-
Notifications
You must be signed in to change notification settings - Fork 14
Intly 3698 - add csv for olm & update oauth image #178
Conversation
e7aa5a9
to
387b8ea
Compare
...y/olm-catalog/mobile-security-service-operator/mobile-security-service-operator.package.yaml
Outdated
Show resolved
Hide resolved
387b8ea
to
a5c3d11
Compare
@@ -0,0 +1,4 @@ | |||
packageName: mobile-security-service |
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.
@damienomurchu I've removed the integreatly-
part of this also. I guess it will still be in your one in integr8ly/manifests, but it's out of place here.
enable and disable specific versions of mobile applications on demand | ||
createdAt: "2019-11-04 12:21:00" | ||
support: Red Hat, Inc. | ||
name: mobile-security-service-operator.v0.4.1 |
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.
@damienomurchu for some reason, when I regenerated with operator-sdk 0.10.1, the version here got a "v" prefix
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.
Yes, same for me, I ended up removing the v
prefix for consistency. This value is just metadata however - the image actually used is referenced further down in the csv
- name: WATCH_NAMESPACE | ||
- name: APP_NAMESPACES | ||
valueFrom: | ||
fieldRef: | ||
fieldPath: metadata.annotations['olm.targetNamespaces'] |
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.
Changed these from their defaults. This project doesn't use the WATCH_NAMESPACE
var for some reason (might change that in future for consistency), and APP_NAMESPACES
was set to a default value that we don't really use.
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.
@grdryn I don't see a value for metadata.annotations['olm.targetNamespaces']
in the file?
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.
@damienomurchu that annotation is populated by the OLM, based on how it's instructed to install the operator (it's the default value of WATCH_NAMESPACE
, when generating a CSV). I haven't actually verified that this works though, I just presume it's a comma-separated list of namespaces.
I've made some changes here. @damienomurchu @laurafitzgerald could you take a quick look? You can validate the bundle locally by checking out this repo and running:
^there will be one warning about a missing icon, but the return value of the command will be zero (success). |
@grdryn changes look fine, but will build an image of this and confirm that it installs ok via operator-hub re: icon, I will fix that, and update the PR |
@damienomurchu I don't think you need to build an image, since none of the changes here go into the image -- you should be able to re-use the existing released image, right Edit: also, we don't have an icon for this service, do we? |
@grdryn its the built manifest image (via make commands here) that gets picked up by operator hub - I'll build a new manifest image from this PR and publish it to my own quay.io and verify it can be deployed correctly from operatorhub. I'm just going to do that now and confirm that none of the differences in the csv vs the mss csv in the manifests repo affect the ability to deploy it via olm |
@damienomurchu ah right. That doesn't build an actual container image though, does it? I think it just uploads the bundle to a quay.io app repo, right? |
👀 |
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.
@grdryn the changes look fine. I verified we wont need the additional crd files as we're using the other ones in the 3.x installer. Ran the command to verify and got the expected results.
I need to make a change here, but it's essentially just that I removed the wrong CRDs, so I need to bring those back & delete the other ones. Edit: it's not as easy as that unfortunately, since they've got different groups. We're in a very disturbing situation where for some reason the |
Actually never mind the above comment, I'll do that in a separate PR. The work to do it is growing exponentially, and isn't really related to this PR. This PR has everything correct AFAIK. |
@damienomurchu I'm going to merge this now before it gets forgotten. I'm not sure if you're still doing some verification on it or not, but either way if you do happen to notice any issues, do let us know! :) |
Motivation
Intly 3698 - add csv for olm & update oauth image for integreatly-operator
What
Add csv for olm & update oauth image
Why
Deploy via olm to rhmi on openshift4
How
Add csv for current version (0.4.1) and update oauth image to quay.io one
Verification Steps
Already verified as part of INTLY-3289
Checklist:
Progress
Additional Notes