Bound each shutdown stage by the stop deadline (closes #100)

Shutdown passes its context to every stage: the streamer, the four
handlers, the WHOIS fetcher, the database maintainer and the HTTP server
each stop waiting when it ends. A handler then skips its last write and
logs how many updates it did not write, and the HTTP server closes the
connections still open. The database is checkpointed and closed last,
and left open once the deadline has passed, since a write may still
hold it. The stop hook gives Shutdown 8 seconds and fx 9, inside the 10
seconds upaas and docker stop allow, so a stop signal exits 0.

Model: opus-5-5

internal/routewatch/peerhandler.go

@@ -5,6 +5,7 @@ package routewatch
 // peer state in the database.
 
 import (
+	"context"
 	"strconv"
 	"sync"
 	"time"
@@ -42,18 +43,24 @@ type PeerHandler struct {
 	db     database.Store
 	logger *logger.Logger
 
-	// Batching. mu guards peerBatch only: the flush goroutine takes the batch
-	// under mu and writes it after releasing mu, so HandleMessage waits for
-	// a database write only while peerBatch is at peerBatchMaxSize.
+	// Batching. mu guards peerBatch and writing only: the flush goroutine
+	// takes the batch under mu and writes it after releasing mu, so
+	// HandleMessage waits for a database write only while peerBatch is at
+	// peerBatchMaxSize.
 	mu        sync.Mutex
 	peerBatch []peerUpdate
+	// writing is how many updates the flush goroutine took from peerBatch
+	// and has not finished writing.
+	writing int
 	// batchTaken is signalled when the flush goroutine takes peerBatch; its L
 	// is &mu.
 	batchTaken  sync.Cond
 	batchFullCh chan struct{} // wakes the flush goroutine when peerBatch is full
 	lastFlush   time.Time     // used only by the flush goroutine
-	stopCh      chan struct{}
-	wg          sync.WaitGroup
+	// stopCh hands Stop's context to the flush goroutine. It holds one, so
+	// Stop does not wait for a write in progress to hand it over.
+	stopCh chan context.Context
+	done   chan struct{} // closed when the flush goroutine returns
 }
 
 type peerUpdate struct {
@@ -74,12 +81,12 @@ func NewPeerHandler(db database.Store, logger *logger.Logger) *PeerHandler {
 		peerBatch:   make([]peerUpdate, 0, peerBatchSize),
 		batchFullCh: make(chan struct{}, 1),
 		lastFlush:   time.Now(),
-		stopCh:      make(chan struct{}),
+		stopCh:      make(chan context.Context, 1),
+		done:        make(chan struct{}),
 	}
 	h.batchTaken.L = &h.mu
 
 	// Start the flush timer goroutine
-	h.wg.Add(1)
 	go h.flushLoop()
 
 	return h
@@ -144,16 +151,32 @@ func (h *PeerHandler) HandleMessage(msg *ristypes.RISMessage) {
 	}
 }
 
-// Stop gracefully stops the handler and flushes remaining batches
-func (h *PeerHandler) Stop() {
-	close(h.stopCh)
-	h.wg.Wait()
+// Stop stops the flush goroutine, which writes the updates the handler holds
+// before it returns, and waits for it until ctx ends. Once ctx has ended the
+// flush goroutine skips that last write, and a write still running is lost
+// when the process exits, so Stop logs how many updates are not written.
+func (h *PeerHandler) Stop(ctx context.Context) {
+	h.stopCh <- ctx
+
+	select {
+	case <-h.done:
+	case <-ctx.Done():
+	}
+
+	h.mu.Lock()
+	unwritten := h.writing + len(h.peerBatch)
+	h.mu.Unlock()
+
+	if unwritten > 0 {
+		h.logger.Error("Stop deadline passed before the peer handler wrote "+
+			"its updates", "updates", unwritten)
+	}
 }
 
 // flushLoop runs in a goroutine, the only one that writes the batch: it
 // flushes when the batch is full, periodically, and once more on Stop.
 func (h *PeerHandler) flushLoop() {
-	defer h.wg.Done()
+	defer close(h.done)
 
 	ticker := time.NewTicker(peerBatchTimeout)
 	defer ticker.Stop()
@@ -166,9 +189,11 @@ func (h *PeerHandler) flushLoop() {
 			}
 		case <-h.batchFullCh:
 			h.flushBatch()
-		case <-h.stopCh:
-			// Final flush
-			h.flushBatch()
+		case ctx := <-h.stopCh:
+			// Final flush, unless the stop deadline has passed
+			if ctx.Err() == nil {
+				h.flushBatch()
+			}
 
 			return
 		}
@@ -176,7 +201,8 @@ func (h *PeerHandler) flushLoop() {
 }
 
 // takeBatch returns the pending updates and leaves an empty batch in their
-// place for HandleMessage to fill.
+// place for HandleMessage to fill. They count in writing until flushBatch has
+// written them.
 func (h *PeerHandler) takeBatch() []peerUpdate {
 	h.mu.Lock()
 	defer h.mu.Unlock()
@@ -186,6 +212,7 @@ func (h *PeerHandler) takeBatch() []peerUpdate {
 	}
 
 	batch := h.peerBatch
+	h.writing = len(batch)
 	h.peerBatch = make([]peerUpdate, 0, peerBatchSize)
 	h.batchTaken.Broadcast()
 
@@ -227,5 +254,9 @@ func (h *PeerHandler) flushBatch() {
 			"error", err, "count", len(dbPeerMap))
 	}
 
+	h.mu.Lock()
+	h.writing = 0
+	h.mu.Unlock()
+
 	h.lastFlush = time.Now()
 }