[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: OpenShift Origin on AWS

Answers inline

On Wed, 10 Oct 2018 at 12:08 am, David Conde <david donedeal ie> wrote:
We have upgraded from the 3.6 reference architecture to the 3.9 aws playbooks in openshift-ansible. There was quite a bit of work in getting nodes ported into the scaling groups. We have upgraded our masters to 3.9 with the BYO playbooks but have not ported them to use scaling groups yet.

Oh wow, I figured we’d have to blow away the cluster to get the scaling groups. Was most of the work on the masters? Because I presume you just deleted the 3.6 nodes and recreated them in the scaling group? 

We'll be sticking with the aws openshift-ansible playbooks in the future over the reference architecture so that we can upgrade easily. 

We came to that conclusion too, it seems like it’ll be more likely to be supported for longer. 

On Tue, Oct 9, 2018 at 1:29 PM Joel Pearson <japearson agiledigital com au> wrote:
There is cloud formation templates as part of the 3.6 reference architecture. But that is now deprecated. I’m using that template at a client site and it worked fine (I’ve adapted it to work with 3.9 by using a static inventory as we didn’t want to revisit our architecture from scratch). We did customise it a fair bit though.


Here is an example of a jinja template that outputs a cloud formation template.

However, you can’t use the playbook as is for 3.9/3.10 because openshift-ansible has breaking changes to the playbooks.

For some reason the new playbooks for 3.9/3.10 don’t use cloud formation, but rather use the amazon ansible plugins instead and directly interact with AWS resources:


That new approach is pretty interesting though as it uses prebuilt AMIs and auto-scaling groups, which make it very quick to add nodes.

Hopefully some of that is useful to you.

On Tue, 9 Oct 2018 at 9:42 pm, Peter Heitman <peter heitman us> wrote:
Thank you for the reminder and the pointer. I know of that document but was too focused on searching for a CloudFormation template. I'll go back to the reference architecture which I'm sure will answer at least some of my questions.

On Sun, Oct 7, 2018 at 4:24 PM Joel Pearson <japearson agiledigital com au> wrote:
Have you seen the AWS reference architecture? https://access.redhat.com/documentation/en-us/reference_architectures/2018/html/deploying_and_managing_openshift_3.9_on_amazon_web_services/index#
On Tue, 2 Oct 2018 at 3:11 am, Peter Heitman <peter heitman us> wrote:
I've created a CloudFormation Stack for simple lab-test deployments of OpenShift Origin on AWS. Now I'd like to understand what would be best for production deployments of OpenShift Origin on AWS. In particular I'd like to create the corresponding CloudFormation Stack.

I've seen the Install Guide page on Configuring for AWS and I've looked through the RedHat QuickStart Guide for OpenShift Enterprise but am still missing information. For example, the RedHat QuickStart Guide creates 3 masters, 3 etcd servers and some number of compute nodes. Where are the routers (infra nodes) located? On the masters or on the etcd servers? How are the ELBs configured to work with those deployed routers? What if some of the traffic you are routing is not http/https? What is required to support that?

I've seen the simple CloudFormation stack (https://sysdig.com/blog/deploy-openshift-aws/) but haven't found anything comparable for something that is closer to production ready (and likely takes advantage of using the AWS VPC QuickStart (https://aws.amazon.com/quickstart/architecture/vpc/). 

Does anyone have any prior work that they could share or point me to? 

Thanks in advance,

Peter Heitman

users mailing list
users lists openshift redhat com
users mailing list
users lists openshift redhat com

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]