Documentation ¶
Index ¶
- Variables
- func GetConnector(host string, port int, username string, password string, database string) (driver.Connector, error)
- func GetConnectorFromConnStr(connStr string, opts ...stdlib.OptionOpenDB) (driver.Connector, error)
- func GetConnectorFromURL(u *url.URL) (driver.Connector, error)
- func PrepareStatements(ctx context.Context, conn *pgx.Conn) error
- type Client
- func (c *Client) Close()
- func (c *Client) Enqueue(j *Job) error
- func (c *Client) EnqueueInTx(j *Job, tx *sql.Tx) error
- func (c *Client) LockJob(queue string) (*Job, error)
- func (c *Client) LockJobContext(ctx context.Context, queue string) (*Job, error)
- func (c *Client) Stats() (results []JobStats, err error)
- type ConnWrapper
- type Conner
- type Job
- func (j *Job) Conn() *sql.Conndeprecated
- func (j *Job) Delete() error
- func (j *Job) DeleteContext(ctx context.Context) error
- func (j *Job) Done()
- func (j *Job) DoneContext(ctx context.Context)
- func (j *Job) Error(msg string) error
- func (j *Job) ErrorContext(ctx context.Context, msg string) error
- func (j *Job) Tx() Txer
- type JobStats
- type Queryer
- type Txer
- type WorkFunc
- type WorkMap
- type Worker
- type WorkerPool
Constants ¶
This section is empty.
Variables ¶
var ErrAgain = errors.New("maximum number of LockJob attempts reached")
ErrAgain Returned by LockJob if a job could not be retrieved from the queue after several attempts because of concurrently running transactions. This error should not be returned unless the queue is under extremely heavy concurrency.
var ErrMissingType = errors.New("job type must be specified")
ErrMissingType is returned when you attempt to enqueue a job with no Type specified.
Functions ¶
func GetConnector ¶
func GetConnectorFromConnStr ¶
func PrepareStatements ¶
PrepareStatements prepar statements
Types ¶
type Client ¶
type Client struct {
// contains filtered or unexported fields
}
Client is a Que client that can add jobs to the queue and remove jobs from the queue.
func MustNewClient ¶
MustNewClient creates a new Client. Panic if the initialization fails.
func (*Client) Close ¶
func (c *Client) Close()
Close disposes all the resources associated to the client
func (*Client) EnqueueInTx ¶
EnqueueInTx adds a job to the queue within the scope of the transaction tx. This allows you to guarantee that an enqueued job will either be committed or rolled back atomically with other changes in the course of this transaction.
It is the caller's responsibility to Commit or Rollback the transaction after this function is called.
func (*Client) LockJobContext ¶
LockJob attempts to retrieve a Job from the database in the specified queue. If a job is found, a session-level Postgres advisory lock is created for the Job's ID. If no job is found, nil will be returned instead of an error.
Because Que uses session-level advisory locks, we have to hold the same connection throughout the process of getting a job, working it, deleting it, and removing the lock.
After the Job has been worked, you must call either Done() or Error() on it in order to return the database connection to the pool and remove the lock.
type Job ¶
type Job struct { // ID is the unique database ID of the Job. It is ignored on job creation. ID int64 // Queue is the name of the queue. It defaults to the empty queue "". Queue string // Priority is the priority of the Job. The default priority is 100, and a // lower number means a higher priority. A priority of 5 would be very // important. Priority int16 // RunAt is the time that this job should be executed. It defaults to now(), // meaning the job will execute immediately. Set it to a value in the future // to delay a job's execution. RunAt time.Time // Type corresponds to the Ruby job_class. If you are interoperating with // Ruby, you should pick suitable Ruby class names (such as MyJob). Type string // Args must be the bytes of a valid JSON string Args []byte // ErrorCount is the number of times this job has attempted to run, but // failed with an error. It is ignored on job creation. ErrorCount int32 // LastError is the error message or stack trace from the last time the job // failed. It is ignored on job creation. LastError sql.NullString // contains filtered or unexported fields }
Job is a single unit of work for Que to perform.
func TestInjectJobConn ¶
// TestInjectJobConn injects *pgx.Conn to Job
func (*Job) DeleteContext ¶
Delete marks this job as complete by deleting it form the database.
You must also later call Done() to return this job's database connection to the pool.
func (*Job) DoneContext ¶
Done releases the Postgres advisory lock on the job and returns the database connection to the pool.
func (*Job) ErrorContext ¶
Error marks the job as failed and schedules it to be reworked. An error message or backtrace can be provided as msg, which will be saved on the job. It will also increase the error count.
You must also later call Done() to return this job's database connection to the pool.
type JobStats ¶
type JobStats struct { Queue string Type string Count int CountWorking int CountErrored int HighestErrorCount int OldestRunAt time.Time }
JobStats stores the statistics information for the queue and type
type Queryer ¶
type Queryer interface { Exec(string, ...interface{}) (sql.Result, error) ExecContext(context.Context, string, ...interface{}) (sql.Result, error) PrepareContext(context.Context, string) (*sql.Stmt, error) Query(string, ...interface{}) (*sql.Rows, error) QueryContext(context.Context, string, ...interface{}) (*sql.Rows, error) QueryRow(string, ...interface{}) *sql.Row QueryRowContext(context.Context, string, ...interface{}) *sql.Row }
Queryer is interface for query
type WorkFunc ¶
WorkFunc is a function that performs a Job. If an error is returned, the job is reenqueued with exponential backoff.
type WorkMap ¶
WorkMap is a map of Job names to WorkFuncs that are used to perform Jobs of a given type.
type Worker ¶
type Worker struct { // Interval is the amount of time that this Worker should sleep before trying // to find another Job. Interval time.Duration // Queue is the name of the queue to pull Jobs off of. The default value, "", // is usable and is the default for both que-go and the ruby que library. Queue string // contains filtered or unexported fields }
Worker is a single worker that pulls jobs off the specified Queue. If no Job is found, the Worker will sleep for Interval seconds.
func NewWorker ¶
NewWorker returns a Worker that fetches Jobs from the Client and executes them using WorkMap. If the type of Job is not registered in the WorkMap, it's considered an error and the job is re-enqueued with a backoff.
Workers default to an Interval of 5 seconds, which can be overridden by setting the environment variable QUE_WAKE_INTERVAL. The default Queue is the nameless queue "", which can be overridden by setting QUE_QUEUE. Either of these settings can be changed on the returned Worker before it is started with Work().
func (*Worker) Shutdown ¶
func (w *Worker) Shutdown()
Shutdown tells the worker to finish processing its current job and then stop. There is currently no timeout for in-progress jobs. This function blocks until the Worker has stopped working. It should only be called on an active Worker.
type WorkerPool ¶
type WorkerPool struct { WorkMap WorkMap Interval time.Duration Queue string // contains filtered or unexported fields }
WorkerPool is a pool of Workers, each working jobs from the queue Queue at the specified Interval using the WorkMap.
func NewWorkerPool ¶
func NewWorkerPool(c *Client, wm WorkMap, count int) *WorkerPool
NewWorkerPool creates a new WorkerPool with count workers using the Client c.
func (*WorkerPool) Shutdown ¶
func (w *WorkerPool) Shutdown()
Shutdown sends a Shutdown signal to each of the Workers in the WorkerPool and waits for them all to finish shutting down.
func (*WorkerPool) Start ¶
func (w *WorkerPool) Start()
Start starts all of the Workers in the WorkerPool.