fix(felhomsshd): keep the persisted port unconditionally (no self-listen flip-flop)

On a continuous reconcile felhom-sshd itself listens on the claimed port, so
re-probing isFree(persisted) found it 'busy' by our own daemon and thrashed to
another candidate every tick. A persisted port is ours — keep it.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
This commit is contained in:
2026-07-05 22:43:22 +02:00
parent d880289b06
commit b99d02b7a9
2 changed files with 14 additions and 15 deletions
+10 -7
View File
@@ -14,15 +14,18 @@ var ErrPortsExhausted = fmt.Errorf("felhomsshd: all candidate OOB ports are busy
// for tests; production impl = probeFree (ss + net.Listen). // for tests; production impl = probeFree (ss + net.Listen).
type portProbe func(port int) bool type portProbe func(port int) bool
// claimPort returns the OOB port, mirroring the spike's shell algorithm: // claimPort returns the OOB port:
// - if a persisted port exists AND is still free → keep it (idempotent, no thrash), // - if a port is already PERSISTED (and != 22) → keep it unconditionally. It is OUR port; on a
// - else the FIRST free candidate → persist + return, // continuous reconcile felhom-sshd is itself LISTENING on it, so re-probing with isFree would
// - else ErrPortsExhausted (LOUD). // (wrongly) find it "busy" by our own daemon and thrash to another candidate every tick. Once
// claimed, the port is stable (the belt @ssh_port and the operator's known port depend on it).
// - else the FIRST free candidate → persist + return (isFree = ss-empty AND a real bind succeeds).
// - else ErrPortsExhausted (LOUD — never :22 or a random port).
// //
// persist writes PortFile; readPersisted reads it. isFree is the probe. All injected for tests. // persist writes the port file; readPersisted reads it. isFree is the probe. All injected for tests.
func claimPort(candidates []int, isFree portProbe, readPersisted func() (int, bool), persist func(int) error) (int, error) { func claimPort(candidates []int, isFree portProbe, readPersisted func() (int, bool), persist func(int) error) (int, error) {
if cur, ok := readPersisted(); ok && cur != 22 && isFree(cur) { if cur, ok := readPersisted(); ok && cur != 22 {
return cur, nil return cur, nil // persisted = ours; keep it (no thrash — felhom-sshd holds it)
} }
for _, p := range candidates { for _, p := range candidates {
if p == 22 { if p == 22 {
+4 -8
View File
@@ -78,10 +78,11 @@ func TestClaimPort_CleanContentionIdempotentExhaustion(t *testing.T) {
t.Fatalf("contention claim = %d / %v, want 2222", p, err) t.Fatalf("contention claim = %d / %v, want 2222", p, err)
} }
// idempotent: persisted 2222 still free (even though 8822 is now free) → keep 2222 (no thrash) // idempotent: a persisted port is KEPT unconditionally (even when isFree(cur) reports busy —
isFree, read, write, _ = claimHarness(map[int]bool{8822: true, 2222: true, 8022: true, 62222: true}, 2222) // that's felhom-sshd itself holding it on a continuous reconcile; re-probing would flip-flop).
isFree, read, write, _ = claimHarness(map[int]bool{8822: true, 2222: false, 8022: true, 62222: true}, 2222)
if p, err := claimPort(cands, isFree, read, write); err != nil || p != 2222 { if p, err := claimPort(cands, isFree, read, write); err != nil || p != 2222 {
t.Fatalf("idempotent claim = %d / %v, want 2222 (kept)", p, err) t.Fatalf("persisted port must be kept even when isFree says busy (own daemon), got %d / %v", p, err)
} }
// exhaustion: all busy → LOUD error, no fallback // exhaustion: all busy → LOUD error, no fallback
@@ -97,9 +98,4 @@ func TestClaimPort_CleanContentionIdempotentExhaustion(t *testing.T) {
t.Fatalf("exhaustion must persist nothing, stored=%d", *stored) t.Fatalf("exhaustion must persist nothing, stored=%d", *stored)
} }
// persisted port that is now BUSY → re-claim a fresh free one (not the stale persisted)
isFree, read, write, _ = claimHarness(map[int]bool{8822: false, 2222: true, 8022: true, 62222: true}, 8822)
if p, err := claimPort(cands, isFree, read, write); err != nil || p != 2222 {
t.Fatalf("stale-persisted re-claim = %d / %v, want 2222", p, err)
}
} }