Open mmoscher opened 2 years ago
Base: 85.05% // Head: 85.05% // No change to project coverage :thumbsup:
Coverage data is based on head (
814ff04
) compared to base (ae23010
). Patch has no changes to coverable lines.
:umbrella: View full report at Codecov.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.
@anishakj seem's that the commit singoffs are broken. IMHO this is wrong.
If this PR is helpful at all, I would create a new one with a single commit introducing these changes.
@mmoscher Have you verified the binary created for arm64
@anishakj nope not yet, since the build can only be done in the emulated docker platform, thus in the container itself. Will check if the container can be validated. Any hints? How to you validate your builds? Thanks in advance!
Status of this?
@janhoy sorry to say, but nothing new yet! Feel free to move on and submit PR's to my fork, if you like. Maybe I will manage to free some time next week.
@mmoscher Have you verified the binary created for arm64
@mmoscher Could you please address the comments provided by @derekm
Change log description
This PR creates mutliarch builds of pravega/zookeeper-operator and pravega/zookeeper using docker buildx when a new GitHub release is crafted.
However... the available
publish
job in theci.yaml
GHA workflow was broken. For example, the job does no code-checkout at all and thus was uncapable of running (finding) the Makefile. Uhrgs.Question: Is the provided GHA worklfow and your
push
Makefile-target used at all? IMHO not :/ :( Because: a) the "ALTREPO" used in the Makefile for pushing image (make push) is pointing to another docker-hub repo where another user is pushing docker-images. For example, in the pravega docker-hub repos a user called pravegaci is pushing images, whereas in the emccorp repos the user is called ecsjenkins. Since the Makefile has only one docker login call, pushing with different users in a CI environment is impossible. b) Additionally, the emccorp/zookeeper-operator:latest tag was created 3 years ago whereas the latest tag is ~5 months old. An automated push on release to the altrepo would have created the tag "latest" as well.All-in-all: it seems there is another CI or manual build-process in place, for what reason I'm uncertain if this PR will help at all :(
@anishakj pretty please clarify how you are publishing your docker-images. Thanks :)
Purpose of the change
What the code does
build-and-push-multiarch-image
andbuild-and-push-multiarch-zk-image
targets in makefile and calls these targets in the GHA CI workflow when a release was createdHow to verify it