show progress bar for downloading kic base image

[medyagh]

we need to treat the kic base image same as VM ISO in the UI

1- download kic base image before creating container (in the same step that we download iso for VM)
2- show progress bar for docker pull (if proress bar not possible, we should show some sort of animation so the user dont be bored and they also know they are waiting on the base image be pulled.

[afbjorklund]

This would be nice to have for all containers (e.g. cache), even though the kindest is the biggest...

There's some discussion upstream, in google/go-containerregistry#670 - it adds a lot of complexity

[afbjorklund]

@medyagh : are you planning to load kic from a tarball too, similar to the preload ?

Or maybe just run docker pull in a separate step, before doing the 'docker run` ?

[medyagh]

@medyagh : are you planning to load kic from a tarball too, similar to the preload ?

Or maybe just run docker pull in a separate step, before doing the 'docker run` ?

yeah I meant that wherever we download the ISO, for kic drivers we just do a docker or podman pull
and show some sort of animation
so when docker run goes, it already has the image.

[afbjorklund]

As workaround, we could load the kicbase image from a tarball that we download ?

313M gcr.io/k8s-minikube/kicbase_v0.0.9

docker load -i kicbase_v0.0.9.tgz

That way, cache would also work with the non-daemon drivers (such as podman).

[afbjorklund]

docker pull gcr.io/k8s-minikube/kicbase:v0.0.9
docker save gcr.io/k8s-minikube/kicbase:v0.0.9 | gzip -9 -c -n > kicbase_v0.0.9.tgz

@medyagh This would also make it available in china #7447

And perhaps as an alternative to using docker hub #7316 #7579

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

[afbjorklund]

Unfortunately progress is only implemented for tarball, but not for daemon.

But that API only supports running Docker anyway, so it is obsolete-by-design.

        if cc.Driver != "docker" {
                // TODO: driver == "podman"
                klog.Info("Driver isn't docker, skipping base image download")
                return
        }
        if image.ExistsImageInDaemon(cc.KicBaseImage) {
                klog.Infof("%s exists in daemon, skipping pull", cc.KicBaseImage)
                return
        }

// ExistsImageInDaemon if img exist in local docker daemon
func ExistsImageInDaemon(img string) bool {
        // Check if image exists locally
        cmd := exec.Command("docker", "images", "--format", "{{.Repository}}:{{.Tag}}@{{.Digest}}")
        if output, err := cmd.Output(); err == nil {
                if strings.Contains(string(output), img) {
                        klog.Infof("Found %s in local docker daemon, skipping pull", img)
                        return true
                }
        }
        // Else, pull it
        return false
}

[afbjorklund]

Turns out that go-containerregistry doesn't do pull, but uses load from archive internally...

[praneshvr00]

yes @medyagh , I thought it was stuck and I just gonna ctrl+c, should upgrade this pulling feature.