Skip to main content

Automated OSP deployments with Tripleo Quickstart

In this article I'm going to show a method for automating OSP (RedHat OpenStack platform) deployments. These automated deployments can be very useful for CI, or simply to experiment and test with the system.

Components involved

  • TOAD: set of playbooks to deploy Jenkins, jenkins-job-builder and an optional ELK stack. This will install a ready to use system with all the preconfigured jobs (including OSP10 deployments and image building).
  • TOAD jenkins-jobs: A set of job templates and macros, using jenkins-job-builder syntax, that get converted into Jenkins jobs for building the OSP base images and for deploying the system.
  • TOAD job-configs: A set of job configurations to be used along with the jenkins-jobs repo. It provides a set of basic configs to build OpenStack using RDO or OSP.
  • TripleO quickstart: set of ansible playbooks, used for building images and for RDO/OSP deployments.

System requirements

  • One VM to run a Jenkins master + nginx provided by TOAD. If you want to deploy additional ELK stack, you will need extra VMS for ElasticSearch, LogStash and Kibana.
  • For virtualized OSP deployments: one baremetal server, with at least 16GB of RAM and 60GB of disk, multi-core. This will act as jenkins slave, and will hold the virtualized undercloud, controller and compute nodes.
  • For baremetal OSP deployments: same requirement as the virtualized one, to hold jenkins slave + undercloud. Two additional baremetal servers, with at least 8GB of RAM, to act as controller and compute nodes.
Please note that the servers acting as slaves will need to have a valid RedHat subscription.

Deployment steps

1. Setup the system

If you rely on TOAD, the initial steps will be completely automated. You will need to follow TOAD documentation for the initial deployment. This will create a Jenkins/nginx VM for you, and will enroll one slave as well, configuring that properly. It will also populate all the Jenkins jobs for you, leaving a system ready to use.
Please note that before running the playbook, you need to create ~/.ansible/vars/toad_vars.yml file properly. The following settings are needed for this use case:
  • rhn_subscription_username: <<your RHN username>>
  • rhn_subscription_password: <<your RHN password>>
  • rhn_subscription_pool_id: <<your RHN pool id>> 
  • slave_mirror_sync: true
This will register the Jenkins slave into the RedHat network, and will create a local mirror, with all the needed RedHat repos, that you can use later on your image builds and deploys.

After the deploy has finished, you can access Jenkins master on the configured IP/hostname, and you will see all the jobs available:

2. Get base images

For OSP 10 deployments, we better need to rely on RHEL 7.3 guest image. However, the packages are not still providing this version, so we'll need to download manually.
In order to do it, open a browser and login into access.redhat.com with your credentials. Then access to https://access.redhat.com/downloads/content/69/ver=/rhel---7/7.3/x86_64/product-software and copy the KVM guest image link into your clipboard.
SSH into your Jenkins slave and execute the following commands:
mkdir /opt/rhel_guest_images
wget -O /opt/rhel_guest_images/rhel-guest-image-7.3.qcow2 <<link_for_7.3_guest>> 

This will make the base image available for the OSP image builds. This process only needs to be done once for each slave you register.

3. Execute image build job

In order to deploy OSP successfully, we first need to generate the undercloud and overcloud base images. This can be done executing a pre-configured job, that takes the previous 7.3 guest image, and composes the desired images, to be reused later by the OSP deployments.
Please note that this job needs to be executed only once, per each enrolled slave on the system.
Running the job is as easy as to login into the Jenkins system, and clicking execute on the build job you need. In our case, you will need to execute the oooq-newton-osp10-build-cira job:


This job will use TripleO quickstart , along with some playbooks that extend it, to compose the undercloud.qcow2 base image. Overcloud and ironic-python-agent images are embedded inside undercloud image, under the /home/stack directory.
When the job finishes, it moves the final images under /home/stack/images/osp<<version>> directory, ready to be reused by the deployment job.

3. Execute deployment job

Once the base images are ready, it's time to launch OSP deployments. Running it is as simple as executing the right job. In that case, for an OSP 10 deployment you can run the oooq-osp10-osp-deploy-cira job.
This will use TripleO quickstart for the deploy, and will rely on the previously generated base images, as well as on the local repo that was created,  to launch a successful OpenStack deployment.
After it finishes, undercloud and overcloud nodes will be available.

4. Accessing the undercloud and overcloud

You then can  SSH into your slave, and switch to stack user. The undercloud can be accessed with the command:
ssh -F $HOME/quickstart/ssh.config.ansible undercloud
Once on the undercloud, you can source the credentials file:
source $HOME/stackrc
 And then execute OpenStack commands. You can get the overcloud nodes by executing nova list command. This will get a list of nodes involved, with their ips:

nova list
+--------------------------------------+-------------------------+--------+------------+-------------+------------------------+
| ID                                   | Name                    | Status | Task State | Power State | Networks               |
+--------------------------------------+-------------------------+--------+------------+-------------+------------------------+
| 3d4a79d1-53ea-4f32-b496-fbdcbbb6a5a3 | overcloud-controller-0  | ACTIVE | -          | Running     | ctlplane=192.168.24.16 |
| 4f8acb6d-6394-4193-a6c6-50d8731fad7d | overcloud-novacompute-0 | ACTIVE | -          | Running     | ctlplane=192.168.24.8  |
+--------------------------------------+-------------------------+--------+------------+-------------+------------------------+

Then you can access the nodes using heat-admin user:

ssh heat-admin@192.168.124.16

5. Accessing the logs

Please note that you can access all the logs for the jobs into the same Jenkins VM. In order to get the logs, you can access to http://<<jenkins_configured_hostname>>/logs. That will show a list of jobs, and for each one, there will be folders storing complete set of logs for each build:


So that's it! After following these steps you will have a working OSP on your systems, that you can use for experimenting, CI, or any other purposes. Enjoy!

Comments

Popular posts from this blog

Describing TOAD - TripleO automated deployer

An overview of the TOAD framework and the advantagesWhat is TOAD? Fully automated deployment using Ansible (single command spin up)Main goal: to automate OSP deployments for continuous integration (CI) and development purposesTOAD is a CI framework using off-the-shelf components that many partners are familiar with:JenkinsJenkins Job Builder (JJB): http://docs.openstack.org/infra/jenkins-job-builder/TripleO Quickstart (oooq): https://www.rdoproject.org/tripleo/Optional ELK Stack (ElasticSearch, Logstash, Kibana) Its core component is TripleO Quickstart, used for TripleO upstream testingFully customizable with oooq settings; can be extendedDeploy environments with one click; trash after finished Of course it’s open source! :)https://github.com/redhat-nfvpe/toadhttps://github.com/redhat-nfvpe/jenkins-jobshttps://github.com/redhat-nfvpe/job-configs What components make up TOAD?Requirements to install TOADTwo different use cases: virtualized and baremetalOnly one Jenkins+Nginx VM needed t…

Start using whole disk images with TripleO

What are the differences between flat partition image and whole disk image? In order to understand this article, you first need to know what a flat partition image and a whole disk image are, and the differences between each other.
flat partition image: disk image that just contains all the desired content in a filesystem, but does not carry any information about partitions on it, and it does not include a bootloader. In order to boot from a whole disk image, the kernel and ramdisk images need to be passed independently when booting, relying on an external system to mount.whole disk image: image that contains all the information about partitions, bootloaders... as well as all the desired content. It can boot independently, without the need of external kernels or systems to mount it. Right now, OpenStack Ironic  supports both kind of images, but OpenStack TripleO was only supporting flat partition images.

TripleO added support for whole disk images Since python-tripleoclient 5.6.0 ver…