Tracking Issue - Conformance Coverage for Networking

[brahmaroutu]

What would you like to be added:
This is a place holder to track Conformance Coverage on the Topic.
This issue should remain open until all related work is accomplished in the k/k repo
This issue will contain analysis of the coverage. Existing tests and additional tests requested with links to those issues/PR's

Why is this needed:
Specifically address CNI networking, identify tests that tests pod-pod communication

/area conformance

[brahmaroutu]

/sig testing

[mgdevstack]

Analyzed e2e from test/e2e/networking/dns.go and needs clarity on promoting existing e2e to Conformance suite keeping in mind windows support.

1. should resolve DNS of partial qualified names for the cluster
   Tracking Issue: Removes PQDNs from conformance tests and places them in their own tests #70156
2. should resolve DNS of partial qualified names for services
   Tracking Issue: Removes PQDNs from conformance tests and places them in their own tests #70156
3. should provide DNS for pods for Hostname and Subdomain
   Tracking Issue: Need Windows DNS configuration tests #73414 , Add dns capabilities to allow for custom dns in Windows Azure/azure-container-networking#305
4. should provide DNS for ExternalName services
   Tracking Issue: Need ExternalName test for mixed-OS clusters #73328
5. should support configurable pod resolv.conf
   Feature addition in azure for custom DNS config: Add dns capabilities to allow for custom dns in Windows Azure/azure-container-networking#305

These tests are failing in [Windows] | Doc Reference

Note:

1. Windows assumes that DNS query containing . in it must be fully qualified.
2. Traking issues are mentioned to track these e2e support for Windows.

Needs clarity On:-

1. Should we label [LinuxOnly] for e2e verifying PQDN resolution, or seperate e2e would be drafted in for windows suite, or any other solution is suggested.
2. Any important aspects to consider while promoting such e2e which are failing on Windows and getting pass on Linux.

[mgdevstack]

cc @BCLAU

[adelina-t]

/cc @PatrickLang @michmike

[smarterclayton]

These are all approved by the SIG?

[michmike]

These are all approved by the SIG?

for the most part yes. is there anyone from sig-network we should be working closely with in the future?

[smarterclayton]

I'd like to see a tech lead from sig-network document that each of these belongs in conformance because they are part of a "required for all distributions API" and are not optional in any case.

[michmike]

@adelina-t and @BCLAU can you please work with sig-network on the ask above by @smarterclayton .
cc: @craiglpeters

[claudiubelu]

Please consult this doc [1] before promoting tests to Conformance. If it's already passing, I'm fine with promoting them to Conformance. If not, ask us to look into them or see what the failure is about. At the very least, if you really want to promote failing tests to Conformance, please add the [LinuxOnly] tag in the test name, so they're automatically skipped by all the Windows jobs in testgrid [2].

[1] https://docs.google.com/spreadsheets/d/1j8fAU0BLgdHWmRW5Ib9LlUydkBZJsghApTercA7WQRM/edit#gid=1365541815
[2] https://testgrid.k8s.io/sig-windows

[johnbelamaric]

sig-network already determined that the PQDN should be LinuxOnly, we moved them out of conformance a couple months ago.

The customDns modifies /etc/resolv.conf and so is LinuxOnly as well.

I'll take a look at the other two.

[johnbelamaric]

#74982 and #74983 do look like they are expected behavior on all clusters. What is the issue with those two on Windows?

[claudiubelu]

So, #74983 (adds should provide DNS for ExternalName services to Conformance), I've tested it using azure-vnet-cni, flannel overlay, flannel l2bridge, and it passed on all 3 of them. I still have to test it using OVN / OVS. Finnally, I would like @benmoss and @yujuhong to confirm that they're also ok with this test. Once they do, we can proceed with the test promotion.

The other one, #74982 (adds should provide DNS for pods for Hostname and Subdomain to Conformance), no. Looking through the code, it seems that it also expects /etc/hosts entries to be validated, but it is a known issue that individual files cannot be mounted in Windows Containers at the moment (see #72729 for precedence). It is planned to be supported once containerd support is added though, so we can revisit this test once that is added.

If you still want to promote something to Conformance, I would suggest splitting the test, as we've done before, and promoting the Subdomain part to Conformance. That part works when using azure-vnet-cni, flannel overlay, flannel l2bridge. See #75591

[mgdevstack]

/reopen
This is a umbrella issue to track related Issues and PRs.

[k8s-ci-robot]

@mgdevstack: Reopened this issue.

In response to this:

/reopen
This is a umbrella issue to track related Issues and PRs.

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.

[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

[johnbelamaric]

/reopen
Umbrella issue

[k8s-ci-robot]

@johnbelamaric: Reopened this issue.

In response to this:

/reopen
Umbrella issue

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.

[johnbelamaric]

/remove-lifecycle stale
/lifecycle frozen

[cmluciano]

@johnbelamaric Do we have a list of items that still need tackled for this?

[shaneutt]

Given the age of this it seems like the reality is this is very stale.

/lifecycle rotten

If not though, let's talk and figure out what's next for this? /cc @aojea, maybe you have some thoughts?

Gonna bring this one up at our next sync as well.

[aojea]

It will be a good exercise going through all the existing e2e test and see the ones we should promote to conformance

https://testgrid.k8s.io/sig-network-kind#sig-network-kind,%20master&include-filter-by-regex=sig-network

/help

[k8s-ci-robot]

@aojea:
This request has been marked as needing help from a contributor.

Guidelines

Please ensure that the issue body includes answers to the following questions:

• Why are we solving this issue?
• To address this issue, are there any code changes? If there are code changes, what needs to be done in the code and what places can the assignee treat as reference points?
• Does this issue have zero to low barrier of entry?
• How can the assignee reach out to you for help?

For more details on the requirements of such an issue, please see here and ensure that they are met.

If this request no longer meets these requirements, the label can be removed
by commenting with the /remove-help command.

In response to this:

It will be a good exercise going through all the existing e2e test and see the ones we should promote to conformance

https://testgrid.k8s.io/sig-network-kind#sig-network-kind,%20master&include-filter-by-regex=sig-network

/help

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.