Merge lp:~themue/juju-core/go-mstate-machine-agent-alive into lp:~juju/juju-core/trunk

[Frank Mueller]

mstate: added agent alive methods

Machine provides AgentAlive(), WaitAgentAlive() and
SetAgentAlive() using the presence package.

https://codereview.appspot.com/6498091/

[Gustavo Niemeyer (niemeyer)]

Nice, LGTM.

https://codereview.appspot.com/6498091/diff/1/mstate/machine.go
File mstate/machine.go (right):

https://codereview.appspot.com/6498091/diff/1/mstate/machine.go#newcode82
mstate/machine.go:82: func (m *Machine) WaitAgentAlive(timeout
time.Duration) error {
Nice implementation.

https://codereview.appspot.com/6498091/diff/1/mstate/machine.go#newcode101
mstate/machine.go:101: panic(fmt.Sprintf("unexpected alive status of
machine %v", m))
"presence reported dead status twice in a row for machine %v"

https://codereview.appspot.com/6498091/diff/1/mstate/machine.go#newcode109
mstate/machine.go:109: // by starting a pinger on its presence node. It
returns the
Please drop the "by starting a pinger on its presence node" part. This
isn't entirely true anymore, and isn't really necessary given the rest
of the doc.

https://codereview.appspot.com/6498091/diff/1/mstate/machine_test.go
File mstate/machine_test.go (right):

https://codereview.appspot.com/6498091/diff/1/mstate/machine_test.go#newcode31
mstate/machine_test.go:31: defer pinger.Kill()
s/Kill/Stop/

We only care that it's not running after the test is done.

https://codereview.appspot.com/6498091/diff/1/mstate/machine_test.go#newcode50
mstate/machine_test.go:50: c.Assert(pinger, Not(IsNil))
This check isn't necessary. We'd go crazy if we had to check everything
that is *not* nil when err is nil.

https://codereview.appspot.com/6498091/diff/1/mstate/machine_test.go#newcode60
mstate/machine_test.go:60: pinger.Kill()
err := pinger.Kill()
c.Assert(err, IsNil)

https://codereview.appspot.com/6498091/

=== modified file 'mstate/machine.go'
--- mstate/machine.go	2012-08-29 10:05:26 +0000
+++ mstate/machine.go	2012-09-05 09:13:24 +0000
@@ -3,8 +3,10 @@
 import (
 	"fmt"
 	"labix.org/v2/mgo/txn"
+	"launchpad.net/juju-core/mstate/presence"
 	"launchpad.net/juju-core/trivial"
 	"strconv"
+	"time"
 )
 
 // Machine represents the state of a machine.
@@ -29,6 +31,11 @@
 	return m.doc.Id
 }
 
+// globalKey returns the global database key for the machine.
+func (m *Machine) globalKey() string {
+	return "m#" + m.String()
+}
+
 // Life returns whether the machine is Alive, Dying or Dead.
 func (m *Machine) Life() Life {
 	return m.doc.Life
@@ -66,6 +73,50 @@
 	return nil
 }
 
+// AgentAlive returns whether the respective remote agent is alive.
+func (m *Machine) AgentAlive() bool {
+	return m.st.presencew.Alive(m.globalKey())
+}
+
+// WaitAgentAlive blocks until the respective agent is alive.
+func (m *Machine) WaitAgentAlive(timeout time.Duration) error {
+	ch := make(chan presence.Change)
+	m.st.presencew.Add(m.globalKey(), ch)
+	defer m.st.presencew.Remove(m.globalKey(), ch)
+	// Initial check.		
+	select {
+	case change := <-ch:
+		if change.Alive {
+			return nil
+		}
+	case <-time.After(timeout):
+		return fmt.Errorf("waiting for agent of machine %v: still not alive after timeout", m)
+	}
+	// Hasn't been alive, so now wait for change.
+	select {
+	case change := <-ch:
+		if change.Alive {
+			return nil
+		}
+		panic(fmt.Sprintf("unexpected alive status of machine %v", m))
+	case <-time.After(timeout):
+		return fmt.Errorf("waiting for agent of machine %v: still not alive after timeout", m)
+	}
+	panic("unreachable")
+}
+
+// SetAgentAlive signals that the agent for machine m is alive
+// by starting a pinger on its presence node. It returns the
+// started pinger.
+func (m *Machine) SetAgentAlive() (*presence.Pinger, error) {
+	p := presence.NewPinger(m.st.presence, m.globalKey())
+	err := p.Start()
+	if err != nil {
+		return nil, err
+	}
+	return p, nil
+}
+
 // InstanceId returns the provider specific machine id for this machine.
 func (m *Machine) InstanceId() (string, error) {
 	if m.doc.InstanceId == "" {

=== modified file 'mstate/machine_test.go'
--- mstate/machine_test.go	2012-08-29 10:05:26 +0000
+++ mstate/machine_test.go	2012-09-05 09:13:24 +0000
@@ -4,6 +4,7 @@
 	. "launchpad.net/gocheck"
 	state "launchpad.net/juju-core/mstate"
 	"sort"
+	"time"
 )
 
 type MachineSuite struct {
@@ -20,6 +21,50 @@
 	c.Assert(err, IsNil)
 }
 
+func (s *MachineSuite) TestMachineSetAgentAlive(c *C) {
+	alive := s.machine.AgentAlive()
+	c.Assert(alive, Equals, false)
+
+	pinger, err := s.machine.SetAgentAlive()
+	c.Assert(err, IsNil)
+	c.Assert(pinger, Not(IsNil))
+	defer pinger.Kill()
+
+	s.State.ForcePresenceRefresh()
+	alive = s.machine.AgentAlive()
+	c.Assert(alive, Equals, true)
+}
+
+func (s *MachineSuite) TestMachineWaitAgentAlive(c *C) {
+	// test -gocheck.f TestMachineWaitAgentAlive
+	timeout := 5 * time.Second
+	alive := s.machine.AgentAlive()
+	c.Assert(alive, Equals, false)
+
+	s.State.ForcePresenceRefresh()
+	err := s.machine.WaitAgentAlive(timeout)
+	c.Assert(err, ErrorMatches, `waiting for agent of machine 0: still not alive after timeout`)
+
+	pinger, err := s.machine.SetAgentAlive()
+	c.Assert(err, IsNil)
+	c.Assert(pinger, Not(IsNil))
+
+	s.State.ForcePresenceRefresh()
+	err = s.machine.WaitAgentAlive(timeout)
+	c.Assert(err, IsNil)
+
+	alive = s.machine.AgentAlive()
+	c.Assert(err, IsNil)
+	c.Assert(alive, Equals, true)
+
+	pinger.Kill()
+
+	s.State.ForcePresenceRefresh()
+	alive = s.machine.AgentAlive()
+	c.Assert(err, IsNil)
+	c.Assert(alive, Equals, false)
+}
+
 func (s *MachineSuite) TestMachineInstanceId(c *C) {
 	machine, err := s.State.AddMachine()
 	c.Assert(err, IsNil)