-
Notifications
You must be signed in to change notification settings - Fork 50
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
🐛 More needed for podman workaround in kind-load (take 2) #794
Conversation
The commit adding the podman workaround for the kind-load target seems to need a bit more work: It needs to have the tag name set before image-archive is saved Signed-off-by: Brett Tofel <btofel@redhat.com>
✅ Deploy Preview for olmv1 ready!
To edit notification comments on pull requests, go to your Netlify site configuration. |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #794 +/- ##
=======================================
Coverage 67.21% 67.21%
=======================================
Files 23 23
Lines 1467 1467
=======================================
Hits 986 986
Misses 415 415
Partials 66 66
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
/assign @perdasilva |
@DEPLOY_TEMPLATE_IMG_NAME=quay.io/operator-framework/operator-controller:devel; \ | ||
TMP_FILE=temp_image.tar; \ | ||
podman tag $(IMG) $$DEPLOY_TEMPLATE_IMG_NAME; \ | ||
echo "Saving image to temporary file $$TMP_FILE"; \ | ||
podman save $$DEPLOY_TEMPLATE_IMG_NAME -o $$TMP_FILE; \ | ||
$(KIND) load image-archive $$TMP_FILE --name $(KIND_CLUSTER_NAME); \ | ||
rm $$TMP_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.
Instead of doing a completely different thing for podman/docker, could we just do this exact procedure for both, where we just use $(CONTAINER_RUNTIME)
instead of hardcoding docker
or podman
?
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.
Meta-comment:
It is really difficult to support every permutation of developer environment, especially if we don't centralize these sorts of task definitions. Whatever we decide here would need to be duplicated in basically all of the active repos and then maintained.
If we don't test each configuration in CI, then they are extremely prone to rot, because the only one guaranteed to work is the one used in CI. If we do test each configuration, then CI becomes more expensive AND it is hard for a developer making this kind of change to fix something about an environment they don't have setup locally.
IMO, if at all possible, it is better for the repo to be very opinionated on the environment and to have developers adapt to the repo instead of vice versa
Even Mac vs Linux is hard enough (and we don't explicitly test Mac in CI). With podman support, the matrix becomes linux/docker, linux/podman, mac/docker, mac/podman.
Could you please explain the issue in more detail? I'm using podman on Mac and do not observe issues with our |
I'd argue our time is more expensive than CI time, most of us have Macs for local dev and I feel like we should be trying to use Podman if its diffs from Docker don't get in our way too badly. Also, CI likely is closer to Podman than Docker internally anyway because CRI-O is using buildah under the hood. If we can't add this conditional to make Podman easy and cheaper for devs to use, I guess I'll work out how to auto-apply/delete before commit these kind of things so I can personally work better, but they're in conditionals, does it matter if the podman version is found to have rotted later? It can get fixed if someone is motivated and all can benefit in practice. |
Ah, well your symlink comment made me go try it. Sure enough the problem here is not what I thought. I thought podman didn't properly support loading a docker-image and thus needed a workaround. But once I added the symlink it's obvious that Closing this PR, thanks @m1kola |
I looked a bit into this - it seems to be a known issue: |
The commit adding the podman workaround for the kind-load target seems to need a bit more work:
It needs to have the tag name set before image-archive is saved
Description
Reviewer Checklist