fix(pd): resolve hostname entries in IpAuthHandler allowlist by bitflicker64 · Pull Request #2962 · apache/hugegraph

bitflicker64 · 2026-03-05T09:03:18Z

Purpose of the PR

fix [Bug] IpAuthHandler blocks all cross-node raft connections in Docker bridge mode — hostname vs IP mismatch #2960

IpAuthHandler only compared the client IP with the allowlist entries directly.
When the allowlist contains hostnames, connections from their resolved IPs could be rejected.

Main Changes

Resolve hostname allowlist entries using InetAddress.getAllByName
Store resolved addresses in a set used for connection validation
Keep runtime validation as a simple Set.contains() lookup

Verifying these changes

Trivial rework / code cleanup without any test coverage. (No Need)
This issue was identified while testing changes in PR refactor(docker): migrate single-node compose from host to bridge networking #2952.

Does this PR potentially affect the following parts?

Documentation Status

Doc - TODO
Doc - Done
Doc - No Need

…at startup to avoid Netty DNS blocking

bitflicker64 · 2026-03-05T21:25:11Z

How I tested:

Built a local Docker image from source with this fix applied (alongside the getLeaderGrpcAddress fix from fix(pd): add timeout and null-safety to getLeaderGrpcAddress() #2961)
Removed static IP workaround from docker-compose — switched back to hostnames (pd0:8610, pd1:8610, pd2:8610) for all raft peers, gRPC hosts, and PD addresses
Brought up the full 3-node cluster (3 PD + 3 Store + 3 Server) in bridge network mode with no ipam static IP config
pd2 won leader election on first boot

Results with pd2 as leader and hostnames only:

partitionCount:12 on all 3 stores ✅
leaderCount:12 on all 3 stores ✅
{"graphs":["hugegraph"]} ✅
All 9 containers healthy ✅
No "Blocked connection" warnings in any PD logs ✅

Before this fix: IpAuthHandler compared raw hostname strings against incoming connection IPs — always failed in bridge mode, requiring static IPs as a workaround.

After this fix: Hostnames are resolved to IPs at startup via InetAddress.getAllByName() — static IP workaround no longer needed.

Related PR: #2952, #2961

Copilot

Pull request overview

Fixes PD Raft inbound connection allowlisting when the configured allowlist contains hostnames (e.g., Docker bridge service names) by resolving those hostnames to IPs up front and validating connections via a simple set lookup.

Changes:

Resolve allowlist entries via InetAddress.getAllByName() at handler construction time
Cache resolved IPs in an immutable Set used by isIpAllowed()
Log a warning when an allowlist hostname can’t be resolved

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

hugegraph-pd/hg-pd-core/src/main/java/org/apache/hugegraph/pd/raft/auth/IpAuthHandler.java

…r changes - Resolve allowlist hostnames to IPs using InetAddress.getAllByName - Add refresh() to update resolved IPs when Raft peer list changes - Wire refresh into RaftEngine.changePeerList() - Add IpAuthHandlerTest covering hostname resolution, refresh behavior, and failure cases

bitflicker64 · 2026-03-11T22:40:15Z

While working on this change, I noticed a couple of related behaviors that appear to predate this PR. The /v1/members/change endpoint relies on RestAuthentication, where Authentication.authenticate() currently validates only the username portion of Basic Auth. I also noticed that GrpcAuthentication does not appear to be wired into the gRPC server. These are outside the scope of this fix, but I thought it might be helpful to mention in case they are worth tracking separately.
Also worth noting: IpAuthHandler.refresh() is wired through RaftEngine.changePeerList(), but PDService.updatePdRaft() calls node.changePeers() directly, so the refresh won't fire for that path.

- Cache leader PeerId after waitingForLeader() and null-check to avoid NPE when leader election times out - Remove incorrect fallback that derived leader gRPC address from local node's port, causing silent misroutes in multi-node clusters - Wire config.getRpcTimeout() into RaftRpcClient's RpcOptions so Bolt transport timeout is consistent with future.get() caller timeout - Replace hardcoded 10000ms in waitingForLeader() with config.getRpcTimeout() - Remove unused RaftOptions variable and dead imports (ReplicatorGroup, ThreadId) Fixes apache#2959 Related to apache#2952, apache#2962

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

hugegraph-pd/hg-pd-test/src/main/java/org/apache/hugegraph/pd/raft/IpAuthHandlerTest.java

hugegraph-pd/hg-pd-core/src/main/java/org/apache/hugegraph/pd/raft/RaftEngine.java

imbajin · 2026-03-13T09:46:36Z

While working on this change, I noticed a couple of related behaviors that appear to predate this PR. The /v1/members/change endpoint relies on RestAuthentication, where Authentication.authenticate() currently validates only the username portion of Basic Auth. I also noticed that GrpcAuthentication does not appear to be wired into the gRPC server. These are outside the scope of this fix, but I thought it might be helpful to mention in case they are worth tracking separately. Also worth noting: IpAuthHandler.refresh() is wired through RaftEngine.changePeerList(), but PDService.updatePdRaft() calls node.changePeers() directly, so the refresh won't fire for that path.

Some context about the auth system in PD/Store:

5eeeb9a

To put it simply, PD/Store didn't have a strict auth mechanism due to legacy design decisions. Since these and other gRPC components were meant to be internal-only, only the graph server was originally built with a formal public-facing auth interface. We later added a basic, quick-and-dirty auth layer for security, but the current implementation remains incomplete. We plan to systematically refactor this later, but for now, let's just add TODOs in the relevant places

bitflicker64 · 2026-03-15T05:26:15Z

We plan to systematically refactor this later, but for now, let's just add TODOs in the relevant places

Thanks for the context, that makes sense. I’ll add the TODOs in the relevant places and work through the reviewed suggestions. Apologies for the delay as my Mac broke so I had to shift my setup to Ubuntu

- Wire IpAuthHandler.refresh() into PDService.updatePdRaft() which calls node.changePeers() directly, bypassing RaftEngine.changePeerList() - Include learner IPs in refresh since updatePdRaft() supports learner peers - Add @before setUp() to IpAuthHandlerTest to prevent stale singleton from earlier suite classes poisoning test assertions - Strengthen testUnresolvableHostnameDoesNotCrash assertions - Replace testGetInstanceReturnsNullBeforeInit with testGetInstanceReturnsSingletonIgnoresNewAllowlist

- Document missing auth check on MemberAPI changePeerList endpoint - Note that password validation is currently skipped in Authentication - Flag that GrpcAuthentication interceptor is not wired into the gRPC server

Copilot

Pull request overview

Copilot reviewed 8 out of 8 changed files in this pull request and generated 3 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

hugegraph-pd/hg-pd-test/src/main/java/org/apache/hugegraph/pd/raft/IpAuthHandlerTest.java

hugegraph-pd/hg-pd-core/src/main/java/org/apache/hugegraph/pd/raft/auth/IpAuthHandler.java

hugegraph-pd/hg-pd-service/src/main/java/org/apache/hugegraph/pd/service/PDService.java

…fresh() wiring - Add RaftEngineIpAuthIntegrationTest to prove the full chain: changePeerList() -> changePeers callback -> handler.refresh() - Add failure-path test to ensure allowlist is not refreshed on failed peer update - Uses mock Node that fires callback synchronously to avoid latch.await() hang - Register in PDCoreSuiteTest suite

Copilot

Pull request overview

Copilot reviewed 9 out of 9 changed files in this pull request and generated 2 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

hugegraph-pd/hg-pd-core/src/main/java/org/apache/hugegraph/pd/raft/RaftEngine.java

hugegraph-pd/hg-pd-test/src/main/java/org/apache/hugegraph/pd/raft/IpAuthHandlerTest.java

- Add timed latch.await() using 3x rpcTimeout to prevent indefinite blocking if changePeers callback is never invoked - Use compareAndSet in both callback and timeout path to avoid race - Restore Thread.currentThread().interrupt() on InterruptedException - Fix testHostnameResolvesToIp() to resolve localhost dynamically instead of hardcoding 127.0.0.1 — avoids flakiness on IPv6 systems

Copilot

Pull request overview

Copilot reviewed 9 out of 9 changed files in this pull request and generated 5 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

hugegraph-pd/hg-pd-test/src/main/java/org/apache/hugegraph/pd/raft/IpAuthHandlerTest.java

hugegraph-pd/hg-pd-core/src/main/java/org/apache/hugegraph/pd/raft/RaftEngine.java

...d/hg-pd-test/src/main/java/org/apache/hugegraph/pd/raft/RaftEngineIpAuthIntegrationTest.java

Fix apache#2960: resolve hostname allowlist entries in IpAuthHandler …

9b9154b

…at startup to avoid Netty DNS blocking

bitflicker64 mentioned this pull request Mar 5, 2026

refactor(docker): migrate single-node compose from host to bridge networking #2952

Open

23 tasks

imbajin requested a review from Copilot March 9, 2026 10:58

Copilot started reviewing on behalf of imbajin March 9, 2026 10:58 View session

Copilot AI reviewed Mar 9, 2026

View reviewed changes

imbajin reviewed Mar 10, 2026

View reviewed changes

hugegraph-pd/hg-pd-core/src/main/java/org/apache/hugegraph/pd/raft/auth/IpAuthHandler.java Show resolved Hide resolved

imbajin reviewed Mar 10, 2026

View reviewed changes

hugegraph-pd/hg-pd-core/src/main/java/org/apache/hugegraph/pd/raft/auth/IpAuthHandler.java Show resolved Hide resolved

bitflicker64 requested a review from imbajin March 12, 2026 07:25

imbajin requested a review from Copilot March 13, 2026 09:36

Copilot started reviewing on behalf of imbajin March 13, 2026 09:37 View session

Copilot AI reviewed Mar 13, 2026

View reviewed changes

imbajin reviewed Mar 13, 2026

View reviewed changes

hugegraph-pd/hg-pd-test/src/main/java/org/apache/hugegraph/pd/raft/IpAuthHandlerTest.java Show resolved Hide resolved

imbajin reviewed Mar 13, 2026

View reviewed changes

hugegraph-pd/hg-pd-core/src/main/java/org/apache/hugegraph/pd/raft/RaftEngine.java Outdated Show resolved Hide resolved

bitflicker64 added 2 commits March 15, 2026 13:20

Add TODO notes documenting current PD authentication limitations

980020a

- Document missing auth check on MemberAPI changePeerList endpoint - Note that password validation is currently skipped in Authentication - Flag that GrpcAuthentication interceptor is not wired into the gRPC server

bitflicker64 requested a review from imbajin March 15, 2026 10:00

imbajin requested a review from Copilot March 15, 2026 13:31

Copilot started reviewing on behalf of imbajin March 15, 2026 13:32 View session

Copilot AI reviewed Mar 15, 2026

View reviewed changes

imbajin reviewed Mar 15, 2026

View reviewed changes

hugegraph-pd/hg-pd-service/src/main/java/org/apache/hugegraph/pd/service/PDService.java Show resolved Hide resolved

bitflicker64 requested a review from imbajin March 15, 2026 20:21

imbajin requested a review from Copilot March 16, 2026 04:34

Copilot started reviewing on behalf of imbajin March 16, 2026 04:34 View session

Copilot AI reviewed Mar 16, 2026

View reviewed changes

hugegraph-pd/hg-pd-core/src/main/java/org/apache/hugegraph/pd/raft/RaftEngine.java Outdated Show resolved Hide resolved

hugegraph-pd/hg-pd-test/src/main/java/org/apache/hugegraph/pd/raft/IpAuthHandlerTest.java Outdated Show resolved Hide resolved

bitflicker64 requested a review from Copilot March 16, 2026 06:28

Copilot started reviewing on behalf of bitflicker64 March 16, 2026 06:28 View session

Copilot AI reviewed Mar 16, 2026

View reviewed changes

fix: avoid changePeerList timeout race and update tests

0aa2d5f

Conversation

bitflicker64 commented Mar 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Purpose of the PR

Main Changes

Verifying these changes

Does this PR potentially affect the following parts?

Documentation Status

Uh oh!

bitflicker64 commented Mar 5, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

bitflicker64 commented Mar 11, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

imbajin commented Mar 13, 2026

Uh oh!

bitflicker64 commented Mar 15, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

bitflicker64 commented Mar 5, 2026 •

edited

Loading