Filter out AI zones when discovering zones in the region. AI zones
do not have quota for general compute resources, so we should not provision
nodes there by default.
The changes done here will update the RHCOS 4.18 bootimage metadata and
address the following issues:
- OCPBUGS-64613: [4.18] coreos-boot-disk link not working with multipath on early boot
This change was generated using:
plume cosa2stream \
--target data/data/coreos/rhcos.json \
--distro rhcos \
--no-signatures \
--name 4.18-9.4 \
--url https://rhcos.mirror.openshift.com/art/storage/prod/streams \
x86_64=418.94.202511191518-0 \
aarch64=418.94.202511191518-0 \
s390x=418.94.202511191518-0 \
ppc64le=418.94.202511191518-0
Signed-off-by: Tiago Bueno <tiago.bueno@gmail.com>
Currently, the approach for removing OpenStack loadbalancers is to look
for appropriate tag (i.e. openshiftClusterID=<cluster_id>) and then
delete it. Issue with this approach is that no such tags are applied on
the loadbalancer resources, and also there is no such tag in description
field. Hence, deleteLoadBalancer function will give 0 results for
existing loadbalancers, as they have no such tag either in tags nor on
description.
With this patch, deleteLoadBalancer has been refactored to get all the
loadbalancers and filter out that resources which have ClusterID in
description, so that they will be safely deleted.
(cherry picked from commit fe679c0d9f)
Notes: there are other yaml-lint warnings but they've been there for a
long time. This commit only focuses on the errors introduced during 4.19
cycle.
Conflicts:
.golangci.yaml
NOTE(stephenfin): Changes to .golangci.yaml were removed because they
don't apply here.
- lb-default-stable: As a default load balancer is now being assigned when one is not provided this test needs to be updated to reflect that change
- lb-unmanaged: I made changes to how the defaults are set. If the load
balancer is user-managed VIPs will not automatically be assigned
anymore. This change needs to be reflected in this test by adding a
apiVIPs and ingressVIPs value to the install-config
Adds unit tests in platform_test.py to verify OpenStack platform defaults.
Covers cases such as:
- Assigning a default OpenShift-managed load balancer when none is specified
- Handling user-managed load balancers with and without VIPs
- Ensuring correct API and Ingress VIP assignments
-If no load balancer is provided, a default OpenShift load balancer is
now assigned automatically. This ensures proper handling and avoids
misconfigurations.
- Previously, when API and Ingress VIPs were not specified, default values were automatically
generated for user-managed load balancers.This was unintended behavior.
Now, if the user does not explicitly provide API and Ingress VIPs, a fatal error is thrown instead.
pkg/destroy/gcp/gcp.go:
** In the case where no error exists but the operation is nil when passed into
gcp.go:handleOperation the destroy process would cause a system panic. Now create an
error in this case so that the destroy process will continue and retry if there are more
pending items.
The changes done here will update the RHCOS 4.18 bootimage metadata and
address the following issues:
- OCPBUGS-62739: Need new CoreOS boot image with nmstate-2.2.50
This change was generated using:
plume cosa2stream \
--target data/data/coreos/rhcos.json \
--distro rhcos \
--no-signatures \
--name 4.18-9.4 \
--url https://rhcos.mirror.openshift.com/art/storage/prod/streams \
x86_64=418.94.202510081222-0 \
aarch64=418.94.202510081222-0 \
s390x=418.94.202510081222-0 \
ppc64le=418.94.202510081222-0
Signed-off-by: Tiago Bueno <tiago.bueno@gmail.com>
With multi-node installations the settings for /var/lib/etcd are
0700. This should be the same from SNO but the bootstrap-in-place
ignition is setting it to 0755.
Previously, we did not set this parameter on v6-primary stacks, wrongly
assuming that the Ironic container would pick up the right stack.
It wouldn't. Presenting with both IP addresses, the networking logic in
the container picks up the first one, which is usually IPv4.
This breaks virtual media on BMC's that cannot access IPv4 addresses.
(cherry picked from commit 4aa85eec23)
When there is OpenStack deployment, which doesn't have swift services,
or there are no containers used at all, listing them will fail the
playbook and leave exit code other than 0, which may interrupt CI.
With this commit, errors from listing containers will be ignored, which
will cover both cases.
** The credentials changed in AWS SDK V2. When the AWS_PROFILE
environment variable was set to an invalid value the install would not
proceed through the failure. The inention was to ask for the aws key and id information.
Now the information will be added to the survey, and the information will be added
to the config in .aws.
The hex based dns service is no longer available
and the replacement service does not take PRs
or looks interested in this feature.
If we want this unit test in the future we will need
to think through how to implement this.
Previously, since zones are extracted from map keys, order is not guaranteed.
This can lead to an issue where master CAPI machine manifest is configured
with a different subnet ID than MAPI machine manifest as they are
handled separately.
This commit ensures the zones are sorted by lexical order before
processing CAPI/MAP machine manifests so that zones are distributed in
the same order.
When using BYO subnets, users might define subnets in aws.vpc.subnets
and define the machineCIDRs in the installconfig from those subnets.
Previously, an SG is attached to api lb that only allows ingress to
tcp/22623 (MCS) from the only the first machineCIDR, which blocks master
nodes from reaching MCS on bootstrap node.
This commit adjusts the source for the SG to allow ingress from control
plane nodes via SG reference instead of relying on the machineCIDR field.
Installer can be used to provision Single Node OpenShift(SNO) only
on AWS, GCP, Azure, PowerVS and Baremetal platforms.
SNO is not supported on all on-prem platforms yet. Even on the
platforms that support SNO installs, the install method currently
supported are assisted/agent/ZTP.
Update test method TestBaremetalGeneratedAssetFiles() to create
a valid Baremetal IPI install config.
