Closed gmolto closed 7 years ago
This does not appear to be related to Ubuntu 16.04 and the lack of Python. I submitted the enclosed template at the bottom, where ami-5ac2cd4d refers to an Ubuntu 14.04 LTS with Python 2.7.6 and the contextualization does not even start. Files are copied to /tmp/.im but no attempt to contextualization appears to occur.
tosca_definitions_version: tosca_simple_yaml_1_0
imports:
description: TOSCA example for specifying a Mesos Cluster
topology_template:
node_templates:
mesos_master:
type: tosca.nodes.indigo.MesosMaster
properties:
marathon_password: test_pass
chronos_password: test_pass
requirements:
- host: mesos_master_server
mesos_slave:
type: tosca.nodes.indigo.MesosSlave
properties:
master_ips: { get_attribute: [ mesos_master_server, public_address ] }
requirements:
- host: mesos_slave_server
mesos_load_balancer:
type: tosca.nodes.indigo.MesosLoadBalancer
properties:
master_ips: { get_attribute: [ mesos_master_server, public_address ] }
requirements:
- host: mesos_lb_server
mesos_master_server:
type: tosca.nodes.indigo.Compute
capabilities:
endpoint:
properties:
network_name: vpc-6366c106.subnet-de5252f6.PUBLIC
scalable:
properties:
count: 1
host:
properties:
num_cpus: 2
mem_size: 2 GB
os:
properties:
type: linux
distribution: ubuntu
version: 14.04
image: aws://us-east-1/ami-5ac2cd4d
credential:
user: ubuntu
token: anything
mesos_slave_server:
type: tosca.nodes.indigo.Compute
capabilities:
endpoint:
properties:
network_name: vpc-6366c106.subnet-de5252f6.PUBLIC
scalable:
properties:
count: 1
host:
properties:
num_cpus: 1
mem_size: 1 GB
os:
properties:
type: linux
distribution: ubuntu
version: 14.04
image: aws://us-east-1/ami-5ac2cd4d
credential:
user: ubuntu
token: anything
mesos_lb_server:
type: tosca.nodes.indigo.Compute
capabilities:
endpoint:
properties:
network_name: vpc-6366c106.subnet-de5252f6.PUBLIC
scalable:
properties:
count: 1
host:
properties:
num_cpus: 1
mem_size: 1 GB
os:
properties:
type: linux
distribution: ubuntu
version: 14.04
image: aws://us-east-1/ami-5ac2cd4d
credential:
user: ubuntu
token: anything
outputs: mesos_lb_ips: value: { get_attribute: [ mesos_lb_server, public_address ] } mesos_master_ips: value: { get_attribute: [ mesos_master_server, public_address ] }
For the sake of completeness, the same behaviour is obtained with simpler templates, different AMIs and deploying on EC2-Classic.
Enclosed a sample template.
tosca_definitions_version: tosca_simple_yaml_1_0
imports:
description: > TOSCA test for launching the Wine Cellar Application (PHP Version) (https://github.com/ccoenraets/wine-cellar-php) using the simple in memory DB with a single web server node.
topology_template:
inputs: download_url: type: string default: https://raw.githubusercontent.com/indigo-dc/tosca-types/master/artifacts/webapps/cellar-webapp-mem.tgz
node_templates: apache: type: tosca.nodes.WebServer.Apache requirements:
host: web_server interfaces: Standard: configure: implementation: https://raw.githubusercontent.com/indigo-dc/tosca-types/master/artifacts/webapps/cellar-mem.yml inputs: download_url: { get_input: download_url }
web_server: type: tosca.nodes.indigo.Compute capabilities: endpoint: properties: network_name: PUBLIC ports: http_port: protocol: tcp source: 80 https_port: protocol: tcp source: 443
host:
properties:
num_cpus: 1
mem_size: 500 MB
# Guest Operating System properties
os:
properties:
# host Operating System image properties
type: linux
distribution: ubuntu
version: 14.04
image: aws://us-east-1/ami-a95044be
credential:
user: ubuntu
token: anything
After rebooting the IM server, the contextualization starts, Ansible gets installed BUT, apparently, the contextualization stops. No other packages are installed and an empty log is obtained when querying: http://servproject.i3m.upv.es:8811/infrastructures/{{IM_INF_ID}}/vms/0/contmsg
I tried with the very same template in my early post. Let's discuss this next week.
This is currently working so I am closing this issue.
I have deployed an instance of Ubuntu 16.04 LTS - Xenial (HVM) on AWS through the IM but, apparently, the contextualization process is stuck, but no error is arisen, probably because the machine does not have python installed.