-
Notifications
You must be signed in to change notification settings - Fork 2.1k
[OCPCLOUD-1201] Add periodics for TechPreviewNoUpgrade clusters #19845
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[OCPCLOUD-1201] Add periodics for TechPreviewNoUpgrade clusters #19845
Conversation
|
/hold This isn't actually doing what I want it to do right now, it is putting the FG in after install where we need it in the manifests before install |
801dfba to
f61962a
Compare
|
/retest |
|
/retest Build01 was broken yesterday so the jobs weren't starting correctly, I'm told this is now fixed |
|
Looks like this is now working as expected. The tests have failed due to a known issue with the MCO, grabbed the below from the I believe the MCO team have a fix in place for this that is currently under review |
|
/hold cancel This is now working as intended, but we know this is broken for now |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
not speaking cron, how frequent is this?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It is every 8 hours, copied from the existing openstack jobs which this would synchronise with. Will talk to the release team about why openstack is so frequent and whether being in sync is going to cause issues
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
|
Not sure I have rights here actually, let's try. /approve |
|
stbenjam
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The comments on each of the new steps, chains and workflows should be updated to reflect their actual purposes
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This comment seems to be incorrect
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This step documentation doesn't seem to be accurate as to the purpose of this chain
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why is parallel in the test name? It's assumed that's what's running unless otherwise specified.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've copied this from the test above, e2e-openstack-parallel on line 221 in this diff. Not sure why it's called that but this is just keeping consistency with the rest of the openstack tests
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
typo: azure?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
No serial test-suite for OpenStack?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
For openstack, this appears to be overriden by an environment variable within the job configuration itself. I'm just trying to be consistent with existing openstack jobs by copying that pattern
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ohh, so we just have periodics for serial suite?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We have serial and the regular tests (parallel?), the block above this should be for the openstack regular tests
JoelSpeed
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Will go through and fix up all the comments on the files today, that was just an omission while I was copy/pasting
Do you only want them running on cron or do you want feedback on each nightly release? If the latter you probably want to add them to the release controller as optional informers: https://docs.ci.openshift.org/docs/architecture/release-gating/
I'll look into that link thank you, I think we want both the cron and release optional informers
The OWNERS files for the steps seem to be missing the PR author, should you be including yourself to maintain these new steps?
I deliberately didn't do that initially since this isn't really my remit. I've been asked to add these by archs because we are making the tech preview more interesting this release (with the CCMs), so it was suggested that it would be good in general to have this, rather than it being something specifically tied to the CCM project which I own. My thought here was that this should be owned in general by the release teams who own the rest of the release informing jobs.
If that's not the correct assumption, I'll update the PR to include myself as an owner on all of these as well
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've copied this from the test above, e2e-openstack-parallel on line 221 in this diff. Not sure why it's called that but this is just keeping consistency with the rest of the openstack tests
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We have serial and the regular tests (parallel?), the block above this should be for the openstack regular tests
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
For openstack, this appears to be overriden by an environment variable within the job configuration itself. I'm just trying to be consistent with existing openstack jobs by copying that pattern
|
@stbenjam @ravisantoshgudimetla I think I've addressed and replied to all the comments as appropriate, PTAL when you have a moment |
|
The MCO failure here will be fixed once openshift/machine-config-operator#2668 is merged |
|
Latest results show that the storage issues are gone now the rebase has landed. But does now highlight that there are two tests that fail because of alerts fired because we are running as TechPreview. I'll raise PRs to follow up on getting those alerts ignored when in tech preview mode |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I thought we had a job already for that:
https://github.com/openshift/release/blob/master/ci-operator/step-registry/openshift/e2e/openstack/ccm/openshift-e2e-openstack-ccm-workflow.yaml
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That one is specific for CCM, this PR is adding generic TechPreviewNoUpgrade tests to prove out TP features in general and in the future, CCM just happens to be a subset of that for now
|
/retest |
1 similar comment
|
/retest |
|
Test failures now seem to be unrelated to the TechPreview testing and may actually be genuine problems/flakes. I think this is ready to go in now. The Azure and OpenStack serials have passed already which is a good sign :) One more test for luck |
|
/lgtm |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I guess inherited from master, but ideally each directory will have more than one approver. Can we find anyone else who can sign up to help maintain this chain going forward?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah I just copied this from the parent folder, perhaps @abhinavdahiya has a suggestion for someone else who can help own the azure configs? Perhaps someone from splat might be able to help?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I feel like there should be a chain gathering up some of this, to reduce the chance that someone adds a new OpenStack knob similar to the existing FIPS, but forgets to update this tech-preview chain to respect that knob here.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'll admit I'm not an expert in these ci-operator configs, but I couldn't see a common chain, I'll have a look and see if there's a chain that contains this stuff, if not, perhaps a separate PR would be appropriate to merge them and update all the affected jobs?
This one is based off the existing chains, just adding the techpreview step
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There doesn't appear to be anything today that I can find that resembles a shared chain for this stuff, have poked the ShiftStack team regarding the issue to see if this is something on their radar
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@pierreprinetti suggested that we can actually avoid this duplcation by not using the ipi-conf like this and just directly adding the techpreview step into the workflow. Since the AWS version of this is already merged, #21227 is a PR which cleans it up to show what Pierre meant, do you think this is a better approach that I should adopt for all of these?
|
Looks like you'll need someone like @jupierce or @bradmwilliams for final approval |
42cd21e to
d8f1ba6
Compare
|
Just rebased to fix conflicts, someone updated the master periodics to move them from build01 to build02 |
d8f1ba6 to
bcb1e12
Compare
|
@JoelSpeed: The following tests failed, say
Full PR test history. Your PR dashboard. DetailsInstructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here. |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: bradmwilliams, deads2k, JoelSpeed, stbenjam The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
@JoelSpeed: Updated the following 3 configmaps:
DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
As part of the Cloud Controller Manager project, we are making some significant changes that will at first, be under the
TechPreviewNoUpgradefeature set.To gain confidence in the new cloud controller managers, we would like to add some signal to the release informing jobs that the TPNU clusters are still viable and healthy.
This PR adds periodics for the parallel and serial jobs for AWS, Azure and OpenStack, the three platforms we are initially targeting for the CCM project.
/assign @deads2k