Merge pull request #12883 from taosdata/feature/async.scheduler

feat: scheduler refactor

source/libs/scheduler/inc/schedulerInt.h

@@ -264,7 +264,7 @@ int32_t schBuildAndSendMsg(SSchJob *job, SSchTask *task, SQueryNodeAddr *addr, i
 SSchJob *schAcquireJob(int64_t refId);
 int32_t schReleaseJob(int64_t refId);
 void schFreeFlowCtrl(SSchJob *pJob);
-int32_t schCheckJobNeedFlowCtrl(SSchJob *pJob, SSchLevel *pLevel);
+int32_t schChkJobNeedFlowCtrl(SSchJob *pJob, SSchLevel *pLevel);
 int32_t schDecTaskFlowQuota(SSchJob *pJob, SSchTask *pTask);
 int32_t schCheckIncTaskFlowQuota(SSchJob *pJob, SSchTask *pTask, bool *enough);
 int32_t schLaunchTasksInFlowCtrlList(SSchJob *pJob, SSchTask *pTask);
@@ -275,6 +275,32 @@ int32_t schBuildAndSendHbMsg(SQueryNodeEpId *nodeEpId);
 int32_t schCloneSMsgSendInfo(void *src, void **dst);
 int32_t schValidateAndBuildJob(SQueryPlan *pDag, SSchJob *pJob);
 void schFreeJobImpl(void *job);
+int32_t schMakeHbCallbackParam(SSchJob *pJob, SSchTask *pTask, void **pParam);
+int32_t schMakeHbRpcCtx(SSchJob *pJob, SSchTask *pTask, SRpcCtx *pCtx);
+int32_t schEnsureHbConnection(SSchJob *pJob, SSchTask *pTask);
+int32_t schUpdateHbConnection(SQueryNodeEpId *epId, SSchTrans *trans);
+int32_t schHandleHbCallback(void *param, const SDataBuf *pMsg, int32_t code);
+void schFreeRpcCtx(SRpcCtx *pCtx);
+int32_t schGetCallbackFp(int32_t msgType, __async_send_cb_fn_t *fp);
+bool schJobNeedToStop(SSchJob *pJob, int8_t *pStatus);
+int32_t schProcessOnTaskSuccess(SSchJob *pJob, SSchTask *pTask);
+int32_t schSaveJobQueryRes(SSchJob *pJob, SResReadyRsp *rsp);
+int32_t schProcessOnExplainDone(SSchJob *pJob, SSchTask *pTask, SRetrieveTableRsp *pRsp);
+void schProcessOnDataFetched(SSchJob *job);
+int32_t schGetTaskFromTaskList(SHashObj *pTaskList, uint64_t taskId, SSchTask **pTask);
+int32_t schUpdateTaskExecNodeHandle(SSchTask *pTask, void *handle, int32_t rspCode);
+void schFreeRpcCtxVal(const void *arg);
+int32_t schMakeBrokenLinkVal(SSchJob *pJob, SSchTask *pTask, SRpcBrokenlinkVal *brokenVal, bool isHb);
+int32_t schRecordTaskExecNode(SSchJob *pJob, SSchTask *pTask, SQueryNodeAddr *addr, void *handle);
+int32_t schExecStaticExplain(void *transport, SArray *pNodeList, SQueryPlan *pDag, int64_t *job, const char *sql,
+                             bool syncSchedule);
+int32_t schExecJobImpl(void *transport, SArray *pNodeList, SQueryPlan *pDag, int64_t *job, const char *sql,
+                              int64_t startTs, bool sync);
+int32_t schChkUpdateJobStatus(SSchJob *pJob, int8_t newStatus);
+int32_t schCancelJob(SSchJob *pJob);
+int32_t schProcessOnJobDropped(SSchJob *pJob, int32_t errCode);
+uint64_t schGenTaskId(void);
+void schCloseJobRef(void);
 
 
 #ifdef __cplusplus

source/libs/scheduler/src/schFlowCtrl.c

@@ -40,7 +40,7 @@ void schFreeFlowCtrl(SSchJob *pJob) {
   pJob->flowCtrl = NULL;
 }
 
-int32_t schCheckJobNeedFlowCtrl(SSchJob *pJob, SSchLevel *pLevel) {
+int32_t schChkJobNeedFlowCtrl(SSchJob *pJob, SSchLevel *pLevel) {
   if (!SCH_IS_QUERY_JOB(pJob)) {
     SCH_JOB_DLOG("job no need flow ctrl, queryJob:%d", SCH_IS_QUERY_JOB(pJob));
     return TSDB_CODE_SUCCESS;

source/libs/scheduler/src/schUtil.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/>.
+ */
+
+#include "catalog.h"
+#include "command.h"
+#include "query.h"
+#include "schedulerInt.h"
+#include "tmsg.h"
+#include "tref.h"
+#include "trpc.h"
+
+void schCloseJobRef(void) {
+  if (!atomic_load_8((int8_t *)&schMgmt.exit)) {
+    return;
+  }
+
+  SCH_LOCK(SCH_WRITE, &schMgmt.lock);
+  if (atomic_load_32(&schMgmt.jobNum) <= 0 && schMgmt.jobRef >= 0) {
+    taosCloseRef(schMgmt.jobRef);
+    schMgmt.jobRef = -1;
+  }
+  SCH_UNLOCK(SCH_WRITE, &schMgmt.lock);
+}
+
+uint64_t schGenTaskId(void) { return atomic_add_fetch_64(&schMgmt.taskId, 1); }
+
+uint64_t schGenUUID(void) {
+  static uint64_t hashId = 0;
+  static int32_t requestSerialId = 0;
+
+  if (hashId == 0) {
+    char    uid[64];
+    int32_t code = taosGetSystemUUID(uid, tListLen(uid));
+    if (code != TSDB_CODE_SUCCESS) {
+      qError("Failed to get the system uid, reason:%s", tstrerror(TAOS_SYSTEM_ERROR(errno)));
+    } else {
+      hashId = MurmurHash3_32(uid, strlen(uid));
+    }
+  }
+
+  int64_t ts      = taosGetTimestampMs();
+  uint64_t pid    = taosGetPId();
+  int32_t val     = atomic_add_fetch_32(&requestSerialId, 1);
+
+  uint64_t id = ((hashId & 0x0FFF) << 52) | ((pid & 0x0FFF) << 40) | ((ts & 0xFFFFFF) << 16) | (val & 0xFFFF);
+  return id;
+}
+
+
+void schFreeRpcCtxVal(const void *arg) {
+  if (NULL == arg) {
+    return;
+  }
+
+  SMsgSendInfo *pMsgSendInfo = (SMsgSendInfo *)arg;
+  taosMemoryFreeClear(pMsgSendInfo->param);
+  taosMemoryFreeClear(pMsgSendInfo);
+}
+
+void schFreeRpcCtx(SRpcCtx *pCtx) {
+  if (NULL == pCtx) {
+    return;
+  }
+  void *pIter = taosHashIterate(pCtx->args, NULL);
+  while (pIter) {
+    SRpcCtxVal *ctxVal = (SRpcCtxVal *)pIter;
+
+    (*ctxVal->freeFunc)(ctxVal->val);
+
+    pIter = taosHashIterate(pCtx->args, pIter);
+  }
+
+  taosHashCleanup(pCtx->args);
+
+  if (pCtx->brokenVal.freeFunc) {
+    (*pCtx->brokenVal.freeFunc)(pCtx->brokenVal.val);
+  }
+}
+
+