fix(axnoded): enforce deny-all network isolation - #183
Merged
Merged
Conversation
Signed-off-by: wayne <rcywjo@gmail.com>
Contributor
Author
|
Rollout boundary: the enforcement change applies when a fresh Allocation is created. A pre-fix Allocation with an already attached interface is not made safe merely by upgrading axnoded in place; operators must cancel/recreate any still-running |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Root cause
Python SDK correctly sends a present empty strict policy; controld canonicalizes and persists it as isolated. Axnoded nevertheless attached the ordinary host interface. Because deny-all correctly omits egressd, outbound HTTPS remained possible.
Change
An isolated Allocation no longer acquires the host network interface, leaving its runsc OCI network namespace private and unconnected. Node startup rejects any accidental connected binding. No DNS-only workaround, public API change, or weaker assertion.
Reproduction and acceptance
run-9051f5e6-900c-46d9-9611-53407bde527d, Allocationalloc-94081338-98bb-42da-a045-3b42802fe3a3received external HTTPS 200; unrestricted control also received 200.5cbd368d9ecdc40435b7c5fed18199fc6a363547, native linux/amd64 source Compose node image IDsha256:795c63c34ee5a5deaeda67b01bf45f993488ad54d5a7311e36558c5339f41c34. A pinned local TLS/HTTP fixture provides actual HTTPS 200 outside the sandbox without relying on public DNS. Deny-all Runrun-ffbd9b14-1327-429a-bb73-dfd90c7b17bb/ Allocationalloc-e17aa6e0-0385-4c11-aaea-b77dd8f3e0a7had HTTPS=false and direct-IP TCP=false. Unrestricted Runrun-a919d192-cadf-4039-a48f-6daa866393b7/ Allocationalloc-454d3a26-a3fc-4863-adf8-97918e9ed773had HTTPS 200 and direct TCP=true.Verification boundaries
The full local Mac gate stopped at the unrelated source Compose DNS doctor fixture issue #182, after the #181 E2E was separately run and passed. The HK host changed profile could not finish proto-generated-check because that host lacks
protoc; its generated-file deletions were restored. PR required CI is the remaining gate. Do not infer broader capacity/qualification from these functional tests.Related: #181. Keep Draft until required CI is complete and reviewed; no merge or release in this task.