file:%" PRId64 ".idx, fsync failed since %s", pWal->cfg.vgId, walGetCurFileFirstVer(pWal),
+             strerror(errno));
+    }
     wTrace("vgId:%d, fileId:%" PRId64 ".log, do fsync", pWal->cfg.vgId, walGetCurFileFirstVer(pWal));
     if (taosFsyncFile(pWal->pLogFile) < 0) {
       wError("vgId:%d, file:%" PRId64 ".log, fsync failed since %s", pWal->cfg.vgId, walGetCurFileFirstVer(pWal),

source/os/src/osEnv.c

@@ -87,11 +87,17 @@ void osUpdate() {
 
 void osCleanup() {}
 
-bool osLogSpaceAvailable() { return tsLogSpace.reserved <= tsLogSpace.size.avail; }
+bool osLogSpaceAvailable() { return tsLogSpace.size.avail > 0; }
 
-bool osDataSpaceAvailable() { return tsDataSpace.reserved <= tsDataSpace.size.avail; }
+bool osDataSpaceAvailable() { return tsDataSpace.size.avail > 0; }
 
-bool osTempSpaceAvailable() { return tsTempSpace.reserved <= tsTempSpace.size.avail; }
+bool osTempSpaceAvailable() { return tsTempSpace.size.avail > 0; }
+
+bool osLogSpaceSufficient() { return tsLogSpace.size.avail > tsLogSpace.reserved; }
+
+bool osDataSpaceSufficient() { return tsDataSpace.size.avail > tsDataSpace.reserved; }
+
+bool osTempSpaceSufficient() { return tsTempSpace.size.avail > tsTempSpace.reserved; }
 
 void osSetTimezone(const char *timezone) { taosSetSystemTimezone(timezone, tsTimezoneStr, &tsDaylight, &tsTimezone); }
 

source/util/src/terror.c

@@ -447,12 +447,13 @@ TAOS_DEFINE_ERROR(TSDB_CODE_TQ_TABLE_SCHEMA_NOT_FOUND,    "TQ table schema not f
 TAOS_DEFINE_ERROR(TSDB_CODE_TQ_NO_COMMITTED_OFFSET,       "TQ no commited offset")
 
 // wal
-TAOS_DEFINE_ERROR(TSDB_CODE_WAL_APP_ERROR,                "Wal unexpected generic error")
+TAOS_DEFINE_ERROR(TSDB_CODE_WAL_APP_ERROR,                "WAL unexpected generic error")
 TAOS_DEFINE_ERROR(TSDB_CODE_WAL_FILE_CORRUPTED,           "WAL file is corrupted")
 TAOS_DEFINE_ERROR(TSDB_CODE_WAL_SIZE_LIMIT,               "WAL size exceeds limit")
 TAOS_DEFINE_ERROR(TSDB_CODE_WAL_INVALID_VER,              "WAL use invalid version")
 TAOS_DEFINE_ERROR(TSDB_CODE_WAL_OUT_OF_MEMORY,            "WAL out of memory")
 TAOS_DEFINE_ERROR(TSDB_CODE_WAL_LOG_NOT_EXIST,            "WAL log not exist")
+TAOS_DEFINE_ERROR(TSDB_CODE_WAL_CHKSUM_MISMATCH,          "WAL checksum mismatch")
 
 // tfs
 TAOS_DEFINE_ERROR(TSDB_CODE_FS_INVLD_CFG,                 "tfs invalid mount config")

tests/develop-test/5-taos-tools/taosbenchmark/default_json.py

@@ -32,6 +32,10 @@ class TDTestCase:
 
         if "community" in selfPath:
             projPath = selfPath[: selfPath.find("community")]
+        elif "src" in selfPath:
+            projPath = selfPath[: selfPath.find("src")]
+        elif "/tools/" in selfPath:
+            projPath = selfPath[: selfPath.find("/tools/")]
         else:
             projPath = selfPath[: selfPath.find("tests")]
 
@@ -55,11 +59,10 @@ class TDTestCase:
         tdLog.info("%s" % cmd)
         os.system("%s" % cmd)
         tdSql.execute("reset query cache")
-        tdSql.query("select count(*) from (select distinct(tbname) from db.stb)")
+        tdSql.query("show db.tables")
-        tdSql.checkData(0, 0, 10)
+        tdSql.checkRows(10)
         tdSql.query("select count(*) from db.stb")
-        if len(tdSql.queryResult):
+        tdSql.checkData(0, 0, 100)
-            tdLog.exit("query result is %d" % len(tdSql.queryResult))
 
     def stop(self):
         tdSql.close()

tests/develop-test/5-taos-tools/taosbenchmark/json/default.json

@@ -20,8 +20,10 @@
     "super_tables": [{
       "name": "stb",
       "childtable_prefix": "stb_",
+      "childtable_count": 10,
+      "insert_rows": 10,
       "columns": [{"type": "INT"}],
       "tags": [{"type": "INT"}]
     }]
   }]
-}
+}

tests/script/coverage_test.sh

@@ -0,0 +1,291 @@
+#!/bin/bash
+
+branch=
+if [ x$1 != x ];then
+        branch=$1
+        echo "Testing branch: $branch"
+else
+        echo "Please enter branch name as a parameter"
+        exit 1
+fi
+
+today=`date +"%Y%m%d"`
+TDENGINE_DIR=/root/pxiao/TDengine
+TDENGINE_COVERAGE_REPORT=$TDENGINE_DIR/tests/coverage-report-$today.log
+
+# Color setting
+RED='\033[0;31m'
+GREEN='\033[1;32m'
+GREEN_DARK='\033[0;32m'
+GREEN_UNDERLINE='\033[4;32m'
+NC='\033[0m'
+
+function buildTDengine {
+	echo "check if TDengine need build"
+	cd $TDENGINE_DIR
+  	git remote prune origin > /dev/null
+	git remote update > /dev/null
+	REMOTE_COMMIT=`git rev-parse --short remotes/origin/$branch`
+	LOCAL_COMMIT=`git rev-parse --short @`
+	echo " LOCAL: $LOCAL_COMMIT"
+	echo "REMOTE: $REMOTE_COMMIT"
+
+	# reset counter
+	lcov -d . --zerocounters
+
+
+	if [ "$LOCAL_COMMIT" == "$REMOTE_COMMIT" ]; then
+		echo "repo up-to-date"
+	else
+		echo "repo need to pull"
+	fi
+
+    	git reset --hard
+    	git checkout -- .
+    	git checkout $branch
+	git clean -dfx
+	git pull
+    	git submodule update --init --recursive -f
+ 	
+  	[ -d $TDENGINE_DIR/debug ] || mkdir $TDENGINE_DIR/debug
+  	cd $TDENGINE_DIR/debug
+
+    	echo "rebuild.."
+	LOCAL_COMMIT=`git rev-parse --short @`
+
+	rm -rf *
+	if [ "$branch" == "3.0" ]; then
+		echo "3.0 ============="
+		cmake -DCOVER=true -DBUILD_TEST=true .. 
+	else
+		cmake -DCOVER=true -DBUILD_TOOLS=true -DBUILD_HTTP=false .. > /dev/null
+	fi
+	make -j4
+	make install
+}
+
+function runGeneralCaseOneByOne {
+	while read -r line; do
+		if [[ $line =~ ^./test.sh* ]]; then
+			case=`echo $line | grep sim$ | awk '{print $NF}'`
+
+			if [ -n "$case" ]; then
+				date +%F\ %T | tee -a  $TDENGINE_COVERAGE_REPORT  && ./test.sh -f $case > /dev/null 2>&1 && \
+				 echo -e "${GREEN}$case success${NC}" | tee -a  $TDENGINE_COVERAGE_REPORT \
+				|| echo -e "${RED}$case failed${NC}" | tee -a  $TDENGINE_COVERAGE_REPORT
+			fi
+		fi
+	done < $1
+}
+
+function runTestNGCaseOneByOne {
+	while read -r line; do
+		if [[ $line =~ ^taostest* ]]; then
+			case=`echo $line | cut -d' ' -f 3 | cut -d'=' -f 2`
+			yaml=`echo $line | cut -d' ' -f 2`
+
+			if [ -n "$case" ]; then
+				date +%F\ %T | tee -a  $TDENGINE_COVERAGE_REPORT  && taostest $yaml --case=$case --keep --disable_collection > /dev/null 2>&1 && \
+				 echo -e "${GREEN}$case success${NC}" | tee -a  $TDENGINE_COVERAGE_REPORT \
+				|| echo -e "${RED}$case failed${NC}" | tee -a  $TDENGINE_COVERAGE_REPORT
+			fi
+		fi
+	done < $1
+}
+
+function runTest {
+	echo "run Test"
+
+	if [ "$branch" == "3.0" ]; then
+		echo "start run unit test case ................"
+		echo " $TDENGINE_DIR/debug "
+		cd $TDENGINE_DIR/debug
+		ctest -j12
+		echo "3.0 unit test done"
+	fi
+
+	cd $TDENGINE_DIR/tests/script
+
+	[ -d ../../sim ] && rm -rf ../../sim
+	[ -f $TDENGINE_COVERAGE_REPORT ] && rm $TDENGINE_COVERAGE_REPORT
+
+	runGeneralCaseOneByOne jenkins/basic.txt
+
+	sed -i "1i\SIM cases test result" $TDENGINE_COVERAGE_REPORT
+
+	totalSuccess=`grep 'success' $TDENGINE_COVERAGE_REPORT | wc -l`
+	if [ "$totalSuccess" -gt "0" ]; then
+		sed -i -e "2i\ ### Total $totalSuccess SIM test case(s) succeed! ###"  $TDENGINE_COVERAGE_REPORT
+	fi
+
+	totalFailed=`grep 'failed\|fault' $TDENGINE_COVERAGE_REPORT | wc -l`
+	if [ "$totalFailed" -ne "0" ]; then
+		sed -i "3i\### Total $totalFailed SIM test case(s) failed! ###"  $TDENGINE_COVERAGE_REPORT
+	fi
+	sed "3G"  $TDENGINE_COVERAGE_REPORT
+
+	stopTaosd
+	echo "run TestNG cases"
+	rm -rf /var/lib/taos/*
+	rm -rf /var/log/taos/*
+        nohup $TDENGINE_DIR/debug/build/bin/taosd -c /etc/taos > /dev/null 2>&1 &
+        sleep 10
+	cd $TDENGINE_DIR/../TestNG/cases
+	runTestNGCaseOneByOne ../scripts/cases.txt
+	echo "TestNG cases done"
+
+	cd $TDENGINE_DIR/tests
+	rm -rf ../sim
+	/root/pxiao/test-all-coverage.sh full python $branch | tee -a $TDENGINE_COVERAGE_REPORT
+	
+	
+	sed -i "4i\Python cases test result" $TDENGINE_COVERAGE_REPORT
+	totalPySuccess=`grep 'python case(s) succeed!' $TDENGINE_COVERAGE_REPORT | awk '{print $4}'`
+	if [ "$totalPySuccess" -gt "0" ]; then
+                sed -i -e "5i\ ### Total $totalPySuccess Python test case(s) succeed! ###"  $TDENGINE_COVERAGE_REPORT
+        fi
+
+	totalPyFailed=`grep 'python case(s) failed!' $TDENGINE_COVERAGE_REPORT | awk '{print $4}'`
+        if [ -z $totalPyFailed ]; then
+		sed -i "6i\\n"  $TDENGINE_COVERAGE_REPORT
+        else
+		sed -i "6i\### Total $totalPyFailed Python test case(s) failed! ###"  $TDENGINE_COVERAGE_REPORT
+        fi
+
+	echo "### run JDBC test cases ###" | tee -a $TDENGINE_COVERAGE_REPORT
+	# Test Connector
+	stopTaosd
+	nohup $TDENGINE_DIR/debug/build/bin/taosd -c /etc/taos > /dev/null 2>&1 &
+	sleep 10
+
+	cd $TDENGINE_DIR/src/connector/jdbc
+	mvn clean package > /dev/null 2>&1
+	mvn test > jdbc-out.log 2>&1
+  	tail -n 20 jdbc-out.log 2>&1 | tee -a $TDENGINE_COVERAGE_REPORT
+
+	# Test C Demo
+	stopTaosd
+	$TDENGINE_DIR/debug/build/bin/taosd -c $TDENGINE_DIR/debug/test/cfg > /dev/null &
+	sleep 10
+	yes | $TDENGINE_DIR/debug/build/bin/demo 127.0.0.1 > /dev/null 2>&1 | tee -a $TDENGINE_COVERAGE_REPORT
+
+	# Test waltest
+	dataDir=`grep dataDir $TDENGINE_DIR/debug/test/cfg/taos.cfg|awk '{print $2}'`
+	walDir=`find $dataDir -name "wal"|head -n1`
+	echo "dataDir: $dataDir" | tee -a $TDENGINE_COVERAGE_REPORT
+	echo "walDir: $walDir" | tee -a $TDENGINE_COVERAGE_REPORT
+	if [ -n "$walDir" ]; then
+		yes | $TDENGINE_DIR/debug/build/bin/waltest -p $walDir > /dev/null 2>&1 | tee -a $TDENGINE_COVERAGE_REPORT
+	fi
+
+  # run Unit Test
+  echo "Run Unit Test: utilTest, queryTest and cliTest"
+  #$TDENGINE_DIR/debug/build/bin/utilTest > /dev/null 2>&1 && echo "utilTest pass!" || echo "utilTest failed!"
+  #$TDENGINE_DIR/debug/build/bin/queryTest > /dev/null 2>&1 && echo "queryTest pass!" || echo "queryTest failed!"
+  #$TDENGINE_DIR/debug/build/bin/cliTest > /dev/null 2>&1 && echo "cliTest pass!" || echo "cliTest failed!"
+
+	stopTaosd
+
+	cd $TDENGINE_DIR/tests/script
+	find . -name '*.sql' | xargs rm -f
+
+	cd $TDENGINE_DIR/tests/pytest
+	find . -name '*.sql' | xargs rm -f
+}
+
+function lcovFunc {
+	echo "collect data by lcov"
+	cd $TDENGINE_DIR
+
+	# collect data
+	lcov -d . --capture --rc lcov_branch_coverage=1 --rc genhtml_branch_coverage=1 --no-external -b $TDENGINE_DIR -o coverage.info
+
+	# remove exclude paths
+	if [ "$branch" == "3.0" ]; then
+		lcov --remove coverage.info \
+			'*/contrib/*' '*/tests/*' '*/test/*'\
+		       	'*/AccessBridgeCalls.c' '*/ttszip.c' '*/dataInserter.c' '*/tlinearhash.c' '*/tsimplehash.c'\
+			'*/texpr.c' '*/runUdf.c' '*/schDbg.c' '*/syncIO.c' '*/tdbOs.c' '*/pushServer.c' '*/osLz4.c'\
+		       	'*/tbase64.c' '*/tbuffer.c' '*/tdes.c' '*/texception.c' '*/tidpool.c' '*/tmempool.c'\
+		       	'*/tthread.c' '*/tversion.c'\
+		       	--rc lcov_branch_coverage=1 -o coverage.info
+	else	
+		lcov --remove coverage.info \
+    		'*/tests/*' '*/test/*' '*/deps/*' '*/plugins/*' '*/taosdef.h' '*/ttype.h' '*/tarithoperator.c' '*/TSDBJNIConnector.c' '*/taosdemo.c'\
+    		--rc lcov_branch_coverage=1 -o coverage.info
+	fi
+
+
+	# generate result
+	echo "generate result"
+	lcov -l --rc lcov_branch_coverage=1 coverage.info | tee -a $TDENGINE_COVERAGE_REPORT
+
+	# push result to coveralls.io
+	echo "push result to coveralls.io"
+	/usr/local/bin/coveralls-lcov coverage.info -t o7uY02qEAgKyJHrkxLGiCOTfL3IGQR2zm | tee -a $TDENGINE_COVERAGE_REPORT
+
+	#/root/pxiao/checkCoverageFile.sh -s $TDENGINE_DIR/source -f $TDENGINE_COVERAGE_REPORT
+	#cat /root/pxiao/fileListNoCoverage.log | tee -a $TDENGINE_COVERAGE_REPORT
+	cat $TDENGINE_COVERAGE_REPORT | grep "| 0.0%" | awk -F "%" '{print $1}' | awk -F "|" '{if($2==0.0)print $1}' | tee -a $TDENGINE_COVERAGE_REPORT
+	
+}
+
+function sendReport {
+	echo "send report"
+	receiver="develop@taosdata.com"
+	mimebody="MIME-Version: 1.0\nContent-Type: text/html; charset=utf-8\n"
+
+	cd $TDENGINE_DIR
+
+	sed -i 's/\x1b\[[0-9;]*m//g' $TDENGINE_COVERAGE_REPORT
+	BODY_CONTENT=`cat $TDENGINE_COVERAGE_REPORT`
+	echo -e "from: <support@taosdata.com>\nto: ${receiver}\nsubject: Coverage test report ${branch} ${today}, commit ID: ${LOCAL_COMMIT}\n\n${today}:\n${BODY_CONTENT}" | \
+	(cat - && uuencode $TDENGINE_COVERAGE_REPORT coverage-report-$today.log) | \
+	/usr/sbin/ssmtp "${receiver}" && echo "Report Sent!"
+}
+
+function stopTaosd {
+	echo "Stop taosd start"
+        systemctl stop taosd
+  	PID=`ps -ef|grep -w taosd | grep -v grep | awk '{print $2}'`
+	while [ -n "$PID" ]
+	do
+    pkill -TERM -x taosd
+    sleep 1
+  	PID=`ps -ef|grep -w taosd | grep -v grep | awk '{print $2}'`
+	done
+	echo "Stop tasod end"
+}
+
+function runTestRandomFail {
+  exec_random_fail_sh=$1
+  default_exec_sh=$TDENGINE_DIR/tests/script/sh/exec.sh
+	[ -f $exec_random_fail_sh ] && cp $exec_random_fail_sh $default_exec_sh || exit 1
+
+  dnodes_random_fail_py=$TDENGINE_DIR/tests/pytest/util/dnodes-no-random-fail.py
+  default_dnodes_py=$TDENGINE_DIR/tests/pytest/util/dnodes.py
+	[ -f $dnodes_random_fail_py ] && cp $dnodes_random_fail_py $default_dnodes_py || exit 1
+
+  runTest NoRandomFail
+}
+
+WORK_DIR=/root/pxiao
+
+date >> $WORK_DIR/cron.log
+echo "Run Coverage Test" | tee -a $WORK_DIR/cron.log
+
+stopTaosd
+buildTDengine
+
+#runTestRandomFail $TDENGINE_DIR/tests/script/sh/exec-random-fail.sh 
+#runTestRandomFail $TDENGINE_DIR/tests/script/sh/exec-default.sh 
+#runTestRandomFail $TDENGINE_DIR/tests/script/sh/exec-no-random-fail.sh 
+
+runTest
+
+lcovFunc
+#sendReport
+stopTaosd
+
+date >> $WORK_DIR/cron.log
+echo "End of Coverage Test" | tee -a $WORK_DIR/cron.log

tests/script/jenkins/basic.txt

@@ -155,7 +155,7 @@
 ./test.sh -f tsim/parser/select_with_tags.sim
 ./test.sh -f tsim/parser/selectResNum.sim
 ./test.sh -f tsim/parser/set_tag_vals.sim
-./test.sh -f tsim/parser/single_row_in_tb.sim
+# TD-19572 ./test.sh -f tsim/parser/single_row_in_tb.sim
 ./test.sh -f tsim/parser/sliding.sim
 ./test.sh -f tsim/parser/slimit_alter_tags.sim
 ./test.sh -f tsim/parser/slimit.sim
@@ -194,7 +194,7 @@
 ./test.sh -f tsim/mnode/basic1.sim
 ./test.sh -f tsim/mnode/basic2.sim
 ./test.sh -f tsim/mnode/basic3.sim
-# TD-17919 ./test.sh -f tsim/mnode/basic4.sim
+./test.sh -f tsim/mnode/basic4.sim
 ./test.sh -f tsim/mnode/basic5.sim
 
 # ---- show ----

tests/script/tsim/mnode/basic1.sim

@@ -42,6 +42,7 @@ sql_error drop mnode on dnode 1
 print =============== create mnode 2
 sql create mnode on dnode 2
 
+print =============== create mnode 2 finished
 $x = 0
 step2: 
 	$x = $x + 1
@@ -69,9 +70,10 @@ if $data(2)[2] != follower then
   goto step2
 endi
 
-sleep 2000
 print ============ drop mnode 2
 sql drop mnode on dnode 2
+
+print ============ drop mnode 2 finished
 sql select * from information_schema.ins_mnodes
 if $rows != 1 then 
   return -1
@@ -109,6 +111,8 @@ sleep 2000
 
 print =============== create mnodes
 sql create mnode on dnode 2
+
+print =============== create mnode 2 finished
 sql select * from information_schema.ins_mnodes
 if $rows != 2 then 
   return -1