馃悰 Fix kubernetes version for e2e tests #960
Conversation
metal3-io-bot
added
the
size/S
|
/test-centos-e2e-integration-main |
|
/test-e2e-upgrade-main-from-release-0-5 |
4 similar comments
|
/test-e2e-upgrade-main-from-release-0-5 |
|
/test-e2e-upgrade-main-from-release-0-5 |
|
/test-e2e-upgrade-main-from-release-0-5 |
|
/test-e2e-upgrade-main-from-release-0-5 |
|
/test-ubuntu-integration-main |
metal3-io-bot
requested review from
furkatgofurov7,
lentzi90,
mboukhalfa,
smoshiur1237 and
tuminoid
|
|
||
| ### Test matrix for k8s version | ||
|
|
||
| Current e2e tests use the following Kubernetes versions for source and target clusters: |
There was a problem hiding this comment.
I know we have this terminology everywhere, but I think we should try to use "bootstrap", "management", "workload" and "self-hosted" terminology instead of "source" and "target". In CAPI "source" and "target" is only used in the context of pivoting. In other contexts it is quite confusing to use these terms.
We start with the bootstrap cluster, we pivot to the target management cluster that is self-hosted.
| Current e2e tests use the following Kubernetes versions for source and target clusters: | |
| Current e2e tests use the following Kubernetes versions for bootstrap and self-hosted management clusters: |
There was a problem hiding this comment.
self-hosted management means the cluster is initialized with control providers however in this context tests like remediation will not initialize the "target" cluster with providers it is a workload cluster.
There was a problem hiding this comment.
Ah even more reason to not say target cluster! What about "CAPI cluster" or "metal3 cluster"? Or we can say "workload/management" cluster
There was a problem hiding this comment.
as management can also refer to bootstrap I suggest to use metal3 cluster or metal3 provisioned cluster
There was a problem hiding this comment.
this is overcomplicating @mboukhalfa , anyways lets take this PR in so that I can test e2e with correct k8s versions and lets debate on the name of the cluster in a followup I would do soon.
|
|
||
| Current e2e tests use the following Kubernetes versions for source and target clusters: | ||
|
|
||
| | tests | source cluster | target init | target final | |
There was a problem hiding this comment.
| | tests | source cluster | target init | target final | | |
| | tests | bootstrap cluster | self-hosted init | self-hosted final | |
There was a problem hiding this comment.
Thanks for the review. I will push a change to this table soon with the uplift to v1.27.1. Can I address the comments in that?
|
/approve |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: mboukhalfa 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 |
metal3-io-bot
added
the
approved
|
/lgtm |
This PR fixes e2e tests with correct k8s versions. Currently we are wrongly initializing the e2e integration tests with one older minor version but we should always initialize it to the latest k8s. This PR also adds a matrix to better understand which k8s version we are using for different e2e tests.