why do we need new identifier in mysql locker

[fudongyingluck]

we already have an identifier for a job.

gocron/job.go

Line 754 in 63f3701

func WithIdentifier(id uuid.UUID) JobOption {

why do we need another one in the locker ? Why don't we just pass that as a parameter ?

gocron/executor.go

Line 387 in 63f3701

lock, err := j.locker.Lock(j.ctx, j.name)

https://github.com/go-co-op/gocron-gorm-lock/blob/c615ac7060c9c09324ea80ed6a68d8c1769e2991/gorm_lock.go#L67

 func (g *GormLocker) Lock(ctx context.Context, key string) (gocron.Lock, error) {
	ji := g.jobIdentifier(ctx, key)  // <- here

	cjb := &CronJobLock{
		JobName:       key,
		JobIdentifier: ji,
		Worker:        g.worker,
		Status:        StatusRunning,
	}
	tx := g.db.WithContext(ctx).Create(cjb)
	if tx.Error != nil {
		return nil, tx.Error
	}
	return &gormLock{db: g.db, id: cjb.GetID()}, nil
}

[JohnRoesler]

@manuelarte ☝️

[manuelarte]

Hi @fudongyingluck,

Maybe the name jobIdentifier is not the most accurate one, but let me explain its goal.
The goal of the jobIdentifier is to identify when the job was run.
So then, the combination of jobName & jobIdentifier uniquely identify a particular execution of a job.

As an example, imagine that you have two instances running (i1 and i2). And you configure a cron job (named test) to run every minute.

At t1, i1 is faster in picking up the job, and creates a record in the database, (jobName: test, jobIdentifier: t1), then, i2 will not pick up the job, because it will try to lock and insert a record with jobName: test, jobIdentifier: t1, but there is a combined unique constraint in the columns jobName and jobIdentifier.

If we would use the jobIdentifier uuid that you mentioned, then another approach than this need to be used, because the different instances would create different uuids, and then both instances would run the job.

[manuelarte]

I created a PR trying to explain it better here.