Release images bundle instead of separate images #11522
Conversation
k8s-ci-robot
added
cncf-cla: yes
Makefile
Outdated
| @@ -769,6 +771,8 @@ bazel-version-dist: bazel-version-dist-linux-amd64 bazel-version-dist-linux-arm6 | |||
| tools/sha256 ${BAZELUPLOAD}/kops/${VERSION}/darwin/amd64/kops ${BAZELUPLOAD}/kops/${VERSION}/darwin/amd64/kops.sha256 | |||
| cp bazel-bin/cmd/kops/windows-amd64/kops ${BAZELUPLOAD}/kops/${VERSION}/windows/amd64/kops.exe | |||
| tools/sha256 ${BAZELUPLOAD}/kops/${VERSION}/windows/amd64/kops.exe ${BAZELUPLOAD}/kops/${VERSION}/windows/amd64/kops.exe.sha256 | |||
| tar cfvz ${BAZELUPLOAD}/kops/${VERSION}/images/images.tar.gz -C ${BAZELIMAGES} ${BAZELIMAGES}/*.tar.gz ${BAZELIMAGES}/*.tar.gz.sha256 | |||
There was a problem hiding this comment.
I don't quite see the point of including the .sha256 of the individual images into the tar file.
There was a problem hiding this comment.
People may want to validate those individually, or compare to what we have in the GCS bucket.
There was a problem hiding this comment.
I'm struggling to figure out a valid use case for comparing a .sha256 in a tar file against another file in the same tar file. What would be the value in validating those individually or comparing to what is in the GCS bucket?
With separate files you can retrieve the small file through a more secure path or use the redundancy to detect cosmic-ray-type errors in the big file. But given the unlikelihood of doing a partial-file download or wanting to salvage part of a slightly corrupted tar file, these use cases don't apply well to checksums embedded in the big file.
There was a problem hiding this comment.
They don't. There is no real use case, but doesn't break anything either.
if you think it would be best to exclude them, I don't mind doing it.
There was a problem hiding this comment.
I think it would be best to exclude them.
hakman
force-pushed
the
release-image-bundle
branch
4 times, most recently
from
43c04a4 to
bebde39
Compare
|
/retest |
|
@johngmyers Is this ok as is now? |
|
I need time to run it in order to see what it produces. It's not obvious by reading the code. |
|
If it helps, I tried to to run this as part of the build to see the output also: |
|
I also was thinking that maybe we start doing |
hakman
force-pushed
the
release-image-bundle
branch
from
bebde39 to
d69ac2f
Compare
k8s-ci-robot
added
the
lgtm
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: johngmyers The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
k8s-ci-robot
added
the
approved
Release images bundle instead of separate images
…-upstream-release-1.21 Automated cherry pick of #11522: Release images bundle instead of separate images
kOps releases contain a considerable number of files, out of which only the binaries are currently used for mirrors.
As agreed during Office Hours, images should be bundled into a single archive file.