Documentation
¶
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type Config ¶
type Config struct {
// Identity records the entity whose connectedness is being
// affirmed by this worker. It's used to create a logger that
// can let us see which agent's pinger is actually failing.
Identity names.Tag
// Start starts a new, running Pinger or returns an error.
Start func() (Pinger, error)
// Clock is used to throttle failed Start attempts.
Clock clock.Clock
// RetryDelay controls by how much we throttle failed Start
// attempts. Note that we only apply the delay when a Start
// fails; if a Pinger ran, however briefly, we'll try to restart
// it immediately, so as to minimise the changes of erroneously
// causing agent-lost to be reported.
RetryDelay time.Duration
}
Config contains the information necessary to drive a Worker.
type Pinger ¶
type Pinger interface {
// Stop kills the pinger, then waits for it to exit.
Stop() error
// Wait waits for the pinger to stop.
Wait() error
}
Pinger exposes some methods implemented by state/presence.Pinger.
type Worker ¶
type Worker struct {
// contains filtered or unexported fields
}
Worker creates a Pinger as configured, and recreates it as it fails until the Worker is stopped; at which point it shuts down any extant Pinger before returning.
func New ¶
New returns a Worker backed by Config. The caller is responsible for Kill()ing the Worker and handling any errors returned from Wait(); but as it happens it's designed to be an apiserver/common.Resource, and never to exit unless Kill()ed, so in practice Stop(), which will call Kill() and Wait() internally, is Good Enough.
Source Files
Click to show internal directories.
Click to hide internal directories.