Ubuntu – Can juju add-machine be used to prepare a machine for service deployment

juju

I'm using juju 1.16.3 and I am attempting to use juju add-machine to get a MaaS node ready for a service deployment. I want to get Ubuntu installed and ready so service deployment takes less time. The machine is started and the OS installed, and the agent-state is "started". However, when I deploy a service, it decideds to start a new machine anyway and ignores the one already started.

When I deploy the service without doing add-machine beforehand, then a new machine is automatically allocated from MaaS and the service deploys to it just fine. The ONLY difference is that when I do juju add-machine prior to deploying the service, it won't deploy to that machine.

Am I misunderstanding the purpose of add-machine? If so, what is the proper method for preparing a machine with a base OS prior to service deployment?

Best Answer

If you're using constraints at all, I think you've hit a consequence of this bug -- none of the existing machines will be matched.

You can work around it with the --to parameter to deploy or add-unit:

$ juju deploy mysql --to 3
$ juju add-unit mysql --to 4

Related Solutions

Juju – How to Fix Juju Services Not Deploying Correctly

Tha it's work! I again install ubuntu server and this time everything goes well but one thing. Now when I juju status I see this:

2012-10-04 13:32:21,110 INFO Connected to environment.
machines:
  0:
    agent-state: running
    dns-name: node-00127968a7be.local
    instance-id: /MAAS/api/1.0/nodes/node-ab7c5eb6-0e08-11e2-bb37-001185e67955/
    instance-state: unknown
  1:
    agent-state: running
    dns-name: node-001185e677fe
    instance-id: /MAAS/api/1.0/nodes/node-82beae92-0e09-11e2-a134-001185e67955/
    instance-state: unknown
  2:
    agent-state: running
    dns-name: node-001185e6772b.local
    instance-id: /MAAS/api/1.0/nodes/node-5c21dc18-0e0a-11e2-a134-001185e67955/
    instance-state: unknown
services:
  mysql:
    charm: cs:precise/mysql-8
    relations:
      db:
      - wordpress
    units:
      mysql/0:
        agent-state: started
        machine: 1
        public-address: node-001185e677fe.localdomain
  wordpress:
    charm: cs:precise/wordpress-9
    exposed: true
    relations:
      db:
      - mysql
      loadbalancer:
      - wordpress
    units:
      wordpress/0:
        agent-state: started
        machine: 2
        open-ports:
        - 80/tcp
        public-address: node-001185e6772b.local

public adress seems be strange because it's name of my nodes. I check on my router adress of node where is wordpress and type on webbrowser but I get

502 Bad Gateway nginx/1.1.19

I don't know what to do. Please help me.

Ubuntu – Can you prevent Juju from destroying machines

I am also new to juju but what I do to prevent this is by deploying an ubuntu charm and then deploying the service on that machine. So, first we should assign a tag to that node in MAAS UI before deploying ubuntu charm. Then

juju deploy ubuntu --constraints tags=<tag that we set>

A new node with ubuntu will come up. Now deploy the service on that machine

juju deploy <charm_url> --to <machine_number>

Best Answer

Related Solutions

Juju – How to Fix Juju Services Not Deploying Correctly

Ubuntu – Can you prevent Juju from destroying machines

Related Question