jujud: run machiner

jujud: run machiner

The test is very slow because of the poll interval.
We'll fix that later, but for the time being this
means that machines will actually transition to
Dead state correctly.

https://code.launchpad.net/~rogpeppe/juju-core/258-jujud-run-machiner/+merge/156867

(do not edit description out of merge proposal)

[rog, 2013-04-03 14:36:41]

Please take a look.

[fwereade, 2013-04-03 14:45:43]

LGTM

https://codereview.appspot.com/8318043/diff/2001/cmd/jujud/machine.go
File cmd/jujud/machine.go (right):

https://codereview.appspot.com/8318043/diff/2001/cmd/jujud/machine.go#newcode107
cmd/jujud/machine.go:107: NewUpgrader(st, m, a.Conf.DataDir),
I'm suddenly freaked out that NewUpgrader might not clone m. This is totally
irrelevant to this CL, though.

https://codereview.appspot.com/8318043/diff/2001/cmd/jujud/machine_test.go
File cmd/jujud/machine_test.go (right):

https://codereview.appspot.com/8318043/diff/2001/cmd/jujud/machine_test.go#ne...
cmd/jujud/machine_test.go:119: // TODO(rog) allow shorter poll interval in
tests.
Please give this a bug in LP and badge it here.

[dimitern, 2013-04-03 14:51:15]

LGTM, just one comment.

https://codereview.appspot.com/8318043/diff/2001/cmd/jujud/machine_test.go
File cmd/jujud/machine_test.go (right):

https://codereview.appspot.com/8318043/diff/2001/cmd/jujud/machine_test.go#ne...
cmd/jujud/machine_test.go:117: case <-time.After(6 * time.Second):
should this be a const somewhere then?

[rog, 2013-04-03 15:34:22]

*** Submitted:

jujud: run machiner

The test is very slow because of the poll interval.
We'll fix that later, but for the time being this
means that machines will actually transition to
Dead state correctly.

R=fwereade, dimitern
CC=
https://codereview.appspot.com/8318043

https://codereview.appspot.com/8318043/diff/2001/cmd/jujud/machine.go
File cmd/jujud/machine.go (right):

https://codereview.appspot.com/8318043/diff/2001/cmd/jujud/machine.go#newcode107
cmd/jujud/machine.go:107: NewUpgrader(st, m, a.Conf.DataDir),
On 2013/04/03 14:45:44, fwereade wrote:
> I'm suddenly freaked out that NewUpgrader might not clone m. This is totally
> irrelevant to this CL, though.

it probably should (let's do a Clone method) but for the time being this is safe
even if it doesn't, because we don't pass m to anything else.

https://codereview.appspot.com/8318043/diff/2001/cmd/jujud/machine_test.go
File cmd/jujud/machine_test.go (right):

https://codereview.appspot.com/8318043/diff/2001/cmd/jujud/machine_test.go#ne...
cmd/jujud/machine_test.go:117: case <-time.After(6 * time.Second):
On 2013/04/03 14:51:15, dimitern wrote:
> should this be a const somewhere then?

Done.

https://codereview.appspot.com/8318043/diff/2001/cmd/jujud/machine_test.go#ne...
cmd/jujud/machine_test.go:119: // TODO(rog) allow shorter poll interval in
tests.
On 2013/04/03 14:45:44, fwereade wrote:
> Please give this a bug in LP and badge it here.

Done.