xxq250
/
homework-jianmu
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

Merge branch '3.0' into feature/queryredirect

This commit is contained in:
dapan1121 2022-06-28 09:31:03 +08:00
parent 62d0c04e0c ddf0ff111d
commit 7da915fa67
135 changed files with 5309 additions and 1250 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
Jenkinsfile2
View File

@ -338,6 +338,14 @@ pipeline {
changeRequest()
}
steps {
script {
def linux_node_ip = sh (
script: 'ip addr|grep 192|grep -v virbr|awk "{print \\\$2}"|sed "s/\\/.*//"',
returnStdout: true
).trim()
echo "${linux_node_ip}"
echo "${WKDIR}/restore.sh -p ${BRANCH_NAME} -n ${BUILD_ID} -c {container name}"
}
catchError(buildResult: 'FAILURE', stageResult: 'FAILURE') {
timeout(time: 120, unit: 'MINUTES'){
pre_test()

2
docs/en/21-tdinternal/01-arch.md
View File

@ -22,7 +22,7 @@ A complete TDengine system runs on one or more physical nodes. Logically, it inc
**Virtual node (vnode)**: To better support data sharding, load balancing and prevent data from overheating or skewing, data nodes are virtualized into multiple virtual nodes (vnode, V2, V3, V4, etc. in the figure). Each vnode is a relatively independent work unit, which is the basic unit of time-series data storage and has independent running threads, memory space and persistent storage path. A vnode contains a certain number of tables (data collection points). When a new table is created, the system checks whether a new vnode needs to be created. The number of vnodes that can be created on a data node depends on the capacity of the hardware of the physical node where the data node is located. A vnode belongs to only one DB, but a DB can have multiple vnodes. In addition to the stored time-series data, a vnode also stores the schema and tag values of the included tables. A virtual node is uniquely identified in the system by the EP of the data node and the VGroup ID to which it belongs and is created and managed by the management node.
**Management node (mnode)**: A virtual logical unit responsible for monitoring and maintaining the running status of all data nodes and load balancing among nodes (M in the figure). At the same time, the management node is also responsible for the storage and management of metadata (including users, databases, tables, static tags, etc.), so it is also called Meta Node. Multiple (up to 5) mnodes can be configured in a TDengine cluster, and they are automatically constructed into a virtual management node group (M0, M1, M2 in the figure). The leader/follower mechanism is adopted for the mnode group and the data synchronization is carried out in a strongly consistent way. Any data update operation can only be executed on the leader. The creation of mnode cluster is completed automatically by the system without manual intervention. There is at most one mnode on each dnode, which is uniquely identified by the EP of the data node to which it belongs. Each dnode automatically obtains the EP of the dnode where all mnodes in the whole cluster are located, through internal messaging interaction.
**Management node (mnode)**: A virtual logical unit responsible for monitoring and maintaining the running status of all data nodes and load balancing among nodes (M in the figure). At the same time, the management node is also responsible for the storage and management of metadata (including users, databases, tables, static tags, etc.), so it is also called Meta Node. Multiple (up to 3) mnodes can be configured in a TDengine cluster, and they are automatically constructed into a virtual management node group (M0, M1, M2 in the figure). The leader/follower mechanism is adopted for the mnode group and the data synchronization is carried out in a strongly consistent way. Any data update operation can only be executed on the leader. The creation of mnode cluster is completed automatically by the system without manual intervention. There is at most one mnode on each dnode, which is uniquely identified by the EP of the data node to which it belongs. Each dnode automatically obtains the EP of the dnode where all mnodes in the whole cluster are located, through internal messaging interaction.
**Virtual node group (VGroup)**: Vnodes on different data nodes can form a virtual node group to ensure the high availability of the system. The virtual node group is managed in a leader/follower mechanism. Write operations can only be performed on the leader vnode, and then replicated to follower vnodes, thus ensuring that one single replica of data is copied on multiple physical nodes. The number of virtual nodes in a vgroup equals the number of data replicas. If the number of replicas of a DB is N, the system must have at least N data nodes. The number of replicas can be specified by the parameter `“replica”` when creating a DB, and the default is 1. Using the multi-replication feature of TDengine, the same high data reliability can be achieved without the need for expensive storage devices such as disk arrays. Virtual node groups are created and managed by the management node, and the management node assigns a system unique ID, aka VGroup ID. If two virtual nodes have the same vnode group ID, it means that they belong to the same group and the data is backed up to each other. The number of virtual nodes in a virtual node group can be dynamically changed, allowing only one, that is, no data replication. VGroup ID is never changed. Even if a virtual node group is deleted, its ID will not be reused.

2
docs/zh/21-tdinternal/01-arch.md
View File

@ -23,7 +23,7 @@ TDengine 分布式架构的逻辑结构图如下：
**虚拟节点（vnode）：** 为更好的支持数据分片、负载均衡，防止数据过热或倾斜，数据节点被虚拟化成多个虚拟节点（vnode，图中 V2，V3，V4 等）。每个 vnode 都是一个相对独立的工作单元，是时序数据存储的基本单元，具有独立的运行线程、内存空间与持久化存储的路径。一个 vnode 包含一定数量的表（数据采集点）。当创建一张新表时，系统会检查是否需要创建新的 vnode。一个数据节点上能创建的 vnode 的数量取决于该数据节点所在物理节点的硬件资源。一个 vnode 只属于一个 DB，但一个 DB 可以有多个 vnode。一个 vnode 除存储的时序数据外，也保存有所包含的表的 schema、标签值等。一个虚拟节点由所属的数据节点的 EP，以及所属的 VGroup ID 在系统内唯一标识，由管理节点创建并管理。
**管理节点（mnode）：** 一个虚拟的逻辑单元，负责所有数据节点运行状态的监控和维护，以及节点之间的负载均衡（图中 M）。同时，管理节点也负责元数据（包括用户、数据库、表、静态标签等）的存储和管理，因此也称为 Meta Node。TDengine 集群中可配置多个（开源版最多不超过 3 个）mnode，它们自动构建成为一个虚拟管理节点组（图中 M0，M1，M2）。mnode 间采用 master/slave 的机制进行管理，而且采取强一致方式进行数据同步，任何数据更新操作只能在 Master 上进行。mnode 集群的创建由系统自动完成，无需人工干预。每个 dnode 上至多有一个 mnode，由所属的数据节点的 EP 来唯一标识。每个 dnode 通过内部消息交互自动获取整个集群中所有 mnode 所在的 dnode 的 EP。
**管理节点（mnode）：** 一个虚拟的逻辑单元，负责所有数据节点运行状态的监控和维护，以及节点之间的负载均衡（图中 M）。同时，管理节点也负责元数据（包括用户、数据库、表、静态标签等）的存储和管理，因此也称为 Meta Node。TDengine 集群中可配置多个（最多不超过 3 个）mnode，它们自动构建成为一个虚拟管理节点组（图中 M0，M1，M2）。mnode 间采用 master/slave 的机制进行管理，而且采取强一致方式进行数据同步，任何数据更新操作只能在 Master 上进行。mnode 集群的创建由系统自动完成，无需人工干预。每个 dnode 上至多有一个 mnode，由所属的数据节点的 EP 来唯一标识。每个 dnode 通过内部消息交互自动获取整个集群中所有 mnode 所在的 dnode 的 EP。
**虚拟节点组（VGroup）：** 不同数据节点上的 vnode 可以组成一个虚拟节点组（vgroup）来保证系统的高可靠。虚拟节点组内采取 master/slave 的方式进行管理。写操作只能在 master vnode 上进行，系统采用异步复制的方式将数据同步到 slave vnode，这样确保了一份数据在多个物理节点上有拷贝。一个 vgroup 里虚拟节点个数就是数据的副本数。如果一个 DB 的副本数为 N，系统必须有至少 N 数据节点。副本数在创建 DB 时通过参数 replica 可以指定，缺省为 1。使用 TDengine 的多副本特性，可以不再需要昂贵的磁盘阵列等存储设备，就可以获得同样的数据高可靠性。虚拟节点组由管理节点创建、管理，并且由管理节点分配一个系统唯一的 ID，VGroup ID。如果两个虚拟节点的 VGroup ID 相同，说明他们属于同一个组，数据互为备份。虚拟节点组里虚拟节点的个数是可以动态改变的，容许只有一个，也就是没有数据复制。VGroup ID 是永远不变的，即使一个虚拟节点组被删除，它的 ID 也不会被收回重复利用。

1
include/client/taos.h
View File

@ -128,6 +128,7 @@ typedef struct setConfRet {
DLL_EXPORT void taos_cleanup(void);
DLL_EXPORT int taos_options(TSDB_OPTION option, const void *arg, ...);
DLL_EXPORT setConfRet taos_set_config(const char *config);
DLL_EXPORT int taos_init(void);
DLL_EXPORT TAOS *taos_connect(const char *ip, const char *user, const char *pass, const char *db, uint16_t port);
DLL_EXPORT TAOS *taos_connect_auth(const char *ip, const char *user, const char *auth, const char *db, uint16_t port);

2
include/common/tdatablock.h
View File

@ -236,7 +236,7 @@ void blockCompressEncode(const SSDataBlock* pBlock, char* data, int32_t*
int8_t needCompress);
const char* blockCompressDecode(SSDataBlock* pBlock, int32_t numOfCols, int32_t numOfRows, const char* pData);
void blockDebugShowData(const SArray* dataBlocks, const char* flag);
void blockDebugShowDataBlocks(const SArray* dataBlocks, const char* flag);
// for debug
char* dumpBlockData(SSDataBlock* pDataBlock, const char* flag, char** dumpBuf);

1
include/common/tdataformat.h
View File

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

2
include/common/tmsg.h
View File

@ -1886,7 +1886,7 @@ typedef struct SVCreateStbReq {
int8_t rollup;
SSchemaWrapper schemaRow;
SSchemaWrapper schemaTag;
SRSmaParam pRSmaParam;
SRSmaParam rsmaParam;
} SVCreateStbReq;
int tEncodeSVCreateStbReq(SEncoder* pCoder, const SVCreateStbReq* pReq);

14
include/libs/nodes/plannodes.h
View File

@ -24,6 +24,8 @@ extern "C" {
#include "querynodes.h"
#include "tname.h"
#define SLOT_NAME_LEN TSDB_TABLE_NAME_LEN + TSDB_COL_NAME_LEN
typedef struct SLogicNode {
ENodeType type;
SNodeList* pTargets; // SColumnNode
@ -74,8 +76,8 @@ typedef struct SScanLogicNode {
int16_t tsColId;
double filesFactor;
SArray* pSmaIndexes;
SNodeList* pPartTags;
bool partSort;
SNodeList* pGroupTags;
bool groupSort;
} SScanLogicNode;
typedef struct SJoinLogicNode {
@ -100,6 +102,7 @@ typedef struct SProjectLogicNode {
typedef struct SIndefRowsFuncLogicNode {
SLogicNode node;
SNodeList* pFuncs;
bool isTailFunc;
} SIndefRowsFuncLogicNode;
typedef struct SInterpFuncLogicNode {
@ -138,6 +141,7 @@ typedef struct SMergeLogicNode {
SNodeList* pInputs;
int32_t numOfChannels;
int32_t srcGroupId;
bool groupSort;
} SMergeLogicNode;
typedef enum EWindowType { WINDOW_TYPE_INTERVAL = 1, WINDOW_TYPE_SESSION, WINDOW_TYPE_STATE } EWindowType;
@ -184,6 +188,7 @@ typedef struct SFillLogicNode {
typedef struct SSortLogicNode {
SLogicNode node;
SNodeList* pSortKeys;
bool groupSort;
} SSortLogicNode;
typedef struct SPartitionLogicNode {
@ -230,6 +235,7 @@ typedef struct SSlotDescNode {
bool reserve;
bool output;
bool tag;
char name[SLOT_NAME_LEN];
} SSlotDescNode;
typedef struct SDataBlockDescNode {
@ -279,7 +285,8 @@ typedef struct STableScanPhysiNode {
double ratio;
int32_t dataRequired;
SNodeList* pDynamicScanFuncs;
SNodeList* pPartitionTags;
SNodeList* pGroupTags;
bool groupSort;
int64_t interval;
int64_t offset;
int64_t sliding;
@ -353,6 +360,7 @@ typedef struct SMergePhysiNode {
SNodeList* pTargets;
int32_t numOfChannels;
int32_t srcGroupId;
bool groupSort;
} SMergePhysiNode;
typedef struct SWinodwPhysiNode {

1
include/libs/nodes/querynodes.h
View File

@ -259,6 +259,7 @@ typedef struct SSelectStmt {
bool hasTailFunc;
bool hasInterpFunc;
bool hasLastRowFunc;
bool groupSort;
} SSelectStmt;
typedef enum ESetOperatorType { SET_OP_TYPE_UNION_ALL = 1, SET_OP_TYPE_UNION } ESetOperatorType;

4
include/libs/sync/sync.h
View File

@ -26,9 +26,9 @@ extern "C" {
extern bool gRaftDetailLog;
#define SYNC_INDEX_BEGIN 0
#define SYNC_INDEX_BEGIN 0
#define SYNC_INDEX_INVALID -1
#define SYNC_TERM_INVALID 0xFFFFFFFFFFFFFFFF
#define SYNC_TERM_INVALID 0xFFFFFFFFFFFFFFFF
typedef uint64_t SyncNodeId;
typedef int32_t SyncGroupId;

4
include/libs/sync/syncTools.h
View File

@ -524,7 +524,7 @@ void syncReconfigFinishLog2(char* s, const SyncReconfigFinish* pMsg);
int32_t syncNodeOnPingCb(SSyncNode* ths, SyncPing* pMsg);
int32_t syncNodeOnPingReplyCb(SSyncNode* ths, SyncPingReply* pMsg);
int32_t syncNodeOnTimeoutCb(SSyncNode* ths, SyncTimeout* pMsg);
int32_t syncNodeOnClientRequestCb(SSyncNode* ths, SyncClientRequest* pMsg);
int32_t syncNodeOnClientRequestCb(SSyncNode* ths, SyncClientRequest* pMsg, SyncIndex* pRetIndex);
int32_t syncNodeOnRequestVoteCb(SSyncNode* ths, SyncRequestVote* pMsg);
int32_t syncNodeOnRequestVoteReplyCb(SSyncNode* ths, SyncRequestVoteReply* pMsg);
int32_t syncNodeOnAppendEntriesCb(SSyncNode* ths, SyncAppendEntries* pMsg);
@ -541,7 +541,7 @@ int32_t syncNodeOnSnapshotRspCb(SSyncNode* ths, SyncSnapshotRsp* pMsg);
// -----------------------------------------
typedef int32_t (*FpOnPingCb)(SSyncNode* ths, SyncPing* pMsg);
typedef int32_t (*FpOnPingReplyCb)(SSyncNode* ths, SyncPingReply* pMsg);
typedef int32_t (*FpOnClientRequestCb)(SSyncNode* ths, SyncClientRequest* pMsg);
typedef int32_t (*FpOnClientRequestCb)(SSyncNode* ths, SyncClientRequest* pMsg, SyncIndex* pRetIndex);
typedef int32_t (*FpOnRequestVoteCb)(SSyncNode* ths, SyncRequestVote* pMsg);
typedef int32_t (*FpOnRequestVoteReplyCb)(SSyncNode* ths, SyncRequestVoteReply* pMsg);
typedef int32_t (*FpOnAppendEntriesCb)(SSyncNode* ths, SyncAppendEntries* pMsg);

2
include/libs/transport/trpc.h
View File

@ -69,7 +69,7 @@ typedef struct SRpcMsg {
} SRpcMsg;
typedef void (*RpcCfp)(void *parent, SRpcMsg *, SEpSet *rf);
typedef bool (*RpcRfp)(int32_t code);
typedef bool (*RpcRfp)(int32_t code, tmsg_t msgType);
typedef struct SRpcInit {
char localFqdn[TSDB_FQDN_LEN];

3
include/os/osSystem.h
View File

@ -29,6 +29,9 @@ extern "C" {
#define tcgetattr TCGETATTR_FUNC_TAOS_FORBID
#endif
#define TAOS_CONSOLE_PROMPT_HEADER "taos> "
#define TAOS_CONSOLE_PROMPT_CONTINUE " -> "
typedef struct TdCmd *TdCmdPtr;
TdCmdPtr taosOpenCmd(const char* cmd);

5
source/client/src/clientEnv.c
View File

@ -88,9 +88,12 @@ void closeTransporter(SAppInstInfo *pAppInfo) {
rpcClose(pAppInfo->pTransporter);
}
static bool clientRpcRfp(int32_t code) {
static bool clientRpcRfp(int32_t code, tmsg_t msgType) {
if (code == TSDB_CODE_RPC_REDIRECT || code == TSDB_CODE_RPC_NETWORK_UNAVAIL || code == TSDB_CODE_NODE_NOT_DEPLOYED ||
code == TSDB_CODE_SYN_NOT_LEADER || code == TSDB_CODE_APP_NOT_READY) {
if (msgType == TDMT_VND_QUERY || msgType == TDMT_VND_FETCH) {
return false;
}
return true;
} else {
return false;

13
source/client/src/clientMain.c
View File

@ -81,6 +81,19 @@ void taos_cleanup(void) {
taosCloseLog();
}
static setConfRet taos_set_config_imp(const char *config){
setConfRet ret = {SET_CONF_RET_SUCC, {0}};
// TODO: need re-implementation
return ret;
}
setConfRet taos_set_config(const char *config){
// TODO pthread_mutex_lock(&setConfMutex);
setConfRet ret = taos_set_config_imp(config);
// pthread_mutex_unlock(&setConfMutex);
return ret;
}
TAOS *taos_connect(const char *ip, const char *user, const char *pass, const char *db, uint16_t port) {
tscDebug("try to connect to %s:%u, user:%s db:%s", ip, port, user, db);
if (user == NULL) {

4
source/common/src/tdatablock.c
View File

@ -1605,7 +1605,7 @@ static char* formatTimestamp(char* buf, int64_t val, int precision) {
return buf;
}
void blockDebugShowData(const SArray* dataBlocks, const char* flag) {
void blockDebugShowDataBlocks(const SArray* dataBlocks, const char* flag) {
char pBuf[128] = {0};
int32_t sz = taosArrayGetSize(dataBlocks);
for (int32_t i = 0; i < sz; i++) {
@ -1613,7 +1613,7 @@ void blockDebugShowData(const SArray* dataBlocks, const char* flag) {
size_t numOfCols = taosArrayGetSize(pDataBlock->pDataBlock);
int32_t rows = pDataBlock->info.rows;
printf("%s |block type %d |child id %d|\n", flag, (int32_t)pDataBlock->info.type, pDataBlock->info.childId);
printf("%s |block type %d |child id %d|group id %zX\n", flag, (int32_t)pDataBlock->info.type, pDataBlock->info.childId, pDataBlock->info.groupId);
for (int32_t j = 0; j < rows; j++) {
printf("%s |", flag);
for (int32_t k = 0; k < numOfCols; k++) {

16
source/common/src/tdataformat.c
View File

@ -862,21 +862,6 @@ void debugPrintSTag(STag *pTag, const char *tag, int32_t ln) {
printf("\n");
}
void debugCheckTags(STag *pTag) {
switch (pTag->flags) {
case 0x0:
case 0x20:
case 0x40:
case 0x60:
break;
default:
ASSERT(0);
}
ASSERT(pTag->nTag <= 128 && pTag->nTag >= 0);
ASSERT(pTag->ver <= 512 && pTag->ver >= 0); // temp condition for pTag->ver
}
static int32_t tPutTagVal(uint8_t *p, STagVal *pTagVal, int8_t isJson) {
int32_t n = 0;
@ -999,7 +984,6 @@ int32_t tTagNew(SArray *pArray, int32_t version, int8_t isJson, STag **ppTag) {
debugPrintSTag(*ppTag, __func__, __LINE__);
#endif
debugCheckTags(*ppTag); // TODO: remove this line after debug
return code;
_err:

14
source/common/src/tmsg.c
View File

@ -3696,7 +3696,7 @@ int32_t tDeserializeSCreateVnodeReq(void *buf, int32_t bufLen, SCreateVnodeReq *
if (tDecodeI8(&decoder, &pReq->isTsma) < 0) return -1;
if (pReq->isTsma) {
if (tDecodeBinaryAlloc(&decoder, &pReq->pTsma, NULL) < 0) return -1;
if (tDecodeBinary(&decoder, (uint8_t **)&pReq->pTsma, NULL) < 0) return -1;
}
tEndDecode(&decoder);
@ -3707,9 +3707,6 @@ int32_t tDeserializeSCreateVnodeReq(void *buf, int32_t bufLen, SCreateVnodeReq *
int32_t tFreeSCreateVnodeReq(SCreateVnodeReq *pReq) {
taosArrayDestroy(pReq->pRetensions);
pReq->pRetensions = NULL;
if (pReq->isTsma) {
taosMemoryFreeClear(pReq->pTsma);
}
return 0;
}
@ -4747,9 +4744,8 @@ int32_t tDecodeSRSmaParam(SDecoder *pCoder, SRSmaParam *pRSmaParam) {
if (tDecodeI64v(pCoder, &pRSmaParam->watermark[i]) < 0) return -1;
if (tDecodeI32v(pCoder, &pRSmaParam->qmsgLen[i]) < 0) return -1;
if (pRSmaParam->qmsgLen[i] > 0) {
uint64_t len;
if (tDecodeBinaryAlloc(pCoder, (void **)&pRSmaParam->qmsg[i], &len) < 0)
return -1; // qmsgLen contains len of '\0'
tDecoderMalloc(pCoder, pRSmaParam->qmsgLen[i]);
if (tDecodeBinary(pCoder, (uint8_t **)&pRSmaParam->qmsg[i], NULL) < 0) return -1; // qmsgLen contains len of '\0'
} else {
pRSmaParam->qmsg[i] = NULL;
}
@ -4767,7 +4763,7 @@ int tEncodeSVCreateStbReq(SEncoder *pCoder, const SVCreateStbReq *pReq) {
if (tEncodeSSchemaWrapper(pCoder, &pReq->schemaRow) < 0) return -1;
if (tEncodeSSchemaWrapper(pCoder, &pReq->schemaTag) < 0) return -1;
if (pReq->rollup) {
if (tEncodeSRSmaParam(pCoder, &pReq->pRSmaParam) < 0) return -1;
if (tEncodeSRSmaParam(pCoder, &pReq->rsmaParam) < 0) return -1;
}
tEndEncode(pCoder);
@ -4783,7 +4779,7 @@ int tDecodeSVCreateStbReq(SDecoder *pCoder, SVCreateStbReq *pReq) {
if (tDecodeSSchemaWrapper(pCoder, &pReq->schemaRow) < 0) return -1;
if (tDecodeSSchemaWrapper(pCoder, &pReq->schemaTag) < 0) return -1;
if (pReq->rollup) {
if (tDecodeSRSmaParam(pCoder, &pReq->pRSmaParam) < 0) return -1;
if (tDecodeSRSmaParam(pCoder, &pReq->rsmaParam) < 0) return -1;
}
tEndDecode(pCoder);

5
source/dnode/mgmt/node_mgmt/src/dmTransport.c
View File

@ -248,9 +248,12 @@ static inline void dmReleaseHandle(SRpcHandleInfo *pHandle, int8_t type) {
}
}
static bool rpcRfp(int32_t code) {
static bool rpcRfp(int32_t code, tmsg_t msgType) {
if (code == TSDB_CODE_RPC_REDIRECT || code == TSDB_CODE_RPC_NETWORK_UNAVAIL || code == TSDB_CODE_NODE_NOT_DEPLOYED ||
code == TSDB_CODE_SYN_NOT_LEADER || code == TSDB_CODE_APP_NOT_READY) {
if (msgType == TDMT_VND_QUERY || msgType == TDMT_VND_FETCH) {
return false;
}
return true;
} else {
return false;

4
source/dnode/mnode/impl/inc/mndTrans.h
View File

@ -39,8 +39,10 @@ typedef struct {
int32_t id;
int32_t errCode;
int32_t acceptableCode;
ETrnStage stage;
int32_t retryCode;
ETrnAct actionType;
ETrnStage stage;
int8_t reserved;
int8_t rawWritten;
int8_t msgSent;
int8_t msgReceived;

4
source/dnode/mnode/impl/src/mndMain.c
View File

@ -442,7 +442,7 @@ int32_t mndProcessSyncMsg(SRpcMsg *pMsg) {
syncPingReplyDestroy(pSyncMsg);
} else if (pMsg->msgType == TDMT_SYNC_CLIENT_REQUEST) {
SyncClientRequest *pSyncMsg = syncClientRequestFromRpcMsg2(pMsg);
code = syncNodeOnClientRequestCb(pSyncNode, pSyncMsg);
code = syncNodeOnClientRequestCb(pSyncNode, pSyncMsg, NULL);
syncClientRequestDestroy(pSyncMsg);
} else if (pMsg->msgType == TDMT_SYNC_REQUEST_VOTE) {
SyncRequestVote *pSyncMsg = syncRequestVoteFromRpcMsg2(pMsg);
@ -491,7 +491,7 @@ int32_t mndProcessSyncMsg(SRpcMsg *pMsg) {
syncPingReplyDestroy(pSyncMsg);
} else if (pMsg->msgType == TDMT_SYNC_CLIENT_REQUEST) {
SyncClientRequest *pSyncMsg = syncClientRequestFromRpcMsg2(pMsg);
code = syncNodeOnClientRequestCb(pSyncNode, pSyncMsg);
code = syncNodeOnClientRequestCb(pSyncNode, pSyncMsg, NULL);
syncClientRequestDestroy(pSyncMsg);
} else if (pMsg->msgType == TDMT_SYNC_REQUEST_VOTE) {
SyncRequestVote *pSyncMsg = syncRequestVoteFromRpcMsg2(pMsg);

2
source/dnode/mnode/impl/src/mndShow.c
View File

@ -234,7 +234,7 @@ static int32_t mndProcessRetrieveSysTableReq(SRpcMsg *pReq) {
if (retrieveReq.user[0] != 0) {
memcpy(pReq->info.conn.user, retrieveReq.user, TSDB_USER_LEN);
} else {
memcpy(pReq->info.conn.user, TSDB_DEFAULT_USER, TSDB_USER_LEN);
memcpy(pReq->info.conn.user, TSDB_DEFAULT_USER, strlen(TSDB_DEFAULT_USER) + 1);
}
if (mndCheckShowPrivilege(pMnode, pReq->info.conn.user, pShow->type, retrieveReq.db) != 0) {
return -1;

20
source/dnode/mnode/impl/src/mndStb.c
View File

@ -427,17 +427,17 @@ static void *mndBuildVCreateStbReq(SMnode *pMnode, SVgObj *pVgroup, SStbObj *pSt
req.schemaTag.pSchema = pStb->pTags;
if (req.rollup) {
req.pRSmaParam.maxdelay[0] = pStb->maxdelay[0];
req.pRSmaParam.maxdelay[1] = pStb->maxdelay[1];
req.rsmaParam.maxdelay[0] = pStb->maxdelay[0];
req.rsmaParam.maxdelay[1] = pStb->maxdelay[1];
if (pStb->ast1Len > 0) {
if (mndConvertRsmaTask(&req.pRSmaParam.qmsg[0], &req.pRSmaParam.qmsgLen[0], pStb->pAst1, pStb->uid,
STREAM_TRIGGER_WINDOW_CLOSE, req.pRSmaParam.watermark[0]) < 0) {
if (mndConvertRsmaTask(&req.rsmaParam.qmsg[0], &req.rsmaParam.qmsgLen[0], pStb->pAst1, pStb->uid,
STREAM_TRIGGER_WINDOW_CLOSE, req.rsmaParam.watermark[0]) < 0) {
goto _err;
}
}
if (pStb->ast2Len > 0) {
if (mndConvertRsmaTask(&req.pRSmaParam.qmsg[1], &req.pRSmaParam.qmsgLen[1], pStb->pAst2, pStb->uid,
STREAM_TRIGGER_WINDOW_CLOSE, req.pRSmaParam.watermark[1]) < 0) {
if (mndConvertRsmaTask(&req.rsmaParam.qmsg[1], &req.rsmaParam.qmsgLen[1], pStb->pAst2, pStb->uid,
STREAM_TRIGGER_WINDOW_CLOSE, req.rsmaParam.watermark[1]) < 0) {
goto _err;
}
}
@ -470,12 +470,12 @@ static void *mndBuildVCreateStbReq(SMnode *pMnode, SVgObj *pVgroup, SStbObj *pSt
tEncoderClear(&encoder);
*pContLen = contLen;
taosMemoryFreeClear(req.pRSmaParam.qmsg[0]);
taosMemoryFreeClear(req.pRSmaParam.qmsg[1]);
taosMemoryFreeClear(req.rsmaParam.qmsg[0]);
taosMemoryFreeClear(req.rsmaParam.qmsg[1]);
return pHead;
_err:
taosMemoryFreeClear(req.pRSmaParam.qmsg[0]);
taosMemoryFreeClear(req.pRSmaParam.qmsg[1]);
taosMemoryFreeClear(req.rsmaParam.qmsg[0]);
taosMemoryFreeClear(req.rsmaParam.qmsg[1]);
return NULL;
}

90
source/dnode/mnode/impl/src/mndTrans.c
View File

@ -15,15 +15,15 @@
#define _DEFAULT_SOURCE
#include "mndTrans.h"
#include "mndPrivilege.h"
#include "mndConsumer.h"
#include "mndDb.h"
#include "mndPrivilege.h"
#include "mndShow.h"
#include "mndSync.h"
#include "mndUser.h"
#define TRANS_VER_NUMBER 1
#define TRANS_ARRAY_SIZE 8
#define TRANS_VER_NUMBER 1
#define TRANS_ARRAY_SIZE 8
#define TRANS_RESERVE_SIZE 64
static SSdbRaw *mndTransActionEncode(STrans *pTrans);
@ -55,7 +55,7 @@ static bool mndTransPerfromFinishedStage(SMnode *pMnode, STrans *pTrans);
static bool mndCannotExecuteTransAction(SMnode *pMnode) { return !pMnode->deploy && !mndIsMaster(pMnode); }
static void mndTransSendRpcRsp(SMnode *pMnode, STrans *pTrans);
static int32_t mndProcessTransReq(SRpcMsg *pReq);
static int32_t mndProcessTransTimer(SRpcMsg *pReq);
static int32_t mndProcessTtl(SRpcMsg *pReq);
static int32_t mndProcessKillTransReq(SRpcMsg *pReq);
@ -73,7 +73,7 @@ int32_t mndInitTrans(SMnode *pMnode) {
.deleteFp = (SdbDeleteFp)mndTransActionDelete,
};
mndSetMsgHandle(pMnode, TDMT_MND_TRANS_TIMER, mndProcessTransReq);
mndSetMsgHandle(pMnode, TDMT_MND_TRANS_TIMER, mndProcessTransTimer);
mndSetMsgHandle(pMnode, TDMT_MND_KILL_TRANS, mndProcessKillTransReq);
mndAddShowRetrieveHandle(pMnode, TSDB_MGMT_TABLE_TRANS, mndRetrieveTrans);
@ -139,8 +139,10 @@ static SSdbRaw *mndTransActionEncode(STrans *pTrans) {
SDB_SET_INT32(pRaw, dataPos, pAction->id, _OVER)
SDB_SET_INT32(pRaw, dataPos, pAction->errCode, _OVER)
SDB_SET_INT32(pRaw, dataPos, pAction->acceptableCode, _OVER)
SDB_SET_INT32(pRaw, dataPos, pAction->retryCode, _OVER)
SDB_SET_INT8(pRaw, dataPos, pAction->actionType, _OVER)
SDB_SET_INT8(pRaw, dataPos, pAction->stage, _OVER)
SDB_SET_INT8(pRaw, dataPos, pAction->reserved, _OVER)
if (pAction->actionType == TRANS_ACTION_RAW) {
int32_t len = sdbGetRawTotalSize(pAction->pRaw);
SDB_SET_INT8(pRaw, dataPos, pAction->rawWritten, _OVER)
@ -163,8 +165,10 @@ static SSdbRaw *mndTransActionEncode(STrans *pTrans) {
SDB_SET_INT32(pRaw, dataPos, pAction->id, _OVER)
SDB_SET_INT32(pRaw, dataPos, pAction->errCode, _OVER)
SDB_SET_INT32(pRaw, dataPos, pAction->acceptableCode, _OVER)
SDB_SET_INT32(pRaw, dataPos, pAction->retryCode, _OVER)
SDB_SET_INT8(pRaw, dataPos, pAction->actionType, _OVER)
SDB_SET_INT8(pRaw, dataPos, pAction->stage, _OVER)
SDB_SET_INT8(pRaw, dataPos, pAction->reserved, _OVER)
if (pAction->actionType == TRANS_ACTION_RAW) {
int32_t len = sdbGetRawTotalSize(pAction->pRaw);
SDB_SET_INT8(pRaw, dataPos, pAction->rawWritten, _OVER)
@ -187,8 +191,10 @@ static SSdbRaw *mndTransActionEncode(STrans *pTrans) {
SDB_SET_INT32(pRaw, dataPos, pAction->id, _OVER)
SDB_SET_INT32(pRaw, dataPos, pAction->errCode, _OVER)
SDB_SET_INT32(pRaw, dataPos, pAction->acceptableCode, _OVER)
SDB_SET_INT32(pRaw, dataPos, pAction->retryCode, _OVER)
SDB_SET_INT8(pRaw, dataPos, pAction->actionType, _OVER)
SDB_SET_INT8(pRaw, dataPos, pAction->stage, _OVER)
SDB_SET_INT8(pRaw, dataPos, pAction->reserved, _OVER)
if (pAction->actionType == TRANS_ACTION_RAW) {
int32_t len = sdbGetRawTotalSize(pAction->pRaw);
SDB_SET_INT8(pRaw, dataPos, pAction->rawWritten, _OVER)
@ -291,10 +297,12 @@ static SSdbRow *mndTransActionDecode(SSdbRaw *pRaw) {
SDB_GET_INT32(pRaw, dataPos, &action.id, _OVER)
SDB_GET_INT32(pRaw, dataPos, &action.errCode, _OVER)
SDB_GET_INT32(pRaw, dataPos, &action.acceptableCode, _OVER)
SDB_GET_INT32(pRaw, dataPos, &action.retryCode, _OVER)
SDB_GET_INT8(pRaw, dataPos, &actionType, _OVER)
action.actionType = actionType;
SDB_GET_INT8(pRaw, dataPos, &stage, _OVER)
action.stage = stage;
SDB_GET_INT8(pRaw, dataPos, &action.reserved, _OVER)
if (action.actionType == TRANS_ACTION_RAW) {
SDB_GET_INT8(pRaw, dataPos, &action.rawWritten, _OVER)
SDB_GET_INT32(pRaw, dataPos, &dataLen, _OVER)
@ -324,10 +332,12 @@ static SSdbRow *mndTransActionDecode(SSdbRaw *pRaw) {
SDB_GET_INT32(pRaw, dataPos, &action.id, _OVER)
SDB_GET_INT32(pRaw, dataPos, &action.errCode, _OVER)
SDB_GET_INT32(pRaw, dataPos, &action.acceptableCode, _OVER)
SDB_GET_INT32(pRaw, dataPos, &action.retryCode, _OVER)
SDB_GET_INT8(pRaw, dataPos, &actionType, _OVER)
action.actionType = actionType;
SDB_GET_INT8(pRaw, dataPos, &stage, _OVER)
action.stage = stage;
SDB_GET_INT8(pRaw, dataPos, &action.reserved, _OVER)
if (action.actionType == TRANS_ACTION_RAW) {
SDB_GET_INT8(pRaw, dataPos, &action.rawWritten, _OVER)
SDB_GET_INT32(pRaw, dataPos, &dataLen, _OVER)
@ -357,10 +367,12 @@ static SSdbRow *mndTransActionDecode(SSdbRaw *pRaw) {
SDB_GET_INT32(pRaw, dataPos, &action.id, _OVER)
SDB_GET_INT32(pRaw, dataPos, &action.errCode, _OVER)
SDB_GET_INT32(pRaw, dataPos, &action.acceptableCode, _OVER)
SDB_GET_INT32(pRaw, dataPos, &action.retryCode, _OVER)
SDB_GET_INT8(pRaw, dataPos, &actionType, _OVER)
action.actionType = actionType;
SDB_GET_INT8(pRaw, dataPos, &stage, _OVER)
action.stage = stage;
SDB_GET_INT8(pRaw, dataPos, &action.reserved, _OVER)
if (action.actionType) {
SDB_GET_INT8(pRaw, dataPos, &action.rawWritten, _OVER)
SDB_GET_INT32(pRaw, dataPos, &dataLen, _OVER)
@ -463,15 +475,25 @@ static int32_t mndTransActionInsert(SSdb *pSdb, STrans *pTrans) {
if (fp) {
(*fp)(pSdb->pMnode, pTrans->param, pTrans->paramLen);
}
pTrans->startFunc = 0;
}
return 0;
}
static void mndTransDropData(STrans *pTrans) {
mndTransDropActions(pTrans->redoActions);
mndTransDropActions(pTrans->undoActions);
mndTransDropActions(pTrans->commitActions);
if (pTrans->redoActions != NULL) {
mndTransDropActions(pTrans->redoActions);
pTrans->redoActions = NULL;
}
if (pTrans->undoActions != NULL) {
mndTransDropActions(pTrans->undoActions);
pTrans->undoActions = NULL;
}
if (pTrans->commitActions != NULL) {
mndTransDropActions(pTrans->commitActions);
pTrans->commitActions = NULL;
}
if (pTrans->rpcRsp != NULL) {
taosMemoryFree(pTrans->rpcRsp);
pTrans->rpcRsp = NULL;
@ -492,6 +514,7 @@ static int32_t mndTransActionDelete(SSdb *pSdb, STrans *pTrans, bool callFunc) {
if (fp) {
(*fp)(pSdb->pMnode, pTrans->param, pTrans->paramLen);
}
pTrans->stopFunc = 0;
}
mndTransDropData(pTrans);
@ -805,7 +828,7 @@ static void mndTransSendRpcRsp(SMnode *pMnode, STrans *pTrans) {
sendRsp = true;
}
} else {
if (pTrans->stage == TRN_STAGE_REDO_ACTION && pTrans->failedTimes > 3) {
if (pTrans->stage == TRN_STAGE_REDO_ACTION && pTrans->failedTimes > 6) {
if (code == 0) code = TSDB_CODE_MND_TRANS_UNKNOW_ERROR;
sendRsp = true;
}
@ -875,8 +898,8 @@ int32_t mndTransProcessRsp(SRpcMsg *pRsp) {
pAction->errCode = pRsp->code;
}
mDebug("trans:%d, %s:%d response is received, code:0x%x, accept:0x%x", transId, mndTransStr(pAction->stage), action,
pRsp->code, pAction->acceptableCode);
mDebug("trans:%d, %s:%d response is received, code:0x%x, accept:0x%x retry:0x%x", transId,
mndTransStr(pAction->stage), action, pRsp->code, pAction->acceptableCode, pAction->retryCode);
mndTransExecute(pMnode, pTrans);
_OVER:
@ -884,6 +907,21 @@ _OVER:
return 0;
}
static void mndTransResetAction(SMnode *pMnode, STrans *pTrans, STransAction *pAction) {
pAction->rawWritten = 0;
pAction->msgSent = 0;
pAction->msgReceived = 0;
if (pAction->errCode == TSDB_CODE_RPC_REDIRECT || pAction->errCode == TSDB_CODE_SYN_NEW_CONFIG_ERROR ||
pAction->errCode == TSDB_CODE_SYN_INTERNAL_ERROR || pAction->errCode == TSDB_CODE_SYN_NOT_LEADER) {
pAction->epSet.inUse = (pAction->epSet.inUse + 1) % pAction->epSet.numOfEps;
mDebug("trans:%d, %s:%d execute status is reset and set epset inuse:%d", pTrans->id, mndTransStr(pAction->stage),
pAction->id, pAction->epSet.inUse);
} else {
mDebug("trans:%d, %s:%d execute status is reset", pTrans->id, mndTransStr(pAction->stage), pAction->id);
}
pAction->errCode = 0;
}
static void mndTransResetActions(SMnode *pMnode, STrans *pTrans, SArray *pArray) {
int32_t numOfActions = taosArrayGetSize(pArray);
@ -894,18 +932,7 @@ static void mndTransResetActions(SMnode *pMnode, STrans *pTrans, SArray *pArray)
continue;
if (pAction->rawWritten && (pAction->errCode == 0 || pAction->errCode == pAction->acceptableCode)) continue;
pAction->rawWritten = 0;
pAction->msgSent = 0;
pAction->msgReceived = 0;
if (pAction->errCode == TSDB_CODE_RPC_REDIRECT || pAction->errCode == TSDB_CODE_SYN_NEW_CONFIG_ERROR ||
pAction->errCode == TSDB_CODE_SYN_INTERNAL_ERROR || pAction->errCode == TSDB_CODE_SYN_NOT_LEADER) {
pAction->epSet.inUse = (pAction->epSet.inUse + 1) % pAction->epSet.numOfEps;
mDebug("trans:%d, %s:%d execute status is reset and set epset inuse:%d", pTrans->id, mndTransStr(pAction->stage),
action, pAction->epSet.inUse);
} else {
mDebug("trans:%d, %s:%d execute status is reset", pTrans->id, mndTransStr(pAction->stage), action);
}
pAction->errCode = 0;
mndTransResetAction(pMnode, pTrans, pAction);
}
}
@ -1112,9 +1139,9 @@ static int32_t mndTransExecuteRedoActionsSerial(SMnode *pMnode, STrans *pTrans)
if (pAction->msgReceived) {
if (pAction->errCode != 0 && pAction->errCode != pAction->acceptableCode) {
code = pAction->errCode;
pAction->msgSent = 0;
pAction->msgReceived = 0;
mDebug("trans:%d, %s:%d execute status is reset", pTrans->id, mndTransStr(pAction->stage), action);
mndTransResetAction(pMnode, pTrans, pAction);
} else {
mDebug("trans:%d, %s:%d execute successfully", pTrans->id, mndTransStr(pAction->stage), action);
}
} else {
code = TSDB_CODE_ACTION_IN_PROGRESS;
@ -1123,6 +1150,8 @@ static int32_t mndTransExecuteRedoActionsSerial(SMnode *pMnode, STrans *pTrans)
if (pAction->rawWritten) {
if (pAction->errCode != 0 && pAction->errCode != pAction->acceptableCode) {
code = pAction->errCode;
} else {
mDebug("trans:%d, %s:%d write successfully", pTrans->id, mndTransStr(pAction->stage), action);
}
}
}
@ -1156,9 +1185,16 @@ static int32_t mndTransExecuteRedoActionsSerial(SMnode *pMnode, STrans *pTrans)
} else if (code == TSDB_CODE_ACTION_IN_PROGRESS) {
mDebug("trans:%d, %s:%d is in progress and wait it finish", pTrans->id, mndTransStr(pAction->stage), pAction->id);
break;
} else if (code == pAction->retryCode) {
mDebug("trans:%d, %s:%d receive code:0x%x and retry", pTrans->id, mndTransStr(pAction->stage), pAction->id, code);
taosMsleep(300);
action--;
continue;
} else {
terrno = code;
pTrans->code = code;
mDebug("trans:%d, %s:%d receive code:0x%x and wait another schedule, failedTimes:%d", pTrans->id,
mndTransStr(pAction->stage), pAction->id, code, pTrans->failedTimes);
break;
}
}
@ -1343,7 +1379,7 @@ void mndTransExecute(SMnode *pMnode, STrans *pTrans) {
mndTransSendRpcRsp(pMnode, pTrans);
}
static int32_t mndProcessTransReq(SRpcMsg *pReq) {
static int32_t mndProcessTransTimer(SRpcMsg *pReq) {
mndTransPullup(pReq->info.node);
return 0;
}

115
source/dnode/mnode/impl/src/mndVgroup.c
View File

@ -15,10 +15,10 @@
#define _DEFAULT_SOURCE
#include "mndVgroup.h"
#include "mndPrivilege.h"
#include "mndDb.h"
#include "mndDnode.h"
#include "mndMnode.h"
#include "mndPrivilege.h"
#include "mndShow.h"
#include "mndTrans.h"
#include "mndUser.h"
@ -896,6 +896,8 @@ int32_t mndAddAlterVnodeConfirmAction(SMnode *pMnode, STrans *pTrans, SDbObj *pD
action.pCont = pHead;
action.contLen = contLen;
action.msgType = TDMT_VND_ALTER_CONFIRM;
// incorrect redirect result will cause this erro
action.retryCode = TSDB_CODE_VND_INVALID_VGROUP_ID;
if (mndTransAppendRedoAction(pTrans, &action) != 0) {
taosMemoryFree(pHead);
@ -942,6 +944,8 @@ static int32_t mndAddSetVnodeStandByAction(SMnode *pMnode, STrans *pTrans, SDbOb
action.contLen = contLen;
action.msgType = TDMT_SYNC_SET_VNODE_STANDBY;
action.acceptableCode = TSDB_CODE_NODE_NOT_DEPLOYED;
// Keep retrying until the target vnode is not the leader
action.retryCode = TSDB_CODE_SYN_IS_LEADER;
if (isRedo) {
if (mndTransAppendRedoAction(pTrans, &action) != 0) {
@ -1003,7 +1007,7 @@ int32_t mndSetMoveVgroupInfoToTrans(SMnode *pMnode, STrans *pTrans, SDbObj *pDb,
mInfo("vgId:%d, will add 1 vnodes", pVgroup->vgId);
if (mndAddVnodeToVgroup(pMnode, &newVg, pArray) != 0) return -1;
if (mndAddCreateVnodeAction(pMnode, pTrans, pDb, &newVg, &newVg.vnodeGid[1], true) != 0) return -1;
if (mndAddCreateVnodeAction(pMnode, pTrans, pDb, &newVg, &newVg.vnodeGid[newVg.replica - 1], true) != 0) return -1;
if (mndAddAlterVnodeAction(pMnode, pTrans, pDb, &newVg, TDMT_VND_ALTER_REPLICA) != 0) return -1;
if (mndAddAlterVnodeConfirmAction(pMnode, pTrans, pDb, &newVg) != 0) return -1;
@ -1017,10 +1021,19 @@ int32_t mndSetMoveVgroupInfoToTrans(SMnode *pMnode, STrans *pTrans, SDbObj *pDb,
if (mndAddDropVnodeAction(pMnode, pTrans, pDb, &newVg, &del, true) != 0) return -1;
if (mndAddAlterVnodeConfirmAction(pMnode, pTrans, pDb, &newVg) != 0) return -1;
SSdbRaw *pRaw = mndVgroupActionEncode(&newVg);
if (pRaw == NULL || mndTransAppendCommitlog(pTrans, pRaw) != 0) return -1;
sdbSetRawStatus(pRaw, SDB_STATUS_READY);
pRaw = NULL;
{
SSdbRaw *pRaw = mndVgroupActionEncode(&newVg);
if (pRaw == NULL || mndTransAppendRedolog(pTrans, pRaw) != 0) return -1;
sdbSetRawStatus(pRaw, SDB_STATUS_READY);
pRaw = NULL;
}
{
SSdbRaw *pRaw = mndVgroupActionEncode(&newVg);
if (pRaw == NULL || mndTransAppendCommitlog(pTrans, pRaw) != 0) return -1;
sdbSetRawStatus(pRaw, SDB_STATUS_READY);
pRaw = NULL;
}
mInfo("vgId:%d, vgroup info after move, replica:%d", newVg.vgId, newVg.replica);
for (int32_t i = 0; i < newVg.replica; ++i) {
@ -1168,10 +1181,19 @@ static int32_t mndRedistributeVgroup(SMnode *pMnode, SRpcMsg *pReq, SDbObj *pDb,
if (mndAddDecVgroupReplicaFromTrans(pMnode, pTrans, pDb, &newVg, pOld3->id) != 0) goto _OVER;
}
pRaw = mndVgroupActionEncode(&newVg);
if (pRaw == NULL || mndTransAppendCommitlog(pTrans, pRaw) != 0) goto _OVER;
sdbSetRawStatus(pRaw, SDB_STATUS_READY);
pRaw = NULL;
{
pRaw = mndVgroupActionEncode(&newVg);
if (pRaw == NULL || mndTransAppendRedolog(pTrans, pRaw) != 0) goto _OVER;
sdbSetRawStatus(pRaw, SDB_STATUS_READY);
pRaw = NULL;
}
{
pRaw = mndVgroupActionEncode(&newVg);
if (pRaw == NULL || mndTransAppendCommitlog(pTrans, pRaw) != 0) goto _OVER;
sdbSetRawStatus(pRaw, SDB_STATUS_READY);
pRaw = NULL;
}
mInfo("vgId:%d, vgroup info after redistribute, replica:%d", newVg.vgId, newVg.replica);
for (int32_t i = 0; i < newVg.replica; ++i) {
@ -1229,7 +1251,8 @@ static int32_t mndProcessRedistributeVgroupMsg(SRpcMsg *pReq) {
}
if (req.dnodeId1 == pVgroup->vnodeGid[0].dnodeId) {
terrno = TSDB_CODE_MND_VGROUP_UN_CHANGED;
// terrno = TSDB_CODE_MND_VGROUP_UN_CHANGED;
code = 0;
goto _OVER;
}
@ -1351,7 +1374,8 @@ static int32_t mndProcessRedistributeVgroupMsg(SRpcMsg *pReq) {
}
if (pNew1 == NULL && pOld1 == NULL && pNew2 == NULL && pOld2 == NULL && pNew3 == NULL && pOld3 == NULL) {
terrno = TSDB_CODE_MND_VGROUP_UN_CHANGED;
// terrno = TSDB_CODE_MND_VGROUP_UN_CHANGED;
code = 0;
goto _OVER;
}
@ -1424,13 +1448,25 @@ int32_t mndBuildAlterVgroupAction(SMnode *pMnode, STrans *pTrans, SDbObj *pDb, S
} else {
}
SSdbRaw *pVgRaw = mndVgroupActionEncode(&newVgroup);
if (pVgRaw == NULL) return -1;
if (mndTransAppendCommitlog(pTrans, pVgRaw) != 0) {
sdbFreeRaw(pVgRaw);
return -1;
{
SSdbRaw *pVgRaw = mndVgroupActionEncode(&newVgroup);
if (pVgRaw == NULL) return -1;
if (mndTransAppendRedolog(pTrans, pVgRaw) != 0) {
sdbFreeRaw(pVgRaw);
return -1;
}
sdbSetRawStatus(pVgRaw, SDB_STATUS_READY);
}
{
SSdbRaw *pVgRaw = mndVgroupActionEncode(&newVgroup);
if (pVgRaw == NULL) return -1;
if (mndTransAppendCommitlog(pTrans, pVgRaw) != 0) {
sdbFreeRaw(pVgRaw);
return -1;
}
sdbSetRawStatus(pVgRaw, SDB_STATUS_READY);
}
sdbSetRawStatus(pVgRaw, SDB_STATUS_READY);
}
return 0;
@ -1538,12 +1574,23 @@ static int32_t mndSetBalanceVgroupInfoToTrans(SMnode *pMnode, STrans *pTrans, SD
if (mndAddIncVgroupReplicaToTrans(pMnode, pTrans, pDb, &newVg, pDst->id) != 0) return -1;
if (mndAddDecVgroupReplicaFromTrans(pMnode, pTrans, pDb, &newVg, pSrc->id) != 0) return -1;
SSdbRaw *pRaw = mndVgroupActionEncode(&newVg);
if (pRaw == NULL || mndTransAppendCommitlog(pTrans, pRaw) != 0) {
sdbFreeRaw(pRaw);
return -1;
{
SSdbRaw *pRaw = mndVgroupActionEncode(&newVg);
if (pRaw == NULL || mndTransAppendRedolog(pTrans, pRaw) != 0) {
sdbFreeRaw(pRaw);
return -1;
}
sdbSetRawStatus(pRaw, SDB_STATUS_READY);
}
{
SSdbRaw *pRaw = mndVgroupActionEncode(&newVg);
if (pRaw == NULL || mndTransAppendCommitlog(pTrans, pRaw) != 0) {
sdbFreeRaw(pRaw);
return -1;
}
sdbSetRawStatus(pRaw, SDB_STATUS_READY);
}
sdbSetRawStatus(pRaw, SDB_STATUS_READY);
mInfo("vgId:%d, vgroup info after balance, replica:%d", newVg.vgId, newVg.replica);
for (int32_t i = 0; i < newVg.replica; ++i) {
@ -1552,7 +1599,8 @@ static int32_t mndSetBalanceVgroupInfoToTrans(SMnode *pMnode, STrans *pTrans, SD
return 0;
}
static int32_t mndBalanceVgroupBetweenDnode(SMnode *pMnode, STrans *pTrans, SDnodeObj *pSrc, SDnodeObj *pDst) {
static int32_t mndBalanceVgroupBetweenDnode(SMnode *pMnode, STrans *pTrans, SDnodeObj *pSrc, SDnodeObj *pDst,
SHashObj *pBalancedVgroups) {
void *pIter = NULL;
int32_t code = -1;
SSdb *pSdb = pMnode->pSdb;
@ -1561,6 +1609,10 @@ static int32_t mndBalanceVgroupBetweenDnode(SMnode *pMnode, STrans *pTrans, SDno
SVgObj *pVgroup = NULL;
pIter = sdbFetch(pSdb, SDB_VGROUP, pIter, (void **)&pVgroup);
if (pIter == NULL) break;
if (taosHashGet(pBalancedVgroups, &pVgroup->vgId, sizeof(int32_t)) != NULL) {
sdbRelease(pSdb, pVgroup);
continue;
}
bool existInSrc = false;
bool existInDst = false;
@ -1577,6 +1629,9 @@ static int32_t mndBalanceVgroupBetweenDnode(SMnode *pMnode, STrans *pTrans, SDno
SDbObj *pDb = mndAcquireDb(pMnode, pVgroup->dbName);
code = mndSetBalanceVgroupInfoToTrans(pMnode, pTrans, pDb, pVgroup, pSrc, pDst);
if (code == 0) {
code = taosHashPut(pBalancedVgroups, &pVgroup->vgId, sizeof(int32_t), &pVgroup->vgId, sizeof(int32_t));
}
mndReleaseDb(pMnode, pDb);
sdbRelease(pSdb, pVgroup);
sdbCancelFetch(pSdb, pIter);
@ -1590,6 +1645,10 @@ static int32_t mndBalanceVgroup(SMnode *pMnode, SRpcMsg *pReq, SArray *pArray) {
int32_t code = -1;
int32_t numOfVgroups = 0;
STrans *pTrans = NULL;
SHashObj *pBalancedVgroups = NULL;
pBalancedVgroups = taosHashInit(16, taosGetDefaultHashFunction(TSDB_DATA_TYPE_INT), false, HASH_NO_LOCK);
if (pBalancedVgroups == NULL) goto _OVER;
pTrans = mndTransCreate(pMnode, TRN_POLICY_RETRY, TRN_CONFLICT_GLOBAL, pReq);
if (pTrans == NULL) goto _OVER;
@ -1613,18 +1672,18 @@ static int32_t mndBalanceVgroup(SMnode *pMnode, SRpcMsg *pReq, SArray *pArray) {
pDst->id, dstScore);
if (srcScore > dstScore - 0.000001) {
code = mndBalanceVgroupBetweenDnode(pMnode, pTrans, pSrc, pDst);
code = mndBalanceVgroupBetweenDnode(pMnode, pTrans, pSrc, pDst, pBalancedVgroups);
if (code == 0) {
pSrc->numOfVnodes--;
pDst->numOfVnodes++;
numOfVgroups++;
continue;
} else {
mError("trans:%d, failed to balance vgroup from dnode:%d to dnode:%d", pTrans->id, pSrc->id, pDst->id);
return -1;
mDebug("trans:%d, no vgroup need to balance from dnode:%d to dnode:%d", pTrans->id, pSrc->id, pDst->id);
break;
}
} else {
mDebug("trans:%d, no vgroup need to balance vgroup any more", pTrans->id);
mDebug("trans:%d, no vgroup need to balance any more", pTrans->id);
break;
}
}

6
source/dnode/vnode/inc/vnode.h
View File

@ -125,6 +125,7 @@ int32_t tsdbGetFileBlocksDistInfo(tsdbReaderT *pReader, STableBlockDistInfo
bool isTsdbCacheLastRow(tsdbReaderT *pReader);
int32_t tsdbGetAllTableList(SMeta *pMeta, uint64_t uid, SArray *list);
int32_t tsdbGetCtbIdList(SMeta *pMeta, int64_t suid, SArray *list);
int32_t tsdbGetStbIdList(SMeta *pMeta, int64_t suid, SArray *list);
void *tsdbGetIdx(SMeta *pMeta);
void *tsdbGetIvtIdx(SMeta *pMeta);
int64_t tsdbGetNumOfRowsInMemTable(tsdbReaderT *pHandle);
@ -198,15 +199,20 @@ typedef struct {
uint64_t groupId;
} STableKeyInfo;
#define TABLE_ROLLUP_ON ((int8_t)0x1)
#define TABLE_IS_ROLLUP(FLG) (((FLG) & (TABLE_ROLLUP_ON)) != 0)
#define TABLE_SET_ROLLUP(FLG) ((FLG) |= TABLE_ROLLUP_ON)
struct SMetaEntry {
int64_t version;
int8_t type;
int8_t flags; // TODO: need refactor?
tb_uid_t uid;
char *name;
union {
struct {
SSchemaWrapper schemaRow;
SSchemaWrapper schemaTag;
SRSmaParam rsmaParam;
} stbEntry;
struct {
int64_t ctime;

1
source/dnode/vnode/src/inc/meta.h
View File

@ -69,6 +69,7 @@ struct SMeta {
TTB* pUidIdx;
TTB* pNameIdx;
TTB* pCtbIdx;
TTB* pSuidIdx;
// ivt idx and idx
void* pTagIvtIdx;
TTB* pTagIdx;

41
source/dnode/vnode/src/inc/sma.h
View File

@ -60,8 +60,9 @@ struct SRSmaStat {
SSma *pSma;
void *tmrHandle;
tmr_h tmrId;
int8_t tmrStat;
int32_t tmrSeconds;
int8_t triggerStat;
int8_t runningStat;
SHashObj *rsmaInfoHash; // key: stbUid, value: SRSmaInfo;
};
@ -72,13 +73,21 @@ struct SSmaStat {
};
T_REF_DECLARE()
};
#define SMA_TSMA_STAT(s) (&(s)->tsmaStat)
#define SMA_RSMA_STAT(s) (&(s)->rsmaStat)
#define SMA_RSMA_INFO_HASH(s) ((s)->rsmaStat.rsmaInfoHash)
#define SMA_RSMA_TMR_HANDLE(s) ((s)->rsmaStat.tmrHandle)
#define SMA_RSMA_TMR_STAT(s) ((s)->rsmaStat.tmrStat)
#define RSMA_INFO_HASH(r) ((r)->rsmaInfoHash)
#define SMA_TSMA_STAT(s) (&(s)->tsmaStat)
#define SMA_RSMA_STAT(s) (&(s)->rsmaStat)
#define RSMA_INFO_HASH(r) ((r)->rsmaInfoHash)
#define RSMA_TMR_ID(r) ((r)->tmrId)
#define RSMA_TMR_HANDLE(r) ((r)->tmrHandle)
#define RSMA_TRIGGER_STAT(r) (&(r)->triggerStat)
#define RSMA_RUNNING_STAT(r) (&(r)->runningStat)
enum {
TASK_TRIGGER_STAT_INIT = 0,
TASK_TRIGGER_STAT_ACTIVE = 1,
TASK_TRIGGER_STAT_INACTIVE = 2,
TASK_TRIGGER_STAT_CANCELLED = 3,
TASK_TRIGGER_STAT_FINISHED = 4,
};
void tdDestroySmaEnv(SSmaEnv *pSmaEnv);
void *tdFreeSmaEnv(SSmaEnv *pSmaEnv);
@ -167,13 +176,18 @@ static FORCE_INLINE void tdSmaStatSetDropped(STSmaStat *pTStat) {
static int32_t tdDestroySmaState(SSmaStat *pSmaStat, int8_t smaType);
void *tdFreeSmaState(SSmaStat *pSmaStat, int8_t smaType);
void *tdFreeRSmaInfo(SRSmaInfo *pInfo);
void *tdFreeRSmaInfo(SRSmaInfo *pInfo);
int32_t tdProcessRSmaCreateImpl(SSma *pSma, SRSmaParam *param, int64_t suid, const char *tbName);
int32_t tdProcessRSmaRestoreImpl(SSma *pSma);
int32_t tdProcessTSmaCreateImpl(SSma *pSma, int64_t version, const char *pMsg);
int32_t tdProcessTSmaInsertImpl(SSma *pSma, int64_t indexUid, const char *msg);
int32_t tdProcessTSmaGetDaysImpl(SVnodeCfg *pCfg, void *pCont, uint32_t contLen, int32_t *days);
// smaFileUtil ================
#define TD_FILE_HEAD_SIZE 512
typedef struct STFInfo STFInfo;
typedef struct STFile STFile;
@ -181,7 +195,7 @@ struct STFInfo {
uint32_t magic;
uint32_t ftype;
uint32_t fver;
uint64_t fsize;
int64_t fsize;
};
struct STFile {
@ -199,11 +213,8 @@ struct STFile {
#define TD_FILE_OPENED(tf) (TD_FILE_PFILE(tf) != NULL)
#define TD_FILE_CLOSED(tf) (!TD_FILE_OPENED(tf))
#define TD_FILE_SET_CLOSED(f) (TD_FILE_PFILE(f) = NULL)
#define TD_FILE_STATE(tf) ((tf)->state)
#define TD_FILE_SET_STATE(tf, s) ((tf)->state = (s))
#define TD_FILE_DID(tf) (TD_FILE_F(tf)->did)
#define TD_FILE_IS_OK(tf) (TD_FILE_STATE(tf) == TD_FILE_STATE_OK)
#define TD_FILE_IS_BAD(tf) (TD_FILE_STATE(tf) == TD_FILE_STATE_BAD)
int32_t tdInitTFile(STFile *pTFile, STfs *pTfs, const char *fname);
int32_t tdCreateTFile(STFile *pTFile, STfs *pTfs, bool updateHeader, int8_t fType);
@ -212,12 +223,14 @@ int64_t tdReadTFile(STFile *pTFile, void *buf, int64_t nbyte);
int64_t tdSeekTFile(STFile *pTFile, int64_t offset, int whence);
int64_t tdWriteTFile(STFile *pTFile, void *buf, int64_t nbyte);
int64_t tdAppendTFile(STFile *pTFile, void *buf, int64_t nbyte, int64_t *offset);
int64_t tdGetTFileSize(STFile *pTFile, int64_t *size);
int32_t tdRemoveTFile(STFile *pTFile);
int32_t tdLoadTFileHeader(STFile *pTFile, STFInfo *pInfo);
int32_t tdUpdateTFileHeader(STFile *pTFile);
void tdUpdateTFileMagic(STFile *pTFile, void *pCksm);
void tdCloseTFile(STFile *pTFile);
void tdGetVndFileName(int32_t vid, const char *dname, const char *fname, char *outputName);
void tdGetVndFileName(int32_t vid, const char *dname, const char *fname, char *outputName);
#ifdef __cplusplus
}

6
source/dnode/vnode/src/inc/vnodeInt.h
View File

@ -76,6 +76,7 @@ void vnodeFree(void* p);
// meta
typedef struct SMCtbCursor SMCtbCursor;
typedef struct SMStbCursor SMStbCursor;
typedef struct STbUidStore STbUidStore;
int metaOpen(SVnode* pVnode, SMeta** ppMeta);
@ -97,6 +98,9 @@ int metaGetTbNum(SMeta* pMeta);
SMCtbCursor* metaOpenCtbCursor(SMeta* pMeta, tb_uid_t uid);
void metaCloseCtbCursor(SMCtbCursor* pCtbCur);
tb_uid_t metaCtbCursorNext(SMCtbCursor* pCtbCur);
SMStbCursor* metaOpenStbCursor(SMeta* pMeta, tb_uid_t uid);
void metaCloseStbCursor(SMStbCursor* pStbCur);
tb_uid_t metaStbCursorNext(SMStbCursor* pStbCur);
STSma* metaGetSmaInfoByIndex(SMeta* pMeta, int64_t indexUid);
STSmaWrapper* metaGetSmaInfoByTable(SMeta* pMeta, tb_uid_t uid, bool deepCopy);
SArray* metaGetSmaIdsByTable(SMeta* pMeta, tb_uid_t uid);
@ -158,6 +162,8 @@ SSubmitReq* tdBlockToSubmit(const SArray* pBlocks, const STSchema* pSchema, bool
// sma
int32_t smaOpen(SVnode* pVnode);
int32_t smaClose(SSma* pSma);
int32_t smaCloseEnv(SSma* pSma);
int32_t smaCloseEx(SSma* pSma);
int32_t tdProcessTSmaCreate(SSma* pSma, int64_t version, const char* msg);
int32_t tdProcessTSmaInsert(SSma* pSma, int64_t indexUid, const char* msg);

18
source/dnode/vnode/src/meta/metaEntry.c
View File

@ -24,22 +24,25 @@ int metaEncodeEntry(SEncoder *pCoder, const SMetaEntry *pME) {
if (tEncodeCStr(pCoder, pME->name) < 0) return -1;
if (pME->type == TSDB_SUPER_TABLE) {
if (tEncodeI8(pCoder, pME->flags) < 0) return -1; // TODO: need refactor?
if (tEncodeSSchemaWrapper(pCoder, &pME->stbEntry.schemaRow) < 0) return -1;
if (tEncodeSSchemaWrapper(pCoder, &pME->stbEntry.schemaTag) < 0) return -1;
if (TABLE_IS_ROLLUP(pME->flags)) {
if (tEncodeSRSmaParam(pCoder, &pME->stbEntry.rsmaParam) < 0) return -1;
}
} else if (pME->type == TSDB_CHILD_TABLE) {
if (tEncodeI64(pCoder, pME->ctbEntry.ctime) < 0) return -1;
if (tEncodeI32(pCoder, pME->ctbEntry.ttlDays) < 0) return -1;
if (tEncodeI32(pCoder, pME->ctbEntry.commentLen) < 0) return -1;
if (tEncodeI32v(pCoder, pME->ctbEntry.commentLen) < 0) return -1;
if (pME->ctbEntry.commentLen > 0){
if (tEncodeCStr(pCoder, pME->ctbEntry.comment) < 0) return -1;
}
if (tEncodeI64(pCoder, pME->ctbEntry.suid) < 0) return -1;
debugCheckTags((STag*)pME->ctbEntry.pTags); // TODO: remove after debug
if (tEncodeTag(pCoder, (const STag *)pME->ctbEntry.pTags) < 0) return -1;
} else if (pME->type == TSDB_NORMAL_TABLE) {
if (tEncodeI64(pCoder, pME->ntbEntry.ctime) < 0) return -1;
if (tEncodeI32(pCoder, pME->ntbEntry.ttlDays) < 0) return -1;
if (tEncodeI32(pCoder, pME->ntbEntry.commentLen) < 0) return -1;
if (tEncodeI32v(pCoder, pME->ntbEntry.commentLen) < 0) return -1;
if (pME->ntbEntry.commentLen > 0){
if (tEncodeCStr(pCoder, pME->ntbEntry.comment) < 0) return -1;
}
@ -64,23 +67,26 @@ int metaDecodeEntry(SDecoder *pCoder, SMetaEntry *pME) {
if (tDecodeCStr(pCoder, &pME->name) < 0) return -1;
if (pME->type == TSDB_SUPER_TABLE) {
if (tDecodeI8(pCoder, &pME->flags) < 0) return -1; // TODO: need refactor?
if (tDecodeSSchemaWrapperEx(pCoder, &pME->stbEntry.schemaRow) < 0) return -1;
if (tDecodeSSchemaWrapperEx(pCoder, &pME->stbEntry.schemaTag) < 0) return -1;
if (TABLE_IS_ROLLUP(pME->flags)) {
if (tDecodeSRSmaParam(pCoder, &pME->stbEntry.rsmaParam) < 0) return -1;
}
} else if (pME->type == TSDB_CHILD_TABLE) {
if (tDecodeI64(pCoder, &pME->ctbEntry.ctime) < 0) return -1;
if (tDecodeI32(pCoder, &pME->ctbEntry.ttlDays) < 0) return -1;
if (tDecodeI32(pCoder, &pME->ctbEntry.commentLen) < 0) return -1;
if (tDecodeI32v(pCoder, &pME->ctbEntry.commentLen) < 0) return -1;
if (pME->ctbEntry.commentLen > 0){
if (tDecodeCStr(pCoder, &pME->ctbEntry.comment) < 0)
return -1;
}
if (tDecodeI64(pCoder, &pME->ctbEntry.suid) < 0) return -1;
if (tDecodeTag(pCoder, (STag **)&pME->ctbEntry.pTags) < 0) return -1; // (TODO)
debugCheckTags((STag*)pME->ctbEntry.pTags); // TODO: remove after debug
} else if (pME->type == TSDB_NORMAL_TABLE) {
if (tDecodeI64(pCoder, &pME->ntbEntry.ctime) < 0) return -1;
if (tDecodeI32(pCoder, &pME->ntbEntry.ttlDays) < 0) return -1;
if (tDecodeI32(pCoder, &pME->ntbEntry.commentLen) < 0) return -1;
if (tDecodeI32v(pCoder, &pME->ntbEntry.commentLen) < 0) return -1;
if (pME->ntbEntry.commentLen > 0){
if (tDecodeCStr(pCoder, &pME->ntbEntry.comment) < 0) return -1;
}

9
source/dnode/vnode/src/meta/metaOpen.c
View File

@ -92,6 +92,13 @@ int metaOpen(SVnode *pVnode, SMeta **ppMeta) {
goto _err;
}
// open pSuidIdx
ret = tdbTbOpen("suid.idx", sizeof(tb_uid_t), 0, uidIdxKeyCmpr, pMeta->pEnv, &pMeta->pSuidIdx);
if (ret < 0) {
metaError("vgId:%d, failed to open meta super table index since %s", TD_VID(pVnode), tstrerror(terrno));
goto _err;
}
// open pTagIdx
// TODO(yihaoDeng), refactor later
char indexFullPath[128] = {0};
@ -141,6 +148,7 @@ _err:
if (pMeta->pTagIvtIdx) indexClose(pMeta->pTagIvtIdx);
if (pMeta->pTagIdx) tdbTbClose(pMeta->pTagIdx);
if (pMeta->pCtbIdx) tdbTbClose(pMeta->pCtbIdx);
if (pMeta->pSuidIdx) tdbTbClose(pMeta->pSuidIdx);
if (pMeta->pNameIdx) tdbTbClose(pMeta->pNameIdx);
if (pMeta->pUidIdx) tdbTbClose(pMeta->pUidIdx);
if (pMeta->pSkmDb) tdbTbClose(pMeta->pSkmDb);
@ -162,6 +170,7 @@ int metaClose(SMeta *pMeta) {
if (pMeta->pTagIdx) tdbTbClose(pMeta->pTagIdx);
#endif
if (pMeta->pCtbIdx) tdbTbClose(pMeta->pCtbIdx);
if (pMeta->pSuidIdx) tdbTbClose(pMeta->pSuidIdx);
if (pMeta->pNameIdx) tdbTbClose(pMeta->pNameIdx);
if (pMeta->pUidIdx) tdbTbClose(pMeta->pUidIdx);
if (pMeta->pSkmDb) tdbTbClose(pMeta->pSkmDb);

66
source/dnode/vnode/src/meta/metaQuery.c
View File

@ -325,6 +325,72 @@ tb_uid_t metaCtbCursorNext(SMCtbCursor *pCtbCur) {
return pCtbIdxKey->uid;
}
struct SMStbCursor {
SMeta *pMeta;
TBC *pCur;
tb_uid_t suid;
void *pKey;
void *pVal;
int kLen;
int vLen;
};
SMStbCursor *metaOpenStbCursor(SMeta *pMeta, tb_uid_t suid) {
SMStbCursor *pStbCur = NULL;
int ret = 0;
int c = 0;
pStbCur = (SMStbCursor *)taosMemoryCalloc(1, sizeof(*pStbCur));
if (pStbCur == NULL) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
return NULL;
}
pStbCur->pMeta = pMeta;
pStbCur->suid = suid;
metaRLock(pMeta);
ret = tdbTbcOpen(pMeta->pSuidIdx, &pStbCur->pCur, NULL);
if (ret < 0) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
metaULock(pMeta);
taosMemoryFree(pStbCur);
return NULL;
}
// move to the suid
tdbTbcMoveTo(pStbCur->pCur, &suid, sizeof(suid), &c);
if (c > 0) {
tdbTbcMoveToNext(pStbCur->pCur);
}
return pStbCur;
}
void metaCloseStbCursor(SMStbCursor *pStbCur) {
if (pStbCur) {
if (pStbCur->pMeta) metaULock(pStbCur->pMeta);
if (pStbCur->pCur) {
tdbTbcClose(pStbCur->pCur);
tdbFree(pStbCur->pKey);
tdbFree(pStbCur->pVal);
}
taosMemoryFree(pStbCur);
}
}
tb_uid_t metaStbCursorNext(SMStbCursor *pStbCur) {
int ret;
ret = tdbTbcNext(pStbCur->pCur, &pStbCur->pKey, &pStbCur->kLen, &pStbCur->pVal, &pStbCur->vLen);
if (ret < 0) {
return 0;
}
return *(tb_uid_t*)pStbCur->pKey;
}
STSchema *metaGetTbTSchema(SMeta *pMeta, tb_uid_t uid, int32_t sver) {
// SMetaReader mr = {0};
STSchema * pTSchema = NULL;

16
source/dnode/vnode/src/meta/metaTable.c
View File

@ -23,6 +23,7 @@ static int metaUpdateNameIdx(SMeta *pMeta, const SMetaEntry *pME);
static int metaUpdateTtlIdx(SMeta *pMeta, const SMetaEntry *pME);
static int metaSaveToSkmDb(SMeta *pMeta, const SMetaEntry *pME);
static int metaUpdateCtbIdx(SMeta *pMeta, const SMetaEntry *pME);
static int metaUpdateSuidIdx(SMeta *pMeta, const SMetaEntry *pME);
static int metaUpdateTagIdx(SMeta *pMeta, const SMetaEntry *pCtbEntry);
static int metaDropTableByUid(SMeta *pMeta, tb_uid_t uid, int *type);
@ -138,6 +139,10 @@ int metaCreateSTable(SMeta *pMeta, int64_t version, SVCreateStbReq *pReq) {
me.name = pReq->name;
me.stbEntry.schemaRow = pReq->schemaRow;
me.stbEntry.schemaTag = pReq->schemaTag;
if (pReq->rollup) {
TABLE_SET_ROLLUP(me.flags);
me.stbEntry.rsmaParam = pReq->rsmaParam;
}
if (metaHandleEntry(pMeta, &me) < 0) goto _err;
@ -209,6 +214,7 @@ _drop_super_table:
&pMeta->txn);
tdbTbDelete(pMeta->pNameIdx, pReq->name, strlen(pReq->name) + 1, &pMeta->txn);
tdbTbDelete(pMeta->pUidIdx, &pReq->suid, sizeof(tb_uid_t), &pMeta->txn);
tdbTbDelete(pMeta->pSuidIdx, &pReq->suid, sizeof(tb_uid_t), &pMeta->txn);
metaULock(pMeta);
@ -436,11 +442,13 @@ static int metaDropTableByUid(SMeta *pMeta, tb_uid_t uid, int *type) {
tdbTbDelete(pMeta->pUidIdx, &uid, sizeof(uid), &pMeta->txn);
if(e.type != TSDB_SUPER_TABLE) metaDeleteTtlIdx(pMeta, &e);
if (e.type == TSDB_CHILD_TABLE) {
tdbTbDelete(pMeta->pCtbIdx, &(SCtbIdxKey){.suid = e.ctbEntry.suid, .uid = uid}, sizeof(SCtbIdxKey), &pMeta->txn);
} else if (e.type == TSDB_NORMAL_TABLE) {
// drop schema.db (todo)
} else if (e.type == TSDB_SUPER_TABLE) {
tdbTbDelete(pMeta->pSuidIdx, &e.uid, sizeof(tb_uid_t), &pMeta->txn);
// drop schema.db (todo)
}
@ -911,6 +919,10 @@ static int metaUpdateUidIdx(SMeta *pMeta, const SMetaEntry *pME) {
return tdbTbInsert(pMeta->pUidIdx, &pME->uid, sizeof(tb_uid_t), &pME->version, sizeof(int64_t), &pMeta->txn);
}
static int metaUpdateSuidIdx(SMeta *pMeta, const SMetaEntry *pME) {
return tdbTbInsert(pMeta->pSuidIdx, &pME->uid, sizeof(tb_uid_t), NULL, 0, &pMeta->txn);
}
static int metaUpdateNameIdx(SMeta *pMeta, const SMetaEntry *pME) {
return tdbTbInsert(pMeta->pNameIdx, pME->name, strlen(pME->name) + 1, &pME->uid, sizeof(tb_uid_t), &pMeta->txn);
}
@ -1081,6 +1093,10 @@ static int metaHandleEntry(SMeta *pMeta, const SMetaEntry *pME) {
} else {
// update schema.db
if (metaSaveToSkmDb(pMeta, pME) < 0) goto _err;
if (pME->type == TSDB_SUPER_TABLE) {
if (metaUpdateSuidIdx(pMeta, pME) < 0) goto _err;
}
}
if (pME->type != TSDB_SUPER_TABLE) {

136
source/dnode/vnode/src/sma/smaEnv.c
View File

@ -126,22 +126,21 @@ static int32_t tdInitSmaStat(SSmaStat **pSmaStat, int8_t smaType, const SSma *pS
}
if (smaType == TSDB_SMA_TYPE_ROLLUP) {
SMA_RSMA_STAT(*pSmaStat)->pSma = (SSma*)pSma;
SRSmaStat *pRSmaStat = (SRSmaStat *)(*pSmaStat);
pRSmaStat->pSma = (SSma *)pSma;
// init timer
SMA_RSMA_TMR_HANDLE(*pSmaStat) = taosTmrInit(10000, 100, 10000, "RSMA_G");
if (!SMA_RSMA_TMR_HANDLE(*pSmaStat)) {
RSMA_TMR_HANDLE(pRSmaStat) = taosTmrInit(10000, 100, 10000, "RSMA");
if (!RSMA_TMR_HANDLE(pRSmaStat)) {
taosMemoryFreeClear(*pSmaStat);
return TSDB_CODE_FAILED;
}
atomic_store_8(&SMA_RSMA_TMR_STAT(*pSmaStat), TASK_TRIGGER_STATUS__ACTIVE);
// init hash
SMA_RSMA_INFO_HASH(*pSmaStat) = taosHashInit(
RSMA_INFO_HASH(pRSmaStat) = taosHashInit(
RSMA_TASK_INFO_HASH_SLOT, taosGetDefaultHashFunction(TSDB_DATA_TYPE_BIGINT), true, HASH_ENTRY_LOCK);
if (!SMA_RSMA_INFO_HASH(*pSmaStat)) {
if (SMA_RSMA_TMR_HANDLE(*pSmaStat)) {
taosTmrCleanUp(SMA_RSMA_TMR_HANDLE(*pSmaStat));
if (!RSMA_INFO_HASH(pRSmaStat)) {
if (RSMA_TMR_HANDLE(pRSmaStat)) {
taosTmrCleanUp(RSMA_TMR_HANDLE(pRSmaStat));
}
taosMemoryFreeClear(*pSmaStat);
return TSDB_CODE_FAILED;
@ -155,12 +154,80 @@ static int32_t tdInitSmaStat(SSmaStat **pSmaStat, int8_t smaType, const SSma *pS
return TSDB_CODE_SUCCESS;
}
static void *tdFreeTSmaStat(STSmaStat *pStat) {
static void tdDestroyTSmaStat(STSmaStat *pStat) {
if (pStat) {
smaDebug("destroy tsma stat");
tDestroyTSma(pStat->pTSma);
taosMemoryFreeClear(pStat->pTSma);
taosMemoryFreeClear(pStat);
taosMemoryFreeClear(pStat->pTSchema);
}
}
static void *tdFreeTSmaStat(STSmaStat *pStat) {
tdDestroyTSmaStat(pStat);
taosMemoryFreeClear(pStat);
return NULL;
}
static void tdDestroyRSmaStat(SRSmaStat *pStat) {
if (pStat) {
smaDebug("vgId:%d destroy rsma stat", SMA_VID(pStat->pSma));
// step 1: set persistence task cancelled
atomic_store_8(RSMA_TRIGGER_STAT(pStat), TASK_TRIGGER_STAT_CANCELLED);
// step 2: stop the persistence timer
taosTmrStopA(&RSMA_TMR_ID(pStat));
// step 3: wait the persistence thread to finish
int32_t nLoops = 0;
if (atomic_load_8(RSMA_RUNNING_STAT(pStat)) == 1) {
while (1) {
if (atomic_load_8(RSMA_TRIGGER_STAT(pStat)) == TASK_TRIGGER_STAT_FINISHED) {
break;
} else {
smaDebug("not destroyed since rsma stat in %" PRIi8, atomic_load_8(RSMA_TRIGGER_STAT(pStat)));
}
++nLoops;
if (nLoops > 1000) {
sched_yield();
nLoops = 0;
}
}
}
// step 4: destroy the rsma info and associated fetch tasks
// TODO: use taosHashSetFreeFp when taosHashSetFreeFp is ready.
void *infoHash = taosHashIterate(RSMA_INFO_HASH(pStat), NULL);
while (infoHash) {
SRSmaInfo *pSmaInfo = *(SRSmaInfo **)infoHash;
tdFreeRSmaInfo(pSmaInfo);
infoHash = taosHashIterate(RSMA_INFO_HASH(pStat), infoHash);
}
taosHashCleanup(RSMA_INFO_HASH(pStat));
// step 5: wait all triggered fetch tasks finished
nLoops = 0;
while (1) {
if (T_REF_VAL_GET((SSmaStat *)pStat) == 0) {
break;
}
++nLoops;
if (nLoops > 1000) {
sched_yield();
nLoops = 0;
}
}
// step 6: cleanup the timer handle
if (RSMA_TMR_HANDLE(pStat)) {
taosTmrCleanUp(RSMA_TMR_HANDLE(pStat));
}
}
}
static void *tdFreeRSmaStat(SRSmaStat *pStat) {
tdDestroyRSmaStat(pStat);
taosMemoryFreeClear(pStat);
return NULL;
}
@ -179,19 +246,9 @@ void *tdFreeSmaState(SSmaStat *pSmaStat, int8_t smaType) {
int32_t tdDestroySmaState(SSmaStat *pSmaStat, int8_t smaType) {
if (pSmaStat) {
if (smaType == TSDB_SMA_TYPE_TIME_RANGE) {
tdFreeTSmaStat(&pSmaStat->tsmaStat);
tdDestroyTSmaStat(SMA_TSMA_STAT(pSmaStat));
} else if (smaType == TSDB_SMA_TYPE_ROLLUP) {
if (SMA_RSMA_TMR_HANDLE(pSmaStat)) {
taosTmrCleanUp(SMA_RSMA_TMR_HANDLE(pSmaStat));
}
// TODO: use taosHashSetFreeFp when taosHashSetFreeFp is ready.
void *infoHash = taosHashIterate(SMA_RSMA_INFO_HASH(pSmaStat), NULL);
while (infoHash) {
SRSmaInfo *pInfoHash = *(SRSmaInfo **)infoHash;
tdFreeRSmaInfo(pInfoHash);
infoHash = taosHashIterate(SMA_RSMA_INFO_HASH(pSmaStat), infoHash);
}
taosHashCleanup(SMA_RSMA_INFO_HASH(pSmaStat));
tdDestroyRSmaStat(SMA_RSMA_STAT(pSmaStat));
} else {
ASSERT(0);
}
@ -259,35 +316,4 @@ int32_t tdCheckAndInitSmaEnv(SSma *pSma, int8_t smaType) {
tdUnLockSma(pSma);
return TSDB_CODE_SUCCESS;
};
int32_t smaTimerInit(void **timer, int8_t *initFlag, const char *label) {
int8_t old;
while (1) {
old = atomic_val_compare_exchange_8(initFlag, 0, 2);
if (old != 2) break;
}
if (old == 0) {
*timer = taosTmrInit(10000, 100, 10000, label);
if (!(*timer)) {
atomic_store_8(initFlag, 0);
return -1;
}
atomic_store_8(initFlag, 1);
}
return 0;
}
void smaTimerCleanUp(void *timer, int8_t *initFlag) {
int8_t old;
while (1) {
old = atomic_val_compare_exchange_8(initFlag, 1, 2);
if (old != 2) break;
}
if (old == 1) {
taosTmrCleanUp(timer);
atomic_store_8(initFlag, 0);
}
}
};

46
source/dnode/vnode/src/sma/smaOpen.c
View File

@ -18,6 +18,7 @@
static int32_t smaEvalDays(SRetention *r, int8_t precision);
static int32_t smaSetKeepCfg(STsdbKeepCfg *pKeepCfg, STsdbCfg *pCfg, int type);
static int32_t rsmaRestore(SSma *pSma);
#define SMA_SET_KEEP_CFG(l) \
do { \
@ -100,6 +101,9 @@ int32_t smaOpen(SVnode *pVnode) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
return -1;
}
pVnode->pSma = pSma;
pSma->pVnode = pVnode;
taosThreadMutexInit(&pSma->mutex, NULL);
pSma->locked = false;
@ -117,37 +121,53 @@ int32_t smaOpen(SVnode *pVnode) {
ASSERT(0);
}
}
// restore the rsma
#if 0
if (rsmaRestore(pSma) < 0) {
goto _err;
}
#endif
}
pVnode->pSma = pSma;
return 0;
_err:
taosMemoryFreeClear(pSma);
return -1;
}
int32_t smaClose(SSma *pSma) {
int32_t smaCloseEnv(SSma *pSma) {
if (pSma) {
SMA_TSMA_ENV(pSma) = tdFreeSmaEnv(SMA_TSMA_ENV(pSma));
SMA_RSMA_ENV(pSma) = tdFreeSmaEnv(SMA_RSMA_ENV(pSma));
}
return 0;
}
int32_t smaCloseEx(SSma *pSma) {
if (pSma) {
taosThreadMutexDestroy(&pSma->mutex);
if SMA_RSMA_TSDB0 (pSma) tsdbClose(&SMA_RSMA_TSDB0(pSma));
if SMA_RSMA_TSDB1 (pSma) tsdbClose(&SMA_RSMA_TSDB1(pSma));
if SMA_RSMA_TSDB2 (pSma) tsdbClose(&SMA_RSMA_TSDB2(pSma));
// SMA_TSMA_ENV(pSma) = tdFreeSmaEnv(SMA_TSMA_ENV(pSma));
// SMA_RSMA_ENV(pSma) = tdFreeSmaEnv(SMA_RSMA_ENV(pSma));
taosMemoryFreeClear(pSma);
}
return 0;
}
int32_t smaClose(SSma *pSma) {
smaCloseEnv(pSma);
smaCloseEx(pSma);
return 0;
}
/**
* @brief rsma env restore
*
* @param pSma
* @return int32_t
*
* @param pSma
* @return int32_t
*/
int32_t smaRestore(SSma *pSma) {
if (!pSma) return 0;
// iterate all stables to restore the rsma env
return TSDB_CODE_SUCCESS;
static int32_t rsmaRestore(SSma *pSma) {
ASSERT(VND_IS_RSMA(pSma->pVnode));
return tdProcessRSmaRestoreImpl(pSma);
}

702
source/dnode/vnode/src/sma/smaRollup.c
View File

@ -14,11 +14,15 @@
*/
#include "sma.h"
#include "tstream.h"
#define RSMA_QTASKINFO_PERSIST_MS 7200000
#define RSMA_QTASKINFO_BUFSIZE 32768
typedef enum { TD_QTASK_TMP_FILE = 0, TD_QTASK_CUR_FILE } TD_QTASK_FILE_T;
static const char *tdQTaskInfoFname[] = {"qtaskinfo.t", "qtaskinfo"};
typedef struct SRSmaQTaskInfoItem SRSmaQTaskInfoItem;
typedef struct SRSmaQTaskFIter SRSmaQTaskFIter;
static int32_t tdUidStorePut(STbUidStore *pStore, tb_uid_t suid, tb_uid_t *uid);
static int32_t tdUpdateTbUidListImpl(SSma *pSma, tb_uid_t *suid, SArray *tbUids);
static int32_t tdSetRSmaInfoItemParams(SSma *pSma, SRSmaParam *param, SRSmaInfo *pRSmaInfo, SReadHandle *handle,
@ -27,17 +31,24 @@ static int32_t tdExecuteRSmaImpl(SSma *pSma, const void *pMsg, int32_t inputType
tb_uid_t suid, int8_t level);
static void tdRSmaFetchTrigger(void *param, void *tmrId);
static void tdRSmaPersistTrigger(void *param, void *tmrId);
static void *tdRSmaPersistExec(void *param);
static void tdRSmaQTaskGetFName(int32_t vid, int8_t ftype, char *outputName);
static int32_t tdRSmaQTaskInfoIterInit(SRSmaQTaskFIter *pIter, STFile *pTFile);
static int32_t tdRSmaQTaskInfoIterNextBlock(SRSmaQTaskFIter *pIter, bool *isFinish);
static int32_t tdRSmaQTaskInfoIterNext(SRSmaQTaskFIter *pIter, SRSmaQTaskInfoItem *pItem, bool *isEnd);
static int32_t tdRSmaQTaskInfoItemRestore(SSma *pSma, const SRSmaQTaskInfoItem *infoItem);
struct SRSmaInfoItem {
SRSmaInfo *pRsmaInfo;
void *taskInfo; // qTaskInfo_t
void *tmrHandle;
tmr_h tmrId;
int8_t level;
int8_t tmrInitFlag;
int8_t triggerStatus; // TASK_TRIGGER_STATUS__IN_ACTIVE/TASK_TRIGGER_STATUS__ACTIVE
int8_t triggerStat;
int32_t maxDelay;
};
struct SRSmaInfo {
STSchema *pTSchema;
SSma *pSma;
@ -45,11 +56,38 @@ struct SRSmaInfo {
SRSmaInfoItem items[TSDB_RETENTION_L2];
};
static FORCE_INLINE void tdFreeTaskHandle(qTaskInfo_t *taskHandle) {
struct SRSmaQTaskInfoItem {
int32_t len;
int8_t type;
int64_t suid;
void *qTaskInfo;
};
struct SRSmaQTaskFIter {
STFile *pTFile;
int64_t offset;
int64_t fsize;
int32_t nBytes;
int32_t nAlloc;
char *buf;
// ------------
int32_t nBufPos;
};
static FORCE_INLINE int32_t tdRSmaQTaskInfoContLen(int32_t lenWithHead) {
return lenWithHead - sizeof(int32_t) - sizeof(int8_t) - sizeof(int64_t);
}
static FORCE_INLINE void tdRSmaQTaskInfoIterDestroy(SRSmaQTaskFIter *pIter) { taosMemoryFreeClear(pIter->buf); }
static FORCE_INLINE void tdFreeTaskHandle(qTaskInfo_t *taskHandle, int32_t vgId, int32_t level) {
// Note: free/kill may in RC
qTaskInfo_t otaskHandle = atomic_load_ptr(taskHandle);
if (otaskHandle && atomic_val_compare_exchange_ptr(taskHandle, otaskHandle, NULL)) {
smaDebug("vgId:%d, free qTaskInfo_t %p of level %d", vgId, otaskHandle, level);
qDestroyTask(otaskHandle);
} else {
smaDebug("vgId:%d, not free qTaskInfo_t %p of level %d", vgId, otaskHandle, level);
}
}
@ -58,14 +96,19 @@ void *tdFreeRSmaInfo(SRSmaInfo *pInfo) {
for (int32_t i = 0; i < TSDB_RETENTION_L2; ++i) {
SRSmaInfoItem *pItem = &pInfo->items[i];
if (pItem->taskInfo) {
tdFreeTaskHandle(pItem->taskInfo);
}
if (pItem->tmrHandle) {
taosTmrCleanUp(pItem->tmrHandle);
smaDebug("vgId:%d, stb %" PRIi64 " stop fetch-timer %p level %d", SMA_VID(pInfo->pSma), pInfo->suid,
pItem->tmrId, i + 1);
taosTmrStopA(&pItem->tmrId);
tdFreeTaskHandle(&pItem->taskInfo, SMA_VID(pInfo->pSma), i + 1);
} else {
smaDebug("vgId:%d, stb %" PRIi64 " no need to destroy rsma info level %d since empty taskInfo",
SMA_VID(pInfo->pSma), pInfo->suid, i + 1);
}
}
taosMemoryFree(pInfo->pTSchema);
taosMemoryFree(pInfo);
} else {
smaDebug("vgId:%d, stb %" PRIi64 " no need to destroy rsma info since empty", SMA_VID(pInfo->pSma), pInfo->suid);
}
return NULL;
@ -81,9 +124,9 @@ static FORCE_INLINE int32_t tdUidStoreInit(STbUidStore **pStore) {
return TSDB_CODE_SUCCESS;
}
static FORCE_INLINE int32_t tdUpdateTbUidListImpl(SSma *pSma, tb_uid_t *suid, SArray *tbUids) {
static int32_t tdUpdateTbUidListImpl(SSma *pSma, tb_uid_t *suid, SArray *tbUids) {
SSmaEnv *pEnv = SMA_RSMA_ENV(pSma);
SSmaStat *pStat = SMA_ENV_STAT(pEnv);
SRSmaStat *pStat = (SRSmaStat *)SMA_ENV_STAT(pEnv);
SRSmaInfo *pRSmaInfo = NULL;
if (!suid || !tbUids) {
@ -92,27 +135,31 @@ static FORCE_INLINE int32_t tdUpdateTbUidListImpl(SSma *pSma, tb_uid_t *suid, SA
return TSDB_CODE_FAILED;
}
pRSmaInfo = taosHashGet(SMA_RSMA_INFO_HASH(pStat), suid, sizeof(tb_uid_t));
pRSmaInfo = taosHashGet(RSMA_INFO_HASH(pStat), suid, sizeof(tb_uid_t));
if (!pRSmaInfo || !(pRSmaInfo = *(SRSmaInfo **)pRSmaInfo)) {
smaError("vgId:%d, failed to get rsma info for uid:%" PRIi64, SMA_VID(pSma), *suid);
terrno = TSDB_CODE_RSMA_INVALID_STAT;
return TSDB_CODE_FAILED;
}
if (pRSmaInfo->items[0].taskInfo && (qUpdateQualifiedTableId(pRSmaInfo->items[0].taskInfo, tbUids, true) < 0)) {
smaError("vgId:%d, update tbUidList failed for uid:%" PRIi64 " since %s", SMA_VID(pSma), *suid, terrstr(terrno));
return TSDB_CODE_FAILED;
} else {
smaDebug("vgId:%d, update tbUidList succeed for qTaskInfo:%p with suid:%" PRIi64 ", uid:%" PRIi64, SMA_VID(pSma),
pRSmaInfo->items[0].taskInfo, *suid, *(int64_t *)taosArrayGet(tbUids, 0));
if (pRSmaInfo->items[0].taskInfo) {
if ((qUpdateQualifiedTableId(pRSmaInfo->items[0].taskInfo, tbUids, true) < 0)) {
smaError("vgId:%d, update tbUidList failed for uid:%" PRIi64 " since %s", SMA_VID(pSma), *suid, terrstr(terrno));
return TSDB_CODE_FAILED;
} else {
smaDebug("vgId:%d, update tbUidList succeed for qTaskInfo:%p with suid:%" PRIi64 ", uid:%" PRIi64, SMA_VID(pSma),
pRSmaInfo->items[0].taskInfo, *suid, *(int64_t *)taosArrayGet(tbUids, 0));
}
}
if (pRSmaInfo->items[1].taskInfo && (qUpdateQualifiedTableId(pRSmaInfo->items[1].taskInfo, tbUids, true) < 0)) {
smaError("vgId:%d, update tbUidList failed for uid:%" PRIi64 " since %s", SMA_VID(pSma), *suid, terrstr(terrno));
return TSDB_CODE_FAILED;
} else {
smaDebug("vgId:%d, update tbUidList succeed for qTaskInfo:%p with suid:%" PRIi64 ", uid:%" PRIi64, SMA_VID(pSma),
pRSmaInfo->items[1].taskInfo, *suid, *(int64_t *)taosArrayGet(tbUids, 0));
if (pRSmaInfo->items[1].taskInfo) {
if ((qUpdateQualifiedTableId(pRSmaInfo->items[1].taskInfo, tbUids, true) < 0)) {
smaError("vgId:%d, update tbUidList failed for uid:%" PRIi64 " since %s", SMA_VID(pSma), *suid, terrstr(terrno));
return TSDB_CODE_FAILED;
} else {
smaDebug("vgId:%d, update tbUidList succeed for qTaskInfo:%p with suid:%" PRIi64 ", uid:%" PRIi64, SMA_VID(pSma),
pRSmaInfo->items[1].taskInfo, *suid, *(int64_t *)taosArrayGet(tbUids, 0));
}
}
return TSDB_CODE_SUCCESS;
@ -159,9 +206,9 @@ int32_t tdFetchTbUidList(SSma *pSma, STbUidStore **ppStore, tb_uid_t suid, tb_ui
return TSDB_CODE_SUCCESS;
}
SSmaStat *pStat = SMA_ENV_STAT(pEnv);
SHashObj *infoHash = NULL;
if (!pStat || !(infoHash = SMA_RSMA_INFO_HASH(pStat))) {
SRSmaStat *pStat = (SRSmaStat *)SMA_ENV_STAT(pEnv);
SHashObj *infoHash = NULL;
if (!pStat || !(infoHash = RSMA_INFO_HASH(pStat))) {
terrno = TSDB_CODE_RSMA_INVALID_STAT;
return TSDB_CODE_FAILED;
}
@ -195,11 +242,11 @@ static int32_t tdSetRSmaInfoItemParams(SSma *pSma, SRSmaParam *param, SRSmaInfo
if (param->qmsg[idx]) {
SRSmaInfoItem *pItem = &(pRSmaInfo->items[idx]);
pItem->pRsmaInfo = pRSmaInfo;
pItem->taskInfo = qCreateStreamExecTaskInfo(param->qmsg[0], pReadHandle);
pItem->taskInfo = qCreateStreamExecTaskInfo(param->qmsg[idx], pReadHandle);
if (!pItem->taskInfo) {
goto _err;
}
pItem->triggerStatus = TASK_TRIGGER_STATUS__IN_ACTIVE;
pItem->triggerStat = TASK_TRIGGER_STAT_INACTIVE;
if (param->maxdelay[idx] < TSDB_MIN_ROLLUP_MAX_DELAY) {
int64_t msInterval =
convertTimeFromPrecisionToUnit(pRetention[idx + 1].freq, pTsdbCfg->precision, TIME_UNIT_MILLISECOND);
@ -211,10 +258,6 @@ static int32_t tdSetRSmaInfoItemParams(SSma *pSma, SRSmaParam *param, SRSmaInfo
pItem->maxDelay = TSDB_MAX_ROLLUP_MAX_DELAY;
}
pItem->level = (idx == 0 ? TSDB_RETENTION_L1 : TSDB_RETENTION_L2);
pItem->tmrHandle = taosTmrInit(10000, 100, 10000, "RSMA");
if (!pItem->tmrHandle) {
goto _err;
}
}
return TSDB_CODE_SUCCESS;
_err:
@ -222,26 +265,21 @@ _err:
}
/**
* @brief Check and init qTaskInfo_t, only applicable to stable with SRSmaParam.
* @brief for rsam create or restore
*
* @param pTsdb
* @param pMeta
* @param pReq
* @param pSma
* @param param
* @param suid
* @param tbName
* @return int32_t
*/
int32_t tdProcessRSmaCreate(SVnode *pVnode, SVCreateStbReq *pReq) {
SSma *pSma = pVnode->pSma;
if (!pReq->rollup) {
smaTrace("vgId:%d, return directly since no rollup for stable %s %" PRIi64, SMA_VID(pSma), pReq->name, pReq->suid);
return TSDB_CODE_SUCCESS;
}
SMeta *pMeta = pVnode->pMeta;
SMsgCb *pMsgCb = &pVnode->msgCb;
SRSmaParam *param = &pReq->pRSmaParam;
int32_t tdProcessRSmaCreateImpl(SSma *pSma, SRSmaParam *param, int64_t suid, const char *tbName) {
SVnode *pVnode = pSma->pVnode;
SMeta *pMeta = pVnode->pMeta;
SMsgCb *pMsgCb = &pVnode->msgCb;
if ((param->qmsgLen[0] == 0) && (param->qmsgLen[1] == 0)) {
smaWarn("vgId:%d, no qmsg1/qmsg2 for rollup stable %s %" PRIi64, SMA_VID(pSma), pReq->name, pReq->suid);
smaDebug("vgId:%d, no qmsg1/qmsg2 for rollup table %s %" PRIi64, SMA_VID(pSma), tbName, suid);
return TSDB_CODE_SUCCESS;
}
@ -251,13 +289,13 @@ int32_t tdProcessRSmaCreate(SVnode *pVnode, SVCreateStbReq *pReq) {
}
SSmaEnv *pEnv = SMA_RSMA_ENV(pSma);
SSmaStat *pStat = SMA_ENV_STAT(pEnv);
SRSmaStat *pStat = (SRSmaStat *)SMA_ENV_STAT(pEnv);
SRSmaInfo *pRSmaInfo = NULL;
pRSmaInfo = taosHashGet(SMA_RSMA_INFO_HASH(pStat), &pReq->suid, sizeof(tb_uid_t));
pRSmaInfo = taosHashGet(RSMA_INFO_HASH(pStat), &suid, sizeof(tb_uid_t));
if (pRSmaInfo) {
ASSERT(0); // TODO: free original pRSmaInfo is exists abnormally
smaWarn("vgId:%d, rsma info already exists for stb: %s, %" PRIi64, SMA_VID(pSma), pReq->name, pReq->suid);
smaDebug("vgId:%d, rsma info already exists for table %s, %" PRIi64, SMA_VID(pSma), tbName, suid);
return TSDB_CODE_SUCCESS;
}
@ -281,26 +319,33 @@ int32_t tdProcessRSmaCreate(SVnode *pVnode, SVCreateStbReq *pReq) {
.vnode = pVnode,
};
STSchema *pTSchema = metaGetTbTSchema(SMA_META(pSma), pReq->suid, -1);
STSchema *pTSchema = metaGetTbTSchema(SMA_META(pSma), suid, -1);
if (!pTSchema) {
terrno = TSDB_CODE_TDB_IVD_TB_SCHEMA_VERSION;
goto _err;
}
pRSmaInfo->pTSchema = pTSchema;
pRSmaInfo->pSma = pSma;
pRSmaInfo->suid = pReq->suid;
pRSmaInfo->suid = suid;
if (tdSetRSmaInfoItemParams(pSma, param, pRSmaInfo, &handle, 0) < 0) {
goto _err;
}
if (tdSetRSmaInfoItemParams(pSma, param, pRSmaInfo, &handle, 1) < 0) {
goto _err;
}
if (taosHashPut(SMA_RSMA_INFO_HASH(pStat), &pReq->suid, sizeof(tb_uid_t), &pRSmaInfo, sizeof(pRSmaInfo)) < 0) {
if (taosHashPut(RSMA_INFO_HASH(pStat), &suid, sizeof(tb_uid_t), &pRSmaInfo, sizeof(pRSmaInfo)) < 0) {
goto _err;
} else {
smaDebug("vgId:%d, register rsma info succeed for suid:%" PRIi64, SMA_VID(pSma), pReq->suid);
smaDebug("vgId:%d, register rsma info succeed for suid:%" PRIi64, SMA_VID(pSma), suid);
}
// start the persist timer
if (TASK_TRIGGER_STAT_INIT ==
atomic_val_compare_exchange_8(RSMA_TRIGGER_STAT(pStat), TASK_TRIGGER_STAT_INIT, TASK_TRIGGER_STAT_ACTIVE)) {
taosTmrStart(tdRSmaPersistTrigger, RSMA_QTASKINFO_PERSIST_MS, pStat, RSMA_TMR_HANDLE(pStat));
}
return TSDB_CODE_SUCCESS;
@ -310,6 +355,24 @@ _err:
return TSDB_CODE_FAILED;
}
/**
* @brief Check and init qTaskInfo_t, only applicable to stable with SRSmaParam.
*
* @param pTsdb
* @param pMeta
* @param pReq
* @return int32_t
*/
int32_t tdProcessRSmaCreate(SVnode *pVnode, SVCreateStbReq *pReq) {
SSma *pSma = pVnode->pSma;
if (!pReq->rollup) {
smaTrace("vgId:%d, return directly since no rollup for stable %s %" PRIi64, SMA_VID(pSma), pReq->name, pReq->suid);
return TSDB_CODE_SUCCESS;
}
return tdProcessRSmaCreateImpl(pSma, &pReq->rsmaParam, pReq->suid, pReq->name);
}
/**
* @brief store suid/[uids], prefer to use array and then hash
*
@ -432,6 +495,16 @@ static int32_t tdFetchSubmitReqSuids(SSubmitReq *pMsg, STbUidStore *pStore) {
return 0;
}
static void tdDestroySDataBlockArray(SArray *pArray) {
#if 0
for (int32_t i = 0; i < taosArrayGetSize(pArray); ++i) {
SSDataBlock *pDataBlock = taosArrayGet(pArray, i);
blockDestroyInner(pDataBlock);
}
#endif
taosArrayDestroy(pArray);
}
static int32_t tdFetchAndSubmitRSmaResult(SRSmaInfoItem *pItem, int8_t blkType) {
SArray *pResult = NULL;
SRSmaInfo *pRSmaInfo = pItem->pRsmaInfo;
@ -462,10 +535,11 @@ static int32_t tdFetchAndSubmitRSmaResult(SRSmaInfoItem *pItem, int8_t blkType)
#if 1
char flag[10] = {0};
snprintf(flag, 10, "level %" PRIi8, pItem->level);
blockDebugShowData(pResult, flag);
blockDebugShowDataBlocks(pResult, flag);
#endif
STsdb *sinkTsdb = (pItem->level == TSDB_RETENTION_L1 ? pSma->pRSmaTsdb1 : pSma->pRSmaTsdb2);
SSubmitReq *pReq = NULL;
// TODO: the schema update should be handled
if (buildSubmitReqFromDataBlock(&pReq, pResult, pRSmaInfo->pTSchema, SMA_VID(pSma), pRSmaInfo->suid) < 0) {
goto _err;
}
@ -476,18 +550,16 @@ static int32_t tdFetchAndSubmitRSmaResult(SRSmaInfoItem *pItem, int8_t blkType)
}
taosMemoryFreeClear(pReq);
} else if (terrno == 0) {
smaDebug("vgId:%d, no rsma %" PRIi8 " data fetched yet", SMA_VID(pSma), pItem->level);
} else {
smaDebug("vgId:%d, no rsma % " PRIi8 " data generated since %s", SMA_VID(pSma), pItem->level, tstrerror(terrno));
smaDebug("vgId:%d, no rsma %" PRIi8 " data fetched since %s", SMA_VID(pSma), pItem->level, tstrerror(terrno));
}
if (blkType == STREAM_DATA_TYPE_SUBMIT_BLOCK) {
atomic_store_8(&pItem->triggerStatus, TASK_TRIGGER_STATUS__ACTIVE);
}
taosArrayDestroy(pResult);
tdDestroySDataBlockArray(pResult);
return TSDB_CODE_SUCCESS;
_err:
taosArrayDestroy(pResult);
tdDestroySDataBlockArray(pResult);
return TSDB_CODE_FAILED;
}
@ -499,26 +571,38 @@ _err:
*/
static void tdRSmaFetchTrigger(void *param, void *tmrId) {
SRSmaInfoItem *pItem = param;
SSma *pSma = pItem->pRsmaInfo->pSma;
SRSmaStat *pStat = (SRSmaStat *)SMA_ENV_STAT((SSmaEnv *)pSma->pRSmaEnv);
if (atomic_load_8(&pItem->triggerStatus) == TASK_TRIGGER_STATUS__ACTIVE) {
smaWarn("%s:%d THREAD:%" PRIi64 " level %" PRIi8 " status is active for tb suid:%" PRIi64, __func__, __LINE__,
taosGetSelfPthreadId(), pItem->level, pItem->pRsmaInfo->suid);
SSDataBlock dataBlock = {.info.type = STREAM_GET_ALL};
atomic_store_8(&pItem->triggerStatus, TASK_TRIGGER_STATUS__IN_ACTIVE);
qSetStreamInput(pItem->taskInfo, &dataBlock, STREAM_DATA_TYPE_SSDATA_BLOCK, false);
tdFetchAndSubmitRSmaResult(pItem, STREAM_DATA_TYPE_SSDATA_BLOCK);
} else {
smaWarn("%s:%d THREAD:%" PRIi64 " level %" PRIi8 " status is inactive for tb suid:%" PRIi64, __func__, __LINE__,
taosGetSelfPthreadId(), pItem->level, pItem->pRsmaInfo->suid);
int8_t rsmaTriggerStat = atomic_load_8(RSMA_TRIGGER_STAT(pStat));
if (rsmaTriggerStat == TASK_TRIGGER_STAT_CANCELLED || rsmaTriggerStat == TASK_TRIGGER_STAT_FINISHED) {
smaDebug("vgId:%d, level %" PRIi8 " not fetch since stat is cancelled for table suid:%" PRIi64, SMA_VID(pSma),
pItem->level, pItem->pRsmaInfo->suid);
return;
}
// taosTmrReset(tdRSmaFetchTrigger, pItem->maxDelay, pItem, pItem->tmrHandle, &pItem->tmrId);
int8_t fetchTriggerStat =
atomic_val_compare_exchange_8(&pItem->triggerStat, TASK_TRIGGER_STAT_ACTIVE, TASK_TRIGGER_STAT_INACTIVE);
if (fetchTriggerStat == TASK_TRIGGER_STAT_ACTIVE) {
smaDebug("vgId:%d, level %" PRIi8 " stat is active for table suid:%" PRIi64, SMA_VID(pSma), pItem->level,
pItem->pRsmaInfo->suid);
tdRefSmaStat(pSma, (SSmaStat *)pStat);
SSDataBlock dataBlock = {.info.type = STREAM_GET_ALL};
qSetStreamInput(pItem->taskInfo, &dataBlock, STREAM_DATA_TYPE_SSDATA_BLOCK, false);
tdFetchAndSubmitRSmaResult(pItem, STREAM_DATA_TYPE_SSDATA_BLOCK);
tdUnRefSmaStat(pSma, (SSmaStat *)pStat);
} else {
smaDebug("vgId:%d, level %" PRIi8 " stat is inactive for table suid:%" PRIi64, SMA_VID(pSma), pItem->level,
pItem->pRsmaInfo->suid);
}
}
static FORCE_INLINE int32_t tdExecuteRSmaImpl(SSma *pSma, const void *pMsg, int32_t inputType, SRSmaInfoItem *pItem,
tb_uid_t suid, int8_t level) {
static int32_t tdExecuteRSmaImpl(SSma *pSma, const void *pMsg, int32_t inputType, SRSmaInfoItem *pItem, tb_uid_t suid,
int8_t level) {
if (!pItem || !pItem->taskInfo) {
smaDebug("vgId:%d, no qTaskInfo to execute rsma %" PRIi8 " task for suid:%" PRIu64, SMA_VID(pSma), level, suid);
return TSDB_CODE_SUCCESS;
@ -533,14 +617,17 @@ static FORCE_INLINE int32_t tdExecuteRSmaImpl(SSma *pSma, const void *pMsg, int3
}
tdFetchAndSubmitRSmaResult(pItem, STREAM_DATA_TYPE_SUBMIT_BLOCK);
atomic_store_8(&pItem->triggerStatus, TASK_TRIGGER_STATUS__ACTIVE);
smaWarn("%s:%d THREAD:%" PRIi64 " process rsma insert", __func__, __LINE__, taosGetSelfPthreadId());
atomic_store_8(&pItem->triggerStat, TASK_TRIGGER_STAT_ACTIVE);
smaDebug("vgId:%d, process rsma insert", SMA_VID(pSma));
SSmaEnv *pEnv = SMA_RSMA_ENV(pSma);
SRSmaStat *pStat = SMA_RSMA_STAT(pEnv->pStat);
taosTmrStart(tdRSmaPersistTrigger, 5000, pStat, pStat->tmrHandle);
taosTmrReset(tdRSmaFetchTrigger, pItem->maxDelay, pItem, pItem->tmrHandle, &pItem->tmrId);
if (pStat->tmrHandle) {
taosTmrReset(tdRSmaFetchTrigger, pItem->maxDelay, pItem, pStat->tmrHandle, &pItem->tmrId);
} else {
ASSERT(0);
}
return TSDB_CODE_SUCCESS;
}
@ -552,10 +639,10 @@ static int32_t tdExecuteRSma(SSma *pSma, const void *pMsg, int32_t inputType, tb
return TSDB_CODE_SUCCESS;
}
SSmaStat *pStat = SMA_ENV_STAT(pEnv);
SRSmaStat *pStat = (SRSmaStat *)SMA_ENV_STAT(pEnv);
SRSmaInfo *pRSmaInfo = NULL;
pRSmaInfo = taosHashGet(SMA_RSMA_INFO_HASH(pStat), &suid, sizeof(tb_uid_t));
pRSmaInfo = taosHashGet(RSMA_INFO_HASH(pStat), &suid, sizeof(tb_uid_t));
if (!pRSmaInfo || !(pRSmaInfo = *(SRSmaInfo **)pRSmaInfo)) {
smaDebug("vgId:%d, return as no rsma info for suid:%" PRIu64, SMA_VID(pSma), suid);
@ -608,7 +695,7 @@ int32_t tdProcessRSmaSubmit(SSma *pSma, void *pMsg, int32_t inputType) {
return TSDB_CODE_SUCCESS;
}
void tdRSmaQTaskGetFName(int32_t vid, int8_t ftype, char* outputName) {
static void tdRSmaQTaskGetFName(int32_t vid, int8_t ftype, char *outputName) {
tdGetVndFileName(vid, "rsma", tdQTaskInfoFname[ftype], outputName);
}
@ -618,6 +705,11 @@ static void *tdRSmaPersistExec(void *param) {
SSma *pSma = pRSmaStat->pSma;
STfs *pTfs = pSma->pVnode->pTfs;
int64_t toffset = 0;
bool isFileCreated = false;
if (TASK_TRIGGER_STAT_CANCELLED == atomic_load_8(RSMA_TRIGGER_STAT(pRSmaStat))) {
goto _end;
}
void *infoHash = taosHashIterate(RSMA_INFO_HASH(pRSmaStat), NULL);
if (!infoHash) {
@ -625,67 +717,143 @@ static void *tdRSmaPersistExec(void *param) {
}
STFile tFile = {0};
int32_t vid = 2;
char qTaskInfoFName[TSDB_FILENAME_LEN];
tdRSmaQTaskGetFName(vid, TD_QTASK_TMP_FILE, qTaskInfoFName);
tdInitTFile(&tFile, pTfs, qTaskInfoFName);
tdCreateTFile(&tFile, pTfs, true, -1);
int32_t vid = SMA_VID(pSma);
while (infoHash) {
SRSmaInfo *pRSmaInfo = *(SRSmaInfo **)infoHash;
char *pOutput = NULL;
int32_t len = 0;
if (qSerializeTaskStatus(pRSmaInfo->items[0].taskInfo, &pOutput, &len) < 0) {
smaError("serialize rsma task for table %" PRIi64 " failed since %s", pRSmaInfo->items[0].pRsmaInfo->suid,
terrstr(terrno));
} else {
smaWarn("serialize rsma task for table %" PRIi64 " success and len is %d", pRSmaInfo->items[0].pRsmaInfo->suid,
len);
}
tdAppendTFile(&tFile, &len, sizeof(len), &toffset);
tdAppendTFile(&tFile, pOutput, len, &toffset);
taosMemoryFree(pOutput);
#if 0
smaDebug("table %" PRIi64 " sleep 15s start ...", pRSmaInfo->items[0].pRsmaInfo->suid);
for (int32_t i = 15; i > 0; --i) {
taosSsleep(1);
smaDebug("table %" PRIi64 " countdown %d", pRSmaInfo->items[0].pRsmaInfo->suid, i);
}
smaDebug("table %" PRIi64 " sleep 15s end ...", pRSmaInfo->items[0].pRsmaInfo->suid);
#endif
for (int32_t i = 0; i < TSDB_RETENTION_L2; ++i) {
qTaskInfo_t taskInfo = pRSmaInfo->items[i].taskInfo;
if (!taskInfo) {
smaDebug("vgId:%d, table %" PRIi64 " level %d qTaskInfo is NULL", vid, pRSmaInfo->suid, i + 1);
continue;
}
char *pOutput = NULL;
int32_t len = 0;
int8_t type = (int8_t)(i + 1);
if (qSerializeTaskStatus(taskInfo, &pOutput, &len) < 0) {
smaError("vgId:%d, table %" PRIi64 " level %d serialize rsma task failed since %s", vid, pRSmaInfo->suid, i + 1,
terrstr(terrno));
goto _err;
} else {
if (!pOutput) {
smaDebug("vgId:%d, table %" PRIi64
" level %d serialize rsma task success but no output(len %d) and no need to persist",
vid, pRSmaInfo->suid, i + 1, len);
continue;
} else if (len <= 0) {
smaDebug("vgId:%d, table %" PRIi64 " level %d serialize rsma task success with len %d and no need to persist",
vid, pRSmaInfo->suid, i + 1, len);
taosMemoryFree(pOutput);
}
smaDebug("vgId:%d, table %" PRIi64 " level %d serialize rsma task success with len %d and need persist", vid,
pRSmaInfo->suid, i + 1, len);
#if 1
if (qDeserializeTaskStatus(taskInfo, pOutput, len) < 0) {
smaError("vgId:%d, table %" PRIi64 "level %d deserialize rsma task failed since %s", vid, pRSmaInfo->suid,
i + 1, terrstr(terrno));
} else {
smaDebug("vgId:%d, table %" PRIi64 " level %d deserialize rsma task success", vid, pRSmaInfo->suid, i + 1);
}
#endif
}
if (!isFileCreated) {
char qTaskInfoFName[TSDB_FILENAME_LEN];
tdRSmaQTaskGetFName(vid, TD_QTASK_TMP_FILE, qTaskInfoFName);
tdInitTFile(&tFile, pTfs, qTaskInfoFName);
tdCreateTFile(&tFile, pTfs, true, -1);
isFileCreated = true;
}
len += (sizeof(len) + sizeof(type) + sizeof(pRSmaInfo->suid));
tdAppendTFile(&tFile, &len, sizeof(len), &toffset);
tdAppendTFile(&tFile, &type, sizeof(type), &toffset);
tdAppendTFile(&tFile, &pRSmaInfo->suid, sizeof(pRSmaInfo->suid), &toffset);
tdAppendTFile(&tFile, pOutput, len, &toffset);
taosMemoryFree(pOutput);
}
infoHash = taosHashIterate(RSMA_INFO_HASH(pRSmaStat), infoHash);
}
_end:
_normal:
if (isFileCreated) {
if (tdUpdateTFileHeader(&tFile) < 0) {
smaError("vgId:%d, failed to update tfile %s header since %s", vid, TD_FILE_FULL_NAME(&tFile), tstrerror(terrno));
tdCloseTFile(&tFile);
tdRemoveTFile(&tFile);
goto _err;
} else {
smaDebug("vgId:%d, succeed to update tfile %s header", vid, TD_FILE_FULL_NAME(&tFile));
}
if (tdUpdateTFileHeader(&tFile) < 0) {
smaError("vgId:%d, failed to update tfile %s header since %s", vid, TD_FILE_FULL_NAME(&tFile), tstrerror(terrno));
tdCloseTFile(&tFile);
tdRemoveTFile(&tFile);
return NULL;
char newFName[TSDB_FILENAME_LEN];
strncpy(newFName, TD_FILE_FULL_NAME(&tFile), TSDB_FILENAME_LEN);
char *pos = strstr(newFName, tdQTaskInfoFname[TD_QTASK_TMP_FILE]);
strncpy(pos, tdQTaskInfoFname[TD_QTASK_CUR_FILE], TSDB_FILENAME_LEN - POINTER_DISTANCE(pos, newFName));
if (taosRenameFile(TD_FILE_FULL_NAME(&tFile), newFName) != 0) {
smaError("vgId:%d, failed to rename %s to %s", vid, TD_FILE_FULL_NAME(&tFile), newFName);
goto _err;
} else {
smaDebug("vgId:%d, succeed to rename %s to %s", vid, TD_FILE_FULL_NAME(&tFile), newFName);
}
}
tdCloseTFile(&tFile);
char newFName[TSDB_FILENAME_LEN];
strncpy(newFName, TD_FILE_FULL_NAME(&tFile), TSDB_FILENAME_LEN);
char *pos = strstr(newFName, tdQTaskInfoFname[TD_QTASK_TMP_FILE]);
strncpy(pos, tdQTaskInfoFname[TD_QTASK_CUR_FILE], TSDB_FILENAME_LEN - POINTER_DISTANCE(pos, newFName));
taosRenameFile(TD_FILE_FULL_NAME(&tFile), newFName);
atomic_store_8(&pRSmaStat->tmrStat, TASK_TRIGGER_STATUS__ACTIVE);
return NULL;
goto _end;
_err:
atomic_store_8(&pRSmaStat->tmrStat, TASK_TRIGGER_STATUS__ACTIVE);
// remove the .tmp file
if (isFileCreated) {
tdRemoveTFile(&tFile);
}
_end:
if (TASK_TRIGGER_STAT_INACTIVE == atomic_val_compare_exchange_8(RSMA_TRIGGER_STAT(pRSmaStat),
TASK_TRIGGER_STAT_INACTIVE,
TASK_TRIGGER_STAT_ACTIVE)) {
smaDebug("vgId:%d, persist task is active again", vid);
} else if (TASK_TRIGGER_STAT_CANCELLED == atomic_val_compare_exchange_8(RSMA_TRIGGER_STAT(pRSmaStat),
TASK_TRIGGER_STAT_CANCELLED,
TASK_TRIGGER_STAT_FINISHED)) {
smaDebug("vgId:%d, persist task is cancelled", vid);
} else {
smaWarn("vgId:%d, persist task in abnormal stat %" PRIi8, vid, atomic_load_8(RSMA_TRIGGER_STAT(pRSmaStat)));
ASSERT(0);
}
atomic_store_8(RSMA_RUNNING_STAT(pRSmaStat), 0);
taosThreadExit(NULL);
return NULL;
}
static void tdRSmaPersistTask(SRSmaStat *pRSmaStat) {
smaWarn("%s:%d entry ", __func__, __LINE__);
TdThread threadId;
TdThreadAttr thAttr;
taosThreadAttrInit(&thAttr);
taosThreadAttrSetDetachState(&thAttr, PTHREAD_CREATE_DETACHED);
TdThread tid;
if (taosThreadCreate(&threadId, &thAttr, tdRSmaPersistExec, pRSmaStat) != 0) {
smaError("failed to create thread to persist rsma qTaskInfo since %s", strerror(errno));
if (taosThreadCreate(&tid, &thAttr, tdRSmaPersistExec, pRSmaStat) != 0) {
if (TASK_TRIGGER_STAT_INACTIVE == atomic_val_compare_exchange_8(RSMA_TRIGGER_STAT(pRSmaStat),
TASK_TRIGGER_STAT_INACTIVE,
TASK_TRIGGER_STAT_ACTIVE)) {
smaDebug("persist task is active again");
} else if (TASK_TRIGGER_STAT_CANCELLED == atomic_val_compare_exchange_8(RSMA_TRIGGER_STAT(pRSmaStat),
TASK_TRIGGER_STAT_CANCELLED,
TASK_TRIGGER_STAT_FINISHED)) {
smaDebug(" persist task is cancelled and set finished");
} else {
smaWarn("persist task in abnormal stat %" PRIi8, atomic_load_8(RSMA_TRIGGER_STAT(pRSmaStat)));
ASSERT(0);
}
atomic_store_8(RSMA_RUNNING_STAT(pRSmaStat), 0);
}
taosThreadAttrDestroy(&thAttr);
smaWarn("%s:%d end ", __func__, __LINE__);
}
/**
@ -696,17 +864,283 @@ static void tdRSmaPersistTask(SRSmaStat *pRSmaStat) {
*/
static void tdRSmaPersistTrigger(void *param, void *tmrId) {
SRSmaStat *pRSmaStat = param;
int8_t tmrStat =
atomic_val_compare_exchange_8(RSMA_TRIGGER_STAT(pRSmaStat), TASK_TRIGGER_STAT_ACTIVE, TASK_TRIGGER_STAT_INACTIVE);
switch (tmrStat) {
case TASK_TRIGGER_STAT_ACTIVE: {
atomic_store_8(RSMA_RUNNING_STAT(pRSmaStat), 1);
if (TASK_TRIGGER_STAT_CANCELLED != atomic_val_compare_exchange_8(RSMA_TRIGGER_STAT(pRSmaStat),
TASK_TRIGGER_STAT_CANCELLED,
TASK_TRIGGER_STAT_FINISHED)) {
smaDebug("rsma persistence start since active");
if (atomic_load_8(&pRSmaStat->tmrStat) == TASK_TRIGGER_STATUS__ACTIVE) {
smaWarn("%s:%d THREAD:%" PRIi64 " rsma persistence start since active", __func__, __LINE__, taosGetSelfPthreadId());
atomic_store_8(&pRSmaStat->tmrStat, TASK_TRIGGER_STATUS__IN_ACTIVE);
// start persist task
tdRSmaPersistTask(pRSmaStat);
// execution
tdRSmaPersistTask(pRSmaStat);
} else {
smaWarn("%s:%d THREAD:%" PRIi64 " rsma persistence not start since inactive", __func__, __LINE__,
taosGetSelfPthreadId());
taosTmrReset(tdRSmaPersistTrigger, RSMA_QTASKINFO_PERSIST_MS, pRSmaStat, pRSmaStat->tmrHandle,
&pRSmaStat->tmrId);
} else {
atomic_store_8(RSMA_RUNNING_STAT(pRSmaStat), 0);
}
} break;
case TASK_TRIGGER_STAT_CANCELLED: {
atomic_store_8(RSMA_TRIGGER_STAT(pRSmaStat), TASK_TRIGGER_STAT_FINISHED);
smaDebug("rsma persistence not start since cancelled and finished");
} break;
case TASK_TRIGGER_STAT_INACTIVE: {
smaDebug("rsma persistence not start since inactive");
} break;
case TASK_TRIGGER_STAT_INIT: {
smaDebug("rsma persistence not start since init");
} break;
default: {
smaWarn("rsma persistence not start since unknown stat %" PRIi8, tmrStat);
ASSERT(0);
} break;
}
}
int32_t tdProcessRSmaRestoreImpl(SSma *pSma) {
SVnode *pVnode = pSma->pVnode;
// step 1: iterate all stables to restore the rsma env
SArray *suidList = taosArrayInit(1, sizeof(tb_uid_t));
if (tsdbGetStbIdList(SMA_META(pSma), 0, suidList) < 0) {
smaError("vgId:%d, failed to restore rsma since get stb id list error: %s", TD_VID(pVnode), terrstr());
return TSDB_CODE_FAILED;
}
taosTmrReset(tdRSmaPersistTrigger, 3600000, pRSmaStat, pRSmaStat->tmrHandle, &pRSmaStat->tmrId);
}
if (taosArrayGetSize(suidList) == 0) {
smaDebug("vgId:%d no need to restore rsma since empty stb id list", TD_VID(pVnode));
return TSDB_CODE_SUCCESS;
}
SMetaReader mr = {0};
metaReaderInit(&mr, SMA_META(pSma), 0);
for (int32_t i = 0; i < taosArrayGetSize(suidList); ++i) {
tb_uid_t suid = *(tb_uid_t *)taosArrayGet(suidList, i);
smaDebug("suid [%d] is %" PRIi64, i, suid);
if (metaGetTableEntryByUid(&mr, suid) < 0) {
smaError("vgId:%d failed to get table meta for %" PRIi64 " since %s", TD_VID(pVnode), suid, terrstr());
goto _err;
}
ASSERT(mr.me.type == TSDB_SUPER_TABLE);
ASSERT(mr.me.uid == suid);
if (TABLE_IS_ROLLUP(mr.me.flags)) {
SRSmaParam *param = &mr.me.stbEntry.rsmaParam;
for (int i = 0; i < 2; ++i) {
smaDebug("vgId: %d table:%" PRIi64 " maxdelay[%d]:%" PRIi64 " watermark[%d]:%" PRIi64, TD_VID(pSma->pVnode),
suid, i, param->maxdelay[i], i, param->watermark[i]);
}
if (tdProcessRSmaCreateImpl(pSma, &mr.me.stbEntry.rsmaParam, suid, mr.me.name) < 0) {
smaError("vgId:%d failed to retore rsma env for %" PRIi64 " since %s", TD_VID(pVnode), suid, terrstr());
goto _err;
}
}
}
// step 2: retrieve qtaskinfo object from the rsma/qtaskinfo file and restore
STFile tFile = {0};
char qTaskInfoFName[TSDB_FILENAME_LEN];
tdRSmaQTaskGetFName(TD_VID(pVnode), TD_QTASK_CUR_FILE, qTaskInfoFName);
if (tdInitTFile(&tFile, pVnode->pTfs, qTaskInfoFName) < 0) {
goto _err;
}
if (tdOpenTFile(&tFile, TD_FILE_READ) < 0) {
goto _err;
}
SRSmaQTaskFIter fIter = {0};
if (tdRSmaQTaskInfoIterInit(&fIter, &tFile) < 0) {
goto _err;
}
SRSmaQTaskInfoItem infoItem = {0};
bool isEnd = false;
int32_t code = 0;
while ((code = tdRSmaQTaskInfoIterNext(&fIter, &infoItem, &isEnd)) == 0) {
if (isEnd) {
break;
}
if ((code = tdRSmaQTaskInfoItemRestore(pSma, &infoItem)) < 0) {
break;
}
}
tdRSmaQTaskInfoIterDestroy(&fIter);
if (code < 0) {
goto _err;
}
metaReaderClear(&mr);
taosArrayDestroy(suidList);
return TSDB_CODE_SUCCESS;
_err:
ASSERT(0);
metaReaderClear(&mr);
taosArrayDestroy(suidList);
smaError("failed to restore rsma info since %s", terrstr());
return TSDB_CODE_FAILED;
}
static int32_t tdRSmaQTaskInfoItemRestore(SSma *pSma, const SRSmaQTaskInfoItem *infoItem) {
SRSmaStat *pStat = (SRSmaStat *)SMA_ENV_STAT((SSmaEnv *)pSma->pRSmaEnv);
SRSmaInfo *pRSmaInfo = NULL;
void *qTaskInfo = NULL;
pRSmaInfo = taosHashGet(RSMA_INFO_HASH(pStat), &infoItem->suid, sizeof(infoItem->suid));
if (!pRSmaInfo || !(pRSmaInfo = *(SRSmaInfo **)pRSmaInfo)) {
smaDebug("vgId:%d, no restore as no rsma info for suid:%" PRIu64, SMA_VID(pSma), infoItem->suid);
return TSDB_CODE_SUCCESS;
}
if (infoItem->type == 1) {
qTaskInfo = pRSmaInfo->items[0].taskInfo;
} else if (infoItem->type == 2) {
qTaskInfo = pRSmaInfo->items[1].taskInfo;
} else {
ASSERT(0);
}
if (!qTaskInfo) {
smaDebug("vgId:%d, no restore as NULL rsma qTaskInfo for suid:%" PRIu64, SMA_VID(pSma), infoItem->suid);
return TSDB_CODE_SUCCESS;
}
if (qDeserializeTaskStatus(qTaskInfo, infoItem->qTaskInfo, infoItem->len) < 0) {
smaError("vgId:%d, restore rsma failed for suid:%" PRIi64 " level %d since %s", SMA_VID(pSma), infoItem->suid,
infoItem->type, terrstr(terrno));
return TSDB_CODE_FAILED;
}
smaDebug("vgId:%d, restore rsma success for suid:%" PRIi64 " level %d", SMA_VID(pSma), infoItem->suid,
infoItem->type);
return TSDB_CODE_SUCCESS;
}
static int32_t tdRSmaQTaskInfoIterInit(SRSmaQTaskFIter *pIter, STFile *pTFile) {
memset(pIter, 0, sizeof(*pIter));
pIter->pTFile = pTFile;
pIter->offset = TD_FILE_HEAD_SIZE;
if (tdGetTFileSize(pTFile, &pIter->fsize) < 0) {
return TSDB_CODE_FAILED;
}
if ((pIter->fsize - TD_FILE_HEAD_SIZE) < RSMA_QTASKINFO_BUFSIZE) {
pIter->nAlloc = pIter->fsize - TD_FILE_HEAD_SIZE;
} else {
pIter->nAlloc = RSMA_QTASKINFO_BUFSIZE;
}
if (pIter->nAlloc < TD_FILE_HEAD_SIZE) {
pIter->nAlloc = TD_FILE_HEAD_SIZE;
}
pIter->buf = taosMemoryMalloc(pIter->nAlloc);
if (!pIter->buf) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
return TSDB_CODE_FAILED;
}
return TSDB_CODE_SUCCESS;
}
static int32_t tdRSmaQTaskInfoIterNextBlock(SRSmaQTaskFIter *pIter, bool *isFinish) {
STFile *pTFile = pIter->pTFile;
int64_t nBytes = RSMA_QTASKINFO_BUFSIZE;
if (pIter->offset >= pIter->fsize) {
*isFinish = true;
return TSDB_CODE_SUCCESS;
}
if ((pIter->fsize - pIter->offset) < RSMA_QTASKINFO_BUFSIZE) {
nBytes = pIter->fsize - pIter->offset;
}
if (tdSeekTFile(pTFile, pIter->offset, SEEK_SET) < 0) {
ASSERT(0);
return TSDB_CODE_FAILED;
}
if (tdReadTFile(pTFile, pIter->buf, nBytes) != nBytes) {
ASSERT(0);
return TSDB_CODE_FAILED;
}
int32_t infoLen = 0;
taosDecodeFixedI32(pIter->buf, &infoLen);
if (infoLen > nBytes) {
ASSERT(infoLen > RSMA_QTASKINFO_BUFSIZE);
pIter->nAlloc = infoLen;
void *pBuf = taosMemoryRealloc(pIter->buf, infoLen);
if (!pBuf) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
return TSDB_CODE_FAILED;
}
pIter->buf = pBuf;
nBytes = infoLen;
if (tdSeekTFile(pTFile, pIter->offset, SEEK_SET)) {
ASSERT(0);
return TSDB_CODE_FAILED;
}
if (tdReadTFile(pTFile, pIter->buf, nBytes) != nBytes) {
ASSERT(0);
return TSDB_CODE_FAILED;
}
}
pIter->offset += nBytes;
pIter->nBytes = nBytes;
pIter->nBufPos = 0;
return TSDB_CODE_SUCCESS;
}
static int32_t tdRSmaQTaskInfoIterNext(SRSmaQTaskFIter *pIter, SRSmaQTaskInfoItem *pItem, bool *isEnd) {
while (1) {
// block iter
bool isFinish = false;
if (tdRSmaQTaskInfoIterNextBlock(pIter, &isFinish) < 0) {
ASSERT(0);
return TSDB_CODE_FAILED;
}
if (isFinish) {
*isEnd = true;
return TSDB_CODE_SUCCESS;
}
// consume the block
int32_t qTaskInfoLenWithHead = 0;
pIter->buf = taosDecodeFixedI32(pIter->buf, &qTaskInfoLenWithHead);
if (qTaskInfoLenWithHead < 0) {
terrno = TSDB_CODE_TDB_FILE_CORRUPTED;
return TSDB_CODE_FAILED;
}
while (1) {
if ((pIter->nBufPos + qTaskInfoLenWithHead) <= pIter->nBytes) {
pIter->buf = taosDecodeFixedI8(pIter->buf, &pItem->type);
pIter->buf = taosDecodeFixedI64(pIter->buf, &pItem->suid);
pItem->qTaskInfo = pIter->buf;
pItem->len = tdRSmaQTaskInfoContLen(qTaskInfoLenWithHead);
// do the restore job
printf("%s:%d ###### restore the qtask info offset:%" PRIi64 "\n", __func__, __LINE__, pIter->offset);
pIter->buf = POINTER_SHIFT(pIter->buf, pItem->len);
pIter->nBufPos += qTaskInfoLenWithHead;
pIter->buf = taosDecodeFixedI32(pIter->buf, &qTaskInfoLenWithHead);
continue;
}
// prepare and load next block in the file
pIter->offset -= (pIter->nBytes - pIter->nBufPos);
break;
}
}
return TSDB_CODE_SUCCESS;
}

16
source/dnode/vnode/src/sma/smaUtil.c
View File

@ -15,7 +15,7 @@
#include "sma.h"
#define TD_FILE_HEAD_SIZE 512
// smaFileUtil ================
#define TD_FILE_STATE_OK 0
#define TD_FILE_STATE_BAD 1
@ -32,7 +32,7 @@ static int32_t tdEncodeTFInfo(void **buf, STFInfo *pInfo) {
tlen += taosEncodeFixedU32(buf, pInfo->magic);
tlen += taosEncodeFixedU32(buf, pInfo->ftype);
tlen += taosEncodeFixedU32(buf, pInfo->fver);
tlen += taosEncodeFixedU64(buf, pInfo->fsize);
tlen += taosEncodeFixedI64(buf, pInfo->fsize);
return tlen;
}
@ -41,7 +41,7 @@ static void *tdDecodeTFInfo(void *buf, STFInfo *pInfo) {
buf = taosDecodeFixedU32(buf, &(pInfo->magic));
buf = taosDecodeFixedU32(buf, &(pInfo->ftype));
buf = taosDecodeFixedU32(buf, &(pInfo->fver));
buf = taosDecodeFixedU64(buf, &(pInfo->fsize));
buf = taosDecodeFixedI64(buf, &(pInfo->fsize));
return buf;
}
@ -69,6 +69,11 @@ int64_t tdSeekTFile(STFile *pTFile, int64_t offset, int whence) {
return loffset;
}
int64_t tdGetTFileSize(STFile *pTFile, int64_t *size) {
ASSERT(TD_FILE_OPENED(pTFile));
return taosFStatFile(pTFile->pFile, size, NULL);
}
int64_t tdReadTFile(STFile *pTFile, void *buf, int64_t nbyte) {
ASSERT(TD_FILE_OPENED(pTFile));
@ -235,4 +240,7 @@ int32_t tdCreateTFile(STFile *pTFile, STfs *pTfs, bool updateHeader, int8_t fTyp
return 0;
}
int32_t tdRemoveTFile(STFile *pTFile) { return tfsRemoveFile(TD_FILE_F(pTFile)); }
int32_t tdRemoveTFile(STFile *pTFile) { return tfsRemoveFile(TD_FILE_F(pTFile)); }
// smaXXXUtil ================
// ...

27
source/dnode/vnode/src/tsdb/tsdbRead.c
View File

@ -2876,6 +2876,33 @@ int32_t tsdbGetCtbIdList(SMeta* pMeta, int64_t suid, SArray* list) {
return TSDB_CODE_SUCCESS;
}
/**
* @brief Get all suids since suid
*
* @param pMeta
* @param suid return all suids in one vnode if suid is 0
* @param list
* @return int32_t
*/
int32_t tsdbGetStbIdList(SMeta* pMeta, int64_t suid, SArray* list) {
SMStbCursor* pCur = metaOpenStbCursor(pMeta, suid);
if(!pCur) {
return TSDB_CODE_FAILED;
}
while (1) {
tb_uid_t id = metaStbCursorNext(pCur);
if (id == 0) {
break;
}
taosArrayPush(list, &id);
}
metaCloseStbCursor(pCur);
return TSDB_CODE_SUCCESS;
}
static void destroyHelper(void* param) {
if (param == NULL) {
return;

7
source/dnode/vnode/src/vnd/vnodeCommit.c
View File

@ -72,7 +72,12 @@ int vnodeBegin(SVnode *pVnode) {
return 0;
}
int vnodeShouldCommit(SVnode *pVnode) { return pVnode->inUse->size > pVnode->config.szBuf / 3; }
int vnodeShouldCommit(SVnode *pVnode) {
if (pVnode->inUse) {
return pVnode->inUse->size > pVnode->config.szBuf / 3;
}
return false;
}
int vnodeSaveInfo(const char *dir, const SVnodeInfo *pInfo) {
char fname[TSDB_FILENAME_LEN];

6
source/dnode/vnode/src/vnd/vnodeOpen.c
View File

@ -152,12 +152,13 @@ SVnode *vnodeOpen(const char *path, STfs *pTfs, SMsgCb msgCb) {
return pVnode;
_err:
if (pVnode->pSma) smaClose(pVnode->pSma);
if (pVnode->pQuery) vnodeQueryClose(pVnode);
if (pVnode->pTq) tqClose(pVnode->pTq);
if (pVnode->pWal) walClose(pVnode->pWal);
if (pVnode->pTsdb) tsdbClose(&pVnode->pTsdb);
if (pVnode->pMeta) metaClose(pVnode->pMeta);
if (pVnode->pSma) smaClose(pVnode->pSma);
tsem_destroy(&(pVnode->canCommit));
taosMemoryFree(pVnode);
@ -166,13 +167,14 @@ _err:
void vnodeClose(SVnode *pVnode) {
if (pVnode) {
smaCloseEnv(pVnode->pSma);
vnodeCommit(pVnode);
vnodeSyncClose(pVnode);
vnodeQueryClose(pVnode);
walClose(pVnode->pWal);
tqClose(pVnode->pTq);
if (pVnode->pTsdb) tsdbClose(&pVnode->pTsdb);
smaClose(pVnode->pSma);
smaCloseEx(pVnode->pSma);
metaClose(pVnode->pMeta);
vnodeCloseBufPool(pVnode);
// destroy handle

2
source/dnode/vnode/src/vnd/vnodeSvr.c
View File

@ -299,7 +299,7 @@ int32_t vnodeProcessWriteMsg(SVnode *pVnode, int64_t version, SRpcMsg *pMsg, SRp
void smaHandleRes(void *pVnode, int64_t smaId, const SArray *data) {
// TODO
blockDebugShowData(data, __func__);
blockDebugShowDataBlocks(data, __func__);
tdProcessTSmaInsert(((SVnode *)pVnode)->pSma, smaId, (const char *)data);
}

69
source/dnode/vnode/src/vnd/vnodeSync.c
View File

@ -137,6 +137,25 @@ void vnodeProposeMsg(SQueueInfo *pInfo, STaosQall *qall, int32_t numOfMsgs) {
vError("vgId:%d, failed to pre-process msg:%p since %s", vgId, pMsg, terrstr());
} else {
code = syncPropose(pVnode->sync, pMsg, vnodeIsMsgWeak(pMsg->msgType));
if (code == 1) {
do {
static int32_t cnt = 0;
if (cnt++ % 1000 == 1) {
vInfo("vgId:%d, msg:%p apply right now, apply index:%ld, msgtype:%s,%d", vgId, pMsg,
pMsg->info.conn.applyIndex, TMSG_INFO(pMsg->msgType), pMsg->msgType);
}
} while (0);
SRpcMsg rsp = {.code = pMsg->code, .info = pMsg->info};
if (vnodeProcessWriteReq(pVnode, pMsg, pMsg->info.conn.applyIndex, &rsp) < 0) {
rsp.code = terrno;
vInfo("vgId:%d, msg:%p failed to apply right now since %s", vgId, pMsg, terrstr());
}
if (rsp.info.handle != NULL) {
tmsgSendRsp(&rsp);
}
}
}
}
@ -163,10 +182,12 @@ void vnodeProposeMsg(SQueueInfo *pInfo, STaosQall *qall, int32_t numOfMsgs) {
SRpcMsg rsp = {.code = TSDB_CODE_RPC_REDIRECT, .info = pMsg->info};
tmsgSendRedirectRsp(&rsp, &newEpSet);
} else {
if (terrno != 0) code = terrno;
vError("vgId:%d, msg:%p failed to propose since %s, code:0x%x", vgId, pMsg, tstrerror(code), code);
SRpcMsg rsp = {.code = code, .info = pMsg->info};
tmsgSendRsp(&rsp);
if (code != 1) {
if (terrno != 0) code = terrno;
vError("vgId:%d, msg:%p failed to propose since %s, code:0x%x", vgId, pMsg, tstrerror(code), code);
SRpcMsg rsp = {.code = code, .info = pMsg->info};
tmsgSendRsp(&rsp);
}
}
vGTrace("vgId:%d, msg:%p is freed, code:0x%x", vgId, pMsg, code);
@ -260,7 +281,7 @@ int32_t vnodeProcessSyncReq(SVnode *pVnode, SRpcMsg *pMsg, SRpcMsg **pRsp) {
SyncClientRequest *pSyncMsg = syncClientRequestFromRpcMsg2(pRpcMsg);
assert(pSyncMsg != NULL);
ret = syncNodeOnClientRequestCb(pSyncNode, pSyncMsg);
ret = syncNodeOnClientRequestCb(pSyncNode, pSyncMsg, NULL);
syncClientRequestDestroy(pSyncMsg);
} else if (pRpcMsg->msgType == TDMT_SYNC_REQUEST_VOTE) {
@ -359,34 +380,18 @@ static void vnodeSyncCommitMsg(SSyncFSM *pFsm, const SRpcMsg *pMsg, SFsmCbMeta c
SyncIndex beginIndex = SYNC_INDEX_INVALID;
char logBuf[256] = {0};
if (pFsm->FpGetSnapshotInfo != NULL) {
(*pFsm->FpGetSnapshotInfo)(pFsm, &snapshot);
beginIndex = snapshot.lastApplyIndex;
}
snprintf(logBuf, sizeof(logBuf),
"==callback== ==CommitCb== execute, pFsm:%p, index:%ld, isWeak:%d, code:%d, state:%d %s, beginIndex :%ld\n",
pFsm, cbMeta.index, cbMeta.isWeak, cbMeta.code, cbMeta.state, syncUtilState2String(cbMeta.state),
beginIndex);
syncRpcMsgLog2(logBuf, (SRpcMsg *)pMsg);
if (cbMeta.index > beginIndex) {
snprintf(
logBuf, sizeof(logBuf),
"==callback== ==CommitCb== execute, pFsm:%p, index:%ld, isWeak:%d, code:%d, state:%d %s, beginIndex :%ld\n",
pFsm, cbMeta.index, cbMeta.isWeak, cbMeta.code, cbMeta.state, syncUtilState2String(cbMeta.state), beginIndex);
syncRpcMsgLog2(logBuf, (SRpcMsg *)pMsg);
SRpcMsg rpcMsg = {.msgType = pMsg->msgType, .contLen = pMsg->contLen};
rpcMsg.pCont = rpcMallocCont(rpcMsg.contLen);
memcpy(rpcMsg.pCont, pMsg->pCont, pMsg->contLen);
syncGetAndDelRespRpc(pVnode->sync, cbMeta.seqNum, &rpcMsg.info);
rpcMsg.info.conn.applyIndex = cbMeta.index;
tmsgPutToQueue(&pVnode->msgCb, APPLY_QUEUE, &rpcMsg);
} else {
char logBuf[256] = {0};
snprintf(logBuf, sizeof(logBuf),
"==callback== ==CommitCb== do not execute, pFsm:%p, index:%ld, isWeak:%d, code:%d, state:%d %s, "
"beginIndex :%ld\n",
pFsm, cbMeta.index, cbMeta.isWeak, cbMeta.code, cbMeta.state, syncUtilState2String(cbMeta.state),
beginIndex);
syncRpcMsgLog2(logBuf, (SRpcMsg *)pMsg);
}
SRpcMsg rpcMsg = {.msgType = pMsg->msgType, .contLen = pMsg->contLen};
rpcMsg.pCont = rpcMallocCont(rpcMsg.contLen);
memcpy(rpcMsg.pCont, pMsg->pCont, pMsg->contLen);
syncGetAndDelRespRpc(pVnode->sync, cbMeta.seqNum, &rpcMsg.info);
rpcMsg.info.conn.applyIndex = cbMeta.index;
tmsgPutToQueue(&pVnode->msgCb, APPLY_QUEUE, &rpcMsg);
}
static void vnodeSyncPreCommitMsg(SSyncFSM *pFsm, const SRpcMsg *pMsg, SFsmCbMeta cbMeta) {

12
source/libs/executor/inc/executorimpl.h
View File

@ -403,7 +403,7 @@ typedef struct SIntervalAggOperatorInfo {
// SOptrBasicInfo should be first, SAggSupporter should be second for stream encode
SOptrBasicInfo binfo; // basic info
SAggSupporter aggSup; // aggregate supporter
SExprSupp scalarSupp; // supporter for perform scalar function
SGroupResInfo groupResInfo; // multiple results build supporter
SInterval interval; // interval info
int32_t primaryTsIndex; // primary time stamp slot id from result of downstream operator.
@ -730,15 +730,12 @@ SOperatorInfo* createAggregateOperatorInfo(SOperatorInfo* downstream, SExprInfo*
SOperatorInfo* createIndefinitOutputOperatorInfo(SOperatorInfo* downstream, SPhysiNode *pNode, SExecTaskInfo* pTaskInfo);
SOperatorInfo* createProjectOperatorInfo(SOperatorInfo* downstream, SProjectPhysiNode* pProjPhyNode, SExecTaskInfo* pTaskInfo);
SOperatorInfo* createSortOperatorInfo(SOperatorInfo* downstream, SSortPhysiNode* pSortPhyNode, SExecTaskInfo* pTaskInfo);
SOperatorInfo* createMultiwaySortMergeOperatorInfo(SOperatorInfo** downStreams, int32_t numStreams, SSDataBlock* pInputBlock,
SSDataBlock* pResBlock, SArray* pSortInfo, SArray* pColMatchColInfo,
SExecTaskInfo* pTaskInfo);
SOperatorInfo* createMultiwayMergeOperatorInfo(SOperatorInfo** dowStreams, size_t numStreams, SMergePhysiNode* pMergePhysiNode, SExecTaskInfo* pTaskInfo);
SOperatorInfo* createSortedMergeOperatorInfo(SOperatorInfo** downstream, int32_t numOfDownstream, SExprInfo* pExprInfo, int32_t num, SArray* pSortInfo, SArray* pGroupInfo, SExecTaskInfo* pTaskInfo);
SOperatorInfo* createIntervalOperatorInfo(SOperatorInfo* downstream, SExprInfo* pExprInfo, int32_t numOfCols,
SSDataBlock* pResBlock, SInterval* pInterval, int32_t primaryTsSlotId,
STimeWindowAggSupp *pTwAggSupp, SExecTaskInfo* pTaskInfo, bool isStream);
STimeWindowAggSupp* pTwAggSupp, SIntervalPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo, bool isStream);
SOperatorInfo* createMergeIntervalOperatorInfo(SOperatorInfo* downstream, SExprInfo* pExprInfo, int32_t numOfCols,
SSDataBlock* pResBlock, SInterval* pInterval, int32_t primaryTsSlotId,
@ -808,9 +805,10 @@ int32_t getMaximumIdleDurationSec();
* ops: root operator
* data: *data save the result of encode, need to be freed by caller
* length: *length save the length of *data
* nOptrWithVal: *nOptrWithVal save the number of optr with value
* return: result code, 0 means success
*/
int32_t encodeOperator(SOperatorInfo* ops, char** data, int32_t *length);
int32_t encodeOperator(SOperatorInfo* ops, char** data, int32_t *length, int32_t *nOptrWithVal);
/*
* ops: root operator, created by caller

7
source/libs/executor/inc/tsort.h
View File

@ -136,6 +136,13 @@ bool tsortIsNullVal(STupleHandle* pVHandle, int32_t colId);
*/
void* tsortGetValue(STupleHandle* pVHandle, int32_t colId);
/**
*
* @param pVHandle
* @return
*/
uint64_t tsortGetGroupId(STupleHandle* pVHandle);
/**
*
* @param pSortHandle

7
source/libs/executor/src/executil.c
View File

@ -159,7 +159,12 @@ int32_t getNumOfTotalRes(SGroupResInfo* pGroupResInfo) {
}
SArray* createSortInfo(SNodeList* pNodeList) {
size_t numOfCols = LIST_LENGTH(pNodeList);
size_t numOfCols = 0;
if (pNodeList != NULL) {
numOfCols = LIST_LENGTH(pNodeList);
} else {
numOfCols = 0;
}
SArray* pList = taosArrayInit(numOfCols, sizeof(SBlockOrderInfo));
if (pList == NULL) {
terrno = TSDB_CODE_OUT_OF_MEMORY;

8
source/libs/executor/src/executorMain.c
View File

@ -222,7 +222,13 @@ int32_t qSerializeTaskStatus(qTaskInfo_t tinfo, char** pOutput, int32_t* len) {
return TSDB_CODE_INVALID_PARA;
}
return encodeOperator(pTaskInfo->pRoot, pOutput, len);
int32_t nOptrWithVal = 0;
int32_t code = encodeOperator(pTaskInfo->pRoot, pOutput, len, &nOptrWithVal);
if ((code == TSDB_CODE_SUCCESS) && (nOptrWithVal = 0)) {
taosMemoryFreeClear(*pOutput);
*len = 0;
}
return code;
}
int32_t qDeserializeTaskStatus(qTaskInfo_t tinfo, const char* pInput, int32_t len) {

37
source/libs/executor/src/executorimpl.c
View File

@ -2929,6 +2929,13 @@ int32_t aggEncodeResultRow(SOperatorInfo* pOperator, char** result, int32_t* len
int32_t totalSize =
sizeof(int32_t) + sizeof(int32_t) + size * (sizeof(int32_t) + keyLen + sizeof(int32_t) + pSup->resultRowSize);
// no result
if (getTotalBufSize(pSup->pResultBuf) == 0) {
*result = NULL;
*length = 0;
return TSDB_CODE_SUCCESS;
}
*result = (char*)taosMemoryCalloc(1, totalSize);
if (*result == NULL) {
return TSDB_CODE_OUT_OF_MEMORY;
@ -4217,8 +4224,8 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
int32_t tsSlotId = ((SColumnNode*)pIntervalPhyNode->window.pTspk)->slotId;
bool isStream = (QUERY_NODE_PHYSICAL_PLAN_STREAM_INTERVAL == type);
pOptr =
createIntervalOperatorInfo(ops[0], pExprInfo, num, pResBlock, &interval, tsSlotId, &as, pTaskInfo, isStream);
pOptr = createIntervalOperatorInfo(ops[0], pExprInfo, num, pResBlock, &interval, tsSlotId, &as, pIntervalPhyNode,
pTaskInfo, isStream);
} else if (QUERY_NODE_PHYSICAL_PLAN_MERGE_ALIGNED_INTERVAL == type) {
SMergeAlignedIntervalPhysiNode* pIntervalPhyNode = (SMergeAlignedIntervalPhysiNode*)pPhyNode;
@ -4262,17 +4269,7 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
pOptr = createGroupSortOperatorInfo(ops[0], (SGroupSortPhysiNode*)pPhyNode, pTaskInfo);
} else if (QUERY_NODE_PHYSICAL_PLAN_MERGE == type) {
SMergePhysiNode* pMergePhyNode = (SMergePhysiNode*)pPhyNode;
SDataBlockDescNode* pDescNode = pPhyNode->pOutputDataBlockDesc;
SSDataBlock* pResBlock = createResDataBlock(pDescNode);
SArray* sortInfo = createSortInfo(pMergePhyNode->pMergeKeys);
int32_t numOfOutputCols = 0;
SArray* pColList =
extractColMatchInfo(pMergePhyNode->pTargets, pDescNode, &numOfOutputCols, COL_MATCH_FROM_SLOT_ID);
SPhysiNode* pChildNode = (SPhysiNode*)nodesListGetNode(pPhyNode->pChildren, 0);
SSDataBlock* pInputDataBlock = createResDataBlock(pChildNode->pOutputDataBlockDesc);
pOptr = createMultiwaySortMergeOperatorInfo(ops, size, pInputDataBlock, pResBlock, sortInfo, pColList, pTaskInfo);
pOptr = createMultiwayMergeOperatorInfo(ops, size, pMergePhyNode, pTaskInfo);
} else if (QUERY_NODE_PHYSICAL_PLAN_MERGE_SESSION == type) {
SSessionWinodwPhysiNode* pSessionNode = (SSessionWinodwPhysiNode*)pPhyNode;
@ -4465,12 +4462,12 @@ int32_t rebuildReader(SOperatorInfo* pOperator, SSubplan* plan, SReadHandle* pHa
return 0;
}
int32_t encodeOperator(SOperatorInfo* ops, char** result, int32_t* length) {
int32_t encodeOperator(SOperatorInfo* ops, char** result, int32_t* length, int32_t* nOptrWithVal) {
int32_t code = TDB_CODE_SUCCESS;
char* pCurrent = NULL;
int32_t currLength = 0;
if (ops->fpSet.encodeResultRow) {
if (result == NULL || length == NULL) {
if (result == NULL || length == NULL || nOptrWithVal == NULL) {
return TSDB_CODE_TSC_INVALID_INPUT;
}
code = ops->fpSet.encodeResultRow(ops, &pCurrent, &currLength);
@ -4481,8 +4478,15 @@ int32_t encodeOperator(SOperatorInfo* ops, char** result, int32_t* length) {
*result = NULL;
}
return code;
} else if (currLength == 0) {
ASSERT(!pCurrent);
goto _downstream;
}
++(*nOptrWithVal);
ASSERT(currLength >= 0);
if (*result == NULL) {
*result = (char*)taosMemoryCalloc(1, currLength + sizeof(int32_t));
if (*result == NULL) {
@ -4508,8 +4512,9 @@ int32_t encodeOperator(SOperatorInfo* ops, char** result, int32_t* length) {
*length = *(int32_t*)(*result);
}
_downstream:
for (int32_t i = 0; i < ops->numOfDownstream; ++i) {
code = encodeOperator(ops->pDownstream[i], result, length);
code = encodeOperator(ops->pDownstream[i], result, length, nOptrWithVal);
if (code != TDB_CODE_SUCCESS) {
return code;
}

10
source/libs/executor/src/scanoperator.c
View File

@ -448,7 +448,7 @@ static SSDataBlock* doTableScanGroup(SOperatorInfo* pOperator) {
qDebug("%s start to repeat ascending order scan data blocks due to query func required", GET_TASKID(pTaskInfo));
for (int32_t i = 0; i < pTableScanInfo->cond.numOfTWindows; ++i) {
STimeWindow* pWin = &pTableScanInfo->cond.twindows[i];
qDebug("%s\t qrange:%" PRId64 "-%" PRId64, GET_TASKID(pTaskInfo), pWin->skey, pWin->ekey);
qDebug("%s qrange:%" PRId64 "-%" PRId64, GET_TASKID(pTaskInfo), pWin->skey, pWin->ekey);
}
// do prepare for the next round table scan operation
tsdbResetReadHandle(pTableScanInfo->dataReader, &pTableScanInfo->cond, 0);
@ -467,7 +467,7 @@ static SSDataBlock* doTableScanGroup(SOperatorInfo* pOperator) {
qDebug("%s start to descending order scan data blocks due to query func required", GET_TASKID(pTaskInfo));
for (int32_t i = 0; i < pTableScanInfo->cond.numOfTWindows; ++i) {
STimeWindow* pWin = &pTableScanInfo->cond.twindows[i];
qDebug("%s\t qrange:%" PRId64 "-%" PRId64, GET_TASKID(pTaskInfo), pWin->skey, pWin->ekey);
qDebug("%s qrange:%" PRId64 "-%" PRId64, GET_TASKID(pTaskInfo), pWin->skey, pWin->ekey);
}
while (pTableScanInfo->scanTimes < total) {
@ -492,7 +492,7 @@ static SSDataBlock* doTableScanGroup(SOperatorInfo* pOperator) {
GET_TASKID(pTaskInfo));
for (int32_t i = 0; i < pTableScanInfo->cond.numOfTWindows; ++i) {
STimeWindow* pWin = &pTableScanInfo->cond.twindows[i];
qDebug("%s\t qrange:%" PRId64 "-%" PRId64, GET_TASKID(pTaskInfo), pWin->skey, pWin->ekey);
qDebug("%s qrange:%" PRId64 "-%" PRId64, GET_TASKID(pTaskInfo), pWin->skey, pWin->ekey);
}
tsdbResetReadHandle(pTableScanInfo->dataReader, &pTableScanInfo->cond, 0);
pTableScanInfo->curTWinIdx = 0;
@ -2053,7 +2053,7 @@ int32_t createScanTableListInfo(STableScanPhysiNode* pTableScanNode, SReadHandle
qDebug("no table qualified for query, TID:0x%" PRIx64 ", QID:0x%" PRIx64, taskId, queryId);
return TSDB_CODE_SUCCESS;
}
code = generateGroupIdMap(pTableListInfo, pHandle, pTableScanNode->pPartitionTags);
code = generateGroupIdMap(pTableListInfo, pHandle, pTableScanNode->pGroupTags);
if (code != TSDB_CODE_SUCCESS) {
return code;
}
@ -2457,7 +2457,7 @@ SOperatorInfo* createTableMergeScanOperatorInfo(STableScanPhysiNode* pTableScanN
if (pInfo == NULL || pOperator == NULL) {
goto _error;
}
if (pTableScanNode->pPartitionTags) {
if (pTableScanNode->pGroupTags) {
taosArraySort(pTableListInfo->pTableList, compareTableKeyInfoByGid);
}

117
source/libs/executor/src/sortoperator.c
View File

@ -142,7 +142,8 @@ SSDataBlock* getSortedBlockData(SSortHandle* pHandle, SSDataBlock* pDataBlock, i
SSDataBlock* loadNextDataBlock(void* param) {
SOperatorInfo* pOperator = (SOperatorInfo*)param;
return pOperator->fpSet.getNextFn(pOperator);
SSDataBlock* pBlock = pOperator->fpSet.getNextFn(pOperator);
return pBlock;
}
// todo refactor: merged with fetch fp
@ -442,7 +443,7 @@ void destroyGroupSortOperatorInfo(void* param, int32_t numOfOutput) {
SOperatorInfo* createGroupSortOperatorInfo(SOperatorInfo* downstream, SGroupSortPhysiNode* pSortPhyNode,
SExecTaskInfo* pTaskInfo) {
SGroupSortOperatorInfo* pInfo = taosMemoryCalloc(1, sizeof(SGroupSortOperatorInfo));
SOperatorInfo* pOperator = taosMemoryCalloc(1, sizeof(SOperatorInfo));
SOperatorInfo* pOperator = taosMemoryCalloc(1, sizeof(SOperatorInfo));
if (pInfo == NULL || pOperator == NULL /* || rowSize > 100 * 1024 * 1024*/) {
goto _error;
}
@ -474,8 +475,8 @@ SOperatorInfo* createGroupSortOperatorInfo(SOperatorInfo* downstream, SGroupSort
pOperator->exprSupp.numOfExprs = numOfCols;
pOperator->pTaskInfo = pTaskInfo;
pOperator->fpSet = createOperatorFpSet(operatorDummyOpenFn, doGroupSort, NULL, NULL, destroyGroupSortOperatorInfo, NULL,
NULL, getGroupSortExplainExecInfo);
pOperator->fpSet = createOperatorFpSet(operatorDummyOpenFn, doGroupSort, NULL, NULL, destroyGroupSortOperatorInfo,
NULL, NULL, getGroupSortExplainExecInfo);
int32_t code = appendDownstream(pOperator, &downstream, 1);
if (code != TSDB_CODE_SUCCESS) {
@ -493,7 +494,7 @@ _error:
//=====================================================================================
// Multiway Sort Merge operator
typedef struct SMultiwaySortMergeOperatorInfo {
typedef struct SMultiwayMergeOperatorInfo {
SOptrBasicInfo binfo;
int32_t bufPageSize;
@ -503,14 +504,17 @@ typedef struct SMultiwaySortMergeOperatorInfo {
SSortHandle* pSortHandle;
SArray* pColMatchInfo; // for index map from table scan output
SSDataBlock* pInputBlock;
int64_t startTs; // sort start time
uint64_t groupId;
} SMultiwaySortMergeOperatorInfo;
SSDataBlock* pInputBlock;
int64_t startTs; // sort start time
bool groupSort;
bool hasGroupId;
uint64_t groupId;
STupleHandle* prefetchedTuple;
} SMultiwayMergeOperatorInfo;
int32_t doOpenMultiwaySortMergeOperator(SOperatorInfo* pOperator) {
SMultiwaySortMergeOperatorInfo* pInfo = pOperator->info;
SExecTaskInfo* pTaskInfo = pOperator->pTaskInfo;
int32_t doOpenMultiwayMergeOperator(SOperatorInfo* pOperator) {
SMultiwayMergeOperatorInfo* pInfo = pOperator->info;
SExecTaskInfo* pTaskInfo = pOperator->pTaskInfo;
if (OPTR_IS_OPENED(pOperator)) {
return TSDB_CODE_SUCCESS;
@ -524,8 +528,9 @@ int32_t doOpenMultiwaySortMergeOperator(SOperatorInfo* pOperator) {
pInfo->pInputBlock, pTaskInfo->id.str);
tsortSetFetchRawDataFp(pInfo->pSortHandle, loadNextDataBlock, NULL, NULL);
tsortSetCompareGroupId(pInfo->pSortHandle, true);
tsortSetCompareGroupId(pInfo->pSortHandle, pInfo->groupSort);
for (int32_t i = 0; i < pOperator->numOfDownstream; ++i) {
SSortSource* ps = taosMemoryCalloc(1, sizeof(SSortSource));
ps->param = pOperator->pDownstream[i];
@ -547,8 +552,8 @@ int32_t doOpenMultiwaySortMergeOperator(SOperatorInfo* pOperator) {
SSDataBlock* getMultiwaySortedBlockData(SSortHandle* pHandle, SSDataBlock* pDataBlock, int32_t capacity,
SArray* pColMatchInfo, SOperatorInfo* pOperator) {
SMultiwaySortMergeOperatorInfo* pInfo = pOperator->info;
SExecTaskInfo* pTaskInfo = pOperator->pTaskInfo;
SMultiwayMergeOperatorInfo* pInfo = pOperator->info;
SExecTaskInfo* pTaskInfo = pOperator->pTaskInfo;
blockDataCleanup(pDataBlock);
@ -560,18 +565,47 @@ SSDataBlock* getMultiwaySortedBlockData(SSortHandle* pHandle, SSDataBlock* pData
blockDataEnsureCapacity(p, capacity);
while (1) {
STupleHandle* pTupleHandle = tsortNextTuple(pHandle);
STupleHandle* pTupleHandle = NULL;
if (pInfo->groupSort) {
if (pInfo->prefetchedTuple == NULL) {
pTupleHandle = tsortNextTuple(pHandle);
} else {
pTupleHandle = pInfo->prefetchedTuple;
pInfo->groupId = tsortGetGroupId(pTupleHandle);
pInfo->prefetchedTuple = NULL;
}
}
else {
pTupleHandle = tsortNextTuple(pHandle);
pInfo->groupId = 0;
}
if (pTupleHandle == NULL) {
break;
}
appendOneRowToDataBlock(p, pTupleHandle);
if (pInfo->groupSort)
{
uint64_t tupleGroupId = tsortGetGroupId(pTupleHandle);
if (!pInfo->hasGroupId) {
pInfo->groupId = tupleGroupId;
pInfo->hasGroupId = true;
appendOneRowToDataBlock(p, pTupleHandle);
} else if (pInfo->groupId == tupleGroupId) {
appendOneRowToDataBlock(p, pTupleHandle);
} else {
pInfo->prefetchedTuple = pTupleHandle;
break;
}
} else {
appendOneRowToDataBlock(p, pTupleHandle);
}
if (p->info.rows >= capacity) {
break;
}
}
if (p->info.rows > 0) {// todo extract method
if (p->info.rows > 0) { // todo extract method
blockDataEnsureCapacity(pDataBlock, p->info.rows);
int32_t numOfCols = taosArrayGetSize(pColMatchInfo);
for (int32_t i = 0; i < numOfCols; ++i) {
@ -593,13 +627,13 @@ SSDataBlock* getMultiwaySortedBlockData(SSortHandle* pHandle, SSDataBlock* pData
return (pDataBlock->info.rows > 0) ? pDataBlock : NULL;
}
SSDataBlock* doMultiwaySortMerge(SOperatorInfo* pOperator) {
SSDataBlock* doMultiwayMerge(SOperatorInfo* pOperator) {
if (pOperator->status == OP_EXEC_DONE) {
return NULL;
}
SExecTaskInfo* pTaskInfo = pOperator->pTaskInfo;
SMultiwaySortMergeOperatorInfo* pInfo = pOperator->info;
SExecTaskInfo* pTaskInfo = pOperator->pTaskInfo;
SMultiwayMergeOperatorInfo* pInfo = pOperator->info;
int32_t code = pOperator->fpSet._openFn(pOperator);
if (code != TSDB_CODE_SUCCESS) {
@ -608,7 +642,6 @@ SSDataBlock* doMultiwaySortMerge(SOperatorInfo* pOperator) {
SSDataBlock* pBlock = getMultiwaySortedBlockData(pInfo->pSortHandle, pInfo->binfo.pRes,
pOperator->resultInfo.capacity, pInfo->pColMatchInfo, pOperator);
if (pBlock != NULL) {
pOperator->resultInfo.totalRows += pBlock->info.rows;
} else {
@ -617,8 +650,8 @@ SSDataBlock* doMultiwaySortMerge(SOperatorInfo* pOperator) {
return pBlock;
}
void destroyMultiwaySortMergeOperatorInfo(void* param, int32_t numOfOutput) {
SMultiwaySortMergeOperatorInfo* pInfo = (SMultiwaySortMergeOperatorInfo*)param;
void destroyMultiwayMergeOperatorInfo(void* param, int32_t numOfOutput) {
SMultiwayMergeOperatorInfo* pInfo = (SMultiwayMergeOperatorInfo*)param;
pInfo->binfo.pRes = blockDataDestroy(pInfo->binfo.pRes);
pInfo->pInputBlock = blockDataDestroy(pInfo->pInputBlock);
@ -626,11 +659,11 @@ void destroyMultiwaySortMergeOperatorInfo(void* param, int32_t numOfOutput) {
taosArrayDestroy(pInfo->pColMatchInfo);
}
int32_t getMultiwaySortMergeExplainExecInfo(SOperatorInfo* pOptr, void** pOptrExplain, uint32_t* len) {
int32_t getMultiwayMergeExplainExecInfo(SOperatorInfo* pOptr, void** pOptrExplain, uint32_t* len) {
ASSERT(pOptr != NULL);
SSortExecInfo* pInfo = taosMemoryCalloc(1, sizeof(SSortExecInfo));
SMultiwaySortMergeOperatorInfo* pOperatorInfo = (SMultiwaySortMergeOperatorInfo*)pOptr->info;
SMultiwayMergeOperatorInfo* pOperatorInfo = (SMultiwayMergeOperatorInfo*)pOptr->info;
*pInfo = tsortGetSortExecInfo(pOperatorInfo->pSortHandle);
*pOptrExplain = pInfo;
@ -638,24 +671,35 @@ int32_t getMultiwaySortMergeExplainExecInfo(SOperatorInfo* pOptr, void** pOptrEx
return TSDB_CODE_SUCCESS;
}
SOperatorInfo* createMultiwaySortMergeOperatorInfo(SOperatorInfo** downStreams, int32_t numStreams,
SSDataBlock* pInputBlock, SSDataBlock* pResBlock, SArray* pSortInfo,
SArray* pColMatchColInfo, SExecTaskInfo* pTaskInfo) {
SMultiwaySortMergeOperatorInfo* pInfo = taosMemoryCalloc(1, sizeof(SMultiwaySortMergeOperatorInfo));
SOperatorInfo* pOperator = taosMemoryCalloc(1, sizeof(SOperatorInfo));
int32_t rowSize = pResBlock->info.rowSize;
SOperatorInfo* createMultiwayMergeOperatorInfo(SOperatorInfo** downStreams, size_t numStreams,
SMergePhysiNode* pMergePhyNode, SExecTaskInfo* pTaskInfo) {
SPhysiNode* pPhyNode = (SPhysiNode*)pMergePhyNode;
SMultiwayMergeOperatorInfo* pInfo = taosMemoryCalloc(1, sizeof(SMultiwayMergeOperatorInfo));
SOperatorInfo* pOperator = taosMemoryCalloc(1, sizeof(SOperatorInfo));
SDataBlockDescNode* pDescNode = pPhyNode->pOutputDataBlockDesc;
SSDataBlock* pResBlock = createResDataBlock(pDescNode);
int32_t rowSize = pResBlock->info.rowSize;
if (pInfo == NULL || pOperator == NULL || rowSize > 100 * 1024 * 1024) {
goto _error;
}
SArray* pSortInfo = createSortInfo(pMergePhyNode->pMergeKeys);
int32_t numOfOutputCols = 0;
SArray* pColMatchColInfo =
extractColMatchInfo(pMergePhyNode->pTargets, pDescNode, &numOfOutputCols, COL_MATCH_FROM_SLOT_ID);
SPhysiNode* pChildNode = (SPhysiNode*)nodesListGetNode(pPhyNode->pChildren, 0);
SSDataBlock* pInputBlock = createResDataBlock(pChildNode->pOutputDataBlockDesc);
initResultSizeInfo(pOperator, 1024);
pInfo->groupSort = pMergePhyNode->groupSort;
pInfo->binfo.pRes = pResBlock;
pInfo->pSortInfo = pSortInfo;
pInfo->pColMatchInfo = pColMatchColInfo;
pInfo->pInputBlock = pInputBlock;
pOperator->name = "MultiwaySortMerge";
pOperator->name = "MultiwayMerge";
pOperator->operatorType = QUERY_NODE_PHYSICAL_PLAN_MERGE;
pOperator->blocking = false;
pOperator->status = OP_NOT_OPENED;
@ -667,9 +711,8 @@ SOperatorInfo* createMultiwaySortMergeOperatorInfo(SOperatorInfo** downStreams,
// one additional is reserved for merged result.
pInfo->sortBufSize = pInfo->bufPageSize * (numStreams + 1);
pOperator->fpSet =
createOperatorFpSet(doOpenMultiwaySortMergeOperator, doMultiwaySortMerge, NULL, NULL,
destroyMultiwaySortMergeOperatorInfo, NULL, NULL, getMultiwaySortMergeExplainExecInfo);
pOperator->fpSet = createOperatorFpSet(doOpenMultiwayMergeOperator, doMultiwayMerge, NULL, NULL,
destroyMultiwayMergeOperatorInfo, NULL, NULL, getMultiwayMergeExplainExecInfo);
int32_t code = appendDownstream(pOperator, downStreams, numStreams);
if (code != TSDB_CODE_SUCCESS) {

32
source/libs/executor/src/timewindowoperator.c
View File

@ -969,6 +969,12 @@ static int32_t doOpenIntervalAgg(SOperatorInfo* pOperator) {
getTableScanInfo(pOperator, &pInfo->order, &scanFlag);
if (pInfo->scalarSupp.pExprInfo != NULL) {
SExprSupp* pExprSup =& pInfo->scalarSupp;
projectApplyFunctions(pExprSup->pExprInfo, pBlock, pBlock, pExprSup->pCtx,
pExprSup->numOfExprs, NULL);
}
// the pDataBlock are always the same one, no need to call this again
setInputDataBlock(pOperator, pSup->pCtx, pBlock, pInfo->order, scanFlag, true);
hashIntervalAgg(pOperator, &pInfo->binfo.resultRowInfo, pBlock, scanFlag, NULL);
@ -1381,6 +1387,11 @@ static SSDataBlock* doStreamIntervalAgg(SOperatorInfo* pOperator) {
continue;
}
if (pInfo->scalarSupp.pExprInfo != NULL) {
SExprSupp* pExprSup = &pInfo->scalarSupp;
projectApplyFunctions(pExprSup->pExprInfo, pBlock, pBlock, pExprSup->pCtx, pExprSup->numOfExprs, NULL);
}
// The timewindow that overlaps the timestamps of the input pBlock need to be recalculated and return to the
// caller. Note that all the time window are not close till now.
// the pDataBlock are always the same one, no need to call this again
@ -1498,18 +1509,27 @@ void increaseTs(SqlFunctionCtx* pCtx) {
SOperatorInfo* createIntervalOperatorInfo(SOperatorInfo* downstream, SExprInfo* pExprInfo, int32_t numOfCols,
SSDataBlock* pResBlock, SInterval* pInterval, int32_t primaryTsSlotId,
STimeWindowAggSupp* pTwAggSupp, SExecTaskInfo* pTaskInfo, bool isStream) {
STimeWindowAggSupp* pTwAggSupp, SIntervalPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo, bool isStream) {
SIntervalAggOperatorInfo* pInfo = taosMemoryCalloc(1, sizeof(SIntervalAggOperatorInfo));
SOperatorInfo* pOperator = taosMemoryCalloc(1, sizeof(SOperatorInfo));
if (pInfo == NULL || pOperator == NULL) {
goto _error;
}
pInfo->win = pTaskInfo->window;
pInfo->order = TSDB_ORDER_ASC;
pInfo->interval = *pInterval;
pInfo->win = pTaskInfo->window;
pInfo->order = TSDB_ORDER_ASC;
pInfo->interval = *pInterval;
pInfo->execModel = pTaskInfo->execModel;
pInfo->twAggSup = *pTwAggSupp;
pInfo->twAggSup = *pTwAggSupp;
if (pPhyNode->window.pExprs != NULL) {
int32_t numOfScalar = 0;
SExprInfo* pScalarExprInfo = createExprInfo(pPhyNode->window.pExprs, NULL, &numOfScalar);
int32_t code = initExprSupp(&pInfo->scalarSupp, pScalarExprInfo, numOfScalar);
if (code != TSDB_CODE_SUCCESS) {
goto _error;
}
}
pInfo->primaryTsIndex = primaryTsSlotId;
@ -2473,7 +2493,7 @@ static SSDataBlock* doStreamFinalIntervalAgg(SOperatorInfo* pOperator) {
if (IS_FINAL_OP(pInfo)) {
int32_t childIndex = getChildIndex(pBlock);
SOperatorInfo* pChildOp = taosArrayGetP(pInfo->pChildren, childIndex);
SIntervalAggOperatorInfo* pChildInfo = pChildOp->info;
SStreamFinalIntervalOperatorInfo* pChildInfo = pChildOp->info;
SExprSupp* pChildSup = &pChildOp->exprSupp;
doClearWindows(&pChildInfo->aggSup, pChildSup, &pChildInfo->interval, pChildInfo->primaryTsIndex,

4
source/libs/executor/src/tsort.c
View File

@ -752,6 +752,10 @@ void* tsortGetValue(STupleHandle* pVHandle, int32_t colIndex) {
}
}
uint64_t tsortGetGroupId(STupleHandle* pVHandle) {
return pVHandle->pBlock->info.groupId;
}
SSortExecInfo tsortGetSortExecInfo(SSortHandle* pHandle) {
SSortExecInfo info = {0};

7
source/libs/function/src/udfd.c
View File

@ -110,7 +110,7 @@ static void udfdProcessRpcRsp(void *parent, SRpcMsg *pMsg, SEpSet *pEpSet);
static int32_t udfdFillUdfInfoFromMNode(void *clientRpc, char *udfName, SUdf *udf);
static int32_t udfdConnectToMnode();
static int32_t udfdLoadUdf(char *udfName, SUdf *udf);
static bool udfdRpcRfp(int32_t code);
static bool udfdRpcRfp(int32_t code, tmsg_t msgType);
static int initEpSetFromCfg(const char *firstEp, const char *secondEp, SCorEpSet *pEpSet);
static int32_t udfdOpenClientRpc();
static int32_t udfdCloseClientRpc();
@ -546,9 +546,12 @@ int32_t udfdLoadUdf(char *udfName, SUdf *udf) {
}
return 0;
}
static bool udfdRpcRfp(int32_t code) {
static bool udfdRpcRfp(int32_t code, tmsg_t msgType) {
if (code == TSDB_CODE_RPC_REDIRECT || code == TSDB_CODE_RPC_NETWORK_UNAVAIL || code == TSDB_CODE_NODE_NOT_DEPLOYED ||
code == TSDB_CODE_SYN_NOT_LEADER || code == TSDB_CODE_APP_NOT_READY) {
if (msgType == TDMT_VND_QUERY || msgType == TDMT_VND_FETCH) {
return false;
}
return true;
} else {
return false;

6
source/libs/nodes/src/nodesCloneFuncs.c
View File

@ -351,7 +351,7 @@ static int32_t logicScanCopy(const SScanLogicNode* pSrc, SScanLogicNode* pDst) {
COPY_SCALAR_FIELD(watermark);
COPY_SCALAR_FIELD(tsColId);
COPY_SCALAR_FIELD(filesFactor);
CLONE_NODE_LIST_FIELD(pPartTags);
CLONE_NODE_LIST_FIELD(pGroupTags);
return TSDB_CODE_SUCCESS;
}
@ -401,6 +401,7 @@ static int32_t logicMergeCopy(const SMergeLogicNode* pSrc, SMergeLogicNode* pDst
CLONE_NODE_LIST_FIELD(pInputs);
COPY_SCALAR_FIELD(numOfChannels);
COPY_SCALAR_FIELD(srcGroupId);
COPY_SCALAR_FIELD(groupSort);
return TSDB_CODE_SUCCESS;
}
@ -436,6 +437,7 @@ static int32_t logicFillCopy(const SFillLogicNode* pSrc, SFillLogicNode* pDst) {
static int32_t logicSortCopy(const SSortLogicNode* pSrc, SSortLogicNode* pDst) {
COPY_BASE_OBJECT_FIELD(node, logicNodeCopy);
CLONE_NODE_LIST_FIELD(pSortKeys);
COPY_SCALAR_FIELD(groupSort);
return TSDB_CODE_SUCCESS;
}
@ -500,7 +502,7 @@ static int32_t physiTableScanCopy(const STableScanPhysiNode* pSrc, STableScanPhy
COPY_SCALAR_FIELD(ratio);
COPY_SCALAR_FIELD(dataRequired);
CLONE_NODE_LIST_FIELD(pDynamicScanFuncs);
CLONE_NODE_LIST_FIELD(pPartitionTags);
CLONE_NODE_LIST_FIELD(pGroupTags);
COPY_SCALAR_FIELD(interval);
COPY_SCALAR_FIELD(offset);
COPY_SCALAR_FIELD(sliding);

37
source/libs/nodes/src/nodesCodeFuncs.c
View File

@ -234,6 +234,8 @@ const char* nodesNodeName(ENodeType type) {
return "PhysiMerge";
case QUERY_NODE_PHYSICAL_PLAN_SORT:
return "PhysiSort";
case QUERY_NODE_PHYSICAL_PLAN_GROUP_SORT:
return "PhysiGroupSort";
case QUERY_NODE_PHYSICAL_PLAN_HASH_INTERVAL:
return "PhysiHashInterval";
case QUERY_NODE_PHYSICAL_PLAN_MERGE_ALIGNED_INTERVAL:
@ -539,7 +541,7 @@ static const char* jkScanLogicPlanScanPseudoCols = "ScanPseudoCols";
static const char* jkScanLogicPlanTableId = "TableId";
static const char* jkScanLogicPlanTableType = "TableType";
static const char* jkScanLogicPlanTagCond = "TagCond";
static const char* jkScanLogicPlanPartTags = "PartTags";
static const char* jkScanLogicPlanGroupTags = "GroupTags";
static int32_t logicScanNodeToJson(const void* pObj, SJson* pJson) {
const SScanLogicNode* pNode = (const SScanLogicNode*)pObj;
@ -561,7 +563,7 @@ static int32_t logicScanNodeToJson(const void* pObj, SJson* pJson) {
code = tjsonAddObject(pJson, jkScanLogicPlanTagCond, nodeToJson, pNode->pTagCond);
}
if (TSDB_CODE_SUCCESS == code) {
code = nodeListToJson(pJson, jkScanLogicPlanPartTags, pNode->pPartTags);
code = nodeListToJson(pJson, jkScanLogicPlanGroupTags, pNode->pGroupTags);
}
return code;
@ -588,7 +590,7 @@ static int32_t jsonToLogicScanNode(const SJson* pJson, void* pObj) {
code = jsonToNodeObject(pJson, jkScanLogicPlanTagCond, &pNode->pTagCond);
}
if (TSDB_CODE_SUCCESS == code) {
code = jsonToNodeList(pJson, jkScanLogicPlanPartTags, &pNode->pPartTags);
code = jsonToNodeList(pJson, jkScanLogicPlanGroupTags, &pNode->pGroupTags);
}
return code;
@ -1430,7 +1432,8 @@ static const char* jkTableScanPhysiPlanTriggerType = "triggerType";
static const char* jkTableScanPhysiPlanWatermark = "watermark";
static const char* jkTableScanPhysiPlanTsColId = "tsColId";
static const char* jkTableScanPhysiPlanFilesFactor = "FilesFactor";
static const char* jkTableScanPhysiPlanPartitionTags = "PartitionTags";
static const char* jkTableScanPhysiPlanGroupTags = "GroupTags";
static const char* jkTableScanPhysiPlanGroupSort = "GroupSort";
static int32_t physiTableScanNodeToJson(const void* pObj, SJson* pJson) {
const STableScanPhysiNode* pNode = (const STableScanPhysiNode*)pObj;
@ -1485,7 +1488,10 @@ static int32_t physiTableScanNodeToJson(const void* pObj, SJson* pJson) {
code = tjsonAddDoubleToObject(pJson, jkTableScanPhysiPlanFilesFactor, pNode->filesFactor);
}
if (TSDB_CODE_SUCCESS == code) {
code = nodeListToJson(pJson, jkTableScanPhysiPlanPartitionTags, pNode->pPartitionTags);
code = nodeListToJson(pJson, jkTableScanPhysiPlanGroupTags, pNode->pGroupTags);
}
if (TSDB_CODE_SUCCESS == code) {
code = tjsonAddBoolToObject(pJson, jkTableScanPhysiPlanGroupSort, pNode->groupSort);
}
return code;
@ -1544,7 +1550,10 @@ static int32_t jsonToPhysiTableScanNode(const SJson* pJson, void* pObj) {
code = tjsonGetDoubleValue(pJson, jkTableScanPhysiPlanFilesFactor, &pNode->filesFactor);
}
if (TSDB_CODE_SUCCESS == code) {
code = jsonToNodeList(pJson, jkTableScanPhysiPlanPartitionTags, &pNode->pPartitionTags);
code = jsonToNodeList(pJson, jkTableScanPhysiPlanGroupTags, &pNode->pGroupTags);
}
if (TSDB_CODE_SUCCESS == code) {
code = tjsonGetBoolValue(pJson, jkTableScanPhysiPlanGroupSort, &pNode->groupSort);
}
return code;
@ -1725,6 +1734,7 @@ static const char* jkMergePhysiPlanMergeKeys = "MergeKeys";
static const char* jkMergePhysiPlanTargets = "Targets";
static const char* jkMergePhysiPlanNumOfChannels = "NumOfChannels";
static const char* jkMergePhysiPlanSrcGroupId = "SrcGroupId";
static const char* jkMergePhysiPlanGroupSort = "GroupSort";
static int32_t physiMergeNodeToJson(const void* pObj, SJson* pJson) {
const SMergePhysiNode* pNode = (const SMergePhysiNode*)pObj;
@ -1742,6 +1752,9 @@ static int32_t physiMergeNodeToJson(const void* pObj, SJson* pJson) {
if (TSDB_CODE_SUCCESS == code) {
code = tjsonAddIntegerToObject(pJson, jkMergePhysiPlanSrcGroupId, pNode->srcGroupId);
}
if (TSDB_CODE_SUCCESS == code) {
code = tjsonAddBoolToObject(pJson, jkMergePhysiPlanGroupSort, pNode->groupSort);
}
return code;
}
@ -1762,6 +1775,9 @@ static int32_t jsonToPhysiMergeNode(const SJson* pJson, void* pObj) {
if (TSDB_CODE_SUCCESS == code) {
code = tjsonGetIntValue(pJson, jkMergePhysiPlanSrcGroupId, &pNode->srcGroupId);
}
if (TSDB_CODE_SUCCESS == code) {
code = tjsonGetBoolValue(pJson, jkMergePhysiPlanGroupSort, &pNode->groupSort);
}
return code;
}
@ -3369,6 +3385,7 @@ static const char* jkSlotDescSlotId = "SlotId";
static const char* jkSlotDescDataType = "DataType";
static const char* jkSlotDescReserve = "Reserve";
static const char* jkSlotDescOutput = "Output";
static const char* jkSlotDescName = "Name";
static int32_t slotDescNodeToJson(const void* pObj, SJson* pJson) {
const SSlotDescNode* pNode = (const SSlotDescNode*)pObj;
@ -3383,6 +3400,9 @@ static int32_t slotDescNodeToJson(const void* pObj, SJson* pJson) {
if (TSDB_CODE_SUCCESS == code) {
code = tjsonAddBoolToObject(pJson, jkSlotDescOutput, pNode->output);
}
if (TSDB_CODE_SUCCESS == code) {
code = tjsonAddStringToObject(pJson, jkSlotDescName, pNode->name);
}
return code;
}
@ -3400,6 +3420,9 @@ static int32_t jsonToSlotDescNode(const SJson* pJson, void* pObj) {
if (TSDB_CODE_SUCCESS == code) {
code = tjsonGetBoolValue(pJson, jkSlotDescOutput, &pNode->output);
}
if (TSDB_CODE_SUCCESS == code) {
code = tjsonGetStringValue(pJson, jkSlotDescName, pNode->name);
}
return code;
}
@ -4137,6 +4160,7 @@ static int32_t specificNodeToJson(const void* pObj, SJson* pJson) {
case QUERY_NODE_PHYSICAL_PLAN_MERGE:
return physiMergeNodeToJson(pObj, pJson);
case QUERY_NODE_PHYSICAL_PLAN_SORT:
case QUERY_NODE_PHYSICAL_PLAN_GROUP_SORT:
return physiSortNodeToJson(pObj, pJson);
case QUERY_NODE_PHYSICAL_PLAN_HASH_INTERVAL:
case QUERY_NODE_PHYSICAL_PLAN_MERGE_ALIGNED_INTERVAL:
@ -4280,6 +4304,7 @@ static int32_t jsonToSpecificNode(const SJson* pJson, void* pObj) {
case QUERY_NODE_PHYSICAL_PLAN_MERGE:
return jsonToPhysiMergeNode(pJson, pObj);
case QUERY_NODE_PHYSICAL_PLAN_SORT:
case QUERY_NODE_PHYSICAL_PLAN_GROUP_SORT:
return jsonToPhysiSortNode(pJson, pObj);
case QUERY_NODE_PHYSICAL_PLAN_HASH_INTERVAL:
case QUERY_NODE_PHYSICAL_PLAN_MERGE_ALIGNED_INTERVAL:

3
source/libs/nodes/src/nodesTraverseFuncs.c
View File

@ -500,7 +500,8 @@ static EDealRes dispatchPhysiPlan(SNode* pNode, ETraversalOrder order, FNodeWalk
}
break;
}
case QUERY_NODE_PHYSICAL_PLAN_SORT: {
case QUERY_NODE_PHYSICAL_PLAN_SORT:
case QUERY_NODE_PHYSICAL_PLAN_GROUP_SORT: {
SSortPhysiNode* pSort = (SSortPhysiNode*)pNode;
res = walkPhysiNode((SPhysiNode*)pNode, order, walker, pContext);
if (DEAL_RES_ERROR != res && DEAL_RES_END != res) {

9
source/libs/nodes/src/nodesUtilFuncs.c
View File

@ -288,6 +288,8 @@ SNode* nodesMakeNode(ENodeType type) {
return makeNode(type, sizeof(SMergePhysiNode));
case QUERY_NODE_PHYSICAL_PLAN_SORT:
return makeNode(type, sizeof(SSortPhysiNode));
case QUERY_NODE_PHYSICAL_PLAN_GROUP_SORT:
return makeNode(type, sizeof(SGroupSortPhysiNode));
case QUERY_NODE_PHYSICAL_PLAN_HASH_INTERVAL:
return makeNode(type, sizeof(SIntervalPhysiNode));
case QUERY_NODE_PHYSICAL_PLAN_MERGE_ALIGNED_INTERVAL:
@ -709,7 +711,7 @@ void nodesDestroyNode(SNode* pNode) {
nodesDestroyNode(pLogicNode->pTagCond);
nodesDestroyNode(pLogicNode->pTagIndexCond);
taosArrayDestroy(pLogicNode->pSmaIndexes);
nodesDestroyList(pLogicNode->pPartTags);
nodesDestroyList(pLogicNode->pGroupTags);
break;
}
case QUERY_NODE_LOGIC_PLAN_JOIN: {
@ -813,7 +815,7 @@ void nodesDestroyNode(SNode* pNode) {
STableScanPhysiNode* pPhyNode = (STableScanPhysiNode*)pNode;
destroyScanPhysiNode((SScanPhysiNode*)pNode);
nodesDestroyList(pPhyNode->pDynamicScanFuncs);
nodesDestroyList(pPhyNode->pPartitionTags);
nodesDestroyList(pPhyNode->pGroupTags);
break;
}
case QUERY_NODE_PHYSICAL_PLAN_PROJECT: {
@ -850,7 +852,8 @@ void nodesDestroyNode(SNode* pNode) {
nodesDestroyList(pPhyNode->pTargets);
break;
}
case QUERY_NODE_PHYSICAL_PLAN_SORT: {
case QUERY_NODE_PHYSICAL_PLAN_SORT:
case QUERY_NODE_PHYSICAL_PLAN_GROUP_SORT: {
SSortPhysiNode* pPhyNode = (SSortPhysiNode*)pNode;
destroyPhysiNode((SPhysiNode*)pPhyNode);
nodesDestroyList(pPhyNode->pExprs);

87
source/libs/parser/src/parTranslater.c
View File

@ -2456,90 +2456,6 @@ static int32_t rewriteUniqueStmt(STranslateContext* pCxt, SSelectStmt* pSelect)
return cxt.pTranslateCxt->errCode;
}
typedef struct SRwriteTailCxt {
STranslateContext* pTranslateCxt;
int64_t limit;
int64_t offset;
} SRwriteTailCxt;
static EDealRes rewriteTailFunc(SNode** pNode, void* pContext) {
SRwriteTailCxt* pCxt = pContext;
if (QUERY_NODE_FUNCTION == nodeType(*pNode)) {
SFunctionNode* pFunc = (SFunctionNode*)*pNode;
if (FUNCTION_TYPE_TAIL == pFunc->funcType) {
pCxt->limit = ((SValueNode*)nodesListGetNode(pFunc->pParameterList, 1))->datum.i;
if (3 == LIST_LENGTH(pFunc->pParameterList)) {
pCxt->offset = ((SValueNode*)nodesListGetNode(pFunc->pParameterList, 2))->datum.i;
}
SNode* pExpr = nodesListGetNode(pFunc->pParameterList, 0);
strcpy(((SExprNode*)pExpr)->aliasName, ((SExprNode*)*pNode)->aliasName);
NODES_CLEAR_LIST(pFunc->pParameterList);
nodesDestroyNode(*pNode);
*pNode = pExpr;
return DEAL_RES_IGNORE_CHILD;
}
}
return DEAL_RES_CONTINUE;
}
static int32_t createLimieNode(SRwriteTailCxt* pCxt, SLimitNode** pOutput) {
*pOutput = (SLimitNode*)nodesMakeNode(QUERY_NODE_LIMIT);
if (NULL == *pOutput) {
return TSDB_CODE_OUT_OF_MEMORY;
}
(*pOutput)->limit = pCxt->limit;
(*pOutput)->offset = pCxt->offset;
return TSDB_CODE_SUCCESS;
}
static SNode* createOrderByExpr(STranslateContext* pCxt) {
SOrderByExprNode* pOrder = (SOrderByExprNode*)nodesMakeNode(QUERY_NODE_ORDER_BY_EXPR);
if (NULL == pOrder) {
return NULL;
}
pCxt->errCode = createPrimaryKeyCol(pCxt, &pOrder->pExpr);
if (TSDB_CODE_SUCCESS != pCxt->errCode) {
nodesDestroyNode((SNode*)pOrder);
return NULL;
}
pOrder->order = ORDER_DESC;
pOrder->nullOrder = NULL_ORDER_FIRST;
return (SNode*)pOrder;
}
/* case 1:
* in: select tail(expr, k, f) from t where_clause
* out: select expr from t where_clause order by _rowts desc limit k offset f
*
* case 2:
* in: select tail(expr, k, f) from t where_clause partition_by_clause
* out: select expr from t where_clause partition_by_clause sort by _rowts desc limit k offset f
*
* case 3:
* in: select tail(expr, k, f) from t where_clause order_by_clause limit_clause
* out: select expr from (
* select expr from t where_clause order by _rowts desc limit k offset f
* ) order_by_clause limit_clause
*
* case 4:
* in: select tail(expr, k, f) from t where_clause partition_by_clause limit_clause
* out:
*/
static int32_t rewriteTailStmt(STranslateContext* pCxt, SSelectStmt* pSelect) {
if (!pSelect->hasTailFunc) {
return TSDB_CODE_SUCCESS;
}
SRwriteTailCxt cxt = {.pTranslateCxt = pCxt, .limit = -1, .offset = -1};
nodesRewriteExprs(pSelect->pProjectionList, rewriteTailFunc, &cxt);
int32_t code = nodesListMakeStrictAppend(&pSelect->pOrderByList, createOrderByExpr(pCxt));
if (TSDB_CODE_SUCCESS == code) {
code = createLimieNode(&cxt, &pSelect->pLimit);
}
pSelect->hasIndefiniteRowsFunc = false;
return code;
}
typedef struct SReplaceOrderByAliasCxt {
STranslateContext* pTranslateCxt;
SNodeList* pProjectionList;
@ -2616,9 +2532,6 @@ static int32_t translateSelectFrom(STranslateContext* pCxt, SSelectStmt* pSelect
if (TSDB_CODE_SUCCESS == code) {
code = rewriteUniqueStmt(pCxt, pSelect);
}
// if (TSDB_CODE_SUCCESS == code) {
// code = rewriteTailStmt(pCxt, pSelect);
// }
if (TSDB_CODE_SUCCESS == code) {
code = rewriteTimelineFunc(pCxt, pSelect);
}

5
source/libs/parser/src/parUtil.c
View File

@ -337,7 +337,11 @@ int32_t trimString(const char* src, int32_t len, char* dst, int32_t dlen) {
static bool isValidateTag(char* input) {
if (!input) return false;
for (size_t i = 0; i < strlen(input); ++i) {
#ifdef WINDOWS
if (input[i] < 0x20 || input[i] > 0x7E) return false;
#else
if (isprint(input[i]) == 0) return false;
#endif
}
return true;
}
@ -377,6 +381,7 @@ int32_t parseJsontoTagData(const char* json, SArray* pTagVals, STag** ppTag, SMs
char* jsonKey = item->string;
if (!isValidateTag(jsonKey)) {
fprintf(stdout,"%s(%d) %s %08" PRId64 "\n", __FILE__, __LINE__,__func__,taosGetSelfPthreadId());fflush(stdout);
retCode = buildSyntaxErrMsg(pMsgBuf, "json key not validate", jsonKey);
goto end;
}

4
source/libs/planner/src/planLogicCreater.c
View File

@ -507,6 +507,8 @@ static int32_t createIndefRowsFuncLogicNode(SLogicPlanContext* pCxt, SSelectStmt
return TSDB_CODE_OUT_OF_MEMORY;
}
pIdfRowsFunc->isTailFunc = pSelect->hasTailFunc;
// indefinite rows functions and _select_values functions
int32_t code = nodesCollectFuncs(pSelect, SQL_CLAUSE_SELECT, fmIsVectorFunc, &pIdfRowsFunc->pFuncs);
if (TSDB_CODE_SUCCESS == code) {
@ -733,6 +735,8 @@ static int32_t createSortLogicNode(SLogicPlanContext* pCxt, SSelectStmt* pSelect
return TSDB_CODE_OUT_OF_MEMORY;
}
pSort->groupSort = pSelect->groupSort;
int32_t code = nodesCollectColumns(pSelect, SQL_CLAUSE_ORDER_BY, NULL, COLLECT_COL_TYPE_ALL, &pSort->node.pTargets);
if (TSDB_CODE_SUCCESS == code && NULL == pSort->node.pTargets) {
code = nodesListMakeStrictAppend(&pSort->node.pTargets,

516
source/libs/planner/src/planOptimizer.c
View File

File diff suppressed because it is too large Load Diff

40
source/libs/planner/src/planPhysiCreater.c
View File

@ -41,8 +41,12 @@ typedef struct SPhysiPlanContext {
static int32_t getSlotKey(SNode* pNode, const char* pStmtName, char* pKey) {
if (QUERY_NODE_COLUMN == nodeType(pNode)) {
SColumnNode* pCol = (SColumnNode*)pNode;
if (NULL != pStmtName && '\0' != pStmtName[0]) {
return sprintf(pKey, "%s.%s", pStmtName, pCol->node.aliasName);
if (NULL != pStmtName) {
if ('\0' != pStmtName[0]) {
return sprintf(pKey, "%s.%s", pStmtName, pCol->node.aliasName);
} else {
return sprintf(pKey, "%s", pCol->node.aliasName);
}
}
if ('\0' == pCol->tableAlias[0]) {
return sprintf(pKey, "%s", pCol->colName);
@ -56,11 +60,13 @@ static int32_t getSlotKey(SNode* pNode, const char* pStmtName, char* pKey) {
return sprintf(pKey, "%s", ((SExprNode*)pNode)->aliasName);
}
static SNode* createSlotDesc(SPhysiPlanContext* pCxt, const SNode* pNode, int16_t slotId, bool output, bool reserve) {
static SNode* createSlotDesc(SPhysiPlanContext* pCxt, const char* pName, const SNode* pNode, int16_t slotId,
bool output, bool reserve) {
SSlotDescNode* pSlot = (SSlotDescNode*)nodesMakeNode(QUERY_NODE_SLOT_DESC);
if (NULL == pSlot) {
return NULL;
}
strcpy(pSlot->name, pName);
pSlot->slotId = slotId;
pSlot->dataType = ((SExprNode*)pNode)->resType;
pSlot->reserve = reserve;
@ -99,10 +105,8 @@ static int32_t putSlotToHashImpl(int16_t dataBlockId, int16_t slotId, const char
return taosHashPut(pHash, pName, len, &index, sizeof(SSlotIndex));
}
static int32_t putSlotToHash(int16_t dataBlockId, int16_t slotId, SNode* pNode, SHashObj* pHash) {
char name[TSDB_TABLE_NAME_LEN + TSDB_COL_NAME_LEN];
int32_t len = getSlotKey(pNode, NULL, name);
return putSlotToHashImpl(dataBlockId, slotId, name, len, pHash);
static int32_t putSlotToHash(const char* pName, int16_t dataBlockId, int16_t slotId, SNode* pNode, SHashObj* pHash) {
return putSlotToHashImpl(dataBlockId, slotId, pName, strlen(pName), pHash);
}
static int32_t createDataBlockDescHash(SPhysiPlanContext* pCxt, int32_t capacity, int16_t dataBlockId,
@ -131,9 +135,11 @@ static int32_t buildDataBlockSlots(SPhysiPlanContext* pCxt, SNodeList* pList, SD
int16_t slotId = 0;
SNode* pNode = NULL;
FOREACH(pNode, pList) {
code = nodesListStrictAppend(pDataBlockDesc->pSlots, createSlotDesc(pCxt, pNode, slotId, true, false));
char name[TSDB_TABLE_NAME_LEN + TSDB_COL_NAME_LEN];
getSlotKey(pNode, NULL, name);
code = nodesListStrictAppend(pDataBlockDesc->pSlots, createSlotDesc(pCxt, name, pNode, slotId, true, false));
if (TSDB_CODE_SUCCESS == code) {
code = putSlotToHash(pDataBlockDesc->dataBlockId, slotId, pNode, pHash);
code = putSlotToHash(name, pDataBlockDesc->dataBlockId, slotId, pNode, pHash);
}
if (TSDB_CODE_SUCCESS == code) {
pDataBlockDesc->totalRowSize += ((SExprNode*)pNode)->resType.bytes;
@ -196,7 +202,8 @@ static int32_t addDataBlockSlotsImpl(SPhysiPlanContext* pCxt, SNodeList* pList,
int32_t len = getSlotKey(pExpr, pStmtName, name);
SSlotIndex* pIndex = taosHashGet(pHash, name, len);
if (NULL == pIndex) {
code = nodesListStrictAppend(pDataBlockDesc->pSlots, createSlotDesc(pCxt, pExpr, nextSlotId, output, reserve));
code =
nodesListStrictAppend(pDataBlockDesc->pSlots, createSlotDesc(pCxt, name, pExpr, nextSlotId, output, reserve));
if (TSDB_CODE_SUCCESS == code) {
code = putSlotToHashImpl(pDataBlockDesc->dataBlockId, nextSlotId, name, len, pHash);
}
@ -513,12 +520,13 @@ static int32_t createTableScanPhysiNode(SPhysiPlanContext* pCxt, SSubplan* pSubp
tNameGetFullDbName(&pScanLogicNode->tableName, pSubplan->dbFName);
pTableScan->dataRequired = pScanLogicNode->dataRequired;
pTableScan->pDynamicScanFuncs = nodesCloneList(pScanLogicNode->pDynamicScanFuncs);
pTableScan->pPartitionTags = nodesCloneList(pScanLogicNode->pPartTags);
pTableScan->pGroupTags = nodesCloneList(pScanLogicNode->pGroupTags);
if ((NULL != pScanLogicNode->pDynamicScanFuncs && NULL == pTableScan->pDynamicScanFuncs) ||
(NULL != pScanLogicNode->pPartTags && NULL == pTableScan->pPartitionTags)) {
(NULL != pScanLogicNode->pGroupTags && NULL == pTableScan->pGroupTags)) {
nodesDestroyNode((SNode*)pTableScan);
return TSDB_CODE_OUT_OF_MEMORY;
}
pTableScan->groupSort = pScanLogicNode->groupSort;
pTableScan->interval = pScanLogicNode->interval;
pTableScan->offset = pScanLogicNode->offset;
pTableScan->sliding = pScanLogicNode->sliding;
@ -1170,8 +1178,9 @@ static int32_t createWindowPhysiNode(SPhysiPlanContext* pCxt, SNodeList* pChildr
static int32_t createSortPhysiNode(SPhysiPlanContext* pCxt, SNodeList* pChildren, SSortLogicNode* pSortLogicNode,
SPhysiNode** pPhyNode) {
SSortPhysiNode* pSort =
(SSortPhysiNode*)makePhysiNode(pCxt, (SLogicNode*)pSortLogicNode, QUERY_NODE_PHYSICAL_PLAN_SORT);
SSortPhysiNode* pSort = (SSortPhysiNode*)makePhysiNode(
pCxt, (SLogicNode*)pSortLogicNode,
pSortLogicNode->groupSort ? QUERY_NODE_PHYSICAL_PLAN_GROUP_SORT : QUERY_NODE_PHYSICAL_PLAN_SORT);
if (NULL == pSort) {
return TSDB_CODE_OUT_OF_MEMORY;
}
@ -1185,7 +1194,7 @@ static int32_t createSortPhysiNode(SPhysiPlanContext* pCxt, SNodeList* pChildren
if (TSDB_CODE_SUCCESS == code && NULL != pPrecalcExprs) {
code = setListSlotId(pCxt, pChildTupe->dataBlockId, -1, pPrecalcExprs, &pSort->pExprs);
if (TSDB_CODE_SUCCESS == code) {
code = addDataBlockSlots(pCxt, pSort->pExprs, pChildTupe);
code = pushdownDataBlockSlots(pCxt, pSort->pExprs, pChildTupe);
}
}
@ -1322,6 +1331,7 @@ static int32_t createMergePhysiNode(SPhysiPlanContext* pCxt, SMergeLogicNode* pM
pMerge->numOfChannels = pMergeLogicNode->numOfChannels;
pMerge->srcGroupId = pMergeLogicNode->srcGroupId;
pMerge->groupSort = pMergeLogicNode->groupSort;
int32_t code = addDataBlockSlots(pCxt, pMergeLogicNode->pInputs, pMerge->node.pOutputDataBlockDesc);

48
source/libs/planner/src/planSpliter.c
View File

@ -362,7 +362,7 @@ static int32_t stbSplGetNumOfVgroups(SLogicNode* pNode) {
}
static int32_t stbSplCreateMergeNode(SSplitContext* pCxt, SLogicSubplan* pSubplan, SLogicNode* pSplitNode,
SNodeList* pMergeKeys, SLogicNode* pPartChild) {
SNodeList* pMergeKeys, SLogicNode* pPartChild, bool groupSort) {
SMergeLogicNode* pMerge = (SMergeLogicNode*)nodesMakeNode(QUERY_NODE_LOGIC_PLAN_MERGE);
if (NULL == pMerge) {
return TSDB_CODE_OUT_OF_MEMORY;
@ -371,6 +371,7 @@ static int32_t stbSplCreateMergeNode(SSplitContext* pCxt, SLogicSubplan* pSubpla
pMerge->srcGroupId = pCxt->groupId;
pMerge->node.precision = pPartChild->precision;
pMerge->pMergeKeys = pMergeKeys;
pMerge->groupSort = groupSort;
int32_t code = TSDB_CODE_SUCCESS;
pMerge->pInputs = nodesCloneList(pPartChild->pTargets);
@ -430,7 +431,7 @@ static int32_t stbSplSplitIntervalForBatch(SSplitContext* pCxt, SStableSplitInfo
SNodeList* pMergeKeys = NULL;
code = stbSplCreateMergeKeysByPrimaryKey(((SWindowLogicNode*)pInfo->pSplitNode)->pTspk, &pMergeKeys);
if (TSDB_CODE_SUCCESS == code) {
code = stbSplCreateMergeNode(pCxt, NULL, pInfo->pSplitNode, pMergeKeys, pPartWindow);
code = stbSplCreateMergeNode(pCxt, NULL, pInfo->pSplitNode, pMergeKeys, pPartWindow, false);
}
if (TSDB_CODE_SUCCESS != code) {
nodesDestroyList(pMergeKeys);
@ -497,12 +498,16 @@ static int32_t stbSplSplitSessionForStream(SSplitContext* pCxt, SStableSplitInfo
return code;
}
static void splSetTableScanType(SLogicNode* pNode, EScanType scanType) {
static void stbSplSetTableMergeScan(SLogicNode* pNode) {
if (QUERY_NODE_LOGIC_PLAN_SCAN == nodeType(pNode)) {
((SScanLogicNode*)pNode)->scanType = scanType;
SScanLogicNode* pScan = (SScanLogicNode*)pNode;
pScan->scanType = SCAN_TYPE_TABLE_MERGE;
if (NULL != pScan->pGroupTags) {
pScan->groupSort = true;
}
} else {
if (1 == LIST_LENGTH(pNode->pChildren)) {
splSetTableScanType((SLogicNode*)nodesListGetNode(pNode->pChildren, 0), scanType);
stbSplSetTableMergeScan((SLogicNode*)nodesListGetNode(pNode->pChildren, 0));
}
}
}
@ -515,7 +520,7 @@ static int32_t stbSplSplitSessionOrStateForBatch(SSplitContext* pCxt, SStableSpl
int32_t code = stbSplCreateMergeKeysByPrimaryKey(((SWindowLogicNode*)pWindow)->pTspk, &pMergeKeys);
if (TSDB_CODE_SUCCESS == code) {
code = stbSplCreateMergeNode(pCxt, pInfo->pSubplan, pChild, pMergeKeys, (SLogicNode*)pChild);
code = stbSplCreateMergeNode(pCxt, pInfo->pSubplan, pChild, pMergeKeys, (SLogicNode*)pChild, true);
}
if (TSDB_CODE_SUCCESS == code) {
@ -524,13 +529,10 @@ static int32_t stbSplSplitSessionOrStateForBatch(SSplitContext* pCxt, SStableSpl
}
if (TSDB_CODE_SUCCESS == code) {
splSetTableScanType(pChild, SCAN_TYPE_TABLE_MERGE);
++(pCxt->groupId);
}
if (TSDB_CODE_SUCCESS == code) {
stbSplSetTableMergeScan(pChild);
pInfo->pSubplan->subplanType = SUBPLAN_TYPE_MERGE;
SPLIT_FLAG_SET_MASK(pInfo->pSubplan->splitFlag, SPLIT_FLAG_STABLE_SPLIT);
++(pCxt->groupId);
} else {
nodesDestroyList(pMergeKeys);
}
@ -560,7 +562,7 @@ static int32_t stbSplSplitState(SSplitContext* pCxt, SStableSplitInfo* pInfo) {
static SNodeList* stbSplGetPartKeys(SLogicNode* pNode) {
if (QUERY_NODE_LOGIC_PLAN_SCAN == nodeType(pNode)) {
return ((SScanLogicNode*)pNode)->pPartTags;
return ((SScanLogicNode*)pNode)->pGroupTags;
} else if (QUERY_NODE_LOGIC_PLAN_PARTITION == nodeType(pNode)) {
return ((SPartitionLogicNode*)pNode)->pPartitionKeys;
} else {
@ -775,6 +777,7 @@ static int32_t stbSplCreatePartSortNode(SSortLogicNode* pSort, SLogicNode** pOut
pPartSort->node.pChildren = pChildren;
splSetParent((SLogicNode*)pPartSort);
pPartSort->pSortKeys = pSortKeys;
pPartSort->groupSort = pSort->groupSort;
code = stbSplCreateMergeKeys(pPartSort->pSortKeys, pPartSort->node.pTargets, &pMergeKeys);
}
@ -789,12 +792,29 @@ static int32_t stbSplCreatePartSortNode(SSortLogicNode* pSort, SLogicNode** pOut
return code;
}
static void stbSplSetScanPartSort(SLogicNode* pNode) {
if (QUERY_NODE_LOGIC_PLAN_SCAN == nodeType(pNode)) {
SScanLogicNode* pScan = (SScanLogicNode*)pNode;
if (NULL != pScan->pGroupTags) {
pScan->groupSort = true;
}
} else {
if (1 == LIST_LENGTH(pNode->pChildren)) {
stbSplSetScanPartSort((SLogicNode*)nodesListGetNode(pNode->pChildren, 0));
}
}
}
static int32_t stbSplSplitSortNode(SSplitContext* pCxt, SStableSplitInfo* pInfo) {
SLogicNode* pPartSort = NULL;
SNodeList* pMergeKeys = NULL;
bool groupSort = ((SSortLogicNode*)pInfo->pSplitNode)->groupSort;
int32_t code = stbSplCreatePartSortNode((SSortLogicNode*)pInfo->pSplitNode, &pPartSort, &pMergeKeys);
if (TSDB_CODE_SUCCESS == code) {
code = stbSplCreateMergeNode(pCxt, pInfo->pSubplan, pInfo->pSplitNode, pMergeKeys, pPartSort);
code = stbSplCreateMergeNode(pCxt, pInfo->pSubplan, pInfo->pSplitNode, pMergeKeys, pPartSort, groupSort);
}
if (TSDB_CODE_SUCCESS == code && groupSort) {
stbSplSetScanPartSort(pPartSort);
}
if (TSDB_CODE_SUCCESS == code) {
code = nodesListMakeStrictAppend(&pInfo->pSubplan->pChildren,
@ -829,7 +849,7 @@ static int32_t stbSplSplitScanNodeForJoin(SSplitContext* pCxt, SLogicSubplan* pS
SNodeList* pMergeKeys = NULL;
int32_t code = stbSplCreateMergeKeysByPrimaryKey(stbSplFindPrimaryKeyFromScan(pScan), &pMergeKeys);
if (TSDB_CODE_SUCCESS == code) {
code = stbSplCreateMergeNode(pCxt, pSubplan, (SLogicNode*)pScan, pMergeKeys, (SLogicNode*)pScan);
code = stbSplCreateMergeNode(pCxt, pSubplan, (SLogicNode*)pScan, pMergeKeys, (SLogicNode*)pScan, false);
}
if (TSDB_CODE_SUCCESS == code) {
code = nodesListMakeStrictAppend(&pSubplan->pChildren,

8
source/libs/planner/test/planBasicTest.cpp
View File

@ -67,6 +67,14 @@ TEST_F(PlanBasicTest, tailFunc) {
run("SELECT TAIL(c1, 10) FROM t1");
run("SELECT TAIL(c2 + 10, 10, 80) FROM t1 WHERE c1 > 10");
run("SELECT TAIL(c2 + 10, 10, 80) FROM t1 WHERE c1 > 10 PARTITION BY c1");
run("SELECT TAIL(c2 + 10, 10, 80) FROM t1 WHERE c1 > 10 ORDER BY 1");
run("SELECT TAIL(c2 + 10, 10, 80) FROM t1 WHERE c1 > 10 LIMIT 5");
run("SELECT TAIL(c2 + 10, 10, 80) FROM t1 WHERE c1 > 10 PARTITION BY c1 LIMIT 5");
}
TEST_F(PlanBasicTest, interpFunc) {

4
source/libs/planner/test/planTestMain.cpp
View File

@ -76,6 +76,7 @@ static void parseArg(int argc, char* argv[]) {
static struct option long_options[] = {
{"dump", optional_argument, NULL, 'd'},
{"skipSql", required_argument, NULL, 's'},
{"limitSql", required_argument, NULL, 'i'},
{"log", required_argument, NULL, 'l'},
{0, 0, 0, 0}
};
@ -88,6 +89,9 @@ static void parseArg(int argc, char* argv[]) {
case 's':
setSkipSqlNum(optarg);
break;
case 'i':
setLimitSqlNum(optarg);
break;
case 'l':
setLogLevel(optarg);
break;

30
source/libs/planner/test/planTestUtil.cpp
View File

@ -51,6 +51,7 @@ enum DumpModule {
DumpModule g_dumpModule = DUMP_MODULE_NOTHING;
int32_t g_skipSql = 0;
int32_t g_limitSql = 0;
int32_t g_logLevel = 131;
void setDumpModule(const char* pModule) {
@ -76,28 +77,33 @@ void setDumpModule(const char* pModule) {
}
void setSkipSqlNum(const char* pNum) { g_skipSql = stoi(pNum); }
void setLimitSqlNum(const char* pNum) { g_limitSql = stoi(pNum); }
void setLogLevel(const char* pLogLevel) { g_logLevel = stoi(pLogLevel); }
int32_t getLogLevel() { return g_logLevel; }
class PlannerTestBaseImpl {
public:
PlannerTestBaseImpl() : sqlNo_(0) {}
PlannerTestBaseImpl() : sqlNo_(0), sqlNum_(0) {}
void useDb(const string& user, const string& db) {
caseEnv_.acctId_ = 0;
caseEnv_.user_ = user;
caseEnv_.db_ = db;
caseEnv_.nsql_ = g_skipSql;
caseEnv_.numOfSkipSql_ = g_skipSql;
caseEnv_.numOfLimitSql_ = g_limitSql;
}
void run(const string& sql) {
++sqlNo_;
if (caseEnv_.nsql_ > 0) {
--(caseEnv_.nsql_);
if (caseEnv_.numOfSkipSql_ > 0) {
--(caseEnv_.numOfSkipSql_);
return;
}
if (caseEnv_.numOfLimitSql_ > 0 && caseEnv_.numOfLimitSql_ == sqlNum_) {
return;
}
++sqlNum_;
reset();
try {
@ -134,7 +140,7 @@ class PlannerTestBaseImpl {
}
void prepare(const string& sql) {
if (caseEnv_.nsql_ > 0) {
if (caseEnv_.numOfSkipSql_ > 0) {
return;
}
@ -148,7 +154,7 @@ class PlannerTestBaseImpl {
}
void bindParams(TAOS_MULTI_BIND* pParams, int32_t colIdx) {
if (caseEnv_.nsql_ > 0) {
if (caseEnv_.numOfSkipSql_ > 0) {
return;
}
@ -161,8 +167,8 @@ class PlannerTestBaseImpl {
}
void exec() {
if (caseEnv_.nsql_ > 0) {
--(caseEnv_.nsql_);
if (caseEnv_.numOfSkipSql_ > 0) {
--(caseEnv_.numOfSkipSql_);
return;
}
@ -197,9 +203,10 @@ class PlannerTestBaseImpl {
int32_t acctId_;
string user_;
string db_;
int32_t nsql_;
int32_t numOfSkipSql_;
int32_t numOfLimitSql_;
caseEnv() : nsql_(0) {}
caseEnv() : numOfSkipSql_(0) {}
};
struct stmtEnv {
@ -401,6 +408,7 @@ class PlannerTestBaseImpl {
stmtEnv stmtEnv_;
stmtRes res_;
int32_t sqlNo_;
int32_t sqlNum_;
};
PlannerTestBase::PlannerTestBase() : impl_(new PlannerTestBaseImpl()) {}

1
source/libs/planner/test/planTestUtil.h
View File

@ -43,6 +43,7 @@ class PlannerTestBase : public testing::Test {
extern void setDumpModule(const char* pModule);
extern void setSkipSqlNum(const char* pNum);
extern void setLimitSqlNum(const char* pNum);
extern void setLogLevel(const char* pLogLevel);
extern int32_t getLogLevel();

12
source/libs/sync/inc/syncEnv.h
View File

@ -28,13 +28,13 @@ extern "C" {
#include "trpc.h"
#include "ttimer.h"
#define TIMER_MAX_MS 0x7FFFFFFF
#define ENV_TICK_TIMER_MS 1000
#define PING_TIMER_MS 1000
#define ELECT_TIMER_MS_MIN 1300
#define ELECT_TIMER_MS_MAX (ELECT_TIMER_MS_MIN * 2)
#define TIMER_MAX_MS 0x7FFFFFFF
#define ENV_TICK_TIMER_MS 1000
#define PING_TIMER_MS 1000
#define ELECT_TIMER_MS_MIN 1300
#define ELECT_TIMER_MS_MAX (ELECT_TIMER_MS_MIN * 2)
#define ELECT_TIMER_MS_RANGE (ELECT_TIMER_MS_MAX - ELECT_TIMER_MS_MIN)
#define HEARTBEAT_TIMER_MS 900
#define HEARTBEAT_TIMER_MS 900
#define EMPTY_RAFT_ID ((SRaftId){.addr = 0, .vgId = 0})

2
source/libs/sync/inc/syncIO.h
View File

@ -50,7 +50,7 @@ typedef struct SSyncIO {
void *pSyncNode;
int32_t (*FpOnSyncPing)(SSyncNode *pSyncNode, SyncPing *pMsg);
int32_t (*FpOnSyncPingReply)(SSyncNode *pSyncNode, SyncPingReply *pMsg);
int32_t (*FpOnSyncClientRequest)(SSyncNode *pSyncNode, SyncClientRequest *pMsg);
int32_t (*FpOnSyncClientRequest)(SSyncNode *pSyncNode, SyncClientRequest *pMsg, SyncIndex *pRetIndex);
int32_t (*FpOnSyncRequestVote)(SSyncNode *pSyncNode, SyncRequestVote *pMsg);
int32_t (*FpOnSyncRequestVoteReply)(SSyncNode *pSyncNode, SyncRequestVoteReply *pMsg);
int32_t (*FpOnSyncAppendEntries)(SSyncNode *pSyncNode, SyncAppendEntries *pMsg);

3
source/libs/sync/inc/syncInt.h
View File

@ -169,7 +169,7 @@ SSyncNode* syncNodeOpen(const SSyncInfo* pSyncInfo);
void syncNodeStart(SSyncNode* pSyncNode);
void syncNodeStartStandBy(SSyncNode* pSyncNode);
void syncNodeClose(SSyncNode* pSyncNode);
int32_t syncNodePropose(SSyncNode* pSyncNode, const SRpcMsg* pMsg, bool isWeak);
int32_t syncNodePropose(SSyncNode* pSyncNode, SRpcMsg* pMsg, bool isWeak);
// option
bool syncNodeSnapshotEnable(SSyncNode* pSyncNode);
@ -233,6 +233,7 @@ SyncIndex syncNodeGetPreIndex(SSyncNode* pSyncNode, SyncIndex index);
SyncTerm syncNodeGetPreTerm(SSyncNode* pSyncNode, SyncIndex index);
int32_t syncNodeGetPreIndexTerm(SSyncNode* pSyncNode, SyncIndex index, SyncIndex* pPreIndex, SyncTerm* pPreTerm);
bool syncNodeIsOptimizedOneReplica(SSyncNode* ths, SRpcMsg* pMsg);
int32_t syncNodeCommit(SSyncNode* ths, SyncIndex beginIndex, SyncIndex endIndex, uint64_t flag);
int32_t syncNodeUpdateNewConfigIndex(SSyncNode* ths, SSyncCfg* pNewCfg);

10
source/libs/sync/inc/syncRaftCfg.h
View File

@ -49,14 +49,14 @@ int32_t raftCfgClose(SRaftCfg *pRaftCfg);
int32_t raftCfgPersist(SRaftCfg *pRaftCfg);
int32_t raftCfgAddConfigIndex(SRaftCfg *pRaftCfg, SyncIndex configIndex);
cJSON *syncCfg2Json(SSyncCfg *pSyncCfg);
char *syncCfg2Str(SSyncCfg *pSyncCfg);
char *syncCfg2SimpleStr(SSyncCfg *pSyncCfg);
cJSON * syncCfg2Json(SSyncCfg *pSyncCfg);
char * syncCfg2Str(SSyncCfg *pSyncCfg);
char * syncCfg2SimpleStr(SSyncCfg *pSyncCfg);
int32_t syncCfgFromJson(const cJSON *pRoot, SSyncCfg *pSyncCfg);
int32_t syncCfgFromStr(const char *s, SSyncCfg *pSyncCfg);
cJSON *raftCfg2Json(SRaftCfg *pRaftCfg);
char *raftCfg2Str(SRaftCfg *pRaftCfg);
cJSON * raftCfg2Json(SRaftCfg *pRaftCfg);
char * raftCfg2Str(SRaftCfg *pRaftCfg);
int32_t raftCfgFromJson(const cJSON *pRoot, SRaftCfg *pRaftCfg);
int32_t raftCfgFromStr(const char *s, SRaftCfg *pRaftCfg);

25
source/libs/sync/inc/syncRaftEntry.h
View File

@ -56,6 +56,31 @@ void syncEntryPrint2(char* s, const SSyncRaftEntry* pObj);
void syncEntryLog(const SSyncRaftEntry* pObj);
void syncEntryLog2(char* s, const SSyncRaftEntry* pObj);
//-----------------------------------
typedef struct SRaftEntryCache {
SHashObj* pEntryHash;
int32_t maxCount;
int32_t currentCount;
TdThreadMutex mutex;
SSyncNode* pSyncNode;
} SRaftEntryCache;
SRaftEntryCache* raftCacheCreate(SSyncNode* pSyncNode, int32_t maxCount);
void raftCacheDestroy(SRaftEntryCache* pCache);
int32_t raftCachePutEntry(struct SRaftEntryCache* pCache, SSyncRaftEntry* pEntry);
int32_t raftCacheGetEntry(struct SRaftEntryCache* pCache, SyncIndex index, SSyncRaftEntry** ppEntry);
int32_t raftCacheGetEntryP(struct SRaftEntryCache* pCache, SyncIndex index, SSyncRaftEntry** ppEntry);
int32_t raftCacheDelEntry(struct SRaftEntryCache* pCache, SyncIndex index);
int32_t raftCacheGetAndDel(struct SRaftEntryCache* pCache, SyncIndex index, SSyncRaftEntry** ppEntry);
int32_t raftCacheClear(struct SRaftEntryCache* pCache);
cJSON* raftCache2Json(SRaftEntryCache* pObj);
char* raftCache2Str(SRaftEntryCache* pObj);
void raftCachePrint(SRaftEntryCache* pObj);
void raftCachePrint2(char* s, SRaftEntryCache* pObj);
void raftCacheLog(SRaftEntryCache* pObj);
void raftCacheLog2(char* s, SRaftEntryCache* pObj);
#ifdef __cplusplus
}
#endif

7
source/libs/sync/inc/syncRaftLog.h
View File

@ -29,9 +29,12 @@ extern "C" {
#include "wal.h"
typedef struct SSyncLogStoreData {
SSyncNode* pSyncNode;
SWal* pWal;
SSyncNode* pSyncNode;
SWal* pWal;
TdThreadMutex mutex;
SWalReadHandle* pWalHandle;
// SyncIndex beginIndex; // valid begin index, default 0, may be set beginIndex > 0
} SSyncLogStoreData;

44
source/libs/sync/inc/syncSnapshot.h
View File

@ -35,12 +35,13 @@ extern "C" {
#define SYNC_SNAPSHOT_RETRY_MS 5000
//---------------------------------------------------
typedef struct SSyncSnapshotSender {
bool start;
int32_t seq;
int32_t ack;
void * pReader;
void * pCurrentBlock;
void *pReader;
void *pCurrentBlock;
int32_t blockLen;
SSnapshot snapshot;
SSyncCfg lastConfig;
@ -59,33 +60,36 @@ void snapshotSenderStart(SSyncSnapshotSender *pSender, SSnapshot
void snapshotSenderStop(SSyncSnapshotSender *pSender);
int32_t snapshotSend(SSyncSnapshotSender *pSender);
int32_t snapshotReSend(SSyncSnapshotSender *pSender);
cJSON * snapshotSender2Json(SSyncSnapshotSender *pSender);
char * snapshotSender2Str(SSyncSnapshotSender *pSender);
char * snapshotSender2SimpleStr(SSyncSnapshotSender *pSender, char *event);
cJSON *snapshotSender2Json(SSyncSnapshotSender *pSender);
char *snapshotSender2Str(SSyncSnapshotSender *pSender);
char *snapshotSender2SimpleStr(SSyncSnapshotSender *pSender, char *event);
//---------------------------------------------------
typedef struct SSyncSnapshotReceiver {
bool start;
int32_t ack;
void * pWriter;
SyncTerm term;
SyncTerm privateTerm;
SSnapshot snapshot;
SSyncNode *pSyncNode;
bool start;
int32_t ack;
void *pWriter;
SyncTerm term;
SyncTerm privateTerm;
SSnapshot snapshot;
SRaftId fromId;
SSyncNode *pSyncNode;
} SSyncSnapshotReceiver;
SSyncSnapshotReceiver *snapshotReceiverCreate(SSyncNode *pSyncNode, SRaftId fromId);
void snapshotReceiverDestroy(SSyncSnapshotReceiver *pReceiver);
void snapshotReceiverStart(SSyncSnapshotReceiver *pReceiver, SyncTerm privateTerm, SyncSnapshotSend *pBeginMsg);
bool snapshotReceiverIsStart(SSyncSnapshotReceiver *pReceiver);
void snapshotReceiverStop(SSyncSnapshotReceiver *pReceiver, bool apply);
cJSON *snapshotReceiver2Json(SSyncSnapshotReceiver *pReceiver);
char * snapshotReceiver2Str(SSyncSnapshotReceiver *pReceiver);
char * snapshotReceiver2SimpleStr(SSyncSnapshotReceiver *pReceiver, char *event);
void snapshotReceiverStart(SSyncSnapshotReceiver *pReceiver, SyncTerm privateTerm, SyncSnapshotSend *pBeginMsg);
bool snapshotReceiverIsStart(SSyncSnapshotReceiver *pReceiver);
void snapshotReceiverStop(SSyncSnapshotReceiver *pReceiver, bool apply);
cJSON *snapshotReceiver2Json(SSyncSnapshotReceiver *pReceiver);
char *snapshotReceiver2Str(SSyncSnapshotReceiver *pReceiver);
char *snapshotReceiver2SimpleStr(SSyncSnapshotReceiver *pReceiver, char *event);
//---------------------------------------------------
// on message
int32_t syncNodeOnSnapshotSendCb(SSyncNode *ths, SyncSnapshotSend *pMsg);
int32_t syncNodeOnSnapshotRspCb(SSyncNode *ths, SyncSnapshotRsp *pMsg);

1
source/libs/sync/src/syncCommit.c
View File

@ -102,6 +102,7 @@ void syncMaybeAdvanceCommitIndex(SSyncNode* pSyncNode) {
}
}
// maybe execute fsm
if (newCommitIndex > pSyncNode->commitIndex) {
SyncIndex beginIndex = pSyncNode->commitIndex + 1;
SyncIndex endIndex = newCommitIndex;

8
source/libs/sync/src/syncIO.c
View File

@ -30,7 +30,7 @@ static int32_t syncIODestroy(SSyncIO *io);
static int32_t syncIOStartInternal(SSyncIO *io);
static int32_t syncIOStopInternal(SSyncIO *io);
static void *syncIOConsumerFunc(void *param);
static void * syncIOConsumerFunc(void *param);
static void syncIOProcessRequest(void *pParent, SRpcMsg *pMsg, SEpSet *pEpSet);
static void syncIOProcessReply(void *pParent, SRpcMsg *pMsg, SEpSet *pEpSet);
static int32_t syncIOAuth(void *parent, char *meterId, char *spi, char *encrypt, char *secret, char *ckey);
@ -242,9 +242,9 @@ static int32_t syncIOStopInternal(SSyncIO *io) {
}
static void *syncIOConsumerFunc(void *param) {
SSyncIO *io = param;
SSyncIO * io = param;
STaosQall *qall;
SRpcMsg *pRpcMsg, rpcMsg;
SRpcMsg * pRpcMsg, rpcMsg;
qall = taosAllocateQall();
while (1) {
@ -281,7 +281,7 @@ static void *syncIOConsumerFunc(void *param) {
if (io->FpOnSyncClientRequest != NULL) {
SyncClientRequest *pSyncMsg = syncClientRequestFromRpcMsg2(pRpcMsg);
ASSERT(pSyncMsg != NULL);
io->FpOnSyncClientRequest(io->pSyncNode, pSyncMsg);
io->FpOnSyncClientRequest(io->pSyncNode, pSyncMsg, NULL);
syncClientRequestDestroy(pSyncMsg);
}

2
source/libs/sync/src/syncIndexMgr.c
View File

@ -126,7 +126,7 @@ cJSON *syncIndexMgr2Json(SSyncIndexMgr *pSyncIndexMgr) {
char *syncIndexMgr2Str(SSyncIndexMgr *pSyncIndexMgr) {
cJSON *pJson = syncIndexMgr2Json(pSyncIndexMgr);
char *serialized = cJSON_Print(pJson);
char * serialized = cJSON_Print(pJson);
cJSON_Delete(pJson);
return serialized;
}

98
source/libs/sync/src/syncMain.c
View File

@ -50,7 +50,7 @@ static int32_t syncNodeAppendNoop(SSyncNode* ths);
// process message ----
int32_t syncNodeOnPingCb(SSyncNode* ths, SyncPing* pMsg);
int32_t syncNodeOnPingReplyCb(SSyncNode* ths, SyncPingReply* pMsg);
int32_t syncNodeOnClientRequestCb(SSyncNode* ths, SyncClientRequest* pMsg);
int32_t syncNodeOnClientRequestCb(SSyncNode* ths, SyncClientRequest* pMsg, SyncIndex* pRetIndex);
// life cycle
static void syncFreeNode(void* param);
@ -627,7 +627,7 @@ int32_t syncPropose(int64_t rid, SRpcMsg* pMsg, bool isWeak) {
return ret;
}
int32_t syncNodePropose(SSyncNode* pSyncNode, const SRpcMsg* pMsg, bool isWeak) {
int32_t syncNodePropose(SSyncNode* pSyncNode, SRpcMsg* pMsg, bool isWeak) {
int32_t ret = 0;
char eventLog[128];
@ -664,13 +664,34 @@ int32_t syncNodePropose(SSyncNode* pSyncNode, const SRpcMsg* pMsg, bool isWeak)
SRpcMsg rpcMsg;
syncClientRequest2RpcMsg(pSyncMsg, &rpcMsg);
if (pSyncNode->FpEqMsg != NULL && (*pSyncNode->FpEqMsg)(pSyncNode->msgcb, &rpcMsg) == 0) {
ret = 0;
// optimized one replica
if (syncNodeIsOptimizedOneReplica(pSyncNode, pMsg)) {
SyncIndex retIndex;
int32_t code = syncNodeOnClientRequestCb(pSyncNode, pSyncMsg, &retIndex);
if (code == 0) {
pMsg->info.conn.applyIndex = retIndex;
rpcFreeCont(rpcMsg.pCont);
syncRespMgrDel(pSyncNode->pSyncRespMgr, seqNum);
ret = 1;
sDebug("vgId:%d optimized index:%ld success, msgtype:%s,%d", pSyncNode->vgId, retIndex,
TMSG_INFO(pMsg->msgType), pMsg->msgType);
} else {
ret = -1;
terrno = TSDB_CODE_SYN_INTERNAL_ERROR;
sError("vgId:%d optimized index:%ld error, msgtype:%s,%d", pSyncNode->vgId, retIndex, TMSG_INFO(pMsg->msgType),
pMsg->msgType);
}
} else {
ret = -1;
terrno = TSDB_CODE_SYN_INTERNAL_ERROR;
sError("syncPropose pSyncNode->FpEqMsg is NULL");
if (pSyncNode->FpEqMsg != NULL && (*pSyncNode->FpEqMsg)(pSyncNode->msgcb, &rpcMsg) == 0) {
ret = 0;
} else {
ret = -1;
terrno = TSDB_CODE_SYN_INTERNAL_ERROR;
sError("enqueue msg error, FpEqMsg is NULL");
}
}
syncClientRequestDestroy(pSyncMsg);
goto _END;
@ -1334,11 +1355,16 @@ void syncNodeEventLog(const SSyncNode* pSyncNode, char* str) {
int32_t userStrLen = strlen(str);
SSnapshot snapshot = {.data = NULL, .lastApplyIndex = -1, .lastApplyTerm = 0};
if (pSyncNode->pFsm->FpGetSnapshotInfo != NULL) {
if (pSyncNode->pFsm != NULL && pSyncNode->pFsm->FpGetSnapshotInfo != NULL) {
pSyncNode->pFsm->FpGetSnapshotInfo(pSyncNode->pFsm, &snapshot);
}
SyncIndex logLastIndex = pSyncNode->pLogStore->syncLogLastIndex(pSyncNode->pLogStore);
SyncIndex logBeginIndex = pSyncNode->pLogStore->syncLogBeginIndex(pSyncNode->pLogStore);
SyncIndex logLastIndex = SYNC_INDEX_INVALID;
SyncIndex logBeginIndex = SYNC_INDEX_INVALID;
if (pSyncNode->pLogStore != NULL) {
logLastIndex = pSyncNode->pLogStore->syncLogLastIndex(pSyncNode->pLogStore);
logBeginIndex = pSyncNode->pLogStore->syncLogBeginIndex(pSyncNode->pLogStore);
}
char* pCfgStr = syncCfg2SimpleStr(&(pSyncNode->pRaftCfg->cfg));
char* printStr = "";
@ -2377,7 +2403,7 @@ int32_t syncNodeOnPingReplyCb(SSyncNode* ths, SyncPingReply* pMsg) {
// /\ UNCHANGED <<messages, serverVars, candidateVars,
// leaderVars, commitIndex>>
//
int32_t syncNodeOnClientRequestCb(SSyncNode* ths, SyncClientRequest* pMsg) {
int32_t syncNodeOnClientRequestCb(SSyncNode* ths, SyncClientRequest* pMsg, SyncIndex* pRetIndex) {
int32_t ret = 0;
syncClientRequestLog2("==syncNodeOnClientRequestCb==", pMsg);
@ -2436,6 +2462,14 @@ int32_t syncNodeOnClientRequestCb(SSyncNode* ths, SyncClientRequest* pMsg) {
rpcFreeCont(rpcMsg.pCont);
}
if (pRetIndex != NULL) {
if (ret == 0 && pEntry != NULL) {
*pRetIndex = pEntry->index;
} else {
*pRetIndex = SYNC_INDEX_INVALID;
}
}
syncEntryDestory(pEntry);
return ret;
}
@ -2600,6 +2634,10 @@ static int32_t syncNodeProposeConfigChangeFinish(SSyncNode* ths, SyncReconfigFin
return 0;
}
bool syncNodeIsOptimizedOneReplica(SSyncNode* ths, SRpcMsg* pMsg) {
return (ths->replicaNum == 1 && syncUtilUserCommit(pMsg->msgType) && ths->vgId != 1);
}
int32_t syncNodeCommit(SSyncNode* ths, SyncIndex beginIndex, SyncIndex endIndex, uint64_t flag) {
int32_t code = 0;
ESyncState state = flag;
@ -2621,19 +2659,33 @@ int32_t syncNodeCommit(SSyncNode* ths, SyncIndex beginIndex, SyncIndex endIndex,
syncEntry2OriginalRpc(pEntry, &rpcMsg);
// user commit
if (ths->pFsm->FpCommitCb != NULL && syncUtilUserCommit(pEntry->originalRpcType)) {
SFsmCbMeta cbMeta = {0};
cbMeta.index = pEntry->index;
cbMeta.lastConfigIndex = syncNodeGetSnapshotConfigIndex(ths, cbMeta.index);
cbMeta.isWeak = pEntry->isWeak;
cbMeta.code = 0;
cbMeta.state = ths->state;
cbMeta.seqNum = pEntry->seqNum;
cbMeta.term = pEntry->term;
cbMeta.currentTerm = ths->pRaftStore->currentTerm;
cbMeta.flag = flag;
if ((ths->pFsm->FpCommitCb != NULL) && syncUtilUserCommit(pEntry->originalRpcType)) {
bool internalExecute = true;
if ((ths->replicaNum == 1) && ths->restoreFinish && (ths->vgId != 1)) {
internalExecute = false;
}
ths->pFsm->FpCommitCb(ths->pFsm, &rpcMsg, cbMeta);
do {
char logBuf[128];
snprintf(logBuf, sizeof(logBuf), "index:%ld, internalExecute:%d", i, internalExecute);
syncNodeEventLog(ths, logBuf);
} while (0);
// execute fsm in apply thread, or execute outside syncPropose
if (internalExecute) {
SFsmCbMeta cbMeta = {0};
cbMeta.index = pEntry->index;
cbMeta.lastConfigIndex = syncNodeGetSnapshotConfigIndex(ths, cbMeta.index);
cbMeta.isWeak = pEntry->isWeak;
cbMeta.code = 0;
cbMeta.state = ths->state;
cbMeta.seqNum = pEntry->seqNum;
cbMeta.term = pEntry->term;
cbMeta.currentTerm = ths->pRaftStore->currentTerm;
cbMeta.flag = flag;
ths->pFsm->FpCommitCb(ths->pFsm, &rpcMsg, cbMeta);
}
}
// config change

8
source/libs/sync/src/syncRaftCfg.c
View File

@ -101,7 +101,7 @@ cJSON *syncCfg2Json(SSyncCfg *pSyncCfg) {
char *syncCfg2Str(SSyncCfg *pSyncCfg) {
cJSON *pJson = syncCfg2Json(pSyncCfg);
char *serialized = cJSON_Print(pJson);
char * serialized = cJSON_Print(pJson);
cJSON_Delete(pJson);
return serialized;
}
@ -109,7 +109,7 @@ char *syncCfg2Str(SSyncCfg *pSyncCfg) {
char *syncCfg2SimpleStr(SSyncCfg *pSyncCfg) {
if (pSyncCfg != NULL) {
int32_t len = 512;
char *s = taosMemoryMalloc(len);
char * s = taosMemoryMalloc(len);
memset(s, 0, len);
snprintf(s, len, "{replica-num:%d, my-index:%d, ", pSyncCfg->replicaNum, pSyncCfg->myIndex);
@ -205,7 +205,7 @@ cJSON *raftCfg2Json(SRaftCfg *pRaftCfg) {
char *raftCfg2Str(SRaftCfg *pRaftCfg) {
cJSON *pJson = raftCfg2Json(pRaftCfg);
char *serialized = cJSON_Print(pJson);
char * serialized = cJSON_Print(pJson);
cJSON_Delete(pJson);
return serialized;
}
@ -271,7 +271,7 @@ int32_t raftCfgFromJson(const cJSON *pRoot, SRaftCfg *pRaftCfg) {
(pRaftCfg->configIndexArr)[i] = atoll(pIndex->valuestring);
}
cJSON *pJsonSyncCfg = cJSON_GetObjectItem(pJson, "SSyncCfg");
cJSON * pJsonSyncCfg = cJSON_GetObjectItem(pJson, "SSyncCfg");
int32_t code = syncCfgFromJson(pJsonSyncCfg, &(pRaftCfg->cfg));
ASSERT(code == 0);

244
source/libs/sync/src/syncRaftEntry.c
View File

@ -180,3 +180,247 @@ void syncEntryLog2(char* s, const SSyncRaftEntry* pObj) {
sTrace("syncEntryLog2 | len:%zu | %s | %s", strlen(serialized), s, serialized);
taosMemoryFree(serialized);
}
//-----------------------------------
SRaftEntryCache* raftCacheCreate(SSyncNode* pSyncNode, int32_t maxCount) {
SRaftEntryCache* pCache = taosMemoryMalloc(sizeof(SRaftEntryCache));
if (pCache == NULL) {
sError("vgId:%d raft cache create error", pSyncNode->vgId);
return NULL;
}
pCache->pEntryHash =
taosHashInit(sizeof(SyncIndex), taosGetDefaultHashFunction(TSDB_DATA_TYPE_BINARY), true, HASH_NO_LOCK);
if (pCache->pEntryHash == NULL) {
sError("vgId:%d raft cache create hash error", pSyncNode->vgId);
return NULL;
}
taosThreadMutexInit(&(pCache->mutex), NULL);
pCache->maxCount = maxCount;
pCache->currentCount = 0;
pCache->pSyncNode = pSyncNode;
return pCache;
}
void raftCacheDestroy(SRaftEntryCache* pCache) {
if (pCache != NULL) {
taosThreadMutexLock(&(pCache->mutex));
taosHashCleanup(pCache->pEntryHash);
taosThreadMutexUnlock(&(pCache->mutex));
taosThreadMutexDestroy(&(pCache->mutex));
taosMemoryFree(pCache);
}
}
// success, return 1
// max count, return 0
// error, return -1
int32_t raftCachePutEntry(struct SRaftEntryCache* pCache, SSyncRaftEntry* pEntry) {
taosThreadMutexLock(&(pCache->mutex));
if (pCache->currentCount >= pCache->maxCount) {
taosThreadMutexUnlock(&(pCache->mutex));
return 0;
}
taosHashPut(pCache->pEntryHash, &(pEntry->index), sizeof(pEntry->index), pEntry, pEntry->bytes);
++(pCache->currentCount);
do {
char eventLog[128];
snprintf(eventLog, sizeof(eventLog), "raft cache add, type:%s,%d, type2:%s,%d, index:%ld, bytes:%d",
TMSG_INFO(pEntry->msgType), pEntry->msgType, TMSG_INFO(pEntry->originalRpcType), pEntry->originalRpcType,
pEntry->index, pEntry->bytes);
syncNodeEventLog(pCache->pSyncNode, eventLog);
} while (0);
taosThreadMutexUnlock(&(pCache->mutex));
return 1;
}
// success, return 0
// error, return -1
// not exist, return -1, terrno = TSDB_CODE_WAL_LOG_NOT_EXIST
int32_t raftCacheGetEntry(struct SRaftEntryCache* pCache, SyncIndex index, SSyncRaftEntry** ppEntry) {
if (ppEntry == NULL) {
return -1;
}
*ppEntry = NULL;
taosThreadMutexLock(&(pCache->mutex));
void* pTmp = taosHashGet(pCache->pEntryHash, &index, sizeof(index));
if (pTmp != NULL) {
SSyncRaftEntry* pEntry = pTmp;
*ppEntry = taosMemoryMalloc(pEntry->bytes);
memcpy(*ppEntry, pTmp, pEntry->bytes);
do {
char eventLog[128];
snprintf(eventLog, sizeof(eventLog), "raft cache get, type:%s,%d, type2:%s,%d, index:%ld",
TMSG_INFO((*ppEntry)->msgType), (*ppEntry)->msgType, TMSG_INFO((*ppEntry)->originalRpcType),
(*ppEntry)->originalRpcType, (*ppEntry)->index);
syncNodeEventLog(pCache->pSyncNode, eventLog);
} while (0);
taosThreadMutexUnlock(&(pCache->mutex));
return 0;
}
taosThreadMutexUnlock(&(pCache->mutex));
terrno = TSDB_CODE_WAL_LOG_NOT_EXIST;
return -1;
}
// success, return 0
// error, return -1
// not exist, return -1, terrno = TSDB_CODE_WAL_LOG_NOT_EXIST
int32_t raftCacheGetEntryP(struct SRaftEntryCache* pCache, SyncIndex index, SSyncRaftEntry** ppEntry) {
if (ppEntry == NULL) {
return -1;
}
*ppEntry = NULL;
taosThreadMutexLock(&(pCache->mutex));
void* pTmp = taosHashGet(pCache->pEntryHash, &index, sizeof(index));
if (pTmp != NULL) {
SSyncRaftEntry* pEntry = pTmp;
*ppEntry = pEntry;
do {
char eventLog[128];
snprintf(eventLog, sizeof(eventLog), "raft cache get, type:%s,%d, type2:%s,%d, index:%ld",
TMSG_INFO((*ppEntry)->msgType), (*ppEntry)->msgType, TMSG_INFO((*ppEntry)->originalRpcType),
(*ppEntry)->originalRpcType, (*ppEntry)->index);
syncNodeEventLog(pCache->pSyncNode, eventLog);
} while (0);
taosThreadMutexUnlock(&(pCache->mutex));
return 0;
}
taosThreadMutexUnlock(&(pCache->mutex));
terrno = TSDB_CODE_WAL_LOG_NOT_EXIST;
return -1;
}
int32_t raftCacheDelEntry(struct SRaftEntryCache* pCache, SyncIndex index) {
taosThreadMutexLock(&(pCache->mutex));
taosHashRemove(pCache->pEntryHash, &index, sizeof(index));
--(pCache->currentCount);
taosThreadMutexUnlock(&(pCache->mutex));
return 0;
}
int32_t raftCacheGetAndDel(struct SRaftEntryCache* pCache, SyncIndex index, SSyncRaftEntry** ppEntry) {
if (ppEntry == NULL) {
return -1;
}
*ppEntry = NULL;
taosThreadMutexLock(&(pCache->mutex));
void* pTmp = taosHashGet(pCache->pEntryHash, &index, sizeof(index));
if (pTmp != NULL) {
SSyncRaftEntry* pEntry = pTmp;
*ppEntry = taosMemoryMalloc(pEntry->bytes);
memcpy(*ppEntry, pTmp, pEntry->bytes);
do {
char eventLog[128];
snprintf(eventLog, sizeof(eventLog), "raft cache get-and-del, type:%s,%d, type2:%s,%d, index:%ld",
TMSG_INFO((*ppEntry)->msgType), (*ppEntry)->msgType, TMSG_INFO((*ppEntry)->originalRpcType),
(*ppEntry)->originalRpcType, (*ppEntry)->index);
syncNodeEventLog(pCache->pSyncNode, eventLog);
} while (0);
taosHashRemove(pCache->pEntryHash, &index, sizeof(index));
--(pCache->currentCount);
taosThreadMutexUnlock(&(pCache->mutex));
return 0;
}
taosThreadMutexUnlock(&(pCache->mutex));
terrno = TSDB_CODE_WAL_LOG_NOT_EXIST;
return -1;
}
int32_t raftCacheClear(struct SRaftEntryCache* pCache) {
taosThreadMutexLock(&(pCache->mutex));
taosHashClear(pCache->pEntryHash);
pCache->currentCount = 0;
taosThreadMutexUnlock(&(pCache->mutex));
return 0;
}
//-----------------------------------
cJSON* raftCache2Json(SRaftEntryCache* pCache) {
char u64buf[128] = {0};
cJSON* pRoot = cJSON_CreateObject();
if (pCache != NULL) {
taosThreadMutexLock(&(pCache->mutex));
snprintf(u64buf, sizeof(u64buf), "%p", pCache->pSyncNode);
cJSON_AddStringToObject(pRoot, "pSyncNode", u64buf);
cJSON_AddNumberToObject(pRoot, "currentCount", pCache->currentCount);
cJSON_AddNumberToObject(pRoot, "maxCount", pCache->maxCount);
cJSON* pEntries = cJSON_CreateArray();
cJSON_AddItemToObject(pRoot, "entries", pEntries);
SSyncRaftEntry* pIter = (SSyncRaftEntry*)taosHashIterate(pCache->pEntryHash, NULL);
if (pIter != NULL) {
SSyncRaftEntry* pEntry = (SSyncRaftEntry*)pIter;
cJSON_AddItemToArray(pEntries, syncEntry2Json(pEntry));
}
while (pIter) {
pIter = taosHashIterate(pCache->pEntryHash, pIter);
if (pIter != NULL) {
SSyncRaftEntry* pEntry = (SSyncRaftEntry*)pIter;
cJSON_AddItemToArray(pEntries, syncEntry2Json(pEntry));
}
}
taosThreadMutexUnlock(&(pCache->mutex));
}
cJSON* pJson = cJSON_CreateObject();
cJSON_AddItemToObject(pJson, "SRaftEntryCache", pRoot);
return pJson;
}
char* raftCache2Str(SRaftEntryCache* pCache) {
cJSON* pJson = raftCache2Json(pCache);
char* serialized = cJSON_Print(pJson);
cJSON_Delete(pJson);
return serialized;
}
void raftCachePrint(SRaftEntryCache* pCache) {
char* serialized = raftCache2Str(pCache);
printf("raftCachePrint | len:%lu | %s \n", strlen(serialized), serialized);
fflush(NULL);
taosMemoryFree(serialized);
}
void raftCachePrint2(char* s, SRaftEntryCache* pCache) {
char* serialized = raftCache2Str(pCache);
printf("raftCachePrint2 | len:%lu | %s | %s \n", strlen(serialized), s, serialized);
fflush(NULL);
taosMemoryFree(serialized);
}
void raftCacheLog(SRaftEntryCache* pCache) {
char* serialized = raftCache2Str(pCache);
sTrace("raftCacheLog | len:%lu | %s", strlen(serialized), serialized);
taosMemoryFree(serialized);
}
void raftCacheLog2(char* s, SRaftEntryCache* pCache) {
if (gRaftDetailLog) {
char* serialized = raftCache2Str(pCache);
sTraceLong("raftCacheLog2 | len:%lu | %s | %s", strlen(serialized), s, serialized);
taosMemoryFree(serialized);
}
}

13
source/libs/sync/src/syncRaftLog.c
View File

@ -257,6 +257,8 @@ static int32_t raftLogGetEntry(struct SSyncLogStore* pLogStore, SyncIndex index,
return -1;
}
taosThreadMutexLock(&(pData->mutex));
code = walReadWithHandle(pWalHandle, index);
if (code != 0) {
int32_t err = terrno;
@ -281,6 +283,7 @@ static int32_t raftLogGetEntry(struct SSyncLogStore* pLogStore, SyncIndex index,
terrno = saveErr;
*/
taosThreadMutexUnlock(&(pData->mutex));
return code;
}
@ -301,6 +304,7 @@ static int32_t raftLogGetEntry(struct SSyncLogStore* pLogStore, SyncIndex index,
terrno = saveErr;
*/
taosThreadMutexUnlock(&(pData->mutex));
return code;
}
@ -364,6 +368,7 @@ SSyncLogStore* logStoreCreate(SSyncNode* pSyncNode) {
pData->pWal = pSyncNode->pWal;
ASSERT(pData->pWal != NULL);
taosThreadMutexInit(&(pData->mutex), NULL);
pData->pWalHandle = walOpenReadHandle(pData->pWal);
ASSERT(pData->pWalHandle != NULL);
@ -408,9 +413,14 @@ SSyncLogStore* logStoreCreate(SSyncNode* pSyncNode) {
void logStoreDestory(SSyncLogStore* pLogStore) {
if (pLogStore != NULL) {
SSyncLogStoreData* pData = pLogStore->data;
taosThreadMutexLock(&(pData->mutex));
if (pData->pWalHandle != NULL) {
walCloseReadHandle(pData->pWalHandle);
pData->pWalHandle = NULL;
}
taosThreadMutexUnlock(&(pData->mutex));
taosThreadMutexDestroy(&(pData->mutex));
taosMemoryFree(pLogStore->data);
taosMemoryFree(pLogStore);
@ -460,6 +470,8 @@ SSyncRaftEntry* logStoreGetEntry(SSyncLogStore* pLogStore, SyncIndex index) {
SWal* pWal = pData->pWal;
if (index >= SYNC_INDEX_BEGIN && index <= logStoreLastIndex(pLogStore)) {
taosThreadMutexLock(&(pData->mutex));
// SWalReadHandle* pWalHandle = walOpenReadHandle(pWal);
SWalReadHandle* pWalHandle = pData->pWalHandle;
ASSERT(pWalHandle != NULL);
@ -503,6 +515,7 @@ SSyncRaftEntry* logStoreGetEntry(SSyncLogStore* pLogStore, SyncIndex index) {
terrno = saveErr;
*/
taosThreadMutexUnlock(&(pData->mutex));
return pEntry;
} else {

2
source/libs/sync/src/syncRaftStore.c
View File

@ -216,7 +216,7 @@ cJSON *raftStore2Json(SRaftStore *pRaftStore) {
char *raftStore2Str(SRaftStore *pRaftStore) {
cJSON *pJson = raftStore2Json(pRaftStore);
char *serialized = cJSON_Print(pJson);
char * serialized = cJSON_Print(pJson);
cJSON_Delete(pJson);
return serialized;
}

12
source/libs/sync/src/syncRespMgr.c
View File

@ -32,11 +32,13 @@ SSyncRespMgr *syncRespMgrCreate(void *data, int64_t ttl) {
}
void syncRespMgrDestroy(SSyncRespMgr *pObj) {
taosThreadMutexLock(&(pObj->mutex));
taosHashCleanup(pObj->pRespHash);
taosThreadMutexUnlock(&(pObj->mutex));
taosThreadMutexDestroy(&(pObj->mutex));
taosMemoryFree(pObj);
if (pObj != NULL) {
taosThreadMutexLock(&(pObj->mutex));
taosHashCleanup(pObj->pRespHash);
taosThreadMutexUnlock(&(pObj->mutex));
taosThreadMutexDestroy(&(pObj->mutex));
taosMemoryFree(pObj);
}
}
int64_t syncRespMgrAdd(SSyncRespMgr *pObj, SRespStub *pStub) {

184
source/libs/sync/src/syncSnapshot.c
View File

@ -21,6 +21,7 @@
#include "syncUtil.h"
#include "wal.h"
//----------------------------------
static void snapshotReceiverDoStart(SSyncSnapshotReceiver *pReceiver, SyncTerm privateTerm,
SyncSnapshotSend *pBeginMsg);
@ -49,7 +50,7 @@ SSyncSnapshotSender *snapshotSenderCreate(SSyncNode *pSyncNode, int32_t replicaI
pSender->pSyncNode->pFsm->FpGetSnapshotInfo(pSender->pSyncNode->pFsm, &(pSender->snapshot));
pSender->finish = false;
} else {
sError("snapshotSenderCreate cannot create sender");
sError("vgId:%d cannot create snapshot sender", pSyncNode->vgId);
}
return pSender;
@ -57,39 +58,59 @@ SSyncSnapshotSender *snapshotSenderCreate(SSyncNode *pSyncNode, int32_t replicaI
void snapshotSenderDestroy(SSyncSnapshotSender *pSender) {
if (pSender != NULL) {
// free current block
if (pSender->pCurrentBlock != NULL) {
taosMemoryFree(pSender->pCurrentBlock);
pSender->pCurrentBlock = NULL;
}
// close reader
if (pSender->pReader != NULL) {
int32_t ret = pSender->pSyncNode->pFsm->FpSnapshotStopRead(pSender->pSyncNode->pFsm, pSender->pReader);
ASSERT(ret == 0);
pSender->pReader = NULL;
}
// free sender
taosMemoryFree(pSender);
}
}
bool snapshotSenderIsStart(SSyncSnapshotSender *pSender) { return pSender->start; }
// begin send snapshot (current term, seq begin)
// begin send snapshot by snapshot, pReader
void snapshotSenderStart(SSyncSnapshotSender *pSender, SSnapshot snapshot, void *pReader) {
ASSERT(!snapshotSenderIsStart(pSender));
pSender->seq = SYNC_SNAPSHOT_SEQ_BEGIN;
pSender->ack = SYNC_SNAPSHOT_SEQ_INVALID;
// init snapshot and reader
ASSERT(pSender->pReader == NULL);
pSender->pReader = pReader;
pSender->snapshot = snapshot;
// init current block
if (pSender->pCurrentBlock != NULL) {
taosMemoryFree(pSender->pCurrentBlock);
}
pSender->blockLen = 0;
// update term
pSender->term = pSender->pSyncNode->pRaftStore->currentTerm;
++(pSender->privateTerm);
// update state
pSender->finish = false;
pSender->start = true;
pSender->seq = SYNC_SNAPSHOT_SEQ_BEGIN;
pSender->ack = SYNC_SNAPSHOT_SEQ_INVALID;
// init last config
if (pSender->snapshot.lastConfigIndex != SYNC_INDEX_INVALID) {
int32_t code = 0;
SSyncRaftEntry *pEntry = NULL;
bool getLastConfig = false;
code = pSender->pSyncNode->pLogStore->syncLogGetEntry(pSender->pSyncNode->pLogStore,
pSender->snapshot.lastConfigIndex, &pEntry);
bool getLastConfig = false;
if (code == 0) {
ASSERT(pEntry != NULL);
@ -112,29 +133,29 @@ void snapshotSenderStart(SSyncSnapshotSender *pSender, SSnapshot snapshot, void
}
}
// last config not found in wal, update to -1
if (!getLastConfig) {
char logBuf[128];
snprintf(logBuf, sizeof(logBuf), "snapshot sender update lcindex from %ld to -1",
pSender->snapshot.lastConfigIndex);
pSender->snapshot.lastConfigIndex = -1;
char *eventLog = snapshotSender2SimpleStr(pSender, logBuf);
syncNodeEventLog(pSender->pSyncNode, eventLog);
taosMemoryFree(eventLog);
SyncIndex oldLastConfigIndex = pSender->snapshot.lastConfigIndex;
SyncIndex newLastConfigIndex = SYNC_INDEX_INVALID;
pSender->snapshot.lastConfigIndex = SYNC_INDEX_INVALID;
memset(&(pSender->lastConfig), 0, sizeof(SSyncCfg));
// event log
do {
char logBuf[128];
snprintf(logBuf, sizeof(logBuf), "snapshot sender update lcindex from %ld to %ld", oldLastConfigIndex,
newLastConfigIndex);
char *eventLog = snapshotSender2SimpleStr(pSender, logBuf);
syncNodeEventLog(pSender->pSyncNode, eventLog);
taosMemoryFree(eventLog);
} while (0);
}
} else {
// no last config
memset(&(pSender->lastConfig), 0, sizeof(SSyncCfg));
}
pSender->sendingMS = SYNC_SNAPSHOT_RETRY_MS;
pSender->term = pSender->pSyncNode->pRaftStore->currentTerm;
++(pSender->privateTerm);
pSender->finish = false;
pSender->start = true;
// build begin msg
SyncSnapshotSend *pMsg = syncSnapshotSendBuild(0, pSender->pSyncNode->vgId);
pMsg->srcId = pSender->pSyncNode->myRaftId;
@ -151,40 +172,39 @@ void snapshotSenderStart(SSyncSnapshotSender *pSender, SSnapshot snapshot, void
SRpcMsg rpcMsg;
syncSnapshotSend2RpcMsg(pMsg, &rpcMsg);
syncNodeSendMsgById(&(pMsg->destId), pSender->pSyncNode, &rpcMsg);
char *eventLog = snapshotSender2SimpleStr(pSender, "snapshot sender send");
syncNodeEventLog(pSender->pSyncNode, eventLog);
taosMemoryFree(eventLog);
syncSnapshotSendDestroy(pMsg);
// event log
do {
char *eventLog = snapshotSender2SimpleStr(pSender, "snapshot sender send");
syncNodeEventLog(pSender->pSyncNode, eventLog);
taosMemoryFree(eventLog);
} while (0);
}
void snapshotSenderStop(SSyncSnapshotSender *pSender) {
// close reader
if (pSender->pReader != NULL) {
int32_t ret = pSender->pSyncNode->pFsm->FpSnapshotStopRead(pSender->pSyncNode->pFsm, pSender->pReader);
ASSERT(ret == 0);
pSender->pReader = NULL;
}
// free current block
if (pSender->pCurrentBlock != NULL) {
taosMemoryFree(pSender->pCurrentBlock);
pSender->pCurrentBlock = NULL;
pSender->blockLen = 0;
}
// update flag
pSender->start = false;
if (gRaftDetailLog) {
char *s = snapshotSender2Str(pSender);
sInfo("snapshotSenderStop %s", s);
taosMemoryFree(s);
}
}
// when sender receiver ack, call this function to send msg from seq
// when sender receive ack, call this function to send msg from seq
// seq = ack + 1, already updated
int32_t snapshotSend(SSyncSnapshotSender *pSender) {
// free memory last time (seq - 1)
// free memory last time (current seq - 1)
if (pSender->pCurrentBlock != NULL) {
taosMemoryFree(pSender->pCurrentBlock);
pSender->pCurrentBlock = NULL;
@ -198,7 +218,7 @@ int32_t snapshotSend(SSyncSnapshotSender *pSender) {
if (pSender->blockLen > 0) {
// has read data
} else {
// read finish
// read finish, update seq to end
pSender->seq = SYNC_SNAPSHOT_SEQ_END;
}
@ -219,25 +239,28 @@ int32_t snapshotSend(SSyncSnapshotSender *pSender) {
SRpcMsg rpcMsg;
syncSnapshotSend2RpcMsg(pMsg, &rpcMsg);
syncNodeSendMsgById(&(pMsg->destId), pSender->pSyncNode, &rpcMsg);
if (pSender->seq == SYNC_SNAPSHOT_SEQ_END) {
char *eventLog = snapshotSender2SimpleStr(pSender, "snapshot sender finish");
syncNodeEventLog(pSender->pSyncNode, eventLog);
taosMemoryFree(eventLog);
} else {
char *eventLog = snapshotSender2SimpleStr(pSender, "snapshot sender sending");
syncNodeEventLog(pSender->pSyncNode, eventLog);
taosMemoryFree(eventLog);
}
syncSnapshotSendDestroy(pMsg);
// event log
do {
char *eventLog = NULL;
if (pSender->seq == SYNC_SNAPSHOT_SEQ_END) {
eventLog = snapshotSender2SimpleStr(pSender, "snapshot sender finish");
} else {
eventLog = snapshotSender2SimpleStr(pSender, "snapshot sender sending");
}
syncNodeEventLog(pSender->pSyncNode, eventLog);
taosMemoryFree(eventLog);
} while (0);
return 0;
}
// send snapshot data from cache
int32_t snapshotReSend(SSyncSnapshotSender *pSender) {
if (pSender->pCurrentBlock != NULL) {
// send current block data
if (pSender->pCurrentBlock != NULL && pSender->blockLen > 0) {
// build msg
SyncSnapshotSend *pMsg = syncSnapshotSendBuild(pSender->blockLen, pSender->pSyncNode->vgId);
pMsg->srcId = pSender->pSyncNode->myRaftId;
pMsg->destId = (pSender->pSyncNode->replicasId)[pSender->replicaIndex];
@ -249,16 +272,20 @@ int32_t snapshotReSend(SSyncSnapshotSender *pSender) {
pMsg->seq = pSender->seq;
memcpy(pMsg->data, pSender->pCurrentBlock, pSender->blockLen);
// send msg
SRpcMsg rpcMsg;
syncSnapshotSend2RpcMsg(pMsg, &rpcMsg);
syncNodeSendMsgById(&(pMsg->destId), pSender->pSyncNode, &rpcMsg);
char *eventLog = snapshotSender2SimpleStr(pSender, "snapshot sender resend");
syncNodeEventLog(pSender->pSyncNode, eventLog);
taosMemoryFree(eventLog);
syncSnapshotSendDestroy(pMsg);
// event log
do {
char *eventLog = snapshotSender2SimpleStr(pSender, "snapshot sender resend");
syncNodeEventLog(pSender->pSyncNode, eventLog);
taosMemoryFree(eventLog);
} while (0);
}
return 0;
}
@ -294,7 +321,6 @@ cJSON *snapshotSender2Json(SSyncSnapshotSender *pSender) {
snprintf(u64buf, sizeof(u64buf), "%lu", pSender->snapshot.lastApplyTerm);
cJSON_AddStringToObject(pSnapshot, "lastApplyTerm", u64buf);
cJSON_AddItemToObject(pRoot, "snapshot", pSnapshot);
snprintf(u64buf, sizeof(u64buf), "%lu", pSender->sendingMS);
cJSON_AddStringToObject(pRoot, "sendingMS", u64buf);
snprintf(u64buf, sizeof(u64buf), "%p", pSender->pSyncNode);
@ -324,7 +350,7 @@ char *snapshotSender2SimpleStr(SSyncSnapshotSender *pSender, char *event) {
char *s = taosMemoryMalloc(len);
SRaftId destId = pSender->pSyncNode->replicasId[pSender->replicaIndex];
char host[128];
char host[64];
uint16_t port;
syncUtilU642Addr(destId.addr, host, sizeof(host), &port);
@ -355,12 +381,12 @@ SSyncSnapshotReceiver *snapshotReceiverCreate(SSyncNode *pSyncNode, SRaftId from
pReceiver->term = pSyncNode->pRaftStore->currentTerm;
pReceiver->privateTerm = 0;
pReceiver->snapshot.data = NULL;
pReceiver->snapshot.lastApplyIndex = -1;
pReceiver->snapshot.lastApplyIndex = SYNC_INDEX_INVALID;
pReceiver->snapshot.lastApplyTerm = 0;
pReceiver->snapshot.lastConfigIndex = -1;
pReceiver->snapshot.lastConfigIndex = SYNC_INDEX_INVALID;
} else {
sInfo("snapshotReceiverCreate cannot create receiver");
sError("vgId:%d cannot create snapshot receiver", pSyncNode->vgId);
}
return pReceiver;
@ -368,29 +394,55 @@ SSyncSnapshotReceiver *snapshotReceiverCreate(SSyncNode *pSyncNode, SRaftId from
void snapshotReceiverDestroy(SSyncSnapshotReceiver *pReceiver) {
if (pReceiver != NULL) {
// close writer
if (pReceiver->pWriter != NULL) {
int32_t ret =
pReceiver->pSyncNode->pFsm->FpSnapshotStopWrite(pReceiver->pSyncNode->pFsm, pReceiver->pWriter, false);
ASSERT(ret == 0);
pReceiver->pWriter = NULL;
}
// free receiver
taosMemoryFree(pReceiver);
}
}
bool snapshotReceiverIsStart(SSyncSnapshotReceiver *pReceiver) { return pReceiver->start; }
// begin receive snapshot msg (current term, seq begin)
// static do start
// receive first snapshot data
// privateTerm, pBeginMsg
static void snapshotReceiverDoStart(SSyncSnapshotReceiver *pReceiver, SyncTerm privateTerm,
SyncSnapshotSend *pBeginMsg) {
// update state
pReceiver->term = pReceiver->pSyncNode->pRaftStore->currentTerm;
pReceiver->privateTerm = privateTerm;
pReceiver->ack = SYNC_SNAPSHOT_SEQ_BEGIN;
pReceiver->fromId = pBeginMsg->srcId;
// update snapshot
pReceiver->snapshot.lastApplyIndex = pBeginMsg->lastIndex;
pReceiver->snapshot.lastApplyTerm = pBeginMsg->lastTerm;
pReceiver->snapshot.lastConfigIndex = pBeginMsg->lastConfigIndex;
// write data
ASSERT(pReceiver->pWriter == NULL);
int32_t ret = pReceiver->pSyncNode->pFsm->FpSnapshotStartWrite(pReceiver->pSyncNode->pFsm, &(pReceiver->pWriter));
ASSERT(ret == 0);
}
static void snapshotReceiverForceStop(SSyncSnapshotReceiver *pReceiver) {
// force close, abandon incomplete data
if (pReceiver->pWriter != NULL) {
int32_t ret =
pReceiver->pSyncNode->pFsm->FpSnapshotStopWrite(pReceiver->pSyncNode->pFsm, pReceiver->pWriter, false);
ASSERT(ret == 0);
pReceiver->pWriter = NULL;
}
pReceiver->start = false;
}
// if receiver receive msg from seq = SYNC_SNAPSHOT_SEQ_BEGIN, start receiver
// if already start, force close, start again
void snapshotReceiverStart(SSyncSnapshotReceiver *pReceiver, SyncTerm privateTerm, SyncSnapshotSend *pBeginMsg) {
@ -413,12 +465,6 @@ void snapshotReceiverStart(SSyncSnapshotReceiver *pReceiver, SyncTerm privateTer
snapshotReceiverDoStart(pReceiver, privateTerm, pBeginMsg);
pReceiver->start = true;
}
if (gRaftDetailLog) {
char *s = snapshotReceiver2Str(pReceiver);
sInfo("snapshotReceiverStart %s", s);
taosMemoryFree(s);
}
}
void snapshotReceiverStop(SSyncSnapshotReceiver *pReceiver, bool apply) {
@ -434,12 +480,6 @@ void snapshotReceiverStop(SSyncSnapshotReceiver *pReceiver, bool apply) {
if (apply) {
// ++(pReceiver->privateTerm);
}
if (gRaftDetailLog) {
char *s = snapshotReceiver2Str(pReceiver);
sInfo("snapshotReceiverStop %s", s);
taosMemoryFree(s);
}
}
cJSON *snapshotReceiver2Json(SSyncSnapshotReceiver *pReceiver) {

4
source/libs/sync/src/syncVoteMgr.c
View File

@ -127,7 +127,7 @@ cJSON *voteGranted2Json(SVotesGranted *pVotesGranted) {
char *voteGranted2Str(SVotesGranted *pVotesGranted) {
cJSON *pJson = voteGranted2Json(pVotesGranted);
char *serialized = cJSON_Print(pJson);
char * serialized = cJSON_Print(pJson);
cJSON_Delete(pJson);
return serialized;
}
@ -256,7 +256,7 @@ cJSON *votesRespond2Json(SVotesRespond *pVotesRespond) {
char *votesRespond2Str(SVotesRespond *pVotesRespond) {
cJSON *pJson = votesRespond2Json(pVotesRespond);
char *serialized = cJSON_Print(pJson);
char * serialized = cJSON_Print(pJson);
cJSON_Delete(pJson);
return serialized;
}

14
source/libs/sync/test/CMakeLists.txt
View File

@ -17,6 +17,7 @@ add_executable(syncVotesRespondTest "")
add_executable(syncIndexMgrTest "")
add_executable(syncLogStoreTest "")
add_executable(syncEntryTest "")
add_executable(syncEntryCacheTest "")
add_executable(syncRequestVoteTest "")
add_executable(syncRequestVoteReplyTest "")
add_executable(syncAppendEntriesTest "")
@ -129,6 +130,10 @@ target_sources(syncEntryTest
PRIVATE
"syncEntryTest.cpp"
)
target_sources(syncEntryCacheTest
PRIVATE
"syncEntryCacheTest.cpp"
)
target_sources(syncRequestVoteTest
PRIVATE
"syncRequestVoteTest.cpp"
@ -362,6 +367,11 @@ target_include_directories(syncEntryTest
"${TD_SOURCE_DIR}/include/libs/sync"
"${CMAKE_CURRENT_SOURCE_DIR}/../inc"
)
target_include_directories(syncEntryCacheTest
PUBLIC
"${TD_SOURCE_DIR}/include/libs/sync"
"${CMAKE_CURRENT_SOURCE_DIR}/../inc"
)
target_include_directories(syncRequestVoteTest
PUBLIC
"${TD_SOURCE_DIR}/include/libs/sync"
@ -610,6 +620,10 @@ target_link_libraries(syncEntryTest
sync
gtest_main
)
target_link_libraries(syncEntryCacheTest
sync
gtest_main
)
target_link_libraries(syncRequestVoteTest
sync
gtest_main

162
source/libs/sync/test/syncEntryCacheTest.cpp Normal file
View File

@ -0,0 +1,162 @@
#include <stdio.h>
#include "syncEnv.h"
#include "syncIO.h"
#include "syncInt.h"
#include "syncRaftLog.h"
#include "syncRaftStore.h"
#include "syncUtil.h"
void logTest() {
sTrace("--- sync log test: trace");
sDebug("--- sync log test: debug");
sInfo("--- sync log test: info");
sWarn("--- sync log test: warn");
sError("--- sync log test: error");
sFatal("--- sync log test: fatal");
}
SSyncRaftEntry* createEntry(int i) {
int32_t dataLen = 20;
SSyncRaftEntry* pEntry = syncEntryBuild(dataLen);
assert(pEntry != NULL);
pEntry->msgType = 88;
pEntry->originalRpcType = 99;
pEntry->seqNum = 3;
pEntry->isWeak = true;
pEntry->term = 100 + i;
pEntry->index = i;
snprintf(pEntry->data, dataLen, "value%d", i);
return pEntry;
}
SSyncNode* createFakeNode() {
SSyncNode* pSyncNode = (SSyncNode*)taosMemoryMalloc(sizeof(SSyncNode));
ASSERT(pSyncNode != NULL);
memset(pSyncNode, 0, sizeof(SSyncNode));
return pSyncNode;
}
SRaftEntryCache* createCache(int maxCount) {
SSyncNode* pSyncNode = createFakeNode();
ASSERT(pSyncNode != NULL);
SRaftEntryCache* pCache = raftCacheCreate(pSyncNode, maxCount);
ASSERT(pCache != NULL);
return pCache;
}
void test1() {
int32_t code = 0;
SRaftEntryCache* pCache = createCache(5);
for (int i = 0; i < 5; ++i) {
SSyncRaftEntry* pEntry = createEntry(i);
code = raftCachePutEntry(pCache, pEntry);
ASSERT(code == 1);
syncEntryDestory(pEntry);
}
raftCacheLog2((char*)"==test1 write 5 entries==", pCache);
SyncIndex index;
index = 1;
code = raftCacheDelEntry(pCache, index);
ASSERT(code == 0);
index = 3;
code = raftCacheDelEntry(pCache, index);
ASSERT(code == 0);
raftCacheLog2((char*)"==test1 delete 1,3==", pCache);
code = raftCacheClear(pCache);
ASSERT(code == 0);
raftCacheLog2((char*)"==clear all==", pCache);
}
void test2() {
int32_t code = 0;
SRaftEntryCache* pCache = createCache(5);
for (int i = 0; i < 5; ++i) {
SSyncRaftEntry* pEntry = createEntry(i);
code = raftCachePutEntry(pCache, pEntry);
ASSERT(code == 1);
syncEntryDestory(pEntry);
}
raftCacheLog2((char*)"==test2 write 5 entries==", pCache);
SyncIndex index;
index = 1;
SSyncRaftEntry* pEntry;
code = raftCacheGetEntry(pCache, index, &pEntry);
ASSERT(code == 0);
syncEntryDestory(pEntry);
syncEntryLog2((char*)"==test2 get entry 1==", pEntry);
index = 2;
code = raftCacheGetEntryP(pCache, index, &pEntry);
ASSERT(code == 0);
syncEntryLog2((char*)"==test2 get entry pointer 2==", pEntry);
// not found
index = 8;
code = raftCacheGetEntry(pCache, index, &pEntry);
ASSERT(code == -1 && terrno == TSDB_CODE_WAL_LOG_NOT_EXIST);
sTrace("==test2 get entry 8 not found==");
// not found
index = 9;
code = raftCacheGetEntryP(pCache, index, &pEntry);
ASSERT(code == -1 && terrno == TSDB_CODE_WAL_LOG_NOT_EXIST);
sTrace("==test2 get entry pointer 9 not found==");
}
void test3() {
int32_t code = 0;
SRaftEntryCache* pCache = createCache(5);
for (int i = 0; i < 5; ++i) {
SSyncRaftEntry* pEntry = createEntry(i);
code = raftCachePutEntry(pCache, pEntry);
ASSERT(code == 1);
syncEntryDestory(pEntry);
}
for (int i = 6; i < 10; ++i) {
SSyncRaftEntry* pEntry = createEntry(i);
code = raftCachePutEntry(pCache, pEntry);
ASSERT(code == 0);
syncEntryDestory(pEntry);
}
raftCacheLog2((char*)"==test3 write 10 entries, max count is 5==", pCache);
}
void test4() {
int32_t code = 0;
SRaftEntryCache* pCache = createCache(5);
for (int i = 0; i < 5; ++i) {
SSyncRaftEntry* pEntry = createEntry(i);
code = raftCachePutEntry(pCache, pEntry);
ASSERT(code == 1);
syncEntryDestory(pEntry);
}
raftCacheLog2((char*)"==test4 write 5 entries==", pCache);
SyncIndex index;
index = 3;
SSyncRaftEntry* pEntry;
code = raftCacheGetAndDel(pCache, index, &pEntry);
ASSERT(code == 0);
syncEntryLog2((char*)"==test4 get-and-del entry 3==", pEntry);
raftCacheLog2((char*)"==test4 after get-and-del entry 3==", pCache);
}
int main(int argc, char** argv) {
gRaftDetailLog = true;
tsAsyncLog = 0;
sDebugFlag = DEBUG_TRACE + DEBUG_SCREEN + DEBUG_FILE + DEBUG_DEBUG;
test1();
test2();
test3();
test4();
return 0;
}

27
source/libs/sync/test/syncRaftLogTest2.cpp
View File

@ -113,7 +113,7 @@ void test2() {
pLogStore = logStoreCreate(pSyncNode);
assert(pLogStore);
pSyncNode->pLogStore = pLogStore;
//pLogStore->syncLogSetBeginIndex(pLogStore, 5);
// pLogStore->syncLogSetBeginIndex(pLogStore, 5);
pLogStore->syncLogRestoreFromSnapshot(pLogStore, 4);
logStoreLog2((char*)"\n\n\ntest2 ----- ", pLogStore);
@ -229,7 +229,7 @@ void test4() {
assert(pLogStore);
pSyncNode->pLogStore = pLogStore;
logStoreLog2((char*)"\n\n\ntest4 ----- ", pLogStore);
//pLogStore->syncLogSetBeginIndex(pLogStore, 5);
// pLogStore->syncLogSetBeginIndex(pLogStore, 5);
pLogStore->syncLogRestoreFromSnapshot(pLogStore, 4);
for (int i = 5; i <= 9; ++i) {
@ -291,7 +291,7 @@ void test5() {
assert(pLogStore);
pSyncNode->pLogStore = pLogStore;
logStoreLog2((char*)"\n\n\ntest5 ----- ", pLogStore);
//pLogStore->syncLogSetBeginIndex(pLogStore, 5);
// pLogStore->syncLogSetBeginIndex(pLogStore, 5);
pLogStore->syncLogRestoreFromSnapshot(pLogStore, 4);
for (int i = 5; i <= 9; ++i) {
@ -412,26 +412,23 @@ void test6() {
do {
SyncIndex firstVer = walGetFirstVer(pWal);
SyncIndex lastVer = walGetLastVer(pWal);
bool isEmpty = walIsEmpty(pWal);
bool isEmpty = walIsEmpty(pWal);
printf("before -------- firstVer:%ld lastVer:%ld isEmpty:%d \n", firstVer, lastVer, isEmpty);
} while (0);
logStoreDestory(pLogStore);
cleanup();
// restart
init();
pLogStore = logStoreCreate(pSyncNode);
assert(pLogStore);
pSyncNode->pLogStore = pLogStore;
do {
SyncIndex firstVer = walGetFirstVer(pWal);
SyncIndex lastVer = walGetLastVer(pWal);
bool isEmpty = walIsEmpty(pWal);
bool isEmpty = walIsEmpty(pWal);
printf("after -------- firstVer:%ld lastVer:%ld isEmpty:%d \n", firstVer, lastVer, isEmpty);
} while (0);
@ -461,13 +458,13 @@ int main(int argc, char** argv) {
}
sTrace("gAssert : %d", gAssert);
/*
test1();
test2();
test3();
test4();
test5();
*/
/*
test1();
test2();
test3();
test4();
test5();
*/
test6();
return 0;

2
source/libs/sync/test/syncRaftLogTest3.cpp
View File

@ -312,7 +312,7 @@ void test5() {
pSyncNode->pLogStore = pLogStore;
logStoreLog2((char*)"\n\n\ntest5 ----- ", pLogStore);
//pSyncNode->pLogStore->syncLogSetBeginIndex(pSyncNode->pLogStore, 6);
// pSyncNode->pLogStore->syncLogSetBeginIndex(pSyncNode->pLogStore, 6);
pLogStore->syncLogRestoreFromSnapshot(pSyncNode->pLogStore, 5);
for (int i = 6; i <= 10; ++i) {
int32_t dataLen = 10;

2
source/libs/transport/inc/transportInt.h
View File

@ -52,7 +52,7 @@ typedef struct {
char user[TSDB_UNI_LEN]; // meter ID
void (*cfp)(void* parent, SRpcMsg*, SEpSet*);
bool (*retry)(int32_t code);
bool (*retry)(int32_t code, tmsg_t msgType);
int index;
void* parent;

2
source/libs/transport/src/transCli.c
View File

@ -1027,7 +1027,7 @@ int cliAppCb(SCliConn* pConn, STransMsg* pResp, SCliMsg* pMsg) {
*/
STransConnCtx* pCtx = pMsg->ctx;
int32_t code = pResp->code;
if (pTransInst->retry != NULL && pTransInst->retry(code)) {
if (pTransInst->retry != NULL && pTransInst->retry(code, pResp->msgType - 1)) {
pMsg->sent = 0;
pCtx->retryCnt += 1;
if (code == TSDB_CODE_RPC_NETWORK_UNAVAIL) {

4
source/os/src/osFile.c
View File

@ -157,14 +157,14 @@ int32_t taosRenameFile(const char *oldName, const char *newName) {
#ifdef WINDOWS
bool code = MoveFileEx(oldName, newName, MOVEFILE_REPLACE_EXISTING | MOVEFILE_COPY_ALLOWED);
if (!code) {
printf("failed to rename file %s to %s, reason:%s", oldName, newName, strerror(errno));
printf("failed to rename file %s to %s, reason:%s\n", oldName, newName, strerror(errno));
}
return !code;
#else
int32_t code = rename(oldName, newName);
if (code < 0) {
printf("failed to rename file %s to %s, reason:%s", oldName, newName, strerror(errno));
printf("failed to rename file %s to %s, reason:%s\n", oldName, newName, strerror(errno));
}
return code;

9
source/os/src/osSystem.c
View File

@ -18,6 +18,10 @@
#include "os.h"
#if defined(WINDOWS)
BOOL WINAPI CtrlHandler(DWORD fdwCtrlType) {
printf("\n" TAOS_CONSOLE_PROMPT_HEADER);
return TRUE;
}
#elif defined(_TD_DARWIN_64)
#else
#include <dlfcn.h>
@ -124,7 +128,7 @@ int taosSetConsoleEcho(bool on) {
void taosSetTerminalMode() {
#if defined(WINDOWS)
// assert(0);
SetConsoleCtrlHandler(CtrlHandler, TRUE);
#else
struct termios newtio;
@ -158,7 +162,6 @@ void taosSetTerminalMode() {
int32_t taosGetOldTerminalMode() {
#if defined(WINDOWS)
// assert(0);
#else
/* Make sure stdin is a terminal. */
if (!isatty(STDIN_FILENO)) {
@ -176,7 +179,7 @@ int32_t taosGetOldTerminalMode() {
void taosResetTerminalMode() {
#if defined(WINDOWS)
// assert(0);
SetConsoleCtrlHandler(CtrlHandler, FALSE);
#else
if (tcsetattr(0, TCSANOW, &oldtio) != 0) {
fprintf(stderr, "Fail to reset the terminal properties!\n");

2
tests/parallel_test/container_build.sh
View File

@ -52,7 +52,7 @@ fi
docker run \
-v $REP_MOUNT_PARAM \
--rm --ulimit core=-1 taos_test:v1.0 sh -c "cd $REP_DIR;rm -rf debug;mkdir -p debug;cd debug;cmake .. -DBUILD_TOOLS=true;make -j $THREAD_COUNT"
--rm --ulimit core=-1 taos_test:v1.0 sh -c "cd $REP_DIR;rm -rf debug;mkdir -p debug;cd debug;cmake .. -DBUILD_HTTP=false -DBUILD_TOOLS=true;make -j $THREAD_COUNT"
ret=$?
exit $ret

107
tests/pytest/util/constant.py
View File

@ -1,5 +1,76 @@
# -*- coding: utf-8 -*-
# basic type
TAOS_DATA_TYPE = [
"INT", "BIGINT", "SMALLINT", "TINYINT", "INT UNSIGNED", "BIGINT UNSIGNED", "SMALLINT UNSIGNED", "TINYINT UNSIGNED",
"FLOAT", "DOUBLE",
"BOOL",
"BINARY", "NCHAR", "VARCHAR",
"TIMESTAMP",
# "MEDIUMBLOB", "BLOB", # add in 3.x
# "DECIMAL", "NUMERIC", # add in 3.x
"JSON", # only for tag
]
TAOS_NUM_TYPE = [
"INT", "BIGINT", "SMALLINT", "TINYINT", "INT UNSIGNED", "BIGINT UNSIGNED", "SMALLINT UNSIGNED", "TINYINT UNSIGNED", "FLOAT", "DOUBLE",
# "DECIMAL", "NUMERIC", # add in 3.x
]
TAOS_CHAR_TYPE = [
"BINARY", "NCHAR", "VARCHAR",
]
TAOS_BOOL_TYPE = ["BOOL",]
TAOS_TS_TYPE = ["TIMESTAMP",]
TAOS_BIN_TYPE = [
"MEDIUMBLOB", "BLOB", # add in 3.x
]
TAOS_TIME_INIT = ["b", "u", "a", "s", "m", "h", "d", "w", "n", "y"]
TAOS_PRECISION = ["ms", "us", "ns"]
PRECISION_DEFAULT = "ms"
PRECISION = PRECISION_DEFAULT
TAOS_KEYWORDS = [
"ABORT", "CREATE", "IGNORE", "NULL", "STAR",
"ACCOUNT", "CTIME", "IMMEDIATE", "OF", "STATE",
"ACCOUNTS", "DATABASE", "IMPORT", "OFFSET", "STATEMENT",
"ADD", "DATABASES", "IN", "OR", "STATE_WINDOW",
"AFTER", "DAYS", "INITIALLY", "ORDER", "STORAGE",
"ALL", "DBS", "INSERT", "PARTITIONS", "STREAM",
"ALTER", "DEFERRED", "INSTEAD", "PASS", "STREAMS",
"AND", "DELIMITERS", "INT", "PLUS", "STRING",
"AS", "DESC", "INTEGER", "PPS", "SYNCDB",
"ASC", "DESCRIBE", "INTERVAL", "PRECISION", "TABLE",
"ATTACH", "DETACH", "INTO", "PREV", "TABLES",
"BEFORE", "DISTINCT", "IS", "PRIVILEGE", "TAG",
"BEGIN", "DIVIDE", "ISNULL", "QTIME", "TAGS",
"BETWEEN", "DNODE", "JOIN", "QUERIES", "TBNAME",
"BIGINT", "DNODES", "KEEP", "QUERY", "TIMES",
"BINARY", "DOT", "KEY", "QUORUM", "TIMESTAMP",
"BITAND", "DOUBLE", "KILL", "RAISE", "TINYINT",
"BITNOT", "DROP", "LE", "REM", "TOPIC",
"BITOR", "EACH", "LIKE", "REPLACE", "TOPICS",
"BLOCKS", "END", "LIMIT", "REPLICA", "TRIGGER",
"BOOL", "EQ", "LINEAR", "RESET", "TSERIES",
"BY", "EXISTS", "LOCAL", "RESTRICT", "UMINUS",
"CACHE", "EXPLAIN", "LP", "ROW", "UNION",
"CACHELAST", "FAIL", "LSHIFT", "RP", "UNSIGNED",
"CASCADE", "FILE", "LT", "RSHIFT", "UPDATE",
"CHANGE", "FILL", "MATCH", "SCORES", "UPLUS",
"CLUSTER", "FLOAT", "MAXROWS", "SELECT", "USE",
"COLON", "FOR", "MINROWS", "SEMI", "USER",
"COLUMN", "FROM", "MINUS", "SESSION", "USERS",
"COMMA", "FSYNC", "MNODES", "SET", "USING",
"COMP", "GE", "MODIFY", "SHOW", "VALUES",
"COMPACT", "GLOB", "MODULES", "SLASH", "VARIABLE",
"CONCAT", "GRANTS", "NCHAR", "SLIDING", "VARIABLES",
"CONFLICT", "GROUP", "NE", "SLIMIT", "VGROUPS",
"CONNECTION", "GT", "NONE", "SMALLINT", "VIEW",
"CONNECTIONS", "HAVING", "NOT", "SOFFSET", "VNODES",
"CONNS", "ID", "NOTNULL", "STABLE", "WAL",
"COPY", "IF", "NOW", "STABLES", "WHERE",
]
# basic data type boundary
TINYINT_MAX = 127
TINYINT_MIN = -128
@ -11,7 +82,7 @@ SMALLINT_MAX = 32767
SMALLINT_MIN = -32768
SMALLINT_UN_MAX = 65535
MALLINT_UN_MIN = 0
SMALLINT_UN_MIN = 0
INT_MAX = 2147483647
INT_MIN = -2147483648
@ -33,8 +104,8 @@ DOUBLE_MIN = -1.7E+308
# schema boundary
BINARY_LENGTH_MAX = 16374
NCAHR_LENGTH_MAX_ = 4093
DBNAME_LENGTH_MAX_ = 64
NCAHR_LENGTH_MAX = 4093
DBNAME_LENGTH_MAX = 64
STBNAME_LENGTH_MAX = 192
STBNAME_LENGTH_MIN = 1
@ -66,4 +137,32 @@ MNODE_SHM_SIZE_DEFAULT = 6292480
VNODE_SHM_SIZE_MAX = 2147483647
VNODE_SHM_SIZE_MIN = 6292480
VNODE_SHM_SIZE_DEFAULT = 31458304
VNODE_SHM_SIZE_DEFAULT = 31458304
# time_init
TIME_MS = 1
TIME_US = TIME_MS/1000
TIME_NS = TIME_US/1000
TIME_S = 1000 * TIME_MS
TIME_M = 60 * TIME_S
TIME_H = 60 * TIME_M
TIME_D = 24 * TIME_H
TIME_W = 7 * TIME_D
TIME_N = 30 * TIME_D
TIME_Y = 365 * TIME_D
# session parameters
INTERVAL_MIN = 1 * TIME_MS if PRECISION == PRECISION_DEFAULT else 1 * TIME_US
# streams and related agg-function
SMA_INDEX_FUNCTIONS = ["MIN", "MAX"]
ROLLUP_FUNCTIONS = ["AVG", "SUM", "MIN", "MAX", "LAST", "FIRST"]
SMA_WATMARK_MAXDELAY_INIT = ['a', "s", "m"]
WATERMARK_MAX = 900000
WATERMARK_MIN = 0
MAX_DELAY_MAX = 900000
MAX_DELAY_MIN = 1

39
tests/pytest/util/sql.py
View File

@ -21,6 +21,7 @@ import psutil
import shutil
import pandas as pd
from util.log import *
from util.constant import *
def _parse_datetime(timestr):
try:
@ -117,8 +118,7 @@ class TDSql:
col_name_list = []
col_type_list = []
self.cursor.execute(sql)
self.queryCols = self.cursor.description
for query_col in self.queryCols:
for query_col in self.cursor.description:
col_name_list.append(query_col[0])
col_type_list.append(query_col[1])
except Exception as e:
@ -301,6 +301,41 @@ class TDSql:
args = (caller.filename, caller.lineno, self.sql, elm, expect_elm)
tdLog.exit("%s(%d) failed: sql:%s, elm:%s == expect_elm:%s" % args)
def get_times(self, time_str, precision="ms"):
caller = inspect.getframeinfo(inspect.stack()[1][0])
if time_str[-1] not in TAOS_TIME_INIT:
tdLog.exit(f"{caller.filename}({caller.lineno}) failed: {time_str} not a standard taos time init")
if precision not in TAOS_PRECISION:
tdLog.exit(f"{caller.filename}({caller.lineno}) failed: {precision} not a standard taos time precision")
if time_str[-1] == TAOS_TIME_INIT[0]:
times = int(time_str[:-1]) * TIME_NS
if time_str[-1] == TAOS_TIME_INIT[1]:
times = int(time_str[:-1]) * TIME_US
if time_str[-1] == TAOS_TIME_INIT[2]:
times = int(time_str[:-1]) * TIME_MS
if time_str[-1] == TAOS_TIME_INIT[3]:
times = int(time_str[:-1]) * TIME_S
if time_str[-1] == TAOS_TIME_INIT[4]:
times = int(time_str[:-1]) * TIME_M
if time_str[-1] == TAOS_TIME_INIT[5]:
times = int(time_str[:-1]) * TIME_H
if time_str[-1] == TAOS_TIME_INIT[6]:
times = int(time_str[:-1]) * TIME_D
if time_str[-1] == TAOS_TIME_INIT[7]:
times = int(time_str[:-1]) * TIME_W
if time_str[-1] == TAOS_TIME_INIT[8]:
times = int(time_str[:-1]) * TIME_N
if time_str[-1] == TAOS_TIME_INIT[9]:
times = int(time_str[:-1]) * TIME_Y
if precision == "ms":
return int(times)
elif precision == "us":
return int(times*1000)
elif precision == "ns":
return int(times*1000*1000)
def taosdStatus(self, state):
tdLog.sleep(5)
pstate = 0

12
tests/script/jenkins/basic.txt
View File

@ -22,15 +22,19 @@
# ---- dnode
./test.sh -f tsim/dnode/balance_replica1.sim
#./test.sh -f tsim/dnode/balance_replica3.sim
./test.sh -f tsim/dnode/create_dnode.sim
./test.sh -f tsim/dnode/drop_dnode_has_mnode.sim
./test.sh -f tsim/dnode/drop_dnode_has_qnode_snode.sim
./test.sh -f tsim/dnode/drop_dnode_has_vnode_replica1.sim
#./test.sh -f tsim/dnode/drop_dnode_has_vnode_replica3.sim
#./test.sh -f tsim/dnode/drop_dnode_has_multi_vnode_replica1.sim
#./test.sh -f tsim/dnode/drop_dnode_has_multi_vnode_replica3.sim
#./test.sh -f tsim/dnode/redistribute_vgroup_replica3_v1_leader.sim
./test.sh -f tsim/dnode/drop_dnode_has_vnode_replica3.sim
./test.sh -f tsim/dnode/drop_dnode_has_multi_vnode_replica1.sim
./test.sh -f tsim/dnode/drop_dnode_has_multi_vnode_replica3.sim
./test.sh -f tsim/dnode/redistribute_vgroup_replica1.sim
./test.sh -f tsim/dnode/redistribute_vgroup_replica3_v1_leader.sim
./test.sh -f tsim/dnode/redistribute_vgroup_replica3_v1_follower.sim
./test.sh -f tsim/dnode/redistribute_vgroup_replica3_v2.sim
./test.sh -f tsim/dnode/redistribute_vgroup_replica3_v3.sim
# ---- insert
./test.sh -f tsim/insert/basic0.sim

26
tests/script/test-all.bat
View File

@ -11,14 +11,17 @@ if not "%2" == "" (
)
for /F "usebackq tokens=*" %%i in (!caseFile!) do (
set line=%%i
if "!line:~,9!" == "./test.sh" (
set /a a+=1
echo !a! Processing %%i
call :GetTimeSeconds !time!
set time1=!_timeTemp!
echo Start at !time!
call !line:./test.sh=wtest.bat! > result_!a!.txt 2>error_!a!.txt
if errorlevel 1 ( call :colorEcho 0c "failed" &echo. && set /a exitNum=8 && echo %%i >>failed.txt ) else ( call :colorEcho 0a "Success" &echo. )
call :CheckSkipCase %%i
if !skipCase! == false (
if "!line:~,9!" == "./test.sh" (
set /a a+=1
echo !a! Processing %%i
call :GetTimeSeconds !time!
set time1=!_timeTemp!
echo Start at !time!
call !line:./test.sh=wtest.bat! > result_!a!.txt 2>error_!a!.txt
if errorlevel 1 ( call :colorEcho 0c "failed" &echo. && set /a exitNum=8 && echo %%i >>failed.txt ) else ( call :colorEcho 0a "Success" &echo. )
)
)
)
exit !exitNum!
@ -56,3 +59,10 @@ for %%a in (%tt%) do (
)
set /a _timeTemp=(%hh%*60+%mm%)*60+%ss%
goto :eof
:CheckSkipCase
set skipCase=false
if "%*" == "./test.sh -f tsim/query/scalarFunction.sim" ( set skipCase=true )
if "%*" == "./test.sh -f tsim/stream/distributeInterval0.sim" ( set skipCase=true )
if "%*" == "./test.sh -f tsim/sma/rsmaCreateInsertQuery.sim" ( set skipCase=true )
:goto eof

1
tests/script/tsim/dnode/balance_replica1.sim
View File

@ -117,7 +117,6 @@ if $rows != 6 then
return -1
endi
return
system sh/exec.sh -n dnode1 -s stop -x SIGINT
system sh/exec.sh -n dnode2 -s stop -x SIGINT
system sh/exec.sh -n dnode3 -s stop -x SIGINT

362
tests/script/tsim/dnode/balance_replica3.sim Normal file
View File

@ -0,0 +1,362 @@
system sh/stop_dnodes.sh
system sh/deploy.sh -n dnode1 -i 1
system sh/deploy.sh -n dnode2 -i 2
system sh/deploy.sh -n dnode3 -i 3
system sh/deploy.sh -n dnode4 -i 4
system sh/deploy.sh -n dnode5 -i 5
system sh/cfg.sh -n dnode1 -c transPullupInterval -v 1
system sh/cfg.sh -n dnode2 -c transPullupInterval -v 1
system sh/cfg.sh -n dnode3 -c transPullupInterval -v 1
system sh/cfg.sh -n dnode4 -c transPullupInterval -v 1
system sh/cfg.sh -n dnode5 -c transPullupInterval -v 1
system sh/cfg.sh -n dnode1 -c supportVnodes -v 0
system sh/exec.sh -n dnode1 -s start
system sh/exec.sh -n dnode2 -s start
system sh/exec.sh -n dnode3 -s start
system sh/exec.sh -n dnode4 -s start
sql connect
print =============== step1 create dnode2
sql create dnode $hostname port 7200
sql create dnode $hostname port 7300
sql create dnode $hostname port 7400
sql create dnode $hostname port 7500
$x = 0
step1:
$x = $x + 1
sleep 1000
if $x == 10 then
print ====> dnode not ready!
return -1
endi
sql show dnodes
print ===> $data00 $data01 $data02 $data03 $data04 $data05
print ===> $data10 $data11 $data12 $data13 $data14 $data15
print ===> $data20 $data21 $data22 $data23 $data24 $data25
print ===> $data30 $data31 $data32 $data33 $data34 $data35
print ===> $data40 $data41 $data42 $data43 $data44 $data45
if $rows != 5 then
return -1
endi
if $data(1)[4] != ready then
goto step1
endi
if $data(2)[4] != ready then
goto step1
endi
if $data(3)[4] != ready then
goto step1
endi
if $data(4)[4] != ready then
goto step1
endi
print =============== step2: create db
sql create database d1 vgroups 4 replica 3
print =============== step32 wait vgroup2
$x = 0
step32:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(2)[4] == leader then
$leaderExist = 1
endi
if $data(2)[6] == leader then
$leaderExist = 1
endi
if $data(2)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step32
endi
print =============== step33 wait vgroup3
$x = 0
step33:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(3)[4] == leader then
$leaderExist = 1
endi
if $data(3)[6] == leader then
$leaderExist = 1
endi
if $data(3)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step33
endi
print =============== step34 wait vgroup4
$x = 0
step34:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(4)[4] == leader then
$leaderExist = 1
endi
if $data(4)[6] == leader then
$leaderExist = 1
endi
if $data(4)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step34
endi
print =============== step35 wait vgroup5
$x = 0
step35:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(4)[4] == leader then
$leaderExist = 1
endi
if $data(4)[6] == leader then
$leaderExist = 1
endi
if $data(4)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step35
endi
print =============== step36: create table
sql use d1
sql create table d1.st (ts timestamp, i int) tags (j int)
sql create table d1.c1 using st tags(1)
sql create table d1.c2 using st tags(1)
sql create table d1.c3 using st tags(1)
sql create table d1.c4 using st tags(1)
sql create table d1.c5 using st tags(1)
sql create table d1.c6 using st tags(1)
sql show d1.tables
if $rows != 6 then
return -1
endi
print =============== step4: start dnode5
system sh/exec.sh -n dnode5 -s start
$x = 0
step4:
$x = $x + 1
sleep 1000
if $x == 10 then
print ====> dnode not ready!
return -1
endi
sql show dnodes
print ===> $data00 $data01 $data02 $data03 $data04 $data05
print ===> $data10 $data11 $data12 $data13 $data14 $data15
print ===> $data20 $data21 $data22 $data23 $data24 $data25
print ===> $data30 $data31 $data32 $data33 $data34 $data35
print ===> $data40 $data41 $data42 $data43 $data44 $data45
if $rows != 5 then
return -1
endi
if $data(1)[4] != ready then
goto step4
endi
if $data(2)[4] != ready then
goto step4
endi
if $data(3)[4] != ready then
goto step4
endi
if $data(4)[4] != ready then
goto step4
endi
if $data(5)[4] != ready then
goto step4
endi
print =============== step5: balance
sql balance vgroup
print =============== step62 wait vgroup2
$x = 0
step62:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(2)[4] == leader then
$leaderExist = 1
endi
if $data(2)[6] == leader then
$leaderExist = 1
endi
if $data(2)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step62
endi
print =============== step63 wait vgroup3
$x = 0
step63:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(3)[4] == leader then
$leaderExist = 1
endi
if $data(3)[6] == leader then
$leaderExist = 1
endi
if $data(3)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step63
endi
print =============== step64 wait vgroup4
$x = 0
step64:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(4)[4] == leader then
$leaderExist = 1
endi
if $data(4)[6] == leader then
$leaderExist = 1
endi
if $data(4)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step64
endi
print =============== step65 wait vgroup5
$x = 0
step65:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(4)[4] == leader then
$leaderExist = 1
endi
if $data(4)[6] == leader then
$leaderExist = 1
endi
if $data(4)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step65
endi
print =============== step7: select data
sql show d1.tables
print rows $rows
if $rows != 6 then
return -1
endi
system sh/exec.sh -n dnode1 -s stop -x SIGINT
system sh/exec.sh -n dnode2 -s stop -x SIGINT
system sh/exec.sh -n dnode3 -s stop -x SIGINT

263
tests/script/tsim/dnode/drop_dnode_has_multi_vnode_replica3.sim
View File

@ -28,7 +28,7 @@ step1:
sql show dnodes
print ===> $data00 $data01 $data02 $data03 $data04 $data05
print ===> $data10 $data11 $data12 $data13 $data14 $data15
if $rows != 3 then
if $rows != 5 then
return -1
endi
if $data(1)[4] != ready then
@ -49,6 +49,128 @@ endi
print =============== step3 create database
sql create database d1 vgroups 4 replica 3
print =============== step32 wait vgroup2
$x = 0
step32:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(2)[4] == leader then
$leaderExist = 1
endi
if $data(2)[6] == leader then
$leaderExist = 1
endi
if $data(2)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step32
endi
print =============== step33 wait vgroup3
$x = 0
step33:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(3)[4] == leader then
$leaderExist = 1
endi
if $data(3)[6] == leader then
$leaderExist = 1
endi
if $data(3)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step33
endi
print =============== step34 wait vgroup4
$x = 0
step34:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(4)[4] == leader then
$leaderExist = 1
endi
if $data(4)[6] == leader then
$leaderExist = 1
endi
if $data(4)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step34
endi
print =============== step35 wait vgroup5
$x = 0
step35:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(4)[4] == leader then
$leaderExist = 1
endi
if $data(4)[6] == leader then
$leaderExist = 1
endi
if $data(4)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step35
endi
print =============== step36: create table
sql use d1
sql create table d1.st (ts timestamp, i int) tags (j int)
sql create table d1.c1 using st tags(1)
@ -57,21 +179,6 @@ if $rows != 1 then
return -1
endi
sql show d1.vgroups
print $data[0][0] $data[0][1] $data[0][2] $data[0][3] $data[0][4]
if $rows != 1 then
return -1
endi
if $data(2)[3] != 2 then
return -1
endi
if $data(2)[5] != 3 then
return -1
endi
if $data(2)[7] != 4 then
return -1
endi
print =============== step4: drop dnode 2
system sh/exec.sh -n dnode5 -s start
$x = 0
@ -85,7 +192,7 @@ step4:
sql show dnodes
print ===> $data00 $data01 $data02 $data03 $data04 $data05
print ===> $data10 $data11 $data12 $data13 $data14 $data15
if $rows != 3 then
if $rows != 5 then
return -1
endi
if $data(1)[4] != ready then
@ -115,21 +222,133 @@ if $rows != 4 then
return -1
endi
print show d1.vgroups
print =============== step62 wait vgroup2
$x = 0
step62:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print $data[0][0] $data[0][1] $data[0][2] $data[0][3] $data[0][4]
if $rows != 1 then
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(2)[4] == leader then
$leaderExist = 1
endi
if $data(2)[6] == leader then
$leaderExist = 1
endi
if $data(2)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step62
endi
print =============== step63 wait vgroup3
$x = 0
step63:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(3)[4] == leader then
$leaderExist = 1
endi
if $data(3)[6] == leader then
$leaderExist = 1
endi
if $data(3)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step63
endi
print =============== step6: select data
print =============== step64 wait vgroup4
$x = 0
step64:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(4)[4] == leader then
$leaderExist = 1
endi
if $data(4)[6] == leader then
$leaderExist = 1
endi
if $data(4)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step64
endi
print =============== step35 wait vgroup5
$x = 0
step65:
$x = $x + 1
sleep 1000
if $x == 60 then
print ====> db not ready!
return -1
endi
sql show d1.vgroups
print ===> $data00 $data01 $data02 $data03 $data04 $data05 $data06 $data07 $data08 $data09
print ===> $data10 $data11 $data12 $data13 $data14 $data15 $data16 $data17 $data18 $data19
print ===> $data20 $data21 $data22 $data23 $data24 $data25 $data26 $data27 $data28 $data29
print ===> $data30 $data31 $data32 $data33 $data34 $data35 $data36 $data37 $data38 $data39
if $rows != 4 then
return -1
endi
if $data(4)[4] == leader then
$leaderExist = 1
endi
if $data(4)[6] == leader then
$leaderExist = 1
endi
if $data(4)[8] == leader then
$leaderExist = 1
endi
if $leaderExist != 1 then
goto step65
endi
print =============== step7: select data
print ===> $data00 $data01
sql show d1.tables
if $rows != 1 then
return -1
endi
return
system sh/exec.sh -n dnode1 -s stop -x SIGINT
system sh/exec.sh -n dnode2 -s stop -x SIGINT
system sh/exec.sh -n dnode3 -s stop -x SIGINT

Some files were not shown because too many files have changed in this diff Show More