Machine identifiers in bundles are 'logical' machines, not concrete machine
identifiers. There is some work being done to allow you to map the logical
identifiers in a bundle to the concrete identifiers in your model.
(Otherwise bundles won't ever work as soon as you deploy 1 extra machine,
etc.), and it would severely limit the ability to share bundles because
your models would always have to match exactly.

On Tue, Jul 11, 2017 at 5:25 PM, Edward Hope-Morley <
<email address hidden>> wrote:

> Public bug reported:
>
> Juju: 2.2.1
> MAAS: 2.2.1
>
> If i create a new model and deploy a bundle as follows:
>
> cat << EOF > bundle.yaml
> machines:
>   '0':
>     series: xenial
>     constraints: 'tags=testnode'
> services:
>   ubuntu:
>     charm: cs:ubuntu
>     num_units: 1
>     to:
>     - lxd:0
> EOF
> juju deploy bundle.yaml
>
> It works just fine:
>
> Model    Controller  Cloud/Region  Version  SLA
> default  mymaas      mymaas        2.2.1    unsupported
>
> App     Version  Status  Scale  Charm   Store       Rev  OS      Notes
> ubuntu  16.04    active      1  ubuntu  jujucharms   10  ubuntu
>
> Unit       Workload  Agent  Machine  Public address   Ports  Message
> ubuntu/0*  active    idle   0/lxd/0  192.168.200.125         ready
>
> Machine  State    DNS              Inst id              Series  AZ
>  Message
> 0        started  192.168.200.124  mr6nmf               xenial  default
> Deployed
> 0/lxd/0  started  192.168.200.125  juju-826296-0-lxd-0  xenial
>  Container started
>
> If i then remove the ubuntu application i deployed it does so cleanly
> and I am left with just machine "0". If I then re-execute juju deploy
> bundle.yaml (unchanged from previous run) it does not use machine "0"
> and instead deploys a new machine onto which it deploys the ubuntu
> application.
>
> Model    Controller  Cloud/Region  Version  SLA
> default  mymaas      mymaas        2.2.1    unsupported
>
> App     Version  Status   Scale  Charm   Store       Rev  OS      Notes
> ubuntu           waiting    0/1  ubuntu  jujucharms   10  ubuntu
>
> Unit      Workload  Agent       Machine  Public address  Ports  Message
> ubuntu/1  waiting   allocating  1/lxd/0                         waiting
> for machine
>
> Machine  State    DNS              Inst id  Series  AZ       Message
> 0        started  192.168.200.124  mr6nmf   xenial  default  Deployed
> 1        pending                   pending  xenial           failed to
> start instance (cannot run instances: cannot run instance: No available
> machine matches constraints: [('agent_name', ['51e51f83-ef54-4cf5-8a48-7de6e3826296']),
> ('zone', ['default']), ('tags', ['testnode']
> )] (resolved to "tags=testnode zone=default")), retrying in 10s (9 more
> attempts)
> 1/lxd/0  pending                   pending  xenial
>
> ** Affects: juju
>      Importance: Undecided
>          Status: New
>
>
> ** Tags: sts
>
> ** Tags added: sts
>
> --
> You received this bug notification because you are subscribed to juju.
> Matching subscriptions: juju bugs
> https://bugs.launchpad.net/bugs/1703610
>
> Title:
>   juju deploy unit placement is broken
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/juju/+bug/1703610/+subscriptions
>