Both images are now sane multi-architecture images and should fix the
kube-proxy container image in the same way.
Signed-off-by: Sascha Grunert <mail@saschagrunert.de>
In downstream contexts, it's extremely useful to be able to combine
all the "testable" images in Kubernetes into a single repo so that
a user could mirror these offline in one chunk, and audit the set of
images for changes. For instance, within OpenShift we would like to
have a single place we can place all the images used by all the tests
with a single authentication scheme. While some images are not "real"
and can't be mirrored (for instance, the images that point to an
auth protected registry), that is not the majority.
This code makes it possible to specify an environment variable
KUBE_TEST_REPO that maps the static strings of the registry to a
single repository by placing the uniqueness in a tag. For instance:
KUBE_TEST_REPO=quay.io/openshift/community-e2e-images
would translate `k8s.gcr.io/prometheus-to-sd:v0.5.0` to `quay.io/openshift/community-e2e-images:e2e-30-k8s-gcr-io-prometheus-to-sd-v0-5-0-6JI59Yih4oaj3oQOjRfhyQ`.
The tag is a safe form of the name, plus the index (the constant within
manifest.go), plus a hash of the full input. The length of the tag is
constrained to the minimum of hash + index + the safe name.
The public method is changed to return two maps - index to original
name and index to test repo name. These maps would be the same if
the env var is not set.
Current logic to check whether a PVC is fully bound are:
1. PVC's volume name is not empty
2. Annotation "pv.kubernetes.io/bind-completed" is properly set
The behavior in the test case only set the annotation, and leave the
volume name to be set by a `FakePVController`.
This will cause a problem for us to run some testcase like scheduler's
perf test, scheduling pod with volume as an example, the first try will
always hit "unbound immediate PersistentVolumeClaims" exception.
As a result, the metric data "schedule_attempts_total", or "scheduling_algorithm_duration_seconds"
will not accurate enough.
Signed-off-by: Dave Chen <dave.chen@arm.com>
For testing certain features, the BusyBox image does not provide all the
tools that are needed. Notably 'dd' from BusyBox does not support
direct-io that is required for skipping caches while doing writes and
reads on a Block-mode PVC attached to different nodes.
The multi-arch container images used in tests live in quay.io which
doesn't support nesting. By making the /volume/ images repo configurable,
we are able to override them despite our current limitation.
Signed-off-by: Rafael Fonseca <r4f4rfs@gmail.com>
Most of these could have been refactored automatically but it wouldn't
have been uglier. The unsophisticated tooling left lots of unnecessary
struct -> pointer -> struct transitions.
Quite a few images are only used a few times in a few tests. Thus,
the images are being centralized into the agnhost image, reducing
the number of images that have to be pulled and used.
This PR replaces the usage of the following images with agnhost:
- mounttest
- mounttest-user
Additionally, removes the usage of the mounttest-user image and removes
it from kubernetes/test/images. RunAsUser is set instead of having that image.
Quite a few images are only used a few times in a few tests. Thus,
the images are being centralized into the agnhost image, reducing
the number of images that have to be pulled and used.
This PR replaces the usage of the following images with agnhost:
- dnsutils
dnsmasq is a Linux specific binary. In order for the tests to also
pass on Windows, CoreDNS should be used instead.
Quite a few images are only used a few times in a few tests. Thus,
the images are being centralized into the agnhost image, reducing
the number of images that have to be pulled and used.
This PR replaces the usage of the following images with agnhost:
- resource-consumer-controller
- test-webserver
The redis version has been bumped to version 5.0.5, but the maximum version supported on
Windows is 3.2. This can lead to failing tests, the output and behaviour can be different
(see #80516). In order to prevent such failures, the amount of times the Redis image is
used can be reduced.
This commit uses the previously added agnhost guestbook subcommand as a replacement for the
Guestbook application created by the test "should create and stop a working application".
Adds AgnhostPrivate to test/utils/image/manifest. Some tests are trying to pull
the agnhost image from the private registry, meaning that we would need to
always build and push the agnhost image to both e2e and private registry
whenever we bump its version. Decoupling them would mean that we only need
to push the image to the e2e registry.
Melds the "pull image from gcr.io" and "pull image from docker hub" tests
into a single test that pulls the agnhost image from the configured
e2eRegistry.
This also removes the need to maintain and update the image
"gcr.io/kubernetes-e2e-test-images/windows-nanoserver:v1". It should have been
a manifest list that also includes future Windows releases, like Windows Server 1903.
Additionally, the image has ~300 MB, meaning that with this change, it won't have to
wait as much to spawn a pod.
Removes "should not be able to pull non-existing image from gcr.io", since
the test "should not be able to pull image from invalid registry" test already
exists, and both of them test the same effect: cannot spawn a pod with an
image that does not exist.
The `err` return value was being overwritten in an unintended way, which
means the function may not return the proper error value. This diff
ensures it does.
A number of tests were using hardcoded image paths instead of
going through the imageutils package. The reason for centralizing
the logic there is to keep an eye on what images we use and where
they come from.
There is strong probabilty that some pods will be deleted when
we are bombarding cluster with high volume of pods. We do that in
Cluster Autoscaler scalability tests and we want to relax check there.
Change-Id: Ib7883666c0c952f61914ab51dcf1f5244e1e7e42
