A tool to deploy kubernetes and OpenStack using K
8000
olla Containers.
This tool provides a method to deploy OpenStack on a Kubernetes
Cluster using Kolla and Kolla-Kubernetes on bare metal servers or
virtual machines. Virtual machines supported are Ubuntu and
Centos.
This tool is designed to be a fun way to play with this technology. In
no way is it supposed to be production ready. It contains both a
verbose mode to see what is going on during deployment and a demo
mode which attempts to walk the user through all the steps.
It contains a lot of options to play with different tool versions
so it can be useful to verify and debug when the environment has
changed - say a kubernetes upgrade.
It currently works only with the kolla-kubernetes methodology but
I intend to expand it to use openstack-helm so the user can
compare and contrast.
All comments, suggestions and improvements welcome.
richwellum@gmail.com
ubuntu@ubuntuk8s:~/os$ ../k8s/ko.py ens3 ens4 -cn -iv pike -dr rwellum -cni weave --logs -h
usage: ko.py [-h] [-mi MGMT_IP] [-vi VIP_IP] [-iv IMAGE_VERSION]
[-hv HELM_VERSION] [-kv K8S_VERSION] [-av ANSIBLE_VERSION]
[-jv JINJA2_VERSION] [-c] [-cc] [-k8s] [-cm] [-os] [-n] [-eg]
[-ec] [-v] [-d] [-f] [-cn] [-dm] [-ng] [-bd BASE_DISTRO]
[-dr DOCKER_REPO] [-cni CNI] [-l] [-drr]
MGMT_INT NEUTRON_INT
This tool provides a method to deploy OpenStack on a Kubernetes Cluster using Kolla
and Kolla-Kubernetes on bare metal servers or virtual machines.
Virtual machines supported are Ubuntu and Centos.
Usage as simple as: "ko.py eth0 eth1"
The host machine must satisfy the following minimum requirements:
- 2 network interfaces
- 8GB min, 16GB preferred - main memory
- 40G min, 80GB preferred - disk space
- 2 CPUs Min, 4 preferred - CPUs
Root access to the deployment host machine is required.
positional arguments:
MGMT_INT The interface to which Kolla binds API services, E.g:
eth0
NEUTRON_INT The interface that will be used for the external
bridge in Neutron, E.g: eth1
optional arguments:
-h, --help show this help message and exit
-mi MGMT_IP, --mgmt_ip MGMT_IP
Provide own MGMT ip address Address, E.g:
10.240.83.111
-vi VIP_IP, --vip_ip VIP_IP
Provide own Keepalived VIP, used with keepalived,
should be an unused IP on management NIC subnet, E.g:
10.240.83.112
-iv IMAGE_VERSION, --image_version IMAGE_VERSION
Specify a different Kolla image version to the default
(ocata)
-hv HELM_VERSION, --helm_version HELM_VERSION
Specify a different helm version to the default(2.7.2)
-kv K8S_VERSION, --k8s_version K8S_VERSION
Specify a different kubernetes version to the
default(1.8.4) - note 1.8.0 is the minimum supported
-av ANSIBLE_VERSION, --ansible_version ANSIBLE_VERSION
Specify a different ansible version to the
default(2.2.0.0)
-jv JINJA2_VERSION, --jinja2_version JINJA2_VERSION
Specify a different jinja2 version to the
default(2.8.1)
-c, --cleanup YMMV: Cleanup existing Kubernetes cluster before
creating a new one. Because LVM is not cleaned up,
space will be used up. "-cc" is far more reliable but
requires a reboot
-cc, --complete_cleanup
Cleanup existing Kubernetes cluster then exit,
rebooting host is advised
-k8s, --kubernetes Stop after bringing up kubernetes, do not install
OpenStack
-cm, --create_minion Do not install Kubernetes or OpenStack, useful for
preparing a multi-node minion
-os, --openstack Build OpenStack on an existing Kubernetes Cluster
-n, --nslookup Pause for the user to manually test nslookup in
kubernetes cluster
-eg, --edit_globals Pause to allow the user to edit the globals.yaml file
- for custom configuration
-ec, --edit_cloud Pause to allow the user to edit the cloud.yaml file -
for custom configuration
-v, --verbose Turn on verbose messages
-d, --demo Display some demo information and offer to move on
-f, --force When used in conjunction with --demo - it will proceed
without user input
-cn, --create_network
Try to create a OpenStack network model, configure
neutron, download and install a test VM
-dm, --dev_mode Adds option to modify kolla and more info
-ng, --no_git Select this to not override git repos previously
downloaded
-bd BASE_DISTRO, --base_distro BASE_DISTRO
Specify a base container image to the default(centos),
like "ubuntu"
-dr DOCKER_REPO, --docker_repo DOCKER_REPO
Specify a different docker repo the default(lokolla),
for exampe "rwellum" has the latest pike images
-cni CNI, --cni CNI Specify a different CNI/SDN tothe default(canal), like
"weave"
-l, --logs Experimental, installs a patch set and runs fluentd
container to gather logs.
-drr, --dry_run Dry run commands only.
E.g.: k8s.py eth0 eth1 -kv 1.6.2 -hv 2.4.2 -iv pike
ubuntu@ubuntuk8s:~/os$
The host machine must satisfy the following minimum requirements:
- 2 network interfaces
- 8GB min, 16GB preferred - main memory
- 40G min, 80GB preferred disk space
- 2 CPU's Min, 4 preferred CPU's
Root access to the deployment host machine is required.
Verify the state of network interfaces. If using a VM spawned on
OpenStack as the host machine, the state of the second interface will be DOWN
on booting the VM.
ip addr show
Bring up the second network interface if it is down.
ip link set ens4 up
Verify if the second interface has an IP address.
ip addr show
This relies heavily on the OpenStack kolla-kubernetes project and in
particular the Bare Metal Deployment Guide:
https://docs.openstack.org/developer/kolla-kubernetes/deployment-guide.html
However support will be added to also install OpenStack with the
openstack-helm project.
The purpose of this tool, when there are so many others out there is:
1. Many tools don't support both Centos and Ubuntu with no input
from the user.
2. I like to play with versions of all the supporting tools, it
helps when users report issues when they upgrade say helm, or
docker, or kubernetes.
3. I like the output of my tool - it's succinct and easy to
follow. Plus the verbose mode is helpful for seeing all the output.
4. Contains a demo mode that walks the user through Kubernetes and OpenStack.
5. This tool verifies it's completeness by generating a VM in the
OpenStack Cluster.
6. Leaves the user with a working OpenStack Cluster with all the
basic services.
7. Very simple to run - just requires two NIC's
8. Lots of options to customize - even edit globals.yaml and cloud.yaml
1. mgmt_int (network_interface):
Name of the interface to be used for management operations.
The `network_interface` variable is the interface to which Kolla binds API
services. For example, when starting Mariadb, it will bind to the IP on the
interface list in the ``network_interface`` variable.
2. neutron_int (neutron_external_interface):
Name of the interface to be used for Neutron operations.
The `neutron_external_interface` variable is the interface that will be used
for the external bridge in Neutron. Without this bridge the deployment instance
traffic will be unable to access the rest of the Internet.
ubuntu@ubuntuk8s:~/os$ ../k8s/ko.py ens3 ens4 -cn -iv pike -dr rwellum -cni weave --logs
Kubernetes - Bring up a Kubernetes Cluster:
Linux Host Info: ('Ubuntu', '16.04', 'xenial')
Networking Info:
Management Int: ens3
Neutron Int: ens4
Management IP: 10.100.100.14
VIP Keepalive: 10.100.100.135
CNI/SDN: weave
Tool Versions:
Docker version: 1.13.1
Helm version: 2.7.2
K8s version: 1.8.4
Ansible version: 2.2.0.0
Jinja2 version: 2.8.1
OpenStack Versions:
Openstack version: pike (5.0.1)
Base image version: centos
Docker repo: rwellum
Other Info:
Logging enabled: True
Dev mode enabled: False
Create Network: True
Demo mode: False
Edit Cloud: False
Edit Globals: False
(01/15) Kubernetes - Installing base tools
(02/15) Kubernetes - Setup NTP
(03/15) Kubernetes - Turn off firewall and ISCSID
(04/15) Kubernetes - Creating Kubernetes repo, installing Kubernetes packages
(05/15) Kubernetes - Start docker and setup the DNS server with the service CIDR
(06/15) Kubernetes - Reload the hand-modified service files
(07/15) Kubernetes - Enable and start kubelet
(08/15) Kubernetes - Fix iptables to enable bridging
(09/15) Kubernetes - Deploying Kubernetes with kubeadm (Slow!)
You can now join any number of machines by running the following on each node as root:
kubeadm join --token 6a3654.dd285959559ce23b 10.100.100.14:6443 --discovery-token-ca-cert-hash sha256:6f73da0c963f6ae6adee483863ed7ceaf3b23730b0d2bfa5cac77b40746fa9b3
(10/15) Kubernetes - Load kubeadm credentials into the system
Note "kubectl get pods --all-namespaces" should work now
(11/15) Kubernetes - Wait for basic Kubernetes (6 pods) infrastructure
*Running pod(s) status after 20 seconds 2:6*
*Running pod(s) status after 60 seconds 5:6*
*All pods 6/6 are started, continuing*
(12/15) Kubernetes - Add API Server
(13/15) Kubernetes - Deploy pod network SDN using Weave CNI
Wait for all pods to be in Running state:
*02 pod(s) are not in Running state*
*01 pod(s) are not in Running state*
*All pods are in Running state*
(14/15) Kubernetes - Mark master node as schedulable by untainting the node
(15/15) Kubernetes - Test 'nslookup kubernetes' - bring up test pod
Wait for all pods to be in Running state:
*01 pod(s) are not in Running state*
*All pods are in Running state*
************************************
Kubernetes Cluster is up and running
************************************
**************************
Kolla - install OpenStack:
**************************
(01/45) Kolla - Overide default RBAC settings
(02/45) Kolla - Install and deploy Helm version 2.7.2 - Tiller pod
Wait for all pods to be in Running state:
*01 pod(s) are not in Running state*
*All pods are in Running state*
(03/45) Kolla - Helm successfully installed
(04/45) Kolla - Clone kolla-ansible
(05/45) Kolla - Clone kolla-kubernetes
(06/45) Kolla - Install kolla-ansible and kolla-kubernetes
(07/45) Kolla - Copy default kolla-ansible configuration to /etc
(08/45) Kolla - Copy default kolla-kubernetes configuration to /etc
(09/45) Kolla - Setup Loopback LVM for Cinder (Slow!)
(10/45) Kolla - Install Python Openstack Client
(11/45) Kolla - Generate default passwords via SPRNG
(12/45) Kolla - Create a Kubernetes namespace "kolla" to isolate this Kolla deployment
(13/45) Kolla - Label Nodes:
Label the AIO node as 'kolla_compute'
Label the AIO node as 'kolla_controller'
(14/45) Kolla - Modify global.yml to setup network_interface and neutron_interface
(15/45) Kolla - Add default config to globals.yml
(16/45) Kolla - Generate the default configuration
(17/45) Kolla - Set libvirt type to QEMU
(18/45) Kolla - Generate the Kubernetes secrets and register them with Kubernetes
(19/45) Kolla - Create and register the Kolla config maps
(20/45) Kolla - Build all Helm microcharts, service charts, and metacharts (Slow!)
(21/45) Kolla - Verify number of helm images
195 Helm images created
(22/45) Kolla - Create a version 5+ cloud.yaml
(23/45) Kolla - Helm Install service chart: \--'openvswitch'--/
Wait for all pods to be in Running state:
*02 pod(s) are not in Running state*
*01 pod(s) are not in Running state*
*All pods are in Running state*
(24/45) Kolla - Helm Install service chart: \--'mariadb'--/
Wait for all pods to be in Running state:
*02 pod(s) are not in Running state*
*01 pod(s) are not in Running state*
*All pods are in Running state*
(25/45) Kolla - Helm Install service chart: \--'rabbitmq'--/
(26/45) Kolla - Helm Install service chart: \--'memcached'--/
(27/45) Kolla - Helm Install service chart: \--'keystone'--/
(28/45) Kolla - Helm Install service chart: \--'glance'--/
(29/45) Kolla - Helm Install service chart: \--'cinder-control'--/
(30/45) Kolla - Helm Install service chart: \--'cinder-volume-lvm'--/
(31/45) Kolla - Helm Install service chart: \--'horizon'--/
(32/45) Kolla - Helm Install service chart: \--'neutron'--/
Wait for all pods to be in Running state:
*45 pod(s) are not in Running state*
*44 pod(s) are not in Running state*
*43 pod(s) are not in Running state*
*42 pod(s) are not in Running state*
*41 pod(s) are not in Running state*
*40 pod(s) are not in Running state*
*39 pod(s) are not in Running state*
*35 pod(s) are not in Running state*
*33 pod(s) are not in Running state*
*32 pod(s) are not in Running state*
*29 pod(s) are not in Running state*
*28 pod(s) are not in Running state*
*27 pod(s) are not in Running state*
*26 pod(s) are not in Running state*
*23 pod(s) are not in Running state*
*20 pod(s) are not in Running state*
*18 pod(s) are not in Running state*
*17 pod(s) are not in Running state*
*15 pod(s) are not in Running state*
*14 pod(s) are not in Running state*
*11 pod(s) are not in Running state*
*10 pod(s) are not in Running state*
*08 pod(s) are not in Running state*
*07 pod(s) are not in Running state*
*06 pod(s) are not in Running state*
*04 pod(s) are not in Running state*
*03 pod(s) are not in Running state*
*01 pod(s) are not in Running state*
*All pods are in Running state*
(33/45) Kolla - Helm Install service chart: \--'nova-control'--/
(34/45) Kolla - Helm Install service chart: \--'nova-compute'--/
Wait for all pods to be in Running state:
*23 pod(s) are not in Running state*
*22 pod(s) are not in Running state*
*20 pod(s) are not in Running state*
*19 pod(s) are not in Running state*
*17 pod(s) are not in Running state*
*16 pod(s) are not in Running state*
*15 pod(s) are not in Running state*
*14 pod(s) are not in Running state*
*13 pod(s) are not in Running state*
*12 pod(s) are not in Running state*
*11 pod(s) are not in Running state*
*10 pod(s) are not in Running state*
*09 pod(s) are not in Running state*
*08 pod(s) are not in Running state*
*04 pod(s) are not in Running state*
*All pods are in Running state*
(35/45) Kolla - Install Fluentd container
(36/45) Kolla - Final Kolla Kubernetes OpenStack pods for namespace kube-system:
NAME READY STATUS RESTARTS AGE
etcd-ubuntuk8s 1/1 Running 0 18m
kube-apiserver-ubuntuk8s 1/1 Running 0 18m
kube-controller-manager-ubuntuk8s 1/1 Running 1 18m
kube-dns-545bc4bfd4-hqr7f 3/3 Running 0 19m
kube-proxy-cwxqj 1/1 Running 0 19m
kube-scheduler-ubuntuk8s 1/1 Running 1 18m
tiller-deploy-5b9d65c7f-kl6hz 1/1 Running 0 16m
weave-net-rjnch 2/2 Running 0 17m
(37/45) Kolla - Final Kolla Kubernetes OpenStack pods for namespace kolla:
NAME READY STATUS RESTARTS AGE
cinder-api-75cf6b6c7b-fblq4 3/3 Running 0 6m
cinder-scheduler-0 1/1 Running 0 6m
cinder-volume-b42fj 1/1 Running 3 6m
glance-api-74896f7757-vgmlx 1/1 Running 0 6m
glance-registry-57886f64c8-zp9dn 3/3 Running 0 6m
horizon-cb5f44bc7-7mg88 1/1 Running 0 6m
iscsid-28qz5 1/1 Running 0 6m
keystone-b94f4dd4-zndqp 1/1 Running 0 6m
mariadb-0 1/1 Running 0 7m
memcached-f855d988d-8c89n 2/2 Running 0 6m
neutron-dhcp-agent-mj7pp 1/1 Running 0 6m
neutron-l3-agent-network-p9swg 1/1 Running 0 6m
neutron-metadata-agent-network-8krdq 1/1 Running 0 6m
neutron-openvswitch-agent-network-8mhgm 1/1 Running 0 6m
neutron-server-79f59644c5-2bc7l 3/3 Running 0 6m
nova-api-64c54b66df-8cjxq 3/3 Running 0 2m
nova-api-create-cell-r7kzw 1/1 Running 0 2m
nova-compute-zl5f9 1/1 Running 0 2m
nova-conductor-0 1/1 Running 0 2m
nova-consoleauth-0 1/1 Running 0 2m
nova-libvirt-k5z95 1/1 Running 0 2m
nova-novncproxy-6ff8fbb6f8-p9zx8 2/3 Running 0 2m
nova-scheduler-0 1/1 Running 0 2m
openvswitch-ovsdb-network-8xv9n 1/1 Running 0 7m
openvswitch-vswitchd-network-rn2mt 1/1 Running 0 7m
placement-api-d998c47c4-4qcbd 1/1 Running 0 2m
rabbitmq-0 1/1 Running 0 6m
tgtd-vzbf7 1/1 Running 0 6m
(38/45) Kolla - Create a keystone admin account and source in to it
(39/45) Kolla - Allow Ingress by changing neutron rules
(40/45) Kolla - Fix Nova, various issues, nova scheduler pod will be restarted
Wait for all pods to be in Running state:
*01 pod(s) are not in Running state*
*All pods are in Running state*
(41/45) Kolla - Configure Neutron, pull images
(42/45) Kolla - Create a demo VM in our OpenStack cluster
To create a demo image VM do:
. ~/keystonerc_admin; openstack server create --image cirros --flavor m1.tiny --key-name mykey --nic net-id=f62cdbad-b8f3-4f6d-8b0e-8f911b6b58a8 test
Kubernetes - Wait for VM demo1 to be in running state:
*Kubernetes - VM demo1 is not Running yet - wait 15s*
*Kubernetes - VM demo1 is not Running yet - wait 15s*
*Kubernetes - VM demo1 is not Running yet - wait 15s*
*Kubernetes - VM demo1 is Running*
(43/45) Kolla - Create floating ip
(44/45) Kolla - nova list to see floating IP and demo VM
+--------------------------------------+-------+--------+------------+-------------+------------------------+
| ID | Name | Status | Task State | Power State | Networks |
+--------------------------------------+-------+--------+------------+-------------+------------------------+
| 64383b56-7785-4d71-83f3-056e0117cf8d | demo1 | ACTIVE | - | Running | public1=192.168.10.166 |
+--------------------------------------+-------+--------+------------+-------------+------------------------+
(45/45) Kolla - To Access Horizon:
Point your browser to: 10.3.3.237
OS_PASSWORD=szkZ2hbOdeX6LgJXqnYXMNfsExVLGXDnyup67MMb
OS_USERNAME=admin
**************************************************************************
Successfully deployed Kolla-Kubernetes. OpenStack Cluster is ready for use
**************************************************************************