diff --git a/include/util/taoserror.h b/include/util/taoserror.h index 842c8a44e17abb55b1796cca5c7090cab89fbadc..450eb432802f5296bb30f97132417bdcbea2c624 100644 --- a/include/util/taoserror.h +++ b/include/util/taoserror.h @@ -185,7 +185,7 @@ int32_t* taosGetErrno(); #define TSDB_CODE_MND_BNODE_NOT_EXIST TAOS_DEF_ERROR_CODE(0, 0x0357) #define TSDB_CODE_MND_TOO_FEW_MNODES TAOS_DEF_ERROR_CODE(0, 0x0358) #define TSDB_CODE_MND_TOO_MANY_MNODES TAOS_DEF_ERROR_CODE(0, 0x0359) -#define TSDB_CODE_MND_CANT_DROP_MASTER TAOS_DEF_ERROR_CODE(0, 0x035A) +#define TSDB_CODE_MND_CANT_DROP_LEADER TAOS_DEF_ERROR_CODE(0, 0x035A) // mnode-acct #define TSDB_CODE_MND_ACCT_ALREADY_EXIST TAOS_DEF_ERROR_CODE(0, 0x0360) diff --git a/source/dnode/mgmt/node_mgmt/src/dmTransport.c b/source/dnode/mgmt/node_mgmt/src/dmTransport.c index 8b939d15ce35e4a8ccb9f032025250476a3b8a82..de7d46ed83b863c22e0d6eca7904115445648c4e 100644 --- a/source/dnode/mgmt/node_mgmt/src/dmTransport.c +++ b/source/dnode/mgmt/node_mgmt/src/dmTransport.c @@ -131,7 +131,8 @@ static void dmProcessRpcMsg(SDnode *pDnode, SRpcMsg *pRpc, SEpSet *pEpSet) { _OVER: if (code != 0) { - dTrace("msg:%p, failed to process since %s, type:%s", pMsg, terrstr(), TMSG_INFO(pRpc->msgType)); + dTrace("failed to process msg:%p since %s, handle:%p", pMsg, terrstr(), pRpc->info.handle); + if (terrno != 0) code = terrno; if (IsReq(pRpc)) { @@ -149,8 +150,10 @@ _OVER: } } - dTrace("msg:%p, is freed", pMsg); - taosFreeQitem(pMsg); + if (pMsg != NULL) { + dTrace("msg:%p, is freed", pMsg); + taosFreeQitem(pMsg); + } rpcFreeCont(pRpc->pCont); } diff --git a/source/dnode/mgmt/node_util/src/dmEps.c b/source/dnode/mgmt/node_util/src/dmEps.c index 937c1ab7faa29191ea65f8770d0ffb2c531b3c35..6f6896ff9f11ae040e765e613e0cd2f7732a599e 100644 --- a/source/dnode/mgmt/node_util/src/dmEps.c +++ b/source/dnode/mgmt/node_util/src/dmEps.c @@ -318,9 +318,9 @@ void dmGetMnodeEpSet(SDnodeData *pData, SEpSet *pEpSet) { void dmGetMnodeEpSetForRedirect(SDnodeData *pData, SRpcMsg *pMsg, SEpSet *pEpSet) { dmGetMnodeEpSet(pData, pEpSet); - dDebug("msg:%p, is redirected, num:%d use:%d", pMsg, pEpSet->numOfEps, pEpSet->inUse); + dTrace("msg is redirected, handle:%p num:%d use:%d", pMsg->info.handle, pEpSet->numOfEps, pEpSet->inUse); for (int32_t i = 0; i < pEpSet->numOfEps; ++i) { - dDebug("mnode index:%d %s:%u", i, pEpSet->eps[i].fqdn, pEpSet->eps[i].port); + dTrace("mnode index:%d %s:%u", i, pEpSet->eps[i].fqdn, pEpSet->eps[i].port); if (strcmp(pEpSet->eps[i].fqdn, tsLocalFqdn) == 0 && pEpSet->eps[i].port == tsServerPort) { pEpSet->inUse = (i + 1) % pEpSet->numOfEps; } diff --git a/source/dnode/mnode/impl/inc/mndTrans.h b/source/dnode/mnode/impl/inc/mndTrans.h index 736fa3b36108e2e840b1f24a25c302262f83e6ca..5ac9d2233f376c4d0435464e6f3032ad105db613 100644 --- a/source/dnode/mnode/impl/inc/mndTrans.h +++ b/source/dnode/mnode/impl/inc/mndTrans.h @@ -36,19 +36,19 @@ typedef enum { } ETrnAct; typedef struct { - int32_t id; - int32_t errCode; - int32_t acceptableCode; - int8_t stage; - ETrnAct actionType; - int8_t rawWritten; - int8_t msgSent; - int8_t msgReceived; - tmsg_t msgType; - SEpSet epSet; - int32_t contLen; - void *pCont; - SSdbRaw *pRaw; + int32_t id; + int32_t errCode; + int32_t acceptableCode; + ETrnStage stage; + ETrnAct actionType; + int8_t rawWritten; + int8_t msgSent; + int8_t msgReceived; + tmsg_t msgType; + SEpSet epSet; + int32_t contLen; + void *pCont; + SSdbRaw *pRaw; } STransAction; typedef void (*TransCbFp)(SMnode *pMnode, void *param, int32_t paramLen); diff --git a/source/dnode/mnode/impl/src/mndDnode.c b/source/dnode/mnode/impl/src/mndDnode.c index c7e347fb8b9573423139565e92bd5e8d59ebf1cd..7be49393372f6d58ac9ce3e1205a52bec07889be 100644 --- a/source/dnode/mnode/impl/src/mndDnode.c +++ b/source/dnode/mnode/impl/src/mndDnode.c @@ -87,6 +87,10 @@ int32_t mndInitDnode(SMnode *pMnode) { void mndCleanupDnode(SMnode *pMnode) {} static int32_t mndCreateDefaultDnode(SMnode *pMnode) { + int32_t code = -1; + SSdbRaw *pRaw = NULL; + STrans *pTrans = NULL; + SDnodeObj dnodeObj = {0}; dnodeObj.id = 1; dnodeObj.createdTime = taosGetTimestampMs(); @@ -95,54 +99,42 @@ static int32_t mndCreateDefaultDnode(SMnode *pMnode) { memcpy(&dnodeObj.fqdn, pMnode->replicas[0].fqdn, TSDB_FQDN_LEN); snprintf(dnodeObj.ep, TSDB_EP_LEN, "%s:%u", dnodeObj.fqdn, dnodeObj.port); - SSdbRaw *pRaw = mndDnodeActionEncode(&dnodeObj); - if (pRaw == NULL) return -1; - if (sdbSetRawStatus(pRaw, SDB_STATUS_READY) != 0) return -1; - - mDebug("dnode:%d, will be created when deploying, raw:%p", dnodeObj.id, pRaw); - - STrans *pTrans = mndTransCreate(pMnode, TRN_POLICY_RETRY, TRN_CONFLICT_GLOBAL, NULL); - if (pTrans == NULL) { - mError("dnode:%s, failed to create since %s", dnodeObj.ep, terrstr()); - return -1; - } - mDebug("trans:%d, used to create dnode:%s", pTrans->id, dnodeObj.ep); + pTrans = mndTransCreate(pMnode, TRN_POLICY_RETRY, TRN_CONFLICT_GLOBAL, NULL); + if (pTrans == NULL) goto _OVER; + mDebug("trans:%d, used to create dnode:%s on first deploy", pTrans->id, dnodeObj.ep); - if (mndTransAppendCommitlog(pTrans, pRaw) != 0) { - mError("trans:%d, failed to append commit log since %s", pTrans->id, terrstr()); - mndTransDrop(pTrans); - return -1; - } + pRaw = mndDnodeActionEncode(&dnodeObj); + if (pRaw == NULL || mndTransAppendCommitlog(pTrans, pRaw) != 0) goto _OVER; sdbSetRawStatus(pRaw, SDB_STATUS_READY); + pRaw = NULL; - if (mndTransPrepare(pMnode, pTrans) != 0) { - mError("trans:%d, failed to prepare since %s", pTrans->id, terrstr()); - mndTransDrop(pTrans); - return -1; - } + if (mndTransPrepare(pMnode, pTrans) != 0) goto _OVER; + code = 0; +_OVER: mndTransDrop(pTrans); - return 0; + sdbFreeRaw(pRaw); + return code; } static SSdbRaw *mndDnodeActionEncode(SDnodeObj *pDnode) { terrno = TSDB_CODE_OUT_OF_MEMORY; SSdbRaw *pRaw = sdbAllocRaw(SDB_DNODE, TSDB_DNODE_VER_NUMBER, sizeof(SDnodeObj) + TSDB_DNODE_RESERVE_SIZE); - if (pRaw == NULL) goto DNODE_ENCODE_OVER; + if (pRaw == NULL) goto _OVER; int32_t dataPos = 0; - SDB_SET_INT32(pRaw, dataPos, pDnode->id, DNODE_ENCODE_OVER) - SDB_SET_INT64(pRaw, dataPos, pDnode->createdTime, DNODE_ENCODE_OVER) - SDB_SET_INT64(pRaw, dataPos, pDnode->updateTime, DNODE_ENCODE_OVER) - SDB_SET_INT16(pRaw, dataPos, pDnode->port, DNODE_ENCODE_OVER) - SDB_SET_BINARY(pRaw, dataPos, pDnode->fqdn, TSDB_FQDN_LEN, DNODE_ENCODE_OVER) - SDB_SET_RESERVE(pRaw, dataPos, TSDB_DNODE_RESERVE_SIZE, DNODE_ENCODE_OVER) - SDB_SET_DATALEN(pRaw, dataPos, DNODE_ENCODE_OVER); + SDB_SET_INT32(pRaw, dataPos, pDnode->id, _OVER) + SDB_SET_INT64(pRaw, dataPos, pDnode->createdTime, _OVER) + SDB_SET_INT64(pRaw, dataPos, pDnode->updateTime, _OVER) + SDB_SET_INT16(pRaw, dataPos, pDnode->port, _OVER) + SDB_SET_BINARY(pRaw, dataPos, pDnode->fqdn, TSDB_FQDN_LEN, _OVER) + SDB_SET_RESERVE(pRaw, dataPos, TSDB_DNODE_RESERVE_SIZE, _OVER) + SDB_SET_DATALEN(pRaw, dataPos, _OVER); terrno = 0; -DNODE_ENCODE_OVER: +_OVER: if (terrno != 0) { mError("dnode:%d, failed to encode to raw:%p since %s", pDnode->id, pRaw, terrstr()); sdbFreeRaw(pRaw); @@ -154,33 +146,32 @@ DNODE_ENCODE_OVER: } static SSdbRow *mndDnodeActionDecode(SSdbRaw *pRaw) { + SSdbRow *pRow = NULL; terrno = TSDB_CODE_OUT_OF_MEMORY; int8_t sver = 0; - if (sdbGetRawSoftVer(pRaw, &sver) != 0) goto DNODE_DECODE_OVER; - + if (sdbGetRawSoftVer(pRaw, &sver) != 0) goto _OVER; if (sver != TSDB_DNODE_VER_NUMBER) { terrno = TSDB_CODE_SDB_INVALID_DATA_VER; - goto DNODE_DECODE_OVER; + goto _OVER; } - SSdbRow *pRow = sdbAllocRow(sizeof(SDnodeObj)); - if (pRow == NULL) goto DNODE_DECODE_OVER; - + pRow = sdbAllocRow(sizeof(SDnodeObj)); + if (pRow == NULL) goto _OVER; SDnodeObj *pDnode = sdbGetRowObj(pRow); - if (pDnode == NULL) goto DNODE_DECODE_OVER; + if (pDnode == NULL) goto _OVER; int32_t dataPos = 0; - SDB_GET_INT32(pRaw, dataPos, &pDnode->id, DNODE_DECODE_OVER) - SDB_GET_INT64(pRaw, dataPos, &pDnode->createdTime, DNODE_DECODE_OVER) - SDB_GET_INT64(pRaw, dataPos, &pDnode->updateTime, DNODE_DECODE_OVER) - SDB_GET_INT16(pRaw, dataPos, &pDnode->port, DNODE_DECODE_OVER) - SDB_GET_BINARY(pRaw, dataPos, pDnode->fqdn, TSDB_FQDN_LEN, DNODE_DECODE_OVER) - SDB_GET_RESERVE(pRaw, dataPos, TSDB_DNODE_RESERVE_SIZE, DNODE_DECODE_OVER) + SDB_GET_INT32(pRaw, dataPos, &pDnode->id, _OVER) + SDB_GET_INT64(pRaw, dataPos, &pDnode->createdTime, _OVER) + SDB_GET_INT64(pRaw, dataPos, &pDnode->updateTime, _OVER) + SDB_GET_INT16(pRaw, dataPos, &pDnode->port, _OVER) + SDB_GET_BINARY(pRaw, dataPos, pDnode->fqdn, TSDB_FQDN_LEN, _OVER) + SDB_GET_RESERVE(pRaw, dataPos, TSDB_DNODE_RESERVE_SIZE, _OVER) terrno = 0; -DNODE_DECODE_OVER: +_OVER: if (terrno != 0) { mError("dnode:%d, failed to decode from raw:%p since %s", pDnode->id, pRaw, terrstr()); taosMemoryFreeClear(pRow); @@ -246,6 +237,7 @@ static SDnodeObj *mndAcquireDnodeByEp(SMnode *pMnode, char *pEpStr) { sdbRelease(pSdb, pDnode); } + terrno = TSDB_CODE_MND_DNODE_NOT_EXIST; return NULL; } @@ -290,25 +282,26 @@ static void mndGetDnodeData(SMnode *pMnode, SArray *pDnodeEps) { } } -static int32_t mndCheckClusterCfgPara(SMnode *pMnode, const SClusterCfg *pCfg) { +static int32_t mndCheckClusterCfgPara(SMnode *pMnode, SDnodeObj *pDnode, const SClusterCfg *pCfg) { if (pCfg->statusInterval != tsStatusInterval) { - mError("statusInterval [%d - %d] cfg inconsistent", pCfg->statusInterval, tsStatusInterval); + mError("dnode:%d, statusInterval:%d inconsistent with cluster:%d", pDnode->id, pCfg->statusInterval, + tsStatusInterval); return DND_REASON_STATUS_INTERVAL_NOT_MATCH; } if ((0 != strcasecmp(pCfg->timezone, tsTimezoneStr)) && (pMnode->checkTime != pCfg->checkTime)) { - mError("timezone [%s - %s] [%" PRId64 " - %" PRId64 "] cfg inconsistent", pCfg->timezone, tsTimezoneStr, - pCfg->checkTime, pMnode->checkTime); + mError("dnode:%d, timezone:%s checkTime:%" PRId64 " inconsistent with cluster %s %" PRId64, pDnode->id, + pCfg->timezone, pCfg->checkTime, tsTimezoneStr, pMnode->checkTime); return DND_REASON_TIME_ZONE_NOT_MATCH; } if (0 != strcasecmp(pCfg->locale, tsLocale)) { - mError("locale [%s - %s] cfg inconsistent", pCfg->locale, tsLocale); + mError("dnode:%d, locale:%s inconsistent with cluster:%s", pDnode->id, pCfg->locale, tsLocale); return DND_REASON_LOCALE_NOT_MATCH; } if (0 != strcasecmp(pCfg->charset, tsCharset)) { - mError("charset [%s - %s] cfg inconsistent.", pCfg->charset, tsCharset); + mError("dnode:%d, charset:%s inconsistent with cluster:%s", pDnode->id, pCfg->charset, tsCharset); return DND_REASON_CHARSET_NOT_MATCH; } @@ -323,15 +316,14 @@ static int32_t mndProcessStatusReq(SRpcMsg *pReq) { if (tDeserializeSStatusReq(pReq->pCont, pReq->contLen, &statusReq) != 0) { terrno = TSDB_CODE_INVALID_MSG; - goto PROCESS_STATUS_MSG_OVER; + goto _OVER; } if (statusReq.dnodeId == 0) { pDnode = mndAcquireDnodeByEp(pMnode, statusReq.dnodeEp); if (pDnode == NULL) { mDebug("dnode:%s, not created yet", statusReq.dnodeEp); - terrno = TSDB_CODE_MND_DNODE_NOT_EXIST; - goto PROCESS_STATUS_MSG_OVER; + goto _OVER; } } else { pDnode = mndAcquireDnode(pMnode, statusReq.dnodeId); @@ -341,13 +333,11 @@ static int32_t mndProcessStatusReq(SRpcMsg *pReq) { pDnode->offlineReason = DND_REASON_DNODE_ID_NOT_MATCH; } mError("dnode:%d, %s not exist", statusReq.dnodeId, statusReq.dnodeEp); - terrno = TSDB_CODE_MND_DNODE_NOT_EXIST; - goto PROCESS_STATUS_MSG_OVER; + goto _OVER; } } - int32_t numOfVloads = (int32_t)taosArrayGetSize(statusReq.pVloads); - for (int32_t v = 0; v < numOfVloads; ++v) { + for (int32_t v = 0; v < taosArrayGetSize(statusReq.pVloads); ++v) { SVnodeLoad *pVload = taosArrayGet(statusReq.pVloads, v); SVgObj *pVgroup = mndAcquireVgroup(pMnode, pVload->vgId); @@ -366,6 +356,7 @@ static int32_t mndProcessStatusReq(SRpcMsg *pReq) { roleChanged = true; } pVgroup->vnodeGid[vg].role = pVload->syncState; + break; } } if (roleChanged) { @@ -405,11 +396,11 @@ static int32_t mndProcessStatusReq(SRpcMsg *pReq) { } mError("dnode:%d, status msg version:%d not match cluster:%d", statusReq.dnodeId, statusReq.sver, tsVersion); terrno = TSDB_CODE_VERSION_NOT_COMPATIBLE; - goto PROCESS_STATUS_MSG_OVER; + goto _OVER; } if (statusReq.dnodeId == 0) { - mDebug("dnode:%d, %s first access, set clusterId %" PRId64, pDnode->id, pDnode->ep, pMnode->clusterId); + mInfo("dnode:%d, %s first access, set clusterId %" PRId64, pDnode->id, pDnode->ep, pMnode->clusterId); } else { if (statusReq.clusterId != pMnode->clusterId) { if (pDnode != NULL) { @@ -418,7 +409,7 @@ static int32_t mndProcessStatusReq(SRpcMsg *pReq) { mError("dnode:%d, clusterId %" PRId64 " not match exist %" PRId64, pDnode->id, statusReq.clusterId, pMnode->clusterId); terrno = TSDB_CODE_MND_INVALID_CLUSTER_ID; - goto PROCESS_STATUS_MSG_OVER; + goto _OVER; } else { pDnode->accessTimes++; mTrace("dnode:%d, status received, access times %d", pDnode->id, pDnode->accessTimes); @@ -426,18 +417,17 @@ static int32_t mndProcessStatusReq(SRpcMsg *pReq) { } // Verify whether the cluster parameters are consistent when status change from offline to ready - int32_t ret = mndCheckClusterCfgPara(pMnode, &statusReq.clusterCfg); - if (0 != ret) { - pDnode->offlineReason = ret; - mError("dnode:%d, cluster cfg inconsistent since:%s", pDnode->id, offlineReason[ret]); + pDnode->offlineReason = mndCheckClusterCfgPara(pMnode, pDnode, &statusReq.clusterCfg); + if (pDnode->offlineReason != 0) { + mError("dnode:%d, cluster cfg inconsistent since:%s", pDnode->id, offlineReason[pDnode->offlineReason]); terrno = TSDB_CODE_MND_INVALID_CLUSTER_CFG; - goto PROCESS_STATUS_MSG_OVER; + goto _OVER; } if (!online) { mInfo("dnode:%d, from offline to online", pDnode->id); } else { - mDebug("dnode:%d, send dnode epset, online:%d ver:% " PRId64 ":%" PRId64 " reboot:%d", pDnode->id, online, + mDebug("dnode:%d, send dnode epset, online:%d dnode_ver:%" PRId64 ":%" PRId64 " reboot:%d", pDnode->id, online, statusReq.dnodeVer, dnodeVer, reboot); } @@ -452,7 +442,7 @@ static int32_t mndProcessStatusReq(SRpcMsg *pReq) { statusRsp.pDnodeEps = taosArrayInit(mndGetDnodeSize(pMnode), sizeof(SDnodeEp)); if (statusRsp.pDnodeEps == NULL) { terrno = TSDB_CODE_OUT_OF_MEMORY; - goto PROCESS_STATUS_MSG_OVER; + goto _OVER; } mndGetDnodeData(pMnode, statusRsp.pDnodeEps); @@ -469,13 +459,17 @@ static int32_t mndProcessStatusReq(SRpcMsg *pReq) { pDnode->lastAccessTime = curMs; code = 0; -PROCESS_STATUS_MSG_OVER: +_OVER: mndReleaseDnode(pMnode, pDnode); taosArrayDestroy(statusReq.pVloads); return code; } static int32_t mndCreateDnode(SMnode *pMnode, SRpcMsg *pReq, SCreateDnodeReq *pCreate) { + int32_t code = -1; + SSdbRaw *pRaw = NULL; + STrans *pTrans = NULL; + SDnodeObj dnodeObj = {0}; dnodeObj.id = sdbGetMaxId(pMnode->pSdb, SDB_DNODE); dnodeObj.createdTime = taosGetTimestampMs(); @@ -484,29 +478,22 @@ static int32_t mndCreateDnode(SMnode *pMnode, SRpcMsg *pReq, SCreateDnodeReq *pC memcpy(dnodeObj.fqdn, pCreate->fqdn, TSDB_FQDN_LEN); snprintf(dnodeObj.ep, TSDB_EP_LEN, "%s:%u", dnodeObj.fqdn, dnodeObj.port); - STrans *pTrans = mndTransCreate(pMnode, TRN_POLICY_ROLLBACK, TRN_CONFLICT_GLOBAL, pReq); - if (pTrans == NULL) { - mError("dnode:%s, failed to create since %s", dnodeObj.ep, terrstr()); - return -1; - } + pTrans = mndTransCreate(pMnode, TRN_POLICY_ROLLBACK, TRN_CONFLICT_GLOBAL, pReq); + if (pTrans == NULL) goto _OVER; mDebug("trans:%d, used to create dnode:%s", pTrans->id, dnodeObj.ep); - SSdbRaw *pCommitRaw = mndDnodeActionEncode(&dnodeObj); - if (pCommitRaw == NULL || mndTransAppendCommitlog(pTrans, pCommitRaw) != 0) { - mError("trans:%d, failed to append commit log since %s", pTrans->id, terrstr()); - mndTransDrop(pTrans); - return -1; - } - sdbSetRawStatus(pCommitRaw, SDB_STATUS_READY); + pRaw = mndDnodeActionEncode(&dnodeObj); + if (pRaw == NULL || mndTransAppendCommitlog(pTrans, pRaw) != 0) goto _OVER; + sdbSetRawStatus(pRaw, SDB_STATUS_READY); + pRaw = NULL; - if (mndTransPrepare(pMnode, pTrans) != 0) { - mError("trans:%d, failed to prepare since %s", pTrans->id, terrstr()); - mndTransDrop(pTrans); - return -1; - } + if (mndTransPrepare(pMnode, pTrans) != 0) goto _OVER; + code = 0; +_OVER: mndTransDrop(pTrans); - return 0; + sdbFreeRaw(pRaw); + return code; } static int32_t mndProcessCreateDnodeReq(SRpcMsg *pReq) { @@ -518,38 +505,37 @@ static int32_t mndProcessCreateDnodeReq(SRpcMsg *pReq) { if (tDeserializeSCreateDnodeReq(pReq->pCont, pReq->contLen, &createReq) != 0) { terrno = TSDB_CODE_INVALID_MSG; - goto CREATE_DNODE_OVER; + goto _OVER; } mDebug("dnode:%s:%d, start to create", createReq.fqdn, createReq.port); if (createReq.fqdn[0] == 0 || createReq.port <= 0 || createReq.port > UINT16_MAX) { terrno = TSDB_CODE_MND_INVALID_DNODE_EP; - goto CREATE_DNODE_OVER; + goto _OVER; } char ep[TSDB_EP_LEN]; snprintf(ep, TSDB_EP_LEN, "%s:%d", createReq.fqdn, createReq.port); pDnode = mndAcquireDnodeByEp(pMnode, ep); if (pDnode != NULL) { - terrno = TSDB_CODE_MND_DNODE_ALREADY_EXIST; - goto CREATE_DNODE_OVER; + goto _OVER; } pUser = mndAcquireUser(pMnode, pReq->conn.user); if (pUser == NULL) { terrno = TSDB_CODE_MND_NO_USER_FROM_CONN; - goto CREATE_DNODE_OVER; + goto _OVER; } if (mndCheckNodeAuth(pUser) != 0) { - goto CREATE_DNODE_OVER; + goto _OVER; } code = mndCreateDnode(pMnode, pReq, &createReq); if (code == 0) code = TSDB_CODE_ACTION_IN_PROGRESS; -CREATE_DNODE_OVER: +_OVER: if (code != 0 && code != TSDB_CODE_ACTION_IN_PROGRESS) { mError("dnode:%s:%d, failed to create since %s", createReq.fqdn, createReq.port, terrstr()); } @@ -559,7 +545,7 @@ CREATE_DNODE_OVER: return code; } -static int32_t mndDropDnode(SMnode *pMnode, SRpcMsg *pReq, SDnodeObj *pDnode, SMnodeObj *pMObj) { +static int32_t mndDropDnode(SMnode *pMnode, SRpcMsg *pReq, SDnodeObj *pDnode, SMnodeObj *pMObj, int32_t numOfVnodes) { int32_t code = -1; SSdbRaw *pRaw = NULL; STrans *pTrans = NULL; @@ -579,8 +565,12 @@ static int32_t mndDropDnode(SMnode *pMnode, SRpcMsg *pReq, SDnodeObj *pDnode, SM sdbSetRawStatus(pRaw, SDB_STATUS_DROPPED); pRaw = NULL; - if (mndSetDropMnodeInfoToTrans(pMnode, pTrans, pMObj) != 0) goto _OVER; - if (mndSetMoveVgroupsInfoToTrans(pMnode, pTrans, pDnode->id) != 0) goto _OVER; + if (pMObj != NULL) { + if (mndSetDropMnodeInfoToTrans(pMnode, pTrans, pMObj) != 0) goto _OVER; + } + if (numOfVnodes > 0) { + if (mndSetMoveVgroupsInfoToTrans(pMnode, pTrans, pDnode->id) != 0) goto _OVER; + } if (mndTransPrepare(pMnode, pTrans) != 0) goto _OVER; code = 0; @@ -617,11 +607,6 @@ static int32_t mndProcessDropDnodeReq(SRpcMsg *pReq) { goto _OVER; } - if (!mndIsDnodeOnline(pDnode, taosGetTimestampMs())) { - terrno = TSDB_CODE_NODE_OFFLINE; - goto _OVER; - } - pMObj = mndAcquireMnode(pMnode, dropReq.dnodeId); if (pMObj != NULL) { if (sdbGetSize(pMnode->pSdb, SDB_MNODE) <= 1) { @@ -629,7 +614,17 @@ static int32_t mndProcessDropDnodeReq(SRpcMsg *pReq) { goto _OVER; } if (pMnode->selfDnodeId == dropReq.dnodeId) { - terrno = TSDB_CODE_MND_CANT_DROP_MASTER; + terrno = TSDB_CODE_MND_CANT_DROP_LEADER; + goto _OVER; + } + } + + int32_t numOfVnodes = mndGetVnodesNum(pMnode, pDnode->id); + if (numOfVnodes > 0 || pMObj != NULL) { + if (!mndIsDnodeOnline(pDnode, taosGetTimestampMs())) { + terrno = TSDB_CODE_NODE_OFFLINE; + mError("dnode:%d, failed to drop since %s, has_mnode:%d numOfVnodes:%d", pDnode->id, terrstr(), pMObj != NULL, + numOfVnodes); goto _OVER; } } @@ -644,7 +639,7 @@ static int32_t mndProcessDropDnodeReq(SRpcMsg *pReq) { goto _OVER; } - code = mndDropDnode(pMnode, pReq, pDnode, pMObj); + code = mndDropDnode(pMnode, pReq, pDnode, pMObj, numOfVnodes); if (code == 0) code = TSDB_CODE_ACTION_IN_PROGRESS; _OVER: @@ -669,7 +664,6 @@ static int32_t mndProcessConfigDnodeReq(SRpcMsg *pReq) { SDnodeObj *pDnode = mndAcquireDnode(pMnode, cfgReq.dnodeId); if (pDnode == NULL) { - terrno = TSDB_CODE_MND_DNODE_NOT_EXIST; mError("dnode:%d, failed to config since %s ", cfgReq.dnodeId, terrstr()); return -1; } @@ -679,17 +673,18 @@ static int32_t mndProcessConfigDnodeReq(SRpcMsg *pReq) { int32_t bufLen = tSerializeSMCfgDnodeReq(NULL, 0, &cfgReq); void *pBuf = rpcMallocCont(bufLen); + + if (pBuf == NULL) return -1; tSerializeSMCfgDnodeReq(pBuf, bufLen, &cfgReq); + mDebug("dnode:%d, send config req to dnode, app:%p", cfgReq.dnodeId, pReq->info.ahandle); SRpcMsg rpcMsg = {.msgType = TDMT_DND_CONFIG_DNODE, .pCont = pBuf, .contLen = bufLen, .info = pReq->info}; - - mDebug("dnode:%d, send config req to dnode, app:%p", cfgReq.dnodeId, rpcMsg.info.ahandle); return tmsgSendReq(&epSet, &rpcMsg); } static int32_t mndProcessConfigDnodeRsp(SRpcMsg *pRsp) { mDebug("config rsp from dnode, app:%p", pRsp->info.ahandle); - return TSDB_CODE_SUCCESS; + return 0; } static int32_t mndRetrieveConfigs(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pBlock, int32_t rows) { @@ -698,7 +693,7 @@ static int32_t mndRetrieveConfigs(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *p int32_t numOfRows = 0; char *cfgOpts[TSDB_CONFIG_NUMBER] = {0}; char cfgVals[TSDB_CONFIG_NUMBER][TSDB_CONIIG_VALUE_LEN + 1] = {0}; - char *pWrite; + char *pWrite = NULL; int32_t cols = 0; cfgOpts[totalRows] = "statusInterval"; @@ -724,7 +719,6 @@ static int32_t mndRetrieveConfigs(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *p cols = 0; STR_WITH_MAXSIZE_TO_VARSTR(buf, cfgOpts[i], TSDB_CONFIG_OPTION_LEN); - SColumnInfoData *pColInfo = taosArrayGet(pBlock->pDataBlock, cols++); colDataAppend(pColInfo, numOfRows, (const char *)buf, false); @@ -774,7 +768,6 @@ static int32_t mndRetrieveDnodes(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pB char b1[9] = {0}; STR_TO_VARSTR(b1, online ? "ready" : "offline"); - pColInfo = taosArrayGet(pBlock->pDataBlock, cols++); colDataAppend(pColInfo, numOfRows, b1, false); @@ -792,7 +785,6 @@ static int32_t mndRetrieveDnodes(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pB } pShow->numOfRows += numOfRows; - return numOfRows; } diff --git a/source/dnode/mnode/impl/src/mndMnode.c b/source/dnode/mnode/impl/src/mndMnode.c index 072ad05f5abbc10cdb93ff36a30e493a570d811e..027de66b42ddf4cf6a42fbf5e02e39e50f479106 100644 --- a/source/dnode/mnode/impl/src/mndMnode.c +++ b/source/dnode/mnode/impl/src/mndMnode.c @@ -588,7 +588,7 @@ static int32_t mndProcessDropMnodeReq(SRpcMsg *pReq) { } if (pMnode->selfDnodeId == dropReq.dnodeId) { - terrno = TSDB_CODE_MND_CANT_DROP_MASTER; + terrno = TSDB_CODE_MND_CANT_DROP_LEADER; goto _OVER; } diff --git a/source/dnode/mnode/impl/src/mndSync.c b/source/dnode/mnode/impl/src/mndSync.c index 245f0938b906300af29bf3f6caf71c834877eaa1..ce025d55478a913681229a0ff6d2563512406711 100644 --- a/source/dnode/mnode/impl/src/mndSync.c +++ b/source/dnode/mnode/impl/src/mndSync.c @@ -193,7 +193,7 @@ int32_t mndInitSync(SMnode *pMnode) { void mndCleanupSync(SMnode *pMnode) { SSyncMgmt *pMgmt = &pMnode->syncMgmt; syncStop(pMgmt->sync); - mDebug("sync:%" PRId64 " is stopped", pMgmt->sync); + mDebug("mnode sync is stopped, id:%" PRId64, pMgmt->sync); tsem_destroy(&pMgmt->syncSem); if (pMgmt->pWal != NULL) { diff --git a/source/dnode/mnode/impl/src/mndTrans.c b/source/dnode/mnode/impl/src/mndTrans.c index 8c0ba65fb295758ed3b0bcbd517259b6ca7846ed..c2097c069b10fe2d7dcb2211d0f93cd5cc9048d7 100644 --- a/source/dnode/mnode/impl/src/mndTrans.c +++ b/source/dnode/mnode/impl/src/mndTrans.c @@ -117,10 +117,10 @@ static SSdbRaw *mndTransActionEncode(STrans *pTrans) { int32_t dataPos = 0; SDB_SET_INT32(pRaw, dataPos, pTrans->id, _OVER) - SDB_SET_INT16(pRaw, dataPos, pTrans->stage, _OVER) - SDB_SET_INT16(pRaw, dataPos, pTrans->policy, _OVER) - SDB_SET_INT16(pRaw, dataPos, pTrans->conflict, _OVER) - SDB_SET_INT16(pRaw, dataPos, pTrans->exec, _OVER) + SDB_SET_INT8(pRaw, dataPos, pTrans->stage, _OVER) + SDB_SET_INT8(pRaw, dataPos, pTrans->policy, _OVER) + SDB_SET_INT8(pRaw, dataPos, pTrans->conflict, _OVER) + SDB_SET_INT8(pRaw, dataPos, pTrans->exec, _OVER) SDB_SET_INT64(pRaw, dataPos, pTrans->createdTime, _OVER) SDB_SET_BINARY(pRaw, dataPos, pTrans->dbname, TSDB_DB_FNAME_LEN, _OVER) SDB_SET_INT32(pRaw, dataPos, pTrans->redoActionPos, _OVER) @@ -256,15 +256,15 @@ static SSdbRow *mndTransActionDecode(SSdbRaw *pRaw) { SDB_GET_INT32(pRaw, dataPos, &pTrans->id, _OVER) - int16_t stage = 0; - int16_t policy = 0; - int16_t conflict = 0; - int16_t exec = 0; - int8_t actionType = 0; - SDB_GET_INT16(pRaw, dataPos, &stage, _OVER) - SDB_GET_INT16(pRaw, dataPos, &policy, _OVER) - SDB_GET_INT16(pRaw, dataPos, &conflict, _OVER) - SDB_GET_INT16(pRaw, dataPos, &exec, _OVER) + int8_t stage = 0; + int8_t policy = 0; + int8_t conflict = 0; + int8_t exec = 0; + int8_t actionType = 0; + SDB_GET_INT8(pRaw, dataPos, &stage, _OVER) + SDB_GET_INT8(pRaw, dataPos, &policy, _OVER) + SDB_GET_INT8(pRaw, dataPos, &conflict, _OVER) + SDB_GET_INT8(pRaw, dataPos, &exec, _OVER) pTrans->stage = stage; pTrans->policy = policy; pTrans->conflict = conflict; @@ -290,7 +290,8 @@ static SSdbRow *mndTransActionDecode(SSdbRaw *pRaw) { SDB_GET_INT32(pRaw, dataPos, &action.acceptableCode, _OVER) SDB_GET_INT8(pRaw, dataPos, &actionType, _OVER) action.actionType = actionType; - SDB_GET_INT8(pRaw, dataPos, &action.stage, _OVER) + SDB_GET_INT8(pRaw, dataPos, &stage, _OVER) + action.stage = stage; if (action.actionType == TRANS_ACTION_RAW) { SDB_GET_INT8(pRaw, dataPos, &action.rawWritten, _OVER) SDB_GET_INT32(pRaw, dataPos, &dataLen, _OVER) @@ -322,7 +323,8 @@ static SSdbRow *mndTransActionDecode(SSdbRaw *pRaw) { SDB_GET_INT32(pRaw, dataPos, &action.acceptableCode, _OVER) SDB_GET_INT8(pRaw, dataPos, &actionType, _OVER) action.actionType = actionType; - SDB_GET_INT8(pRaw, dataPos, &action.stage, _OVER) + SDB_GET_INT8(pRaw, dataPos, &stage, _OVER) + action.stage = stage; if (action.actionType == TRANS_ACTION_RAW) { SDB_GET_INT8(pRaw, dataPos, &action.rawWritten, _OVER) SDB_GET_INT32(pRaw, dataPos, &dataLen, _OVER) @@ -354,7 +356,8 @@ static SSdbRow *mndTransActionDecode(SSdbRaw *pRaw) { SDB_GET_INT32(pRaw, dataPos, &action.acceptableCode, _OVER) SDB_GET_INT8(pRaw, dataPos, &actionType, _OVER) action.actionType = actionType; - SDB_GET_INT8(pRaw, dataPos, &action.stage, _OVER) + SDB_GET_INT8(pRaw, dataPos, &stage, _OVER) + action.stage = stage; if (action.actionType) { SDB_GET_INT8(pRaw, dataPos, &action.rawWritten, _OVER) SDB_GET_INT32(pRaw, dataPos, &dataLen, _OVER) @@ -878,7 +881,6 @@ static void mndTransResetActions(SMnode *pMnode, STrans *pTrans, SArray *pArray) mDebug("trans:%d, %s:%d execute status is reset", pTrans->id, mndTransStr(pAction->stage), action); } pAction->errCode = 0; - } } @@ -890,11 +892,12 @@ static int32_t mndTransWriteSingleLog(SMnode *pMnode, STrans *pTrans, STransActi pAction->rawWritten = true; pAction->errCode = 0; code = 0; - mDebug("trans:%d, %s:%d write to sdb", pTrans->id, mndTransStr(pAction->stage), pAction->id); + mDebug("trans:%d, %s:%d write to sdb, type:%s status:%s", pTrans->id, mndTransStr(pAction->stage), pAction->id, + sdbTableName(pAction->pRaw->type), sdbStatusName(pAction->pRaw->status)); } else { pAction->errCode = (terrno != 0) ? terrno : code; - mError("trans:%d, %s:%d failed to write sdb since %s", pTrans->id, mndTransStr(pAction->stage), pAction->id, - terrstr()); + mError("trans:%d, %s:%d failed to write sdb since %s, type:%s status:%s", pTrans->id, mndTransStr(pAction->stage), + pAction->id, terrstr(), sdbTableName(pAction->pRaw->type), sdbStatusName(pAction->pRaw->status)); } return code; @@ -916,18 +919,26 @@ static int32_t mndTransSendSingleMsg(SMnode *pMnode, STrans *pTrans, STransActio } memcpy(rpcMsg.pCont, pAction->pCont, pAction->contLen); + char detail[1024] = {0}; + int32_t len = snprintf(detail, sizeof(detail), "msgType:%s numOfEps:%d inUse:%d", TMSG_INFO(pAction->msgType), + pAction->epSet.numOfEps, pAction->epSet.inUse); + for (int32_t i = 0; i < pTrans->lastErrorEpset.numOfEps; ++i) { + len += snprintf(detail + len, sizeof(detail) - len, " ep:%d-%s:%u", i, pAction->epSet.eps[i].fqdn, + pAction->epSet.eps[i].port); + } + int32_t code = tmsgSendReq(&pAction->epSet, &rpcMsg); if (code == 0) { pAction->msgSent = 1; pAction->msgReceived = 0; pAction->errCode = 0; - mDebug("trans:%d, %s:%d is sent to %s:%u", pTrans->id, mndTransStr(pAction->stage), pAction->id, - pAction->epSet.eps[pAction->epSet.inUse].fqdn, pAction->epSet.eps[pAction->epSet.inUse].port); + mDebug("trans:%d, %s:%d is sent, %s", pTrans->id, mndTransStr(pAction->stage), pAction->id, detail); } else { pAction->msgSent = 0; pAction->msgReceived = 0; pAction->errCode = (terrno != 0) ? terrno : code; - mError("trans:%d, %s:%d not send since %s", pTrans->id, mndTransStr(pAction->stage), pAction->id, terrstr()); + mError("trans:%d, %s:%d not send since %s, %s", pTrans->id, mndTransStr(pAction->stage), pAction->id, terrstr(), + detail); } return code; @@ -1424,9 +1435,9 @@ static int32_t mndRetrieveTrans(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock *pBl if (epset.numOfEps > 0) { len += snprintf(detail + len, sizeof(detail) - len, "msgType:%s numOfEps:%d inUse:%d ", TMSG_INFO(pTrans->lastErrorMsgType), epset.numOfEps, epset.inUse); - } - for (int32_t i = 0; i < pTrans->lastErrorEpset.numOfEps; ++i) { - len += snprintf(detail + len, sizeof(detail) - len, "ep:%d-%s:%u ", i, epset.eps[i].fqdn, epset.eps[i].port); + for (int32_t i = 0; i < pTrans->lastErrorEpset.numOfEps; ++i) { + len += snprintf(detail + len, sizeof(detail) - len, "ep:%d-%s:%u ", i, epset.eps[i].fqdn, epset.eps[i].port); + } } } STR_WITH_MAXSIZE_TO_VARSTR(lastError, detail, pShow->pMeta->pSchemas[cols].bytes); diff --git a/source/dnode/mnode/sdb/inc/sdb.h b/source/dnode/mnode/sdb/inc/sdb.h index 4a00befa1e8ec1b4ef4ff20a51a066ed08cf1883..944454380433ed1ee4944269ba6ef864e33742ed 100644 --- a/source/dnode/mnode/sdb/inc/sdb.h +++ b/source/dnode/mnode/sdb/inc/sdb.h @@ -392,6 +392,7 @@ int32_t sdbStopWrite(SSdb *pSdb, SSdbIter *pIter, bool isApply); int32_t sdbDoWrite(SSdb *pSdb, SSdbIter *pIter, void *pBuf, int32_t len); const char *sdbTableName(ESdbType type); +const char *sdbStatusName(ESdbStatus status); void sdbPrintOper(SSdb *pSdb, SSdbRow *pRow, const char *oper); int32_t sdbGetIdFromRaw(SSdb *pSdb, SSdbRaw *pRaw); diff --git a/source/dnode/mnode/sdb/src/sdbHash.c b/source/dnode/mnode/sdb/src/sdbHash.c index 162da2bd0aaa3e2400f14cefa0596b5022e7afbe..d1b1e3163574efe1152c282455af19ad1ff569f1 100644 --- a/source/dnode/mnode/sdb/src/sdbHash.c +++ b/source/dnode/mnode/sdb/src/sdbHash.c @@ -65,7 +65,7 @@ const char *sdbTableName(ESdbType type) { } } -static const char *sdbStatusName(ESdbStatus status) { +const char *sdbStatusName(ESdbStatus status) { switch (status) { case SDB_STATUS_CREATING: return "creating"; diff --git a/source/util/src/terror.c b/source/util/src/terror.c index c14c279d8151d80a5009f70a1c3ca0b1142e794f..0357ce7b4b1a934dd2a380c0cc2e2e5a2cc6708b 100644 --- a/source/util/src/terror.c +++ b/source/util/src/terror.c @@ -188,9 +188,9 @@ TAOS_DEFINE_ERROR(TSDB_CODE_MND_SNODE_ALREADY_EXIST, "Snode already exists" TAOS_DEFINE_ERROR(TSDB_CODE_MND_SNODE_NOT_EXIST, "Snode not there") TAOS_DEFINE_ERROR(TSDB_CODE_MND_BNODE_ALREADY_EXIST, "Bnode already exists") TAOS_DEFINE_ERROR(TSDB_CODE_MND_BNODE_NOT_EXIST, "Bnode not there") -TAOS_DEFINE_ERROR(TSDB_CODE_MND_TOO_FEW_MNODES, "The replicas of mnode cannot less than 1") -TAOS_DEFINE_ERROR(TSDB_CODE_MND_TOO_MANY_MNODES, "The replicas of mnode cannot exceed 3") -TAOS_DEFINE_ERROR(TSDB_CODE_MND_CANT_DROP_MASTER, "Can't drop mnode which is leader") +TAOS_DEFINE_ERROR(TSDB_CODE_MND_TOO_FEW_MNODES, "The replica of mnode cannot less than 1") +TAOS_DEFINE_ERROR(TSDB_CODE_MND_TOO_MANY_MNODES, "The replica of mnode cannot exceed 3") +TAOS_DEFINE_ERROR(TSDB_CODE_MND_CANT_DROP_LEADER, "Cannot drop mnode which is leader") // mnode-acct TAOS_DEFINE_ERROR(TSDB_CODE_MND_ACCT_ALREADY_EXIST, "Account already exists") diff --git a/tests/script/jenkins/basic.txt b/tests/script/jenkins/basic.txt index 32056e8d0ceb2d128b25b86e24ce9d42dfa6ba15..3249c02e8869e51a738914fb531a9e315d3e7c41 100644 --- a/tests/script/jenkins/basic.txt +++ b/tests/script/jenkins/basic.txt @@ -21,7 +21,7 @@ ./test.sh -f tsim/db/taosdlog.sim # ---- dnode -./test.sh -f tsim/dnode/basic1.sim +./test.sh -f tsim/dnode/create_dnode.sim # ---- insert ./test.sh -f tsim/insert/basic0.sim diff --git a/tests/script/tsim/dnode/basic1.sim b/tests/script/tsim/dnode/create_dnode.sim similarity index 100% rename from tests/script/tsim/dnode/basic1.sim rename to tests/script/tsim/dnode/create_dnode.sim