Files
smarm/src/cluster/discovery.rs
T
claude-asm-audit 4c0e42152f feat(cluster): RFC 010 c10–c16, follow-ups and Phase 6 (squash of 16ef583..d9c62a8)
Tree snapshot of d9c62a8 (2026-08-18). The 20 source commits between
16ef583 (c9) and d9c62a8 were never pushed and the clone that held them
was lost; this commit carries their combined tree verbatim so the build
history stays auditable from the c1–c9 commits below it. Original
hashes as recorded in the session handoff:

  c10  f03e94d  pid targeting + auto-serialization (RemotePid, D14 name
                on the wire); Phase 3 gate
  c11  7ef4bad  DownReason::Disconnected, wire tag 5
  c12  d124162  remote monitors (Monitor/Demonitor/Down frames)
  c13  9de967b  connection-loss synthesis (A+B: Monitors::teardown +
                unread-command Disconnected); Phase 4 gate
  c14  7e822b7  eager pg eviction (reaper actor, ReaperInboxes)
       dbe1a22  InboundVerdict::label(), trace::Event::ClusterInbound
       31a9877  tests/channel.rs monitor-churn target gated on `go`
       653559e  Discovery::Withdrawn{name, addr}
  c15  b41d76e  distributed pg: Sync on NodeUp, Join/Leave broadcast,
                NodeDown sweep, members_all; PgMsg wire type
  c16  fafa881  pick_any / dispatch_any; Phase 5 complete
  Phase 6 Tier A:
       195c73e  p4  NodeEvent::NodeDown(NodeInfo)
       48fd766  p1  connector Candidate{name, addr, state}
       ce8cf99  p2+p7 conn.rs select arms as Vec<Arm>; Outbound::Drained
       bf24988  p6  RemotePid::from_local -> Option
       9ae0380  p3  PeerStanding{Free, Claimed, Dialing}
       c7d62a1  p11 cluster::Timing knobs, threaded by value
       46f171d  p11 cluster_disconnect un-ignored on SMARM_FAST_TIMING
  Phase 6 Tier B:
       391a9ae  p5  cluster::RemoteDownReason{Local, Disconnected};
                    DownReason::Disconnected removed from core
       7ddd908  p9  pg ctl channel unconditional, one cfg seam at spawn
       d9c62a8      PeerNameMismatch parks the candidate; ClusterDial trace

Verified at d9c62a8: default 361/0, cluster 448/0, clippy --lib on
default / cluster / cluster+smarm-trace, fmt, 10x flake on
cluster_dial_mismatch, 5x on cluster_pg.
2026-08-18 16:00:00 +02:00

77 lines
3.2 KiB
Rust

//! RFC 010 c7b — peer discovery: the [`Strategy`] seam and the static-seeds
//! implementation.
//!
//! A strategy is **push-based and runs as its own actor**: the
//! [`connector`](crate::cluster::connector) spawns it with the sending end of
//! a channel, and the strategy emits [`Discovery`] events whenever it learns
//! something — once at startup for a static list, continuously for a future
//! mDNS/DNS strategy — for as long as it cares to run. Returning ends the
//! strategy actor; the candidates it pushed live on in the connector (the
//! connector owns all retry/backoff state, so a strategy never re-announces).
//!
//! A candidate is a **`(node_name, addr)` pair**, not a bare address: the
//! dial path and the D7 tie-break are keyed by peer *name* (the dial intent
//! must be registered before connecting so a crossing inbound `Hello` sees
//! it), so an anonymous dial would reintroduce exactly the
//! simultaneous-connect flap D7 exists to prevent. Discovery mechanisms know
//! names — that is what they discover.
use crate::channel::Sender;
/// A discovery event, as pushed by a [`Strategy`].
///
/// `Candidate` announces, `Withdrawn` retracts — the primitive pair. A
/// strategy that wants TTL semantics builds them on top (track its own
/// last-seen times, emit `Withdrawn` on expiry); the connector deliberately
/// has no clock of its own for candidates (D11: strategies never
/// re-announce, the connector owns retry). `#[non_exhaustive]` so more can
/// land without breaking strategies.
#[derive(Debug, Clone, PartialEq, Eq)]
#[non_exhaustive]
pub enum Discovery {
/// A peer worth dialing: its claimed node name and a dialable address.
Candidate { name: String, addr: String },
/// Stop dialing this `(name, addr)`. Dial-set only: a connection that
/// is already up is membership's business and is left alone; an
/// attempt in flight completes on its own; a later `Candidate` for the
/// same pair re-adds it with fresh backoff. Unknown pairs are ignored.
Withdrawn { name: String, addr: String },
}
/// A source of peers to dial. Implementations are spawned as actors by the
/// connector — see the module docs for the contract.
pub trait Strategy: Send + 'static {
/// Run the strategy: push [`Discovery`] events into `out` as they are
/// learned; return when done discovering (or when `out` reports closed —
/// the connector is gone). Runs inside an actor, so blocking
/// cooperatively is fine.
fn run(self: Box<Self>, out: Sender<Discovery>);
}
/// The static-seeds strategy: a fixed `(name, addr)` list, announced once.
#[derive(Debug, Clone, Default)]
pub struct StaticSeeds {
seeds: Vec<(String, String)>,
}
impl StaticSeeds {
pub fn new(seeds: impl IntoIterator<Item = (impl Into<String>, impl Into<String>)>) -> Self {
StaticSeeds {
seeds: seeds
.into_iter()
.map(|(n, a)| (n.into(), a.into()))
.collect(),
}
}
}
impl Strategy for StaticSeeds {
fn run(self: Box<Self>, out: Sender<Discovery>) {
for (name, addr) in self.seeds {
if out.send(Discovery::Candidate { name, addr }).is_err() {
return; // connector gone; nobody to discover for
}
}
}
}