1

This is my first go at containers, so please bear with me. I tried to deploy openstack using MAAS2.0.

I setup the required machines ( 4 nodes + Juju Controller) and I followed the instructions on this blog which was quite thorough

Deploying OpenStack on MAAS 1.9+ with Juju

Now there were some differences from Juju 2.0 and the Openstack bundle currently available and that of the writing of the blog but generally it all seems ok until the nova-compute configuration. I always get stuck here. In juju gui it only asks for the bridge network and os-internal-network. I know os-internal-network is the internal-api subnet, but I am not sure what to make of the bridge

Machine  State    DNS           Inst id  Series  AZ
0        started  10.100.0.111  t66637   xenial  default
0/lxd/0  down                   pending  xenial  
0/lxd/1  down                   pending  xenial  
0/lxd/2  down                   pending  xenial  
1        started  10.100.0.114  ctf87e   xenial  default
1/lxd/0  down                   pending  xenial  
1/lxd/1  down                   pending  xenial  
1/lxd/2  down                   pending  xenial  
2        started  10.100.0.112  a4medb   xenial  default
2/lxd/0  down                   pending  xenial  
2/lxd/1  down                   pending  xenial  
2/lxd/2  down                   pending  xenial  
3        started  10.100.0.113  fa3xwa   xenial  default
3/lxd/0  down                   pending  xenial  
3/lxd/1  down                   pending  xenial  
3/lxd/2  down                   pending  xenial  

and here is the output for juju show-machine

model: openstack-base
machines:
  "0":
    juju-status:
      current: started
      since: 14 Apr 2017 01:12:02Z
      version: 2.1.2
    dns-name: 10.100.0.111
    ip-addresses:
    - 10.100.0.111
    - 10.14.0.111
    - 10.150.0.111
    - 10.200.0.111
    - 10.250.0.111
    - 10.50.0.111
    instance-id: t66637
    machine-status:
      current: running
      message: Deployed
      since: 14 Apr 2017 01:11:13Z
    series: xenial
    containers:
      0/lxd/0:
        juju-status:
          current: down
          message: agent is not communicating with the server
          since: 14 Apr 2017 01:12:49Z
        instance-id: pending
        machine-status:
          current: provisioning error
          message: 'unable to setup network: no obvious space for container "0/lxd/0",
        host machine has spaces: "admin-api", "compute-data", "internal-api",
        "maas-mgmt", "public-api", "storage-data"'
      since: 14 Apr 2017 01:12:49Z
    series: xenial
  0/lxd/1:
    juju-status:
      current: down
      message: agent is not communicating with the server
      since: 14 Apr 2017 01:13:25Z
    instance-id: pending
    machine-status:
      current: provisioning error
      message: 'unable to setup network: no obvious space for container "0/lxd/1",
        host machine has spaces: "admin-api", "compute-data", "internal-api",
        "maas-mgmt", "public-api", "storage-data"'
      since: 14 Apr 2017 01:13:25Z
    series: xenial
  0/lxd/2:
    juju-status:
      current: down
      message: agent is not communicating with the server
      since: 14 Apr 2017 01:14:01Z
    instance-id: pending
    machine-status:
      current: provisioning error
      message: 'unable to setup network: no obvious space for container "0/lxd/2",
        host machine has spaces: "admin-api", "compute-data", "internal-api",
        "maas-mgmt", "public-api", "storage-data"'
      since: 14 Apr 2017 01:14:01Z
    series: xenial
hardware: arch=amd64 cores=12 mem=12288M tags=virtual availability-zone=default

Same output for the other 3 physical machines.

Last thing, (which i am sure its not related, but for completeness) the whole setup is not on actual machine, they are 5 VMs in vSphere all connected using single NICs to a vSwitch (4096 all) so its forwarding all vlans. Each VM got its ips assigned statically in MAAS.

Thanks alot

$juju spaces
Space             Subnets
admin-api         10.150.0.0/20
compute-data      10.250.0.0/20
compute-external  10.99.0.0/20
internal-api      10.100.0.0/20
maas-mgmt         10.14.0.0/20
                  192.168.122.0/24
public-api        10.50.0.0/20
storage-cluster   10.30.0.0/20
storage-data      10.200.0.0/20
unused            192.168.140.0/24
8
  • can you share the output of $juju spaces? you're deploying openstack base bundle, right?
    – Junaid
    Apr 14, 2017 at 13:38
  • I added the output, yes I am using the openstack base bundle ... any ideas ?
    – a.atlam
    Apr 14, 2017 at 21:16
  • It's been a while I haven't used MAAS and Juju. I think you've deleted the 'default' space, have you tried creating the space 'default' again and add maas-mgmt network 10.14.0.0/20 to that space?
    – Junaid
    Apr 15, 2017 at 3:16
  • yes it worked ... LXD containers got started and bridged ... now I am stuck on how to bridge a trunk to get all the required vlan networks to the lxd container ... any thoughts ? is this even the right approach? I am almost certain I am doing something wrong, otherwise how come all these people deploy successfully ??? any ideas ?
    – a.atlam
    Apr 25, 2017 at 16:17
  • 1
    For example, using network spaces, you can deploy 'nova-cloud-controller' charm using $ juju deploy nova-cloud-controller --to lxd:<machine-number-where-lxd-is-gonna-created> --bind "public=<your-public-space> internal=<internal-space-name> admin=<admin-space-name> shared-db=<internal-space-name>"
    – Junaid
    Apr 25, 2017 at 18:08

0

You must log in to answer this question.

Browse other questions tagged .