Build has been canceled!

#5096

Committed 17 May 2026 01:15AM UTC coverage: 73.276% (-0.006%) from 73.282%

Build # #5096

Build Type

push

travis-ci

Committed by

web-flow

Commit Message

feat (TDgpt): Dynamic Model Synchronization Enhancements (#35344)

* refactor: do some internal refactor.

* fix: fix multiprocess sync issue.

* feat: add dynamic anomaly detection and forecasting services

* fix: log error message for undeploying model in exception handling

* Potential fix for pull request finding

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

* Potential fix for pull request finding

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

* Potential fix for pull request finding

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

* Potential fix for pull request finding

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

* fix: handle undeploy when model exists only on disk

Agent-Logs-Url: https://github.com/taosdata/TDengine/sessions/286aafa0-c3ce-4c27-b803-2707571e9dc1

Co-authored-by: hjxilinx <8252296+hjxilinx@users.noreply.github.com>

* fix: guard dynamic registry concurrent access

Agent-Logs-Url: https://github.com/taosdata/TDengine/sessions/5e4db858-6458-40f4-ac28-d1b1b7f97c18

Co-authored-by: hjxilinx <8252296+hjxilinx@users.noreply.github.com>

* fix: tighten service list locking scope

Agent-Logs-Url: https://github.com/taosdata/TDengine/sessions/5e4db858-6458-40f4-ac28-d1b1b7f97c18

Co-authored-by: hjxilinx <8252296+hjxilinx@users.noreply.github.com>

* fix: restore prophet support and update tests per review feedback

Agent-Logs-Url: https://github.com/taosdata/TDengine/sessions/92298ae1-7da6-4d07-b20e-101c7cd0b26b

Co-authored-by: hjxilinx <8252296+hjxilinx@users.noreply.github.com>

* fix: improve test name and move copy inside lock scope

Agent-Logs-Url: https://github.com/taosdata/TDengine/sessions/92298ae1-7da6-4d07-b20e-101c7cd0b26b

Co-authored-by: hjxilinx <8252296+hjxilinx@users.noreply.github.com>

* Potential fix for pull request finding

Co-au... (continued)

Coverage Stats

281228 of 383795 relevant lines covered (73.28%)

132339998.33 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

79.25

/source/libs/sync/src/syncTimeout.c

/*
 * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
 *
 * This program is free software: you can use, redistribute, and/or modify
 * it under the terms of the GNU Affero General Public License, version 3
 * or later ("AGPL"), as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
 * FITNESS FOR A PARTICULAR PURPOSE.
 *
 * You should have received a copy of the GNU Affero General Public License
 * along with this program. If not, see <http://www.gnu.org/licenses/>.
 */

#define _DEFAULT_SOURCE
#include "syncTimeout.h"
#include "syncElection.h"
#include "syncRaftCfg.h"
#include "syncRaftLog.h"
#include "syncReplication.h"
#include "syncRespMgr.h"
#include "syncSnapshot.h"
#include "syncUtil.h"

static void syncNodeCleanConfigIndex(SSyncNode* ths) {
#if 0
  int32_t   newArrIndex = 0;
  SyncIndex newConfigIndexArr[MAX_CONFIG_INDEX_COUNT] = {0};
  SSnapshot snapshot = {0};

  ths->pFsm->FpGetSnapshotInfo(ths->pFsm, &snapshot);
  if (snapshot.lastApplyIndex != SYNC_INDEX_INVALID) {
    for (int32_t i = 0; i < ths->raftCfg.configIndexCount; ++i) {
      if (ths->raftCfg.configIndexArr[i] < snapshot.lastConfigIndex) {
        // pass
      } else {
        // save
        newConfigIndexArr[newArrIndex] = ths->raftCfg.configIndexArr[i];
        ++newArrIndex;
      }
    }

    int32_t oldCnt = ths->raftCfg.configIndexCount;
    ths->raftCfg.configIndexCount = newArrIndex;
    memcpy(ths->raftCfg.configIndexArr, newConfigIndexArr, sizeof(newConfigIndexArr));

    int32_t code = syncWriteCfgFile(ths);
    if (code != 0) {
      sNFatal(ths, "failed to persist cfg");
    } else {
      sNTrace(ths, "clean config index arr, old-cnt:%d, new-cnt:%d", oldCnt, ths->raftCfg.configIndexCount);
    }
  }
#endif
}

static int32_t syncNodeTimerRoutine(SSyncNode* ths) {
  ths->tmrRoutineNum++;

  sDebug("vgId:%d, timer routine, status report", ths->vgId);
  if (ths->tmrRoutineNum % (tsRoutineReportInterval / (ths->pingTimerMS / 1000)) == 0) {
    sNInfo(ths, "timer routines");
  } else {
    sNTrace(ths, "timer routines");
  }

  // timer replicate
  sDebug("vgId:%d, timer routine, node replicate", ths->vgId);
  int32_t ret = 0;
  if((ret = syncNodeReplicate(ths)) != 0){
    sWarn("vgId:%d, failed to replicate node since %s", ths->vgId, tstrerror(ret));
  }

  // clean mnode index
  sDebug("vgId:%d, timer routine, clean config index", ths->vgId);
  if (syncNodeIsMnode(ths)) {
    syncNodeCleanConfigIndex(ths);
  }

  sDebug("vgId:%d, timer routine, snapshot resend", ths->vgId);
  int64_t timeNow = taosGetTimestampMs();

  for (int i = 0; i < ths->peersNum; ++i) {
    SSyncSnapshotSender* pSender = syncNodeGetSnapshotSender(ths, &(ths->peersId[i]));
    if (pSender != NULL) {
      if (ths->isStart && (ths->state == TAOS_SYNC_STATE_LEADER || ths->state == TAOS_SYNC_STATE_ASSIGNED_LEADER) &&
          pSender->start) {
        int64_t elapsedMs = timeNow - pSender->lastSendTime;
        if (elapsedMs < SYNC_SNAP_RESEND_MS) {
          continue;
        }

        if (elapsedMs > SYNC_SNAP_TIMEOUT_MS) {
          sSError(pSender, "snap replication timeout, terminate.");
          snapshotSenderStop(pSender, false);
        } else {
          sSWarn(pSender, "snap replication resend.");
          if((ret = snapshotReSend(pSender)) != 0){
            sWarn("vgId:%d, failed to snapshot resend in timer routine since %s", ths->vgId, tstrerror(ret));
          }
        }
      }
    }
  }

  sDebug("vgId:%d, timer routine, resp clean", ths->vgId);
  if (!syncNodeIsMnode(ths)) {
    syncRespClean(ths->pSyncRespMgr);
  }

  return 0;
}

int32_t syncNodeOnTimeout(SSyncNode* ths, const SRpcMsg* pRpc) {
  int32_t      ret = 0;
  SyncTimeout* pMsg = pRpc->pCont;

  syncLogRecvTimer(ths, pMsg, &pRpc->info.traceId);

  if (pMsg->timeoutType == SYNC_TIMEOUT_PING) {
    if (atomic_load_64(&ths->pingTimerLogicClockUser) <= pMsg->logicClock) {
      ++(ths->pingTimerCounter);

      TAOS_CHECK_RETURN(syncNodeTimerRoutine(ths));
    }

  } else if (pMsg->timeoutType == SYNC_TIMEOUT_ELECTION) {
    if (atomic_load_64(&ths->electTimerLogicClock) <= pMsg->logicClock) {
      ++(ths->electTimerCounter);

      TAOS_CHECK_RETURN(syncNodeElect(ths));
    }

  } else if (pMsg->timeoutType == SYNC_TIMEOUT_HEARTBEAT) {
    if (atomic_load_64(&ths->heartbeatTimerLogicClockUser) <= pMsg->logicClock) {
      ++(ths->heartbeatTimerCounter);
      sTrace("vgId:%d, sync timer, type:replicate count:%" PRIu64 ", lc-user:%" PRIu64, ths->vgId,
             ths->heartbeatTimerCounter, ths->heartbeatTimerLogicClockUser);
    }

  } else {
    sError("vgId:%d, recv unknown timer-type:%d", ths->vgId, pMsg->timeoutType);
  }

  return ret;
}

1	/*
2	* Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
3	*
4	* This program is free software: you can use, redistribute, and/or modify
5	* it under the terms of the GNU Affero General Public License, version 3
6	* or later ("AGPL"), as published by the Free Software Foundation.
7	*
8	* This program is distributed in the hope that it will be useful, but WITHOUT
9	* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
10	* FITNESS FOR A PARTICULAR PURPOSE.
11	*
12	* You should have received a copy of the GNU Affero General Public License
13	* along with this program. If not, see <http://www.gnu.org/licenses/>.
14	*/
15
16	#define _DEFAULT_SOURCE
17	#include "syncTimeout.h"
18	#include "syncElection.h"
19	#include "syncRaftCfg.h"
20	#include "syncRaftLog.h"
21	#include "syncReplication.h"
22	#include "syncRespMgr.h"
23	#include "syncSnapshot.h"
24	#include "syncUtil.h"
25
26	static void syncNodeCleanConfigIndex(SSyncNode* ths) {	8,037,184✔
27	#if 0
28	int32_t newArrIndex = 0;
29	SyncIndex newConfigIndexArr[MAX_CONFIG_INDEX_COUNT] = {0};
30	SSnapshot snapshot = {0};
31
32	ths->pFsm->FpGetSnapshotInfo(ths->pFsm, &snapshot);
33	if (snapshot.lastApplyIndex != SYNC_INDEX_INVALID) {
34	for (int32_t i = 0; i < ths->raftCfg.configIndexCount; ++i) {
35	if (ths->raftCfg.configIndexArr[i] < snapshot.lastConfigIndex) {
36	// pass
37	} else {
38	// save
39	newConfigIndexArr[newArrIndex] = ths->raftCfg.configIndexArr[i];
40	++newArrIndex;
41	}
42	}
43
44	int32_t oldCnt = ths->raftCfg.configIndexCount;
45	ths->raftCfg.configIndexCount = newArrIndex;
46	memcpy(ths->raftCfg.configIndexArr, newConfigIndexArr, sizeof(newConfigIndexArr));
47
48	int32_t code = syncWriteCfgFile(ths);
49	if (code != 0) {
50	sNFatal(ths, "failed to persist cfg");
51	} else {
52	sNTrace(ths, "clean config index arr, old-cnt:%d, new-cnt:%d", oldCnt, ths->raftCfg.configIndexCount);
53	}
54	}
55	#endif
56	}	8,037,184✔
57
58	static int32_t syncNodeTimerRoutine(SSyncNode* ths) {	44,724,225✔
59	ths->tmrRoutineNum++;	44,724,225✔
60
61	sDebug("vgId:%d, timer routine, status report", ths->vgId);	44,725,398✔
62	if (ths->tmrRoutineNum % (tsRoutineReportInterval / (ths->pingTimerMS / 1000)) == 0) {	44,725,398✔
63	sNInfo(ths, "timer routines");	229,156✔
64	} else {
65	sNTrace(ths, "timer routines");	44,486,311✔
66	}
67
68	// timer replicate
69	sDebug("vgId:%d, timer routine, node replicate", ths->vgId);	44,715,467✔
70	int32_t ret = 0;	44,715,467✔
71	if((ret = syncNodeReplicate(ths)) != 0){	44,715,467✔
72	sWarn("vgId:%d, failed to replicate node since %s", ths->vgId, tstrerror(ret));	45,147✔
73	}
74
75	// clean mnode index
76	sDebug("vgId:%d, timer routine, clean config index", ths->vgId);	44,719,419✔
77	if (syncNodeIsMnode(ths)) {	44,719,419✔
78	syncNodeCleanConfigIndex(ths);	8,037,184✔
79	}
80
81	sDebug("vgId:%d, timer routine, snapshot resend", ths->vgId);	44,720,381✔
82	int64_t timeNow = taosGetTimestampMs();	44,714,871✔
83
84	for (int i = 0; i < ths->peersNum; ++i) {	65,826,915✔
85	SSyncSnapshotSender* pSender = syncNodeGetSnapshotSender(ths, &(ths->peersId[i]));	21,112,291✔
86	if (pSender != NULL) {	21,112,774✔
87	if (ths->isStart && (ths->state == TAOS_SYNC_STATE_LEADER \|\| ths->state == TAOS_SYNC_STATE_ASSIGNED_LEADER) &&	21,112,774✔
88	pSender->start) {	6,678,995✔
89	int64_t elapsedMs = timeNow - pSender->lastSendTime;	278✔
90	if (elapsedMs < SYNC_SNAP_RESEND_MS) {	278✔
91	continue;	278✔
92	}
93
94	if (elapsedMs > SYNC_SNAP_TIMEOUT_MS) {	×
95	sSError(pSender, "snap replication timeout, terminate.");	×
96	snapshotSenderStop(pSender, false);	×
97	} else {
98	sSWarn(pSender, "snap replication resend.");	×
99	if((ret = snapshotReSend(pSender)) != 0){	×
100	sWarn("vgId:%d, failed to snapshot resend in timer routine since %s", ths->vgId, tstrerror(ret));	×
101	}
102	}
103	}
104	}
105	}
106
107	sDebug("vgId:%d, timer routine, resp clean", ths->vgId);	44,718,599✔
108	if (!syncNodeIsMnode(ths)) {	44,718,599✔
109	syncRespClean(ths->pSyncRespMgr);	36,676,191✔
110	}
111
112	return 0;	44,718,863✔
113	}
114
115	int32_t syncNodeOnTimeout(SSyncNode* ths, const SRpcMsg* pRpc) {	45,341,070✔
116	int32_t ret = 0;	45,341,070✔
117	SyncTimeout* pMsg = pRpc->pCont;	45,341,070✔
118
119	syncLogRecvTimer(ths, pMsg, &pRpc->info.traceId);	45,343,202✔
120
121	if (pMsg->timeoutType == SYNC_TIMEOUT_PING) {	45,331,693✔
122	if (atomic_load_64(&ths->pingTimerLogicClockUser) <= pMsg->logicClock) {	44,726,409✔
123	++(ths->pingTimerCounter);	44,725,382✔
124
125	TAOS_CHECK_RETURN(syncNodeTimerRoutine(ths));	44,725,000✔
126	}
127
128	} else if (pMsg->timeoutType == SYNC_TIMEOUT_ELECTION) {	614,609✔
129	if (atomic_load_64(&ths->electTimerLogicClock) <= pMsg->logicClock) {	614,609✔
130	++(ths->electTimerCounter);	612,274✔
131
132	TAOS_CHECK_RETURN(syncNodeElect(ths));	612,274✔
133	}
134
135	} else if (pMsg->timeoutType == SYNC_TIMEOUT_HEARTBEAT) {	×
136	if (atomic_load_64(&ths->heartbeatTimerLogicClockUser) <= pMsg->logicClock) {	×
137	++(ths->heartbeatTimerCounter);	×
138	sTrace("vgId:%d, sync timer, type:replicate count:%" PRIu64 ", lc-user:%" PRIu64, ths->vgId,	×
139	ths->heartbeatTimerCounter, ths->heartbeatTimerLogicClockUser);
140	}
141
142	} else {
143	sError("vgId:%d, recv unknown timer-type:%d", ths->vgId, pMsg->timeoutType);	×
144	}
145
146	return ret;	45,321,743✔
147	}

taosdata / TDengine / #5096

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous