kolla ansible deploy openstack high availability cluster

Keywords: network ansible pip OpenStack

kolla ansibel deploy openstack high availability cluster

reference resources: https://docs.openstack.org/kolla-ansible/latest/user/quickstart.html

kolla installation node requirements:

  • 2 network interfaces
  • 8GB main memory
  • 40GB disk space

For the installation of kolla, the target machine is required to have two network cards. Add a new network card in vmware workstation:

  • ens33, NAT mode, management network, tenant network can be reused with the network, and static IP can be configured normally.
  • ens37, bridge mode, does not need to configure the IP address. In fact, this is used by the br ex binding of neutron. The virtual machine accesses the external network through this network card.

Node planning

  • Three control nodes (network node and control node are combined), one computing node + one storage node (cinder lvm)
  • The first control node is also the kolla ansible deployment node
  • The control node and the computing node have two network cards, the storage node has one network card, the network card 1NAT mode serves as the management network, the network card 2 bridge mode serves as the external network.
  • The operating system uses CentOS7.8 minimal, and the CPU memory of all nodes is configured as 2C4G.

Node and IP planning

Node name IP address role
control01 ens33: 192.168.93.60
ens37: 		Control node
control02 ens33: 192.168.93.61
ens37: 		Control node
control03 ens33: 192.168.93.62
ens37: 		Control node
compute01 ens33: 192.168.93.63
ens37: 		Calculation node
storage01 ens33: 192.168.93.64
Storage node

If you need to add an additional disk to the storage01 node to enable the cinder, take / dev/sdb as an example, and execute on the storage01 node

pvcreate /dev/sdb
vgcreate cinder-volumes /dev/sdb

Note that the volume group name is cinder volumes, and the default is globals.yml agreement.

[root@kolla ~]# cat /etc/kolla/globals.yml | grep cinder_volume_group
#cinder_volume_group: "cinder-volumes"

All nodes configure IP addresses according to the plan

[root@localhost ~]# cat /etc/sysconfig/network-scripts/ifcfg-ens33
TYPE=Ethernet
PROXY_METHOD=none
BROWSER_ONLY=no
BOOTPROTO=none
DEFROUTE=yes
IPV4_FAILURE_FATAL=no
IPV6INIT=yes
IPV6_AUTOCONF=yes
IPV6_DEFROUTE=yes
IPV6_FAILURE_FATAL=no
IPV6_ADDR_GEN_MODE=stable-privacy
NAME=ens33
UUID=752fa178-bbb6-4ab3-84ee-aa86a34a16b4
DEVICE=ens33
ONBOOT=yes
IPADDR=192.168.93.61
PREFIX=24
GATEWAY=192.168.93.2
DNS1=114.114.114.114
DNS2=8.8.8.8
IPV6_PRIVACY=no
[root@localhost ~]# cat /etc/sysconfig/network-scripts/ifcfg-ens37
TYPE=Ethernet
PROXY_METHOD=none
BROWSER_ONLY=no
BOOTPROTO=none
DEFROUTE=yes
IPV4_FAILURE_FATAL=no
IPV6INIT=yes
IPV6_AUTOCONF=yes
IPV6_DEFROUTE=yes
IPV6_FAILURE_FATAL=no
IPV6_ADDR_GEN_MODE=stable-privacy
NAME=ens37
UUID=c1513435-9706-3d76-802b-00c4e901ee82
DEVICE=ens37
ONBOOT=yes
IPV6_PRIVACY=no

Basic configuration

Without special instructions, all the following operations are performed in the kolla ansible deployment node, namely the control01 node.

Installation dependency

yum install -y python-devel libffi-devel gcc openssl-devel libselinux-python

Install Ansible

yum install -y ansible

Configure alicloud pip source

mkdir ~/.pip
cat > ~/.pip/pip.conf << EOF 
[global]
trusted-host=mirrors.aliyun.com
index-url=https://mirrors.aliyun.com/pypi/simple/
EOF

Install kolla ansible

Correspondence between kolla version and openstack version: https://releases.openstack.org/teams/kolla.html

yum install -y epel-release
yum install -y python-pip
pip install -U pip
pip install kolla-ansible==9.1.0 --ignore-installed PyYAML

Copy kolla ansible configuration

mkdir -p /etc/kolla
chown $USER:$USER /etc/kolla

##Copy globals.yml and passwords.yml
cp -r /usr/share/kolla-ansible/etc_examples/kolla/* /etc/kolla

##Copy all-in-one and multinode inventory files
cp /usr/share/kolla-ansible/ansible/inventory/* .

Modify ansible configuration file

$ vim /etc/ansible/ansible.cfg
[defaults]
host_key_checking=False
pipelining=True
forks=100

Modify the multi node inventory file, and other defaults are OK

$ cat multinode
[control]
# These hostname must be resolvable from your deployment host
control01
control02
control03

[network]
control01
control02
control03

[compute]
compute01

[monitoring]
control01

[storage]
storage01
...

Configure ssh security free

ssh-keygen
ssh-copy-id 192.168.93.61
ssh-copy-id 192.168.93.62
ssh-copy-id 192.168.93.63
ssh-copy-id 192.168.93.64
ssh-copy-id 192.168.93.65

Configure host name (all nodes)

hostnamectl set-hostname control01
hostnamectl set-hostname control02
hostnamectl set-hostname control03
hostnamectl set-hostname compute01
hostnamectl set-hostname storage01

Configure host name resolution (kolla will configure host name resolution in pre configuration, and only operate in deployment node here)

cat > /etc/hosts <<EOF
192.168.93.61 control01
192.168.93.62 control02
192.168.93.63 control03
192.168.93.64 compute01
192.168.93.65 storage01
EOF

Check whether the inventory configuration is correct. Execute:

ansible -i multinode all -m ping

Generate kolla password

kolla-genpwd

Modify keystone_admin_password can be modified to user-defined, which is used in the login dashboard

sed -i 's#keystone_admin_password:.*#keystone_admin_password: kolla#g' /etc/kolla/passwords.yml 

$ cat /etc/kolla/passwords.yml | grep keystone_admin_password
keystone_admin_password: kolla

Modify global profile globals.yml

cp /etc/kolla/globals.yml{,.bak}
cat >> /etc/kolla/globals.yml <<EOF

#version
kolla_base_distro: "centos"
kolla_install_type: "binary"
openstack_release: "train"

#vip
kolla_internal_vip_address: "192.168.93.200"

#docker registry
docker_registry: "registry.cn-shenzhen.aliyuncs.com"
docker_namespace: "kollaimage"

#network
network_interface: "ens33"
neutron_external_interface: "ens37"
neutron_plugin_agent: "openvswitch"
enable_neutron_provider_networks: "yes"

#storage
enable_cinder: "yes"
enable_cinder_backend_lvm: "yes"

#virt_type
nova_compute_virt_type: "qemu"
EOF

Modify the official installation yum source of docker as Alibaba cloud yum source, and configure docker image acceleration. Specify to use Alibaba cloud image acceleration.

$ vim /usr/share/kolla-ansible/ansible/roles/baremetal/defaults/main.yaml
docker_yum_url: "https://mirrors.aliyun.com/docker-ce/linux/{{ ansible_distribution | lower }}"
docker_custom_config: {"registry-mirrors": ["https://uyah70su.mirror.aliyuncs.com"]}

Deploy openstack

Easy is executed when bootstrap servers are running_ The install PIP command installs pip. If the network is slow and the task is stuck here all the time, you can manually install PIP in all nodes in advance (optional).

mkdir ~/.pip
cat > ~/.pip/pip.conf << EOF 
[global]
trusted-host=mirrors.aliyun.com
index-url=https://mirrors.aliyun.com/pypi/simple/
EOF

yum install -y epel-release
yum install -y python-pip
pip install -U pip

Start deploying openstack

kolla-ansible -i ./multinode bootstrap-servers

#Deployment check
kolla-ansible -i ./multinode prechecks

#Pull image
kolla-ansible -i ./multinode pull

#Execute deployment
kolla-ansible -i ./multinode deploy

OpenStack client

Generate an openrc file with administrator user credentials set

kolla-ansible post-deploy

cat /etc/kolla/admin-openrc.sh

Using docker as the openstack client

docker pull registry.cn-shanghai.aliyuncs.com/davycloud/centos-binary-openstack-base:train

The kolla ansible deployment node starts a temporary container:

docker run -d --name client \
  --restart always \
  -v /etc/kolla/admin-openrc.sh:/admin-openrc.sh:ro \
  -v /usr/share/kolla-ansible/init-runonce:/init-runonce:rw \
  registry.cn-shenzhen.aliyuncs.com/kollaimage/centos-binary-openstack-base:train sleep infinity

Enter the container to execute the openstack command

[root@control01 ~]# docker exec -it client bash
()[root@7863d16d6839 /]#  source /admin-openrc.sh
()[root@7863d16d6839 /]# openstack service list
+----------------------------------+-------------+----------------+
| ID                               | Name        | Type           |
+----------------------------------+-------------+----------------+
| 2ab6a571b71444adb69f10d18a876c55 | keystone    | identity       |
| 4b0ea157a8dc4ef5adc720dda56902ce | cinderv2    | volumev2       |
| 57b38f38444949f98e6f3ed46b5a24e6 | cinderv3    | volumev3       |
| 84431d0e3727426f82faf4a9bb898045 | heat        | orchestration  |
| 8a8675bc425a45ac9e83140b18a8c864 | nova_legacy | compute_legacy |
| b54c859fa8214ca68bb5002e5e58ade2 | glance      | image          |
| b9adda688c024c56aace1902340cadf7 | heat-cfn    | cloudformation |
| c301474681f74271a39b74d29899d7ee | nova        | compute        |
| c3fb659dea224e3ab8655f77976cc98f | placement   | placement      |
| edfa32d4099a435397497b19b055fd10 | neutron     | network        |
+----------------------------------+-------------+----------------+

kolla provides a script for quickly creating demo instances, modifying the external network configuration of init RunOnce sample scripts.

()[root@7863d16d6839 /]#  vi init-runonce
# This EXT_NET_CIDR is your public network,that you want to connect to the internet via.
ENABLE_EXT_NET=${ENABLE_EXT_NET:-1}
EXT_NET_CIDR=${EXT_NET_CIDR:-'192.168.1.0/24'}
EXT_NET_RANGE=${EXT_NET_RANGE:-'start=192.168.1.200,end=192.168.1.250'}
EXT_NET_GATEWAY=${EXT_NET_GATEWAY:-'192.168.1.1'}

()[root@7863d16d6839 /]#  bash init-runonce

The script will first download the cirros image from github. If the network is slow, you can download it to the cache directory in advance

docker exec -it client mkdir -p /opt/cache/files/
wget https://github.com/cirros-dev/cirros/releases/download/0.4.0/cirros-0.4.0-x86_64-disk.img
docker cp cirros-0.4.0-x86_64-disk.img client:/opt/cache/files/

Run a cirros instance according to the prompt after the script is executed

openstack server create \
    --image cirros \
    --flavor m1.tiny \
    --key-name mykey \
    --network demo-net \
    demo1

Visit horizon

Find vip address in 3 control nodes

[root@control03 ~]# ip a | grep ens33
2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
    inet 192.168.93.63/24 brd 192.168.93.255 scope global noprefixroute ens33
    inet 192.168.93.200/32 scope global ens33

Browser access: http://192.168.93.200 , user name and password are in admin-openrc.sh It can be found in. The default is admin/kolla

cat /admin-openrc.sh

View network topology

Assign floating IP to an instance

Connect to the instance and verify the access to the Internet. The default account password of cirros image is cirros/gocubsgo

[root@control01 ~]# ssh cirros@192.168.1.224
cirros@192.168.1.224's password:
$
$
$ ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
       valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc pfifo_fast qlen 1000
    link/ether fa:16:3e:e4:16:00 brd ff:ff:ff:ff:ff:ff
    inet 10.0.0.74/24 brd 10.0.0.255 scope global eth0
       valid_lft forever preferred_lft forever
    inet6 fe80::f816:3eff:fee4:1600/64 scope link
       valid_lft forever preferred_lft forever
$
$ ping -c 4 www.openstack.org
PING www.openstack.org (104.20.110.33): 56 data bytes
64 bytes from 104.20.110.33: seq=1 ttl=53 time=178.064 ms
64 bytes from 104.20.110.33: seq=2 ttl=53 time=177.797 ms
64 bytes from 104.20.110.33: seq=3 ttl=53 time=178.392 ms

--- www.openstack.org ping statistics ---
4 packets transmitted, 3 packets received, 25% packet loss
round-trip min/avg/max = 177.797/178.084/178.392 ms

Profile directory

Directory of configuration files of each component / etc/kolla/

Log file directory of each component / var/log/kolla/

Clean up cluster

kolla-ansible destroy --include-images --yes-i-really-really-mean-it
vgremove cinder-volume

Posted by jimmy patel on Sat, 13 Jun 2020 23:52:24 -0700

Hot Keywords

©2024 Programmer Group