kubectl proxy --accept-hosts appears to filter destination, not source.

[wrossmann]

Is this a BUG REPORT or FEATURE REQUEST?: BUG REPORT

Kubernetes version:

• Client: v1.8.3
• Server: v1.7.8-gke.0

Environment:

• Cloud provider: Google [GKE]
• OS:
  • Client: CentOS 6/7 [client]
  • Server: cos-stable-61-9765-66-0
• Install tools: gcloud

What happened:
In my development VM [10.0.1.100] I started kubectl proxy to provide full access to the Dashboard UI to my workstation [10.0.2.200], but not everyone else in the office. The command used was:

kubectl proxy --port=8080 --address=10.0.1.100 --accept-hosts='^10\.0\.2\.200$'

as well as many variants of the regular expression with/without \., ^, and $

Which yielded the following for all endpoints:

<h3>Unauthorized</h3>

It was suggested in the #kubernetes-users Slack channel that it may be running the regex against a hostname, but I've double-checked and we do not have reverse lookups enabled for our network, and I don't think that kubectl would be looking up NetBIOS info.

It can be made to work with --accept-hosts='^.*$' as you would expect, however I also decided to try --accept-hosts='^10\.0\.1\.100$' on a hunch, which ended up working the same as the wildcard, allowing anyone on the network access to the proxy.

What you expected to happen:

1. --accept-hosts='^10\.0\.2\.200$' should allow only the stated IP access to the proxy.
   • actual: universal inaccessibility
2. --accept-hosts='^10\.0\.1\.100$' should allow only local requests, similar to the 127.0.0.1 defaults.
   • actual: open access to all

How to reproduce it:

With two hosts attached to the same network:

• Host A: kubectl proxy --port=8080 --address=HOST_A_IP --accept-hosts='^HOST_B_IP$'
• Host B: curl http://HOST_A_IP:8080/

Anything else we need to know:

This has wider implications than me simply being paranoid about my local network. In searching for solutions to this problem I found pages upon pages of people stating --accept-hosts='^.*$' and --disable-filter=true as general solutions, even in the face of machines with public-facing interfaces. The supposed workarounds for this issue are causing people to unwittingly expose very sensitive information and APIs to potentially the entire world unless they take additional action to prevent it, eg: firewall rules.

eg: See the discussion in kubernetes/dashboard#692

[apelisse]

I don't know who knows about proxy. @liggitt ?

[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]

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
/remove-lifecycle stale

[nikhita]

/remove-lifecycle rotten

[seans3]

/sig cli
/area kubectl
/priority P2

[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

[nikhita]

/remove-lifecycle stale

[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

[george-angel]

/remove-lifecycle stale

[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

[fejta-bot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

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

[k8s-ci-robot]

@fejta-bot: Closing this issue.

In response to this:

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

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

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.