feat: Raft clustering with TLS, streaming snapshots, and production hardening#74
Merged
ApiliumDevTeam merged 14 commits intodevfrom Mar 13, 2026
Merged
feat: Raft clustering with TLS, streaming snapshots, and production hardening#74ApiliumDevTeam merged 14 commits intodevfrom
ApiliumDevTeam merged 14 commits intodevfrom
Conversation
New crate `aingle_wal` with segment-based WAL, hash chain integrity, thread-safe writer, reader with replay/verification, and segment rotation. WAL integrated into AppState and mutation paths (triples, memory) behind `#[cfg(feature = "cluster")]`. All 20 WAL tests pass. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
New crate `aingle_raft` with openraft TypeConfig, log store, state machine, network layer, and consistency levels. Cluster REST endpoints added (status, join, leave, members, WAL stats/verify). P2pMessage extended with Raft + cluster variants. CLI flags for cluster mode. All 15 raft tests pass. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
ConsistencyLevel enum (Local/Quorum/Linearizable) with header parsing. Read endpoints (get_triple, list_triples) now accept X-Consistency header and route through appropriate consistency logic when cluster feature is enabled. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
LwwTriple (Last-Writer-Wins with deterministic tie-break by node_id) and OrSet (Observed-Remove Set for triple existence) implemented in aingle_graph behind `#[cfg(feature = "crdt")]`. Merge is commutative, associative, and idempotent. All 9 CRDT tests pass. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…sfer) ClusterSnapshot with TripleSnapshot wire format for full state transfer. STM explicitly excluded (node-local). HNSW index rebuilt locally from replicated LTM. LTM WAL entry kinds (LtmEntityCreate, LtmLinkCreate, LtmEntityDelete) already present from Phase 1. Snapshot serialization roundtrip tested. All 18 raft tests pass. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Bump all 10 product crates from 0.4.2 → 0.5.0 and update internal dependency version ranges from "0.4" → "0.5" to match. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Implement RaftLogReader and RaftLogStorage for CortexLogStore with WAL-backed persistence. Vote and committed state persisted to JSON files. Recovery on restart reads WAL segments to rebuild the in-memory BTreeMap. Add RaftEntry and Noop variants to WalEntryKind. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Connect CortexStateMachine to real GraphDB and IneruMemory so Raft-committed mutations are applied: TripleInsert/Delete to graph, MemoryStore/Forget to Ineru LTM. Add CortexSnapshotBuilder for full-state snapshots. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Implement RaftNetworkFactory and RaftNetworkV2 for CortexNetworkConnection. Add RaftRpcSender trait to decouple from QUIC transport, enabling stub senders during bootstrap and real P2P transport at runtime. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Bootstrap openraft::Raft in main.rs with CortexLogStore, CortexStateMachine, and CortexNetworkFactory. Add raft and cluster_node_id fields to AppState. Single-node cluster auto-initializes when no peers are configured. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Route triple and memory writes through Raft in cluster mode. Add ensure_linearizable guards to GET handlers honoring X-Consistency header (linearizable via ReadIndex, quorum via LeaseRead, local passthrough). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Rewrite cluster status/join/leave/members endpoints to use real Raft metrics (role, term, leader, membership). Join adds learner then promotes to voter; leave removes node from voter set via change_membership. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…d Mayros badge - Add Clustering section with 3-node quickstart, TLS, and endpoint reference - Add Consensus Layer (Raft, WAL, Streaming Snapshots, TLS) to architecture diagram - Add aingle_raft and aingle_wal to platform components table - Add "Powers Mayros AI" badge linking to ApiliumCode/mayros - Update Rust version badge and prerequisites from 1.70 to 1.83 - Add cluster build command to quickstart
Refactors cluster initialization into a dedicated, reusable module. Implements robust HTTP-based Raft RPC with TLS encryption, shared secret authentication, and exponential backoff for inter-node communication. Adds automatic leader redirection (HTTP 307) for client requests and cluster management operations to improve client routing and cluster availability. Introduces chunked snapshot transfers with Blake3 integrity checksums for efficient and reliable state replication, especially for large datasets. Improves WAL durability by persisting purge and truncation boundaries. Ensures data consistency by routing all write operations through Raft when clustering is enabled, preventing direct writes and potential split-brain. Includes comprehensive integration tests for cluster functionality.
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
1 participant
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.
Summary
aingle_walcrate with segment rotation, CRC32 checksums, and crash-safe durabilityaingle_raftcrate — leader election, log replication, streaming snapshots (512KB chunks with ACK), blake3 snapshot checksums, atomic dual-lock state restorecluster_initmodule with TLS support (self-signed via rcgen or custom PEM certs), HMAC-authenticated inter-node RPCs, exponential backoff join retries/cluster/status,/members,/join,/leave,/wal/stats,/wal/verifyX-Consistency: quorumheader for strong consistencysubtle), Raft shutdown timeout, config validation, learner rollback on failed membership changes, snapshot buffer TTL eviction (5min/256MB cap)Crates added
aingle_walaingle_raftTest plan
cargo check --workspacepassescargo check -p aingle_cortex --features clusterpassescargo test -p aingle_raft— 33/33 tests passcargo test -p aingle_cortex --features cluster --lib— 144/144 tests passcargo test -p aingle_cortex --features cluster --test cluster_integration_test— 3/3 integration tests pass (single-node bootstrap, 3-node replication, WAL stats/verify)