🌱 run replication tests on a shared kcp instance #2620
Conversation
|
/test e2e-sharded |
| ctx, cancel := context.WithCancel(context.Background()) | ||
| t.Cleanup(cancel) | ||
|
|
||
| // TODO (p0lyn0mial): detect the type of the env we are running on (single vs multi-shard) |
There was a problem hiding this comment.
well, ci/prow/e2e-sharded keeps failing so I need to fix it in this PR :)
p0lyn0mial
force-pushed
the
run-replication-tests-on-shared-instance
branch
2 times, most recently
from
83f84c1 to
a18bd4a
Compare
|
/retest |
a standalone version of the cache server is tested on a multi-shard environment(in TestReplication test).
a helper function for creating a rest config for the cache server depending on the underlying test environment.
p0lyn0mial
force-pushed
the
run-replication-tests-on-shared-instance
branch
from
a18bd4a to
7702dde
Compare
the CI jobs wait for the admin kubeconfig before running the tests. creating the kubeconfig after the shards are ready makes sure that the shards are registered and that the env is stable.
| "Shard", | ||
| corev1alpha1.SchemeGroupVersion.WithResource("shards"), | ||
| &corev1alpha1.Shard{ | ||
| ObjectMeta: metav1.ObjectMeta{Name: "test-shard"}, | ||
| ObjectMeta: metav1.ObjectMeta{Name: withPseudoRandomSuffix("test-shard")}, |
There was a problem hiding this comment.
nit: can all or most of these use GenerateName?
There was a problem hiding this comment.
yes, we could do that but that would require some/bigger changes because the tests need stable name so that they can find resources for modification/deletion.
I can prepare a follow-up PR for that^
| err := <-terminatedCh | ||
| shardsErrCh <- indexErrTuple{i, err} | ||
| }(i, terminatedCh) | ||
| } |
There was a problem hiding this comment.
I explained in the commit msg, have a look eb96d8b
| // intended to be common between fixture for servers whose lifecycle | ||
| // is test-managed and fixture for servers whose lifecycle is managed | ||
| // separately from a test run. | ||
| func loadKubeConfig(kubeconfigPath string) (clientcmd.ClientConfig, error) { | ||
| func LoadKubeConfig(kubeconfigPath, contextName string) (clientcmd.ClientConfig, error) { |
There was a problem hiding this comment.
why do we need this? We shouldn't build a helper library for kube.
There was a problem hiding this comment.
I wanted to avoid copying of code that loads a kubeconfig
There was a problem hiding this comment.
There was a problem hiding this comment.
@sttts k8s libraries have just a little too much power, every project I've ever seen has their own simplification to not copy-paste ...
|
/lgtm |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: stevekuznetsov The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
openshift-ci
bot
added
the
approved
kcp-ci-bot
added
the
size/L
Summary
previously all scenarios were run in a private environment. After the change only the scenarios that are disruptive run in a sandbox, the rest is run on a shared kcp instance (either on a single or multi-shard env)
This change allows us to run non-disruptive scenarios on a multi-shard cluster.
Related issue(s)
#2596
kcp-dev/contrib-tmc#84