今天就跟大家聊聊有关dkron的fsm是怎样的,可能很多人都不太了解,为了让大家更加了解,小编给大家总结了以下内容,希望大家根据这篇文章可以有所收获。
// MessageType is the type to encode FSM commands. type MessageType uint8 const ( // SetJobType is the command used to store a job in the store. SetJobType MessageType = iota // DeleteJobType is the command used to delete a Job from the store. DeleteJobType // SetExecutionType is the command used to store an Execution to the store. SetExecutionType // DeleteExecutionsType is the command used to delete executions from the store. DeleteExecutionsType // ExecutionDoneType is the command to perform the logic needed once an exeuction // is done. ExecutionDoneType )
MessageType可以分为SetJobType、DeleteJobType、SetExecutionType、DeleteExecutionsType、ExecutionDoneType
type dkronFSM struct { store Storage // proAppliers holds the set of pro only LogAppliers proAppliers LogAppliers } // NewFSM is used to construct a new FSM with a blank state func newFSM(store Storage, logAppliers LogAppliers) *dkronFSM { return &dkronFSM{ store: store, proAppliers: logAppliers, } } // Apply applies a Raft log entry to the key-value store. func (d *dkronFSM) Apply(l *raft.Log) interface{} { buf := l.Data msgType := MessageType(buf[0]) log.WithField("command", msgType).Debug("fsm: received command") switch msgType { case SetJobType: return d.applySetJob(buf[1:]) case DeleteJobType: return d.applyDeleteJob(buf[1:]) case ExecutionDoneType: return d.applyExecutionDone(buf[1:]) case SetExecutionType: return d.applySetExecution(buf[1:]) } // Check enterprise only message types. if applier, ok := d.proAppliers[msgType]; ok { return applier(buf[1:], l.Index) } return nil } func (d *dkronFSM) applySetJob(buf []byte) interface{} { var pj dkronpb.Job if err := proto.Unmarshal(buf, &pj); err != nil { return err } job := NewJobFromProto(&pj) if err := d.store.SetJob(job, false); err != nil { return err } return nil } func (d *dkronFSM) applyDeleteJob(buf []byte) interface{} { var djr dkronpb.DeleteJobRequest if err := proto.Unmarshal(buf, &djr); err != nil { return err } job, err := d.store.DeleteJob(djr.GetJobName()) if err != nil { return err } return job } func (d *dkronFSM) applyExecutionDone(buf []byte) interface{} { var execDoneReq dkronpb.ExecutionDoneRequest if err := proto.Unmarshal(buf, &execDoneReq); err != nil { return err } execution := NewExecutionFromProto(execDoneReq.Execution) log.WithField("execution", execution.Key()). WithField("output", string(execution.Output)). Debug("fsm: Setting execution") _, err := d.store.SetExecutionDone(execution) return err } func (d *dkronFSM) applySetExecution(buf []byte) interface{} { var pbex dkronpb.Execution if err := proto.Unmarshal(buf, &pbex); err != nil { return err } execution := NewExecutionFromProto(&pbex) key, err := d.store.SetExecution(execution) if err != nil { return err } return key } // Snapshot returns a snapshot of the key-value store. We wrap // the things we need in dkronSnapshot and then send that over to Persist. // Persist encodes the needed data from dkronSnapshot and transport it to // Restore where the necessary data is replicated into the finite state machine. // This allows the consensus algorithm to truncate the replicated log. func (d *dkronFSM) Snapshot() (raft.FSMSnapshot, error) { return &dkronSnapshot{store: d.store}, nil } // Restore stores the key-value store to a previous state. func (d *dkronFSM) Restore(r io.ReadCloser) error { defer r.Close() return d.store.Restore(r) } // LogApplier is the definition of a function that can apply a Raft log type LogApplier func(buf []byte, index uint64) interface{}
dkronFSM定义了store、proAppliers属性;Apply方法将raft的log保存到KV存储中,具体分不同msgType做不同处理;最后根据msgType查找LogAppliers
dkron的FSM根据不同msgType做不同处理,具体有applySetJob、applyDeleteJob、applyExecutionDone、applySetExecution方法。
看完上述内容,你们对dkron的fsm是怎样的有进一步的了解吗?如果还想了解更多知识或者相关内容,请关注亿速云行业资讯频道,感谢大家的支持。
免责声明:本站发布的内容(图片、视频和文字)以原创、转载和分享为主,文章观点不代表本网站立场,如果涉及侵权请联系站长邮箱:is@yisu.com进行举报,并提供相关证据,一经查实,将立刻删除涉嫌侵权内容。