When add capacity blocks as nodepool, the nodepool doesn't turn to ready with error "failed to describe capacityReservation cr-011db017388a95436, code UnauthorizedOperation"
    

4.19.0-0.nightly-multi-2025-05-28-053245
    

always
    

    1.Purchase a trn1.32xlarge type capacity blocks with 1 instance
    2.Add it to nodepool
/usr/bin/hypershift create nodepool aws   --cluster-name  ac67f3afb8e9780db634   --name additional-ac67f3afb8e9780db634   --node-count 1   --instance-type trn1.32xlarge   --arch arm64   --release-image registry.build09.ci.openshift.org/ci-op-p2kzgzsy/release@sha256:129b08aee145503a30433c417d9d475f17d1999785007cf31a86e0e21952acc0  --render > /tmp/np.yaml  && /tmp/yq e -i '\''.spec.platform.aws.placement.capacityReservation = {"id": "cr-011db017388a95436", "marketType": "CapacityBlocks"}'\'' /tmp/np.yaml && oc apply -f /tmp/np.yaml
    3.
    

The nodepool does not go to ready.

  - lastTransitionTime: "2025-05-29T11:55:47Z"
    message: failed to describe capacityReservation cr-011db017388a95436, code UnauthorizedOperation
    observedGeneration: 1
    reason: AWSError
    status: "False"
    type: ValidPlatformConfig


    

Could add capacity blocks as nodepool.
    

Must-gather log
https://drive.google.com/drive/folders/1xg2r3mtG-vE92Zcz-ZvAYj-OvvY-_P8u?usp=drive_link 
Or you can get must gather log from
https://gcsweb-qe-private-deck-ci.apps.ci.l2s4.p1.openshiftapps.com/gcs/qe-private-deck/pr-logs/pull/openshift_release/64957/rehearse-64957-periodic-ci-openshift-openshift-tests-private-release-4.19-multi-nightly-aws-ipi-ovn-hypershift-capacity-reservation-gpu-guest-f999/1928035351234875392/artifacts/aws-ipi-ovn-hypershift-capacity-reservation-gpu-guest-f999/dump/