未验证 提交 7467a54a 编写于 作者: L Liu Jicong 提交者: GitHub

Merge pull request #17396 from taosdata/fix/client_mem_leak

fix: client mem leak
...@@ -709,6 +709,7 @@ void tmqSendHbReq(void* param, void* tmrId) { ...@@ -709,6 +709,7 @@ void tmqSendHbReq(void* param, void* tmrId) {
int64_t refId = *(int64_t*)param; int64_t refId = *(int64_t*)param;
tmq_t* tmq = taosAcquireRef(tmqMgmt.rsetId, refId); tmq_t* tmq = taosAcquireRef(tmqMgmt.rsetId, refId);
if (tmq == NULL) { if (tmq == NULL) {
taosMemoryFree(param);
return; return;
} }
int64_t consumerId = tmq->consumerId; int64_t consumerId = tmq->consumerId;
...@@ -939,10 +940,9 @@ tmq_t* tmq_consumer_new(tmq_conf_t* conf, char* errstr, int32_t errstrLen) { ...@@ -939,10 +940,9 @@ tmq_t* tmq_consumer_new(tmq_conf_t* conf, char* errstr, int32_t errstrLen) {
return NULL; return NULL;
} }
int64_t* pRefId = taosMemoryMalloc(sizeof(int64_t));
*pRefId = pTmq->refId;
if (pTmq->hbBgEnable) { if (pTmq->hbBgEnable) {
int64_t* pRefId = taosMemoryMalloc(sizeof(int64_t));
*pRefId = pTmq->refId;
pTmq->hbLiveTimer = taosTmrStart(tmqSendHbReq, 1000, pRefId, tmqMgmt.timer); pTmq->hbLiveTimer = taosTmrStart(tmqSendHbReq, 1000, pRefId, tmqMgmt.timer);
} }
......
...@@ -115,7 +115,6 @@ typedef struct { ...@@ -115,7 +115,6 @@ typedef struct {
typedef struct { typedef struct {
SMqDataRsp dataRsp; SMqDataRsp dataRsp;
SMqRspHead rspHead;
char subKey[TSDB_SUBSCRIBE_KEY_LEN]; char subKey[TSDB_SUBSCRIBE_KEY_LEN];
SRpcHandleInfo pInfo; SRpcHandleInfo pInfo;
} STqPushEntry; } STqPushEntry;
...@@ -183,6 +182,7 @@ int32_t tqOffsetCommitFile(STqOffsetStore* pStore); ...@@ -183,6 +182,7 @@ int32_t tqOffsetCommitFile(STqOffsetStore* pStore);
// tqSink // tqSink
void tqTableSink(SStreamTask* pTask, void* vnode, int64_t ver, void* data); void tqTableSink(SStreamTask* pTask, void* vnode, int64_t ver, void* data);
void tqTableSink1(SStreamTask* pTask, void* vnode, int64_t ver, void* data);
// tqOffset // tqOffset
char* tqOffsetBuildFName(const char* path, int32_t ver); char* tqOffsetBuildFName(const char* path, int32_t ver);
......
...@@ -192,7 +192,7 @@ int32_t tqPushDataRsp(STQ* pTq, STqPushEntry* pPushEntry) { ...@@ -192,7 +192,7 @@ int32_t tqPushDataRsp(STQ* pTq, STqPushEntry* pPushEntry) {
return -1; return -1;
} }
memcpy(buf, &pPushEntry->rspHead, sizeof(SMqRspHead)); memcpy(buf, &pPushEntry->dataRsp.head, sizeof(SMqRspHead));
void* abuf = POINTER_SHIFT(buf, sizeof(SMqRspHead)); void* abuf = POINTER_SHIFT(buf, sizeof(SMqRspHead));
...@@ -215,7 +215,7 @@ int32_t tqPushDataRsp(STQ* pTq, STqPushEntry* pPushEntry) { ...@@ -215,7 +215,7 @@ int32_t tqPushDataRsp(STQ* pTq, STqPushEntry* pPushEntry) {
tFormatOffset(buf1, 80, &pRsp->reqOffset); tFormatOffset(buf1, 80, &pRsp->reqOffset);
tFormatOffset(buf2, 80, &pRsp->rspOffset); tFormatOffset(buf2, 80, &pRsp->rspOffset);
tqDebug("vgId:%d, from consumer:%" PRId64 ", (epoch %d) push rsp, block num: %d, reqOffset:%s, rspOffset:%s", tqDebug("vgId:%d, from consumer:%" PRId64 ", (epoch %d) push rsp, block num: %d, reqOffset:%s, rspOffset:%s",
TD_VID(pTq->pVnode), pPushEntry->rspHead.consumerId, pRsp->head.epoch, pRsp->blockNum, buf1, buf2); TD_VID(pTq->pVnode), pRsp->head.consumerId, pRsp->head.epoch, pRsp->blockNum, buf1, buf2);
return 0; return 0;
} }
...@@ -560,9 +560,9 @@ int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg) { ...@@ -560,9 +560,9 @@ int32_t tqProcessPollReq(STQ* pTq, SRpcMsg* pMsg) {
memcpy(pPushEntry->subKey, pHandle->subKey, TSDB_SUBSCRIBE_KEY_LEN); memcpy(pPushEntry->subKey, pHandle->subKey, TSDB_SUBSCRIBE_KEY_LEN);
dataRsp.withTbName = 0; dataRsp.withTbName = 0;
memcpy(&pPushEntry->dataRsp, &dataRsp, sizeof(SMqDataRsp)); memcpy(&pPushEntry->dataRsp, &dataRsp, sizeof(SMqDataRsp));
pPushEntry->rspHead.consumerId = consumerId; pPushEntry->dataRsp.head.consumerId = consumerId;
pPushEntry->rspHead.epoch = reqEpoch; pPushEntry->dataRsp.head.epoch = reqEpoch;
pPushEntry->rspHead.mqMsgType = TMQ_MSG_TYPE__POLL_RSP; pPushEntry->dataRsp.head.mqMsgType = TMQ_MSG_TYPE__POLL_RSP;
taosHashPut(pTq->pPushMgr, pHandle->subKey, strlen(pHandle->subKey) + 1, &pPushEntry, sizeof(void*)); taosHashPut(pTq->pPushMgr, pHandle->subKey, strlen(pHandle->subKey) + 1, &pPushEntry, sizeof(void*));
tqDebug("tmq poll: consumer %ld, subkey %s, vg %d save handle to push mgr", consumerId, pHandle->subKey, tqDebug("tmq poll: consumer %ld, subkey %s, vg %d save handle to push mgr", consumerId, pHandle->subKey,
TD_VID(pTq->pVnode)); TD_VID(pTq->pVnode));
...@@ -924,7 +924,7 @@ int32_t tqExpandTask(STQ* pTq, SStreamTask* pTask) { ...@@ -924,7 +924,7 @@ int32_t tqExpandTask(STQ* pTq, SStreamTask* pTask) {
pTask->smaSink.smaSink = smaHandleRes; pTask->smaSink.smaSink = smaHandleRes;
} else if (pTask->outputType == TASK_OUTPUT__TABLE) { } else if (pTask->outputType == TASK_OUTPUT__TABLE) {
pTask->tbSink.vnode = pTq->pVnode; pTask->tbSink.vnode = pTq->pVnode;
pTask->tbSink.tbSinkFunc = tqTableSink; pTask->tbSink.tbSinkFunc = tqTableSink1;
ASSERT(pTask->tbSink.pSchemaWrapper); ASSERT(pTask->tbSink.pSchemaWrapper);
ASSERT(pTask->tbSink.pSchemaWrapper->pSchema); ASSERT(pTask->tbSink.pSchemaWrapper->pSchema);
......
...@@ -284,6 +284,212 @@ SSubmitReq* tqBlockToSubmit(SVnode* pVnode, const SArray* pBlocks, const STSchem ...@@ -284,6 +284,212 @@ SSubmitReq* tqBlockToSubmit(SVnode* pVnode, const SArray* pBlocks, const STSchem
return ret; return ret;
} }
void tqTableSink1(SStreamTask* pTask, void* vnode, int64_t ver, void* data) {
const SArray* pBlocks = (const SArray*)data;
SVnode* pVnode = (SVnode*)vnode;
int64_t suid = pTask->tbSink.stbUid;
char* stbFullName = pTask->tbSink.stbFullName;
STSchema* pTSchema = pTask->tbSink.pTSchema;
SSchemaWrapper* pSchemaWrapper = pTask->tbSink.pSchemaWrapper;
int32_t blockSz = taosArrayGetSize(pBlocks);
bool createTb = true;
SArray* tagArray = taosArrayInit(1, sizeof(STagVal));
if (!tagArray) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
return;
}
tqDebug("vgId:%d, task %d write into table, block num: %d", TD_VID(pVnode), pTask->taskId, blockSz);
for (int32_t i = 0; i < blockSz; i++) {
SSDataBlock* pDataBlock = taosArrayGet(pBlocks, i);
if (pDataBlock->info.type == STREAM_DELETE_RESULT) {
SBatchDeleteReq deleteReq = {0};
deleteReq.deleteReqs = taosArrayInit(0, sizeof(SSingleDeleteReq));
deleteReq.suid = suid;
tqBuildDeleteReq(pVnode, stbFullName, pDataBlock, &deleteReq);
int32_t len;
int32_t code;
tEncodeSize(tEncodeSBatchDeleteReq, &deleteReq, len, code);
if (code < 0) {
//
ASSERT(0);
}
SEncoder encoder;
void* serializedDeleteReq = rpcMallocCont(len + sizeof(SMsgHead));
void* abuf = POINTER_SHIFT(serializedDeleteReq, sizeof(SMsgHead));
tEncoderInit(&encoder, abuf, len);
tEncodeSBatchDeleteReq(&encoder, &deleteReq);
tEncoderClear(&encoder);
taosArrayDestroy(deleteReq.deleteReqs);
((SMsgHead*)serializedDeleteReq)->vgId = pVnode->config.vgId;
SRpcMsg msg = {
.msgType = TDMT_VND_BATCH_DEL,
.pCont = serializedDeleteReq,
.contLen = len + sizeof(SMsgHead),
};
if (tmsgPutToQueue(&pVnode->msgCb, WRITE_QUEUE, &msg) != 0) {
rpcFreeCont(serializedDeleteReq);
tqDebug("failed to put delete req into write-queue since %s", terrstr());
}
} else {
SVCreateTbReq createTbReq = {0};
// set const
createTbReq.flags = 0;
createTbReq.type = TSDB_CHILD_TABLE;
createTbReq.ctb.suid = suid;
// set super table name
SName name = {0};
tNameFromString(&name, stbFullName, T_NAME_ACCT | T_NAME_DB | T_NAME_TABLE);
createTbReq.ctb.stbName = strdup((char*)tNameGetTableName(&name)); // strdup(stbFullName);
// set tag content
taosArrayClear(tagArray);
STagVal tagVal = {
.cid = taosArrayGetSize(pDataBlock->pDataBlock) + 1,
.type = TSDB_DATA_TYPE_UBIGINT,
.i64 = (int64_t)pDataBlock->info.groupId,
};
taosArrayPush(tagArray, &tagVal);
createTbReq.ctb.tagNum = taosArrayGetSize(tagArray);
STag* pTag = NULL;
tTagNew(tagArray, 1, false, &pTag);
if (pTag == NULL) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
taosArrayDestroy(tagArray);
tdDestroySVCreateTbReq(&createTbReq);
return;
}
createTbReq.ctb.pTag = (uint8_t*)pTag;
// set tag name
SArray* tagName = taosArrayInit(1, TSDB_COL_NAME_LEN);
char tagNameStr[TSDB_COL_NAME_LEN] = {0};
strcpy(tagNameStr, "group_id");
taosArrayPush(tagName, tagNameStr);
createTbReq.ctb.tagName = tagName;
// set table name
if (pDataBlock->info.parTbName[0]) {
createTbReq.name = strdup(pDataBlock->info.parTbName);
} else {
createTbReq.name = buildCtbNameByGroupId(stbFullName, pDataBlock->info.groupId);
}
int32_t schemaLen;
int32_t code;
tEncodeSize(tEncodeSVCreateTbReq, &createTbReq, schemaLen, code);
if (code < 0) {
tdDestroySVCreateTbReq(&createTbReq);
taosArrayDestroy(tagArray);
return;
}
// save schema str
void* schemaStr = taosMemoryMalloc(schemaLen);
if (schemaStr == NULL) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
tdDestroySVCreateTbReq(&createTbReq);
taosArrayDestroy(tagArray);
return;
}
SEncoder encoder = {0};
tEncoderInit(&encoder, schemaStr, schemaLen);
code = tEncodeSVCreateTbReq(&encoder, &createTbReq);
if (code < 0) {
terrno = TSDB_CODE_OUT_OF_MEMORY;
tdDestroySVCreateTbReq(&createTbReq);
taosArrayDestroy(tagArray);
tEncoderClear(&encoder);
taosMemoryFree(schemaStr);
return;
}
tEncoderClear(&encoder);
tdDestroySVCreateTbReq(&createTbReq);
int32_t cap = sizeof(SSubmitReq);
int32_t rows = pDataBlock->info.rows;
int32_t maxLen = TD_ROW_MAX_BYTES_FROM_SCHEMA(pTSchema);
cap += sizeof(SSubmitBlk) + schemaLen + rows * maxLen;
SSubmitReq* ret = rpcMallocCont(cap);
ret->header.vgId = pVnode->config.vgId;
ret->length = sizeof(SSubmitReq);
ret->numOfBlocks = htonl(1);
SSubmitBlk* blkHead = POINTER_SHIFT(ret, sizeof(SSubmitReq));
blkHead->numOfRows = htonl(pDataBlock->info.rows);
blkHead->sversion = htonl(pTSchema->version);
blkHead->suid = htobe64(suid);
// uid is assigned by vnode
blkHead->uid = 0;
blkHead->schemaLen = 0;
tqDebug("tq sink, convert block %d, rows: %d", i, rows);
int32_t dataLen = 0;
void* blkSchema = POINTER_SHIFT(blkHead, sizeof(SSubmitBlk));
STSRow* rowData = blkSchema;
if (createTb) {
memcpy(blkSchema, schemaStr, schemaLen);
blkHead->schemaLen = htonl(schemaLen);
rowData = POINTER_SHIFT(blkSchema, schemaLen);
}
taosMemoryFree(schemaStr);
for (int32_t j = 0; j < rows; j++) {
SRowBuilder rb = {0};
tdSRowInit(&rb, pTSchema->version);
tdSRowSetTpInfo(&rb, pTSchema->numOfCols, pTSchema->flen);
tdSRowResetBuf(&rb, rowData);
for (int32_t k = 0; k < pTSchema->numOfCols; k++) {
const STColumn* pColumn = &pTSchema->columns[k];
SColumnInfoData* pColData = taosArrayGet(pDataBlock->pDataBlock, k);
if (colDataIsNull_s(pColData, j)) {
tdAppendColValToRow(&rb, pColumn->colId, pColumn->type, TD_VTYPE_NULL, NULL, false, pColumn->offset, k);
} else {
void* colData = colDataGetData(pColData, j);
tdAppendColValToRow(&rb, pColumn->colId, pColumn->type, TD_VTYPE_NORM, colData, true, pColumn->offset, k);
}
}
tdSRowEnd(&rb);
int32_t rowLen = TD_ROW_LEN(rowData);
rowData = POINTER_SHIFT(rowData, rowLen);
dataLen += rowLen;
}
blkHead->dataLen = htonl(dataLen);
ret->length += sizeof(SSubmitBlk) + schemaLen + dataLen;
ret->length = htonl(ret->length);
SRpcMsg msg = {
.msgType = TDMT_VND_SUBMIT,
.pCont = ret,
.contLen = ntohl(ret->length),
};
if (tmsgPutToQueue(&pVnode->msgCb, WRITE_QUEUE, &msg) != 0) {
rpcFreeCont(ret);
tqDebug("failed to put into write-queue since %s", terrstr());
}
}
}
taosArrayDestroy(tagArray);
}
void tqTableSink(SStreamTask* pTask, void* vnode, int64_t ver, void* data) { void tqTableSink(SStreamTask* pTask, void* vnode, int64_t ver, void* data) {
const SArray* pRes = (const SArray*)data; const SArray* pRes = (const SArray*)data;
SVnode* pVnode = (SVnode*)vnode; SVnode* pVnode = (SVnode*)vnode;
......
...@@ -52,7 +52,7 @@ class TDTestCase: ...@@ -52,7 +52,7 @@ class TDTestCase:
paraDict['rowsPerTbl'] = self.rowsPerTbl paraDict['rowsPerTbl'] = self.rowsPerTbl
tmqCom.initConsumerTable() tmqCom.initConsumerTable()
tdCom.create_database(tdSql, paraDict["dbName"],paraDict["dropFlag"], vgroups=paraDict["vgroups"],replica=1) tdCom.create_database(tdSql, paraDict["dbName"],paraDict["dropFlag"], vgroups=paraDict["vgroups"],replica=1, wal_retention_size=-1,wal_retention_period=-1)
tdLog.info("create stb") tdLog.info("create stb")
tmqCom.create_stable(tdSql, dbName=paraDict["dbName"],stbName=paraDict["stbName"]) tmqCom.create_stable(tdSql, dbName=paraDict["dbName"],stbName=paraDict["stbName"])
# tdLog.info("create ctb") # tdLog.info("create ctb")
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册