github.com/matrixorigin/matrixone@v0.7.0/pkg/logservice/service_bootstrap.go (about)

     1  // Copyright 2021 - 2022 Matrix Origin
     2  //
     3  // Licensed under the Apache License, Version 2.0 (the "License");
     4  // you may not use this file except in compliance with the License.
     5  // You may obtain a copy of the License at
     6  //
     7  //      http://www.apache.org/licenses/LICENSE-2.0
     8  //
     9  // Unless required by applicable law or agreed to in writing, software
    10  // distributed under the License is distributed on an "AS IS" BASIS,
    11  // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    12  // See the License for the specific language governing permissions and
    13  // limitations under the License.
    14  
    15  package logservice
    16  
    17  import (
    18  	"context"
    19  	"time"
    20  
    21  	"github.com/lni/dragonboat/v4"
    22  	"github.com/matrixorigin/matrixone/pkg/common/runtime"
    23  	"github.com/matrixorigin/matrixone/pkg/pb/task"
    24  	"go.uber.org/zap"
    25  )
    26  
    27  func (s *Service) BootstrapHAKeeper(ctx context.Context, cfg Config) error {
    28  	replicaID, bootstrapping := cfg.Bootstrapping()
    29  	if !bootstrapping {
    30  		return nil
    31  	}
    32  	members, err := cfg.GetInitHAKeeperMembers()
    33  	if err != nil {
    34  		return err
    35  	}
    36  	if err := s.store.startHAKeeperReplica(replicaID, members, false); err != nil {
    37  		// let's be a little less strict, when HAKeeper replica is already
    38  		// running as a result of store.startReplicas(), we just ignore the
    39  		// dragonboat.ErrShardAlreadyExist error below.
    40  		if err != dragonboat.ErrShardAlreadyExist {
    41  			s.runtime.SubLogger(runtime.SystemInit).Error("failed to start hakeeper replica", zap.Error(err))
    42  			return err
    43  		}
    44  	}
    45  	numOfLogShards := cfg.BootstrapConfig.NumOfLogShards
    46  	numOfDNShards := cfg.BootstrapConfig.NumOfDNShards
    47  	numOfLogReplicas := cfg.BootstrapConfig.NumOfLogShardReplicas
    48  	for i := 0; i < checkBootstrapCycles; i++ {
    49  		select {
    50  		case <-ctx.Done():
    51  			return nil
    52  		default:
    53  		}
    54  		if err := s.store.setInitialClusterInfo(numOfLogShards,
    55  			numOfDNShards, numOfLogReplicas); err != nil {
    56  			s.runtime.SubLogger(runtime.SystemInit).Error("failed to set initial cluster info", zap.Error(err))
    57  			if err == dragonboat.ErrShardNotFound {
    58  				return nil
    59  			}
    60  			time.Sleep(time.Second)
    61  			continue
    62  		}
    63  		s.runtime.SubLogger(runtime.SystemInit).Info("initial cluster info set")
    64  		break
    65  	}
    66  	for i := 0; i < checkBootstrapCycles; i++ {
    67  		select {
    68  		case <-ctx.Done():
    69  			return nil
    70  		default:
    71  		}
    72  		if err := s.createInitTasks(ctx); err == nil {
    73  			break
    74  		}
    75  		time.Sleep(time.Second)
    76  	}
    77  	return nil
    78  }
    79  
    80  func (s *Service) createInitTasks(ctx context.Context) error {
    81  	if err := s.store.taskScheduler.Create(ctx, []task.TaskMetadata{{
    82  		ID:       task.TaskCode_SystemInit.String(),
    83  		Executor: task.TaskCode_SystemInit,
    84  	}}); err != nil {
    85  		s.runtime.SubLogger(runtime.SystemInit).Error("failed to create init tasks", zap.Error(err))
    86  		return err
    87  	}
    88  	s.runtime.SubLogger(runtime.SystemInit).Info("init tasks created")
    89  	return nil
    90  }