[e2e/node] update base image from busybox to agnhost

Since around 2024/11/27, a Node E2E test that repeatedly creates Pods
and checks whether they terminate with the expected exit code has
occasionally ended with exit code 2, making it flaky.

There have been no changes in kubelet that would modify the exit code
before 2024/11/27, so the cause is likely elsewhere.

Rather than an issue with the container runtime, it is possible that
the problem lies in the BusyBox base image. To narrow down the cause,
we will replace the base image from BusyBox to Agnhost.

Signed-off-by: Tsubasa Nagasawa <toversus2357@gmail.com>
This commit is contained in:
Tsubasa Nagasawa 2025-02-24 11:14:14 +09:00
parent 7f3d23554b
commit 2feecd46bd

View File

@ -610,7 +610,7 @@ func (s podFastDeleteScenario) Pod(worker, attempt int) *v1.Pod {
InitContainers: []v1.Container{
{
Name: "fail",
Image: imageutils.GetE2EImage(imageutils.BusyBox),
Image: imageutils.GetE2EImage(imageutils.Agnhost),
Command: []string{
"/bin/false",
},
@ -625,7 +625,7 @@ func (s podFastDeleteScenario) Pod(worker, attempt int) *v1.Pod {
Containers: []v1.Container{
{
Name: "blocked",
Image: imageutils.GetE2EImage(imageutils.BusyBox),
Image: imageutils.GetE2EImage(imageutils.Agnhost),
Command: []string{
"/bin/true",
},
@ -654,7 +654,7 @@ func (s podFastDeleteScenario) Pod(worker, attempt int) *v1.Pod {
Containers: []v1.Container{
{
Name: "fail",
Image: imageutils.GetE2EImage(imageutils.BusyBox),
Image: imageutils.GetE2EImage(imageutils.Agnhost),
Command: []string{
"/bin/false",
},