On 10/11/2022 00:57, Camila Granella wrote:

Hello everyone,

This is a friendly reminder of CentOS CI changes' current status and upcoming action items (check [0]) in addition to a quick update on our new OpenShift instance.

According to our previous communications [0], we are approaching the last phase of the changes that have been implemented to CentOS CI. We are currently closing phase 2, which began in October. In December, we will move to *phase 3*:

*Phase 3 - Decommission*

• Legacy/compatibility API deprecated and requests (even for EC2 instances) will no longer be accepted
• All tenants that opted in will be using only EC2 for aarch64/x86_64 and on-premise cloud for ppc64le

For a full description of the project phases, you can check [0].

Currently, our Infra & Releng team is working to deploy our new OpenShift instance (AWS) in production after some improvements that have been done in staging. When OpenShift gets fully set up, the team will work to migrate the Pagure project. This will then be documented and passed on to you (users) as a guide in order to *facilitate the workload migration that is expected to be done on your side*. We are expecting this to becompleted by the end of the year.

We will be sharing further information on this at the beginning of Q1CY2023. A reasonable timeline will be provided so that you have enough bandwidth to work on the migration of your workloads.

Please, feel free to reach out to us in case you have any questions or concerns.

Regards,

Small update on something we just changed in existing setup (before being able to deploy the Openshift cluster in AWS) : we initially decided to use the .cloud.ci.centos.org zones for everything hosted in aws but we were using it internally for some ppc64le VMs. Starting from now (was just pushed in git and applied by ansible), ppc64le nodes will be using the .one. (for opennebula) part in their fqdn. That means that for x86_64 and aarch64, nothing changes (it was in the .pool.ci.centos.org route53 hosted zone) but for ppc64le you'll just see a fqdn name change (nothing else, same cloud images provisioned but different names assigned, while the pool will be drained and reprovisionned)

WRT openshift, we have it fully deployed in .stg. (all under ansible control) So that will be now easy to reprovision the real CI prod cluster (in its own dedicated VPC, using EFS as storage for PersistentVolumes in openshift) in the following days/weeks. This new one will also get all groups/users synced directly from FAS so once a project is accepted, it will be just a matter or creating a specific group and that's it. People (like in SIGs) with chair/sponsor rights will be able to grant/remove access themselves through https://accounts.centos.org