Merge "Document how to install and prepare OpenStack VIM"
This commit is contained in:
Zuul
Gerrit Code Review
commit
caf389a6aa
|
|
@ -41,8 +41,19 @@ and another is VIM systems. Tacker system can be installed
|
|||
install/kolla.rst
|
||||
install/devstack.rst
|
||||
install/manual_installation.rst
|
||||
|
||||
Target VIM installation
|
||||
=======================
|
||||
|
||||
Most of time, the target VIM existed for Tacker to manage. This section shows
|
||||
us how to prepare a target VIM for Tacker.
|
||||
|
||||
.. toctree::
|
||||
:maxdepth: 1
|
||||
|
||||
install/openstack_vim_installation.rst
|
||||
|
||||
|
||||
Getting Started
|
||||
===============
|
||||
|
||||
|
|
|
|||
|
|
@ -30,18 +30,13 @@ In Tacker MANO system, the VNF can be onboarded to one target OpenStack, which
|
|||
is also called VIM. Get one account on this OpenStack. For example, the below
|
||||
is the account information collected in file vim-config.yaml::
|
||||
|
||||
auth_url: 'http://10.18.112.10:5000'
|
||||
auth_url: 'http://10.1.0.5:5000'
|
||||
username: 'nfv_user'
|
||||
password: 'mySecretPW'
|
||||
project_name: 'nfv'
|
||||
project_domain_name: 'Default'
|
||||
user_domain_name: 'Default'
|
||||
|
||||
.. note::
|
||||
|
||||
Here nfv_user need have 'admin' and 'advsvc' role on the
|
||||
project that will be used to deploy VNFs according to the VNF descriptor.
|
||||
|
||||
|
||||
2.) Register the VIM that will be used as a default VIM for VNF deployments.
|
||||
This will be required when the optional argument --vim-id is not provided by
|
||||
|
|
@ -56,7 +51,7 @@ the user during vnf-create.
|
|||
|
||||
|
||||
Onboarding sample VNF
|
||||
=======================
|
||||
=====================
|
||||
|
||||
1). Create a sample-vnfd.yaml file with the following content:
|
||||
|
||||
|
|
@ -112,15 +107,6 @@ Onboarding sample VNF
|
|||
You can find more sample tosca templates at
|
||||
https://github.com/openstack/tacker/tree/master/samples/tosca-templates/vnfd.
|
||||
|
||||
.. note::
|
||||
|
||||
For this VNFD to work, the VIM account should be able to create Flavor via
|
||||
heat on target VIM. Please refer to the OpenStack HEAT documentation about
|
||||
how to enable Flavor creation.
|
||||
|
||||
The target OpenStack should have cirros-0.3.5-x86_64-disk glance image and
|
||||
net_mgmt virtual network created for the VIM account.
|
||||
|
||||
|
||||
2). Create a sample vnfd.
|
||||
|
||||
|
|
|
|||
Binary file not shown.
|
After Width: | Height: | Size: 79 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 146 KiB |
|
|
@ -20,6 +20,241 @@ OpenStack VIM Installation
|
|||
==========================
|
||||
|
||||
This document describes how to install a simple OpenStack which can be
|
||||
a Tacker VIM.
|
||||
a Tacker VIM. Since the kolla-ansible provides a good and stable way to deploy
|
||||
an OpenStack environment, this document adopts it to finish this task.
|
||||
|
||||
TODO
|
||||
The target OpenStack consists of three nodes. Besides the OpenStack role, the
|
||||
first one plays kolla-ansible deploy role, which will run a local registry,
|
||||
and be used to run kolla-ansible tool.
|
||||
|
||||
The basic information and the topology of these nodes is like this:
|
||||
|
||||
.. image:: openstack_nodes.png
|
||||
:scale: 50 %
|
||||
|
||||
|
||||
Prepare kolla-ansible
|
||||
~~~~~~~~~~~~~~~~~~~~~
|
||||
|
||||
About how to prepare Docker and kolla-ansible environment,
|
||||
please refer to
|
||||
https://docs.openstack.org/kolla-ansible/latest/user/quickstart.html
|
||||
|
||||
|
||||
Set up local kolla-ansible docker registry
|
||||
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
||||
|
||||
Kolla-ansible is publishing the packaged Docker images at
|
||||
http://tarballs.openstack.org/kolla/images/. This document will use
|
||||
centos-source-registry-pike.tar.gz. So Download it:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
# wget http://tarballs.openstack.org/kolla/images/centos-source-registry-pike.tar.gz
|
||||
|
||||
..
|
||||
|
||||
|
||||
And unpack it:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
# tar xzvf centos-source-registry-pike.tar.gz -C /opt/registry/
|
||||
|
||||
..
|
||||
|
||||
And start Docker registry container:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
# docker run -d -v /opt/registry:/var/lib/registry -p 4000:5000 --restart=always --name registry registry:2
|
||||
|
||||
..
|
||||
|
||||
|
||||
And set Docker to access local registry via insecure channel:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
# cat /usr/lib/systemd/system/docker.service | grep insecure
|
||||
ExecStart=/usr/bin/dockerd --insecure-registry 10.1.0.6:4000
|
||||
# systemctl daemon-reload
|
||||
# systemctl restart docker
|
||||
|
||||
..
|
||||
|
||||
.. note::
|
||||
|
||||
The way to set up Docker to access insecure registry depends on operating
|
||||
system and Docker version, above way is just an example.
|
||||
|
||||
|
||||
And verify the local registry contains the needed images:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
# curl -k localhost:4000/v2/_catalog
|
||||
# curl -k localhost:4000/v2/lokolla/centos-source-fluentd/tags/list
|
||||
{"name":"lokolla/centos-source-fluentd","tags":["5.0.1"]}
|
||||
|
||||
..
|
||||
|
||||
|
||||
Install OpenStack
|
||||
~~~~~~~~~~~~~~~~~
|
||||
|
||||
1. Edit kolla ansible's configuration file /etc/kolla/globals.yml:
|
||||
|
||||
.. code-block:: ini
|
||||
|
||||
---
|
||||
kolla_install_type: "source"
|
||||
openstack_release: "5.0.1"
|
||||
kolla_internal_vip_address: "10.1.0.5"
|
||||
docker_registry: "10.1.0.6:4000"
|
||||
docker_namespace: "lokolla"
|
||||
api_interface: "eth0"
|
||||
tunnel_interface: "eth1"
|
||||
neutron_external_interface: "eth2"
|
||||
enable_glance: "yes"
|
||||
enable_haproxy: "yes"
|
||||
enable_keystone: "yes"
|
||||
enable_mariadb: "yes"
|
||||
enable_memcached: "yes"
|
||||
enable_neutron: "yes"
|
||||
enable_nova: "yes"
|
||||
enable_rabbitmq: "yes"
|
||||
enable_aodh: "yes"
|
||||
enable_ceilometer: "yes"
|
||||
enable_gnocchi: "yes"
|
||||
enable_heat: "yes"
|
||||
enable_horizon: "yes"
|
||||
enable_neutron_sfc: "yes"
|
||||
|
||||
|
||||
.. note::
|
||||
|
||||
If nodes are using different network interface names to connect each other,
|
||||
please define them in inventory file.
|
||||
|
||||
"10.1.0.5" is an un-used ip address, will be used as VIP address, realized
|
||||
by keepalived container.
|
||||
|
||||
|
||||
2. Run kolla-genpwd to generate system passwords:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
$ sudo cp etc/kolla/passwords.yml /etc/kolla/passwords.yml
|
||||
$ sudo kolla-genpwd
|
||||
|
||||
..
|
||||
|
||||
.. note::
|
||||
|
||||
If the pypi version is used to install kolla-ansible the skeleton passwords
|
||||
file may be under '/usr/share/kolla-ansible/etc_examples/kolla'.
|
||||
|
||||
|
||||
With this command, /etc/kolla/passwords.yml will be populated with
|
||||
generated passwords.
|
||||
|
||||
|
||||
3. Editor inventory:
|
||||
|
||||
First copy the sample multinode inventory file from kolla-ansible:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
# cp inventory/multinode ~/
|
||||
|
||||
..
|
||||
|
||||
|
||||
Then edit it to contain all of the OpenStack nodes.
|
||||
|
||||
.. code-block:: ini
|
||||
|
||||
[all_vim_nodes]
|
||||
10.1.0.8
|
||||
10.1.0.7
|
||||
10.1.0.6
|
||||
|
||||
[control:children]
|
||||
all_vim_nodes
|
||||
|
||||
[network:children]
|
||||
all_vim_nodes
|
||||
|
||||
[compute:children]
|
||||
all_vim_nodes
|
||||
|
||||
[monitoring:children]
|
||||
all_vim_nodes
|
||||
|
||||
[storage:children]
|
||||
#if the tacker needs volume feature, put related nodes here
|
||||
|
||||
4. Run kolla ansible deploy to install OpenStack system:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
# kolla-ansible deploy -i ~/multinode
|
||||
|
||||
..
|
||||
|
||||
|
||||
5. Run kolla ansible post-deploy to generate tacker access environment file:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
# kolla-ansible post-deploy
|
||||
|
||||
..
|
||||
|
||||
With this command, the "admin-openrc.sh" will be generated at
|
||||
/etc/kolla/admin-openrc.sh.
|
||||
|
||||
|
||||
Prepare OpenStack
|
||||
~~~~~~~~~~~~~~~~~
|
||||
|
||||
After installation, OpenStack administrator needs to:
|
||||
|
||||
* Initialized the flavors that the platform will support. Most
|
||||
Tacker sample TOSCA templates will ask Tacker to create Flavor
|
||||
on demand. If not, the specified flavor in templates must exist
|
||||
in OpenStack.
|
||||
|
||||
* Upload related images. Tacker repo's sample TOSCA templates are
|
||||
referring to cirros image named 'cirros-0.3.5-x86_64-disk', so
|
||||
this image should uploaded into OpenStack before Tacker uses it.
|
||||
|
||||
In additions, following steps are needed:
|
||||
|
||||
1. Create projects and users which can be used by Tacker:
|
||||
|
||||
This is a simple task for any OpenStack administrator, but one thing to pay
|
||||
attention to is that the user must have 'admin' and 'heat_stack_owner'
|
||||
roles on the user's project.
|
||||
|
||||
.. image:: openstack_role.png
|
||||
:scale: 50 %
|
||||
|
||||
|
||||
2. Create Neutron networks:
|
||||
|
||||
Most sample TOSCA templates assume there are three Neutron networks in
|
||||
target OpenStack that the VIM user can use:
|
||||
|
||||
* net_mgmt, which is a network Tacker system can access to. Some Tacker
|
||||
features, such as monitor policies, need Tacker to access started VNF
|
||||
virtual machines. For Tacker to access VNF via net_mgmt, net_mgmt can
|
||||
be a provider network.
|
||||
|
||||
* net0 and net1, which are two business networks which VNFs will use.
|
||||
How to connected them depends on the VNFs' business.
|
||||
|
||||
So create these three networks accordingly. For commands to create Neutron
|
||||
networks, please refer to
|
||||
https://docs.openstack.org/python-openstackclient/latest/cli/command-objects/network.html
|
||||
|
|
|
|||
Loading…
Reference in New Issue