Add CI periodic Windows Jobs.#5165
Conversation
|
Hi @adelina-t. Thanks for your PR. I'm waiting for a containerd member to verify that this patch is reasonable to test. If it is, they should reply with Once the patch is verified, the new status will be reflected by the I understand the commands that are listed here. DetailsInstructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
0f68ddf to
c5e3099
Compare
c5e3099 to
34abcfc
Compare
There was a problem hiding this comment.
I'm bit confused regarding the lines there. We cloned the latest hcsshim, but checked out the version on containerd's go.mod. What is the version we'd like to test against?
There was a problem hiding this comment.
Yes. Basically I'm doing what the windows integration PR test does now. The goal is to periodically test the HEAD version of containerD and as such, to avoid any issue that might be introduced by using latest hcsshim , we use the version indicated by containerD. I am actually thinking of adding some flexibility here that would allow for creation of test runs that would say: validate latest hcsshim with a stable containerD version ( or whatever else combination that there might be a need for ), but that is something I'm considering down the line.
Also, in doing it as such, this job can easily be used to test other containerD branches and being sure that it will build hcsshim properly for that contaienrD version / branch. ( Which reminds me, I need to use the github env params for the containerD checkout as well so it doesn't pull master all the time ).
There was a problem hiding this comment.
I wonder if it would be good in the future to disentangle the version of the hcsshim shim binary that is used, from the version of the other hcsshim package dependencies that are used. That is probably a better discussion to be had elsewhere though. :)
863262b to
213c169
Compare
77c7024 to
6610152
Compare
There was a problem hiding this comment.
if we add a creationTimestamp to the RG it will be cleaned up in case of pipeline failure: https://github.com/Azure/rg-cleanup/blob/23be71c1aec703ee3fd60c2e8549ef7ecc5f17b3/main.go#L20
6610152 to
fdfc23d
Compare
fdfc23d to
308797e
Compare
Signed-off-by: Adelina Tuvenie <[email protected]>
308797e to
421fc6e
Compare
| with: | ||
| azcliversion: latest | ||
| inlinescript: | | ||
| az group create -n ${{ matrix.AZURE_RESOURCE_GROUP }} -l ${{ env.AZURE_DEFAULT_LOCATION }} |
There was a problem hiding this comment.
Can you add `--tags creationTimestamp=$(date +%Y-%m-%dT%T%z) so the RG will get cleaned up by the automation in event anything goes wrong?
7665d96 to
b94b9d4
Compare
|
This job refers to various secrets/credentials; I assume those are not set up yet in the containerd/containerd repo? Do you have a run of this in a branch to validate it's working properly? |
No, they are not set up yet, we have to coordinate with a maintainer with access to the repo to set those secrets.
My branch https://github.com/adelina-t/containerd/tree/azure_ci_testing contains the exact same code as https://github.com/adelina-t/containerd/tree/azure_ci_workflow (the one from which the PR originated) + a commit to enable workflow_dispatch trigger. Corresponding github actions job runs can be seen here. This workflow is also set up to publish results to testgrid, you can see them on the dashboard here to get an ideea: https://testgrid.k8s.io/sig-windows-containerd-runtime-signal#win-2004-containerd-master-integration . All were ran from my fork when testing. Corresponding artifacts are published here, you can see a folder for the Windows sac2004. |
cpuguy83
left a comment
There was a problem hiding this comment.
This looks like a good start.
LGTM.
Thanks!
|
a lot of work stitching things together, GCS upload / Windows Job / Github Actions .... heads's spinning :) |
|
LGTM let's kick the tires. |
C'est la vie if we want test grid. Good part is that at least some bits shouldn't need to change all that much or at least aren't that brittle. |
b94b9d4 to
fd7b07c
Compare
|
Build succeeded.
|
fd7b07c to
1b207a6
Compare
|
Build succeeded.
|
|
/retest |
|
@adelina-t: Cannot trigger testing until a trusted user reviews the PR and leaves an DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
1b207a6 to
e5a0c47
Compare
Add 2019 to matrix Signed-off-by: Adelina Tuvenie <[email protected]>
Signed-off-by: Adelina Tuvenie <[email protected]>
e5a0c47 to
77285e3
Compare
|
Build succeeded.
|
estesp
left a comment
There was a problem hiding this comment.
LGTM, looks like this is ready.
Before merge, I believe we will have to add the following secrets to our containerd org settings; will need to work offline to get that set up:
AZURE_SUB_ID
AZURE_CREDS
GCP_PROJECT_ID
GCP_SA_KEY
Excellent :) @estesp How can we coordinate so I can send the secrets? I've already gotten all I need in terms of service accounts for Azure internally, I can provide them. |
|
@adelina-t you can send to any of our emails or the security mailing list. You can use the gpg keys registered on Github (estesp, dmcgowan) |
Done |
Part of the ongoing work described here: #5039
This introduces a periodic Windows 1909 testing job. This will be expanded to multiple Windows Versions. The goal is to improve testing volume & coverage on multiple Windows versions in order to pick up any regression or intermittent issue with containerD for Windows.
TO-DOs: