OpenStack Cloud High Availability Cluster Deployment using Kolla Ansible & External Shared Ceph Storage

OpenStack is a collection of open-source tools.

Different OpenStack cloud deployment methods are available.

Kolla Ansible is one of the recommended deployment method.

Deployment Overview: –

3 Node Ceph Cluster Deployed. Please check the following link – How to deploy the ceph storage cluster – https://kdinesh.in/deploying-ceph-storage-cluster-using-cephadm/

Machine	IP	Hostname	Comments
Deployment	172.90.0.33	OpenStackcloud	For Deploying and Managing the Cloud
Controller0	172.90.0.34	Controller0	VIP – 172.90.0.201, VIP DNS – OpenStack
Controller1	172.90.0.35	Controller1
Controller2	172.90.0.36	Controller2
Compute0	172.90.0.37	Compute0
Compute1	172.90.0.38	Compute1

Add a second nic for the controller and compute nodes (Don’t configure any IP for Second NIC). For the Production environment use network bonding.

Following Openstack services will be deployed

Keystone – Authentication service, Glance – Image Service

Nova – Compute Service, Neutron – Networking Service

Horizon – Dashboard Service, SKYLINE – New Dashboard

Heat – Orchestration Service, Magnum – Container Infra Service

Masakari – High Availability service, Trove – Database service

Sahara – Bigdata service – Hadoop …etc., Freezer – Backup and Disaster recovery service

Deployment Steps: –

OpenStack Version – Antelope

Create 6 VMs – One deployment,Three controllers, Two Computes.

Configure DNS records for the Controllers. Computes and VIP IP’s.

Check DNS Resolution

Download or copy the sshandntp script file and run it on the Deployment machine. You can download the script file using the following link – https://github.com/Dineshk1205/openstackmultinode/blob/main/sshandntp.sh

Note:- Update NTP and Hostname according to your Infra

Download or copy the openstackmultinodeprep script file and run it on the Deployment machine. You can download the script file using the following link – https://github.com/Dineshk1205/openstackmultinode/blob/main/openstackmultinodeprep.sh

In the Script update IP’S, NIC, Hostnames and other information according to your infra.

Download or Copy the cephconfigopenstack script file and run it on the Ceph0 machine. You can download the script file using the following link – https://github.com/Dineshk1205/openstackmultinode/blob/main/Cephconfigopenstack.sh

Script Configured– Pools, Passwordless ssh to the deployment node, and Copies cep. conf and ceph keyrings to the deployment machine.

You can check Ceph Dashboard – In the pools – backups, images, VMS, and volumes pools.

Download or copy the cephconfigopenstack script file and run it on the Deployment machine. You can download the script file using the following link – https://github.com/Dineshk1205/openstackmultinode/blob/main/OpenstackDeployment.sh

It will take 15 to 20 min based on your environment.

Once Deployment is completed. In /etc/kolla Directory check admin-openrc.sh file and view file(cat admin-openrc.sh) Copy the admin password.

Log in and check the Service status on Dashboard.

Update network CIDR details and other details on /usr/local/share/kola-ansible/init-runonce file and run .

Instance High Availability Testing: –

In Dashboard > Instance-ha. Create a segment.

Enter the Segment name, Select Recovery method, and click on create.

Add Hosts to Segment.

Add Compute 0 and compute1 as a host.

Create an instance.

ADD metadata – HA_Enabled True. Otherwise instance HA will not work.

Instance created and running Instance IP – 172.90.0.43. also, check the ceph dashboard.

Instance Running on Compute1

For testing the HA service , shutdown compute1 node because currently instance running on compute1 .

The instance evacuated to compute0.

The instance evacuation process may take 5 to 10 min.

You can see the instance overview – instance running on compute0.