[upgrade test] set common CPU Model for aws and azure (#3314)

* [upgrade test] set common CPU Model for aws and azure

When performing a live migration, virt-controller picks a target node that has all of the cpu features as the source node.
if the cluster nodes have different CPU features, the VMIM will be stuck at Pending.
setting defaultCPUModel for AWS and Azure to ensure all of the cpu features in all nodes are aligned and live migration will be done.

Signed-off-by: Oren Cohen <[email protected]>

* Fix issue in AWS lanes

If the worker is with ARM architecture, CI failes to build the index
image, as it explicitly uses AMD64 base image.

This PR change the base image to be a multi-arch manifest, to allow ARM
build.

Signed-off-by: Nahshon Unna-Tsameret <[email protected]>

---------

Signed-off-by: Oren Cohen <[email protected]>
Signed-off-by: Nahshon Unna-Tsameret <[email protected]>
Co-authored-by: Nahshon Unna-Tsameret <[email protected]>

Author: orenc1

deploy/index-image/Dockerfile.bundle.ci-index-image-upgrade

@@ -1,4 +1,4 @@
-FROM quay.io/fedora/fedora:40-x86_64 AS builder
+FROM quay.io/fedora/fedora:42 AS builder
 ARG INITIAL_VERSION=1.15.0
 ARG INITIAL_VERSION_SED="1\.15\.0"
 ARG TARGET_VERSION=100.0.0

deploy/olm-catalog/Dockerfile.bundle.ci-index-image-upgrade

@@ -1,4 +1,4 @@
-FROM quay.io/fedora/fedora:40-x86_64 AS builder
+FROM quay.io/fedora/fedora:42 AS builder
 ARG INITIAL_VERSION=1.15.0
 ARG INITIAL_VERSION_SED="1\.15\.0"
 ARG TARGET_VERSION=100.0.0

hack/consecutive-upgrades-test.sh

@@ -217,6 +217,9 @@ END
 
 CSV=$( ${CMD} get csv -o name -n ${HCO_NAMESPACE} | grep "kubevirt-hyperconverged-operator")
 
+# Patch the default CPU model to ensure a successful live migration
+${CMD} patch hco kubevirt-hyperconverged -n ${HCO_NAMESPACE} --type=json -p='[{"op": "add", "path": "/spec/defaultCPUModel", "value": "Westmere"}]'
+
 Msg "operator conditions before upgrade"
 source ./hack/check_operator_condition.sh
 KUBECTL_BINARY=${CMD} INSTALLED_NAMESPACE=${HCO_NAMESPACE} printOperatorCondition "${INITIAL_VERSION}"
@@ -237,6 +240,8 @@ upgrade $MID_VERSION $TARGET_VERSION $OO_LAST_BUNDLE
 
 Msg "make sure that we don't have outdated VMs"
 
+${CMD} get vmim -n ${VMS_NAMESPACE} -o yaml
+
 INFRASTRUCTURETOPOLOGY=$(${CMD} get infrastructure.config.openshift.io cluster -o json | jq -j '.status.infrastructureTopology')
 UPDATE_METHODS=$(${CMD} get hco ${HCO_RESOURCE_NAME} -n ${HCO_NAMESPACE} -o jsonpath='{.spec .workloadUpdateStrategy .workloadUpdateMethods}')