move images from gcr.io to quay.io#79390
Conversation
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: coffeepac The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
/ok-to-test |
|
/test pull-kubernetes-e2e-gce-100-performance |
|
/hold @coffeepac Can we deprecate this code completely? Is it used somewhere? It's concerning that binary images in k/k are in repos that aren't in the project's control. cc: @kubernetes/sig-cluster-lifecycle-pr-reviews |
|
What code do you want to deprecate? The whole monitoring stuff? Imho this is used a lot. The image is also used in the fluentd-elasticsearch helm chart: https://hub.helm.sh/charts/kiwigrid/fluentd-elasticsearch Also the current gcr.io docker repo (gcr.io/fluentd-elasticsearch/fluentd) is not in the projects control. The reason for the move away from the the original repo (gcr.io/google-containers/fluentd-elasticsearch) has been, that only Google employees we're able to build the images manually and we sometimes had to wait for several weeks to get this done, when we liked to release a new version of the image. I also would like to have a repo where community members and or maintainers of the docker images have easy push access. An automatic build & push pipeline would be also nice. /test pull-kubernetes-e2e-gce-100-performance |
| docker build --tag ${PREFIX}/${IMAGE}:${TAG} . | ||
|
|
||
| push: | ||
| docker push ${PREFIX}/${IMAGE}:${TAG} |
There was a problem hiding this comment.
please add push with latest tag too...
| docker build --tag ${PREFIX}/${IMAGE}:${TAG} . | ||
|
|
||
| push: | ||
| docker push ${PREFIX}/${IMAGE}:${TAG} No newline at end of file |
There was a problem hiding this comment.
please add push with latest tag too...
|
What would be the right directory now? |
juju moved out of k/k. |
|
@piosz what are your thoughts? |
|
Assuming there is a SIG to provide sponsorship, |
|
a. this is owned by sig-instrumentation. however, I request both @cblecker and @neolit123 release their holds on this PR as I am spending money I shouldn't be to continue to host these images. I do not have stats for how large an impact it would be if I cut off access to these images but it wouldn't be non-zero. please see: #78684 |
|
@coffeepac I'm poking some folks to see what we can do here. I don't want you to keep having to pay for things, but I also don't want to thrash where this image is located. I understand this is urgent for cost reasons, and I'll be circling back with an answer ASAP. |
|
thanks @cblecker. this is entirely my own fault for incurring these charges as well. I'm bad at reading fine print apparently. EDIT re: impact, rough math suggests there are/projected to be about ~1350 downloads of these images for June based on size of image and actual/projected costs. |
|
/lgtm |
|
@cblecker any progress on this or can we do a bit of location thrash while we sort out a long term solution? |
|
/hold cancel This would also result in location thrash though, as we don't currently promote from staging repos to |
|
@coffeepac Immediate term, yeah, we can just move the reference location. The next step after this is I'd like to have a plan for moving this code out, I think. The manifests look stale, this is just the image generation piece, while other pieces like the helm chart for installation https://github.com/kiwigrid/helm-charts/blob/master/charts/fluentd-elasticsearch/values.yaml#L2 exist elsewhere. There's also a couple other references to the old image location in our code base: https://cs.k8s.io/?q=gcr.io%2Ffluentd-elasticsearch WDYT? |
|
Happy to move the code into a sig-inst repo. I tried doing that 6 months ago and it wasn't possible for some reason, I think it was tied to the build process for the add-on manager? But k8s infra has come along since then so I'm happy to do it. I'll get a plan onto the feature ticket early next week for discussion. Thanks for the help friends, I know this is a poor solution and I appreciate you taking it to help out my old poor decision. |
|
If it helps I'm willing to move fluentd-elasticsearch helm chart out of the Kiwigrid repo. Is there already an official k8s org helm registry available somewhere or would we need to create one? |
|
Where is this image on quay.io? I've tried the following: docker pull quay.io/fluentd-elasticsearch/fluentd:v2.5.1 and all three return the following error: Error response from daemon: unauthorized: access to the requested resource is not authorized |
The base image moved from gcr.io to quay.io. Ref kubernetes/kubernetes#79390 Signed-off-by: ialidzhikov <i.alidjikov@gmail.com>
What type of PR is this?
/kind feature
What this PR does / why we need it:
moves images from being hosted in
gcr.iotoquay.io. The reason for this move is thatgcr.ioisn't free,quay.iois and I don't have a personal budget to host these images.Once this is merged I'm going to turn off access to the existing images. If people need older images I can rebuild specific versions when asked.
Which issue(s) this PR fixes:
Fixes #
Special notes for your reviewer:
Does this PR introduce a user-facing change?: