+++ This bug was initially created as a clone of Bug #1705753 +++
+++ This bug was initially created as a clone of Bug #1705752 +++
The following PR https://github.com/openshift/origin/pull/22696 introduced an e2e which is verifying that all the operators provide descriptive information as part of their schema which is then available through oc explain command.
Your component has a few missing bits, see the commented out lines from the PR.
This should be fixed similarly to https://github.com/openshift/cluster-version-operator/pull/168 and then update origin's e2e test and uncomment the lines from your components.
--- Additional comment from Maciej Szulik on 2019-05-02 21:10:49 UTC ---
Adam, I couldn't figure out a component for cluster-samples-operator but I've created just one issue wrt to imageregistry and sample operator configs.
I'd like to see both addressed.
--- Additional comment from Ben Parees on 2019-05-02 21:16:08 UTC ---
Created this clone to address the samples-operator aspect of Maciej's comment above.
--- Additional comment from Gabe Montero on 2019-05-02 23:26:30 UTC ---
so I see this targeted at 4.1.0, but the priority/severity are unspecified
should I assume this has to get into 4.1.0 GA ?
--- Additional comment from Ben Parees on 2019-05-02 23:59:19 UTC ---
unless Maciej says otherwise, yes. You can ask him what the importance is since he originated them.
--- Additional comment from Adam Kaplan on 2019-05-03 16:55:13 UTC ---
@Maciej this is a bit of a riskier proposition, since adding library-go requires bumping k8s.io deps to 1.13. Can we address this as tech debt in 4.2 (with a potential z-stream update for 4.1)?
--- Additional comment from Gabe Montero on 2019-05-03 18:25:25 UTC ---
fyi https://github.com/openshift/cluster-openshift-controller-manager-operator appears to implement the crd codegen defs and is glide based
--- Additional comment from Gabe Montero on 2019-05-03 19:36:41 UTC ---
In addition to the 1.13 bump, samples operator is going to have to be added to openshift/api a la https://github.com/openshift/api/pull/183
in order for
go run ./vendor/github.com/openshift/library-go/cmd/crd-schema-gen/main.go --domain openshift.io --apis-dir vendor/github.com/openshift/api --manifests-dir manifests/
Without that update, I get
no API type for found CRD manifests
exit status 1
--- Additional comment from Gabe Montero on 2019-05-03 19:39:21 UTC ---
Paul just told me that Maciej says z-stream is fine
--- Additional comment from Gabe Montero on 2019-05-03 20:17:02 UTC ---
sorted out with some Ben assistance the following to bypass openshift/api updates:
go run ./vendor/github.com/openshift/library-go/cmd/crd-schema-gen/main.go --domain operator.openshift.io --apis-dir pkg/apis --manifests-dir manifests/
--- Additional comment from Maciej Szulik on 2019-05-03 20:18:37 UTC ---
Yeah, that should work. The main gist is just the crd-schema-gen call. 4.1.x is perfect to land this.
--- Additional comment from Gabe Montero on 2019-05-03 21:01:21 UTC ---
PR https://github.com/openshift/cluster-samples-operator/pull/144 is up for review and regression testing
while we wait for 4.1.x to open up.
--- Additional comment from Paul Weil on 2019-05-10 12:07:50 UTC ---
--- Additional comment from Ben Parees on 2019-05-13 13:58:10 UTC ---
moving to 4.2.0, if you decide to backport it to 4.1.z create a clone.
--- Additional comment from Gabe Montero on 2019-05-16 19:56:57 UTC ---
master branch / 4.2.0 PR has merged
moving to ON_QA to await build
--- Additional comment from Gabe Montero on 2019-05-17 11:43:57 UTC ---
Turns out 4.2 builds are not ready yet on the OSBS/brew side. I'm in contact with the ART to track when they become available.
Moving to Modified
--- Additional comment from Gabe Montero on 2019-05-23 18:57:31 UTC ---
After running `oc explain` with the changes from https://github.com/openshift/cluster-samples-operator/pull/144 I'm going to update some godoc
and clean up a few things, enhance a few things.
New PR coming
Moving back to ASSIGNED until new PR comes up.
Also, I think there is enough there though to re-enable the test in openshift/origin/test/extended/cli/explain.go
--- Additional comment from Gabe Montero on 2019-05-23 20:27:51 UTC ---
PR https://github.com/openshift/cluster-samples-operator/pull/148 is up
FYI ... run oc explain configs --api-version=samples.operator.openshift.io/v1 to see spec/status updates
--- Additional comment from Gabe Montero on 2019-05-23 20:55:03 UTC ---
Also, https://github.com/openshift/origin/pull/22899 re-enables samples operator in the extended test Maciej referenced in the description
--- Additional comment from Gabe Montero on 2019-05-23 22:04:32 UTC ---
PR has merged
--- Additional comment from XiuJuan Wang on 2019-05-24 03:40:26 UTC ---
Check with 4.2.0-0.ci-2019-05-23-230227
The DESCRIPTION part is empty, and inconsistent space between apiVersion, metadata
$ oc explain configs --api-version=samples.operator.openshift.io/v1
APIVersion defines the versioned schema of this representation of an
object. Servers should convert recognized schemas to the latest internal
value, and may reject unrecognized values. More info:
Kind is a string value representing the REST resource this object
represents. Servers may infer this from the endpoint the client submits
requests to. Cannot be updated. In CamelCase. More info:
Standard object's metadata. More info:
Spec contains the desired configuration and state for the Samples Operator,
controlling various behavior around the imagestreams and templates it
creates/updates in the openshift namespace.
Status contains the actual configuration in effect, as well as various
details that describe the state of the Samples Operator.
--- Additional comment from Gabe Montero on 2019-05-24 14:23:44 UTC ---
Those are generic `oc explain` items, not anything to do with the changes around this bug.
You'll see the same thing if you run `oc explain` against the other CRDs. That is what I saw
when I tested this.
Open a bug against the CLI component if you want those addressed.
The spec/status you noted above is valid verification of this bug.
--- Additional comment from XiuJuan Wang on 2019-05-27 02:03:12 UTC ---
Move to verified as comment #20
*** Bug 1734402 has been marked as a duplicate of this bug. ***
@Wenjing - can you be more specific regarding "DESCRIPTION part needs to modified according to https://bugzilla.redhat.com/show_bug.cgi?id=1732148" ... there is a lot of dialogue in that bug.
The whitespace is whatever the crd tooling and oc explain does at this point. I don't think we should hold verification on that. Looks fine to me. And consistent
with what I see with the other `oc explain` invocation on CRDs I see in a cluster.
Do you have an example of an `oc explain` where you like the whitespace? If so, I can compare our invocation of the crd tooling with those resources, in case there
is some option to use that I am unaware of.
I belive Wenjing said as this https://bugzilla.redhat.com/show_bug.cgi?id=1732148#c16 , but for the orignal task to generate samples opearor definitions, I think we could mark this as verified.
For whitespace, it exists in all CRDs
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.