Fix integration test timeout with VM image pull optimization #178
Conversation
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: yalzhang The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
yalzhang
force-pushed
the
update_pre_pull_image
branch
from
5ea1bb4 to
a616774
Compare
- Change VM imagePullPolicy to "IfNotPresent" for image caching - Increase test_attestation VM startup timeout from 300s to 900s - Update pre-pull-images.sh to use $APPROVED_IMAGE from Makefile - Delete the kubevirt images in the pre-pull-images.sh since the multi-arch load issues Fixes first integration test timeout while speeding up subsequent tests. Signed-off-by: Yalan Zhang <yalzhang@redhat.com>
yalzhang
force-pushed
the
update_pre_pull_image
branch
from
a616774 to
653ef93
Compare
| "quay.io/kubevirt/virt-launcher:${KV_VERSION}" | ||
| "quay.io/kubevirt/virt-handler:${KV_VERSION}" | ||
| "quay.io/kubevirt/virt-api:${KV_VERSION}" | ||
| "quay.io/kubevirt/virt-controller:${KV_VERSION}" | ||
| "quay.io/kubevirt/virt-operator:${KV_VERSION}" |
There was a problem hiding this comment.
multi-arch load issues
which issues?
| echo "Error: Failed to pull $IMAGE" | ||
| fi | ||
| kind load docker-image $IMAGE | ||
| docker pull "$IMAGE" && kind load docker-image "$IMAGE" |
There was a problem hiding this comment.
+1 but maybe
| docker pull "$IMAGE" && kind load docker-image "$IMAGE" | |
| docker pull "$IMAGE" && kind load docker-image "$IMAGE" || echo "Error: Failed to pull $IMAGE" |
| container_disk: Some(VirtualMachineTemplateSpecVolumesContainerDisk { | ||
| image: image.to_string(), | ||
| image_pull_policy: Some("Always".to_string()), | ||
| image_pull_policy: Some("IfNotPresent".to_string()), |
There was a problem hiding this comment.
Does this increase waits a lot? As per the k8s docs, even for Always, if the digest matches, the image does not get downloaded again. My worry is that some new image that doesn't receive a new tag is not used.
|
|
||
| test_ctx.info(format!("Waiting for VM {} to reach Running state", vm_name)); | ||
| virt::wait_for_vm_running(client, namespace, vm_name, 300).await?; | ||
| virt::wait_for_vm_running(client, namespace, vm_name, 900).await?; |
There was a problem hiding this comment.
Not for this PR but maybe we should have configurable multipliers for these.
2 participants
Fixes first integration test timeout while speeding up subsequent tests.