Skip to content

Conversation

markgoddard
Copy link
Contributor

  • CI: Update aio jobs to run on ARC
  • Support credentials for custom DNF repositories
  • CI: Use Ark package repositories to install packages
  • CI: Switch from sms-lab-release to openstack cloud
  • CI: Update aio VM flavor, network and subnet for Leafcloud
  • CI: Update GitHub action versions for node20
  • Revert hashicorp/setup-terraform action to v2 due to failure
  • CI: Revert to ens3 interface for Leafcloud
  • Add Apt auth configuration
  • CI: Increase flavor to en1.large - some jobs were failing Tempest jobs
  • CI: Run container image build job on ARC cluster
  • Add package repository credentials to container image build
  • Don't verify Apt repo CA initially when using HTTPS in container build

MaxBed4d and others added 15 commits February 7, 2024 14:24
The jobs now run in containers, so some changes are necessary.
This allows us to use packages from authenticated repositories such as
those on Ark. This patch depends on
https://review.opendev.org/c/openstack/kayobe/+/908142/
Previously we were using Test Pulp on SMS lab, but this is out of
action. Switching to Ark allows CI jobs to run on Leafcloud (or anywhere
with Internet access).
This more generic name should be more portable.
node16 support is deprecated and will be dropped.
Leafcloud is not using the q35 machine type yet.
This allows us to use packages from authenticated repositories such as
those on Ark. This patch depends on
https://review.opendev.org/c/openstack/kayobe/+/908413
The Actions Runner Controller (ARC) cluster runs GitHub Actions jobs on
a Kubernetes cluster. This allows us to dynamically provision resources
for our CI workflows.

This change switches the container image build jobs to use the ARC
cluster. The job now runs in a container, with access to a Docker in
Docker container for building images. This allows us to skip the host
configuration part of the workflow.

See https://github.com/stackhpc/ARC-Installer/
This is necessary to access package repositories hosted on Ark.
The Ubuntu base image doesn't contain the ca-certificates package, so we
can't verify an HTTPS package mirror to download the ca-certificates
package. Using the upstream repos may result in a version conflict when
we switch to another mirror.
CI: Use ARC runners for aio jobs, move CI to Leafcloud
CI: Use ARC runners for container image build jobs
@markgoddard markgoddard requested a review from a team as a code owner February 9, 2024 09:17
@markgoddard markgoddard self-assigned this Feb 9, 2024
@Alex-Welsh
Copy link
Member

There I was thinking our days of TF errors were over 😞

We sometimes hit temporary failures during VM creation. Adding a retry
may help to keep things running.
This job is not present in yoga, so was not tested there.
@markgoddard
Copy link
Contributor Author

This is mostly working now, just waiting on upgrade jobs. The check container image tags job needs some fixes because the stackhpc-dev container repositories on Ark are "push repositories". It's non-voting though, so I suggest we merge and fix that later.

Copy link
Member

@Alex-Welsh Alex-Welsh left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Couple of threads to follow up on but LGTM as a fix for our AIO issues

@markgoddard markgoddard merged commit 59ce72a into stackhpc/zed Feb 9, 2024
@markgoddard markgoddard deleted the zed-yoga-merge branch February 9, 2024 14:05
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants