swarm/storage: improve instrumentation

swarm/storage/localstore: fix broken metric (#1373)

p2p/protocols: count different messages (#1374)

cmd/swarm: disable snapshot create test due to constant flakes (#1376)

swarm/network: remove redundant goroutine (#1377)

cmd/swarm/swarm-snapshot/create_test.go

@@ -21,7 +21,6 @@ import (
 	"fmt"
 	"io/ioutil"
 	"os"
-	"runtime"
 	"sort"
 	"strconv"
 	"strings"
@@ -34,9 +33,7 @@ import (
 // It runs a few "create" commands with different flag values and loads generated
 // snapshot files to validate their content.
 func TestSnapshotCreate(t *testing.T) {
-	if runtime.GOOS == "windows" {
-		t.Skip()
-	}
+	t.Skip("todo: fix this")
 
 	for _, v := range []struct {
 		name     string

p2p/protocols/protocol.go

@@ -254,6 +254,7 @@ func (p *Peer) Drop() {
 func (p *Peer) Send(ctx context.Context, msg interface{}) error {
 	defer metrics.GetOrRegisterResettingTimer("peer.send_t", nil).UpdateSince(time.Now())
 	metrics.GetOrRegisterCounter("peer.send", nil).Inc(1)
+	metrics.GetOrRegisterCounter(fmt.Sprintf("peer.send.%T", msg), nil).Inc(1)
 
 	var b bytes.Buffer
 	if tracing.Enabled {

swarm/network/stream/messages.go

@@ -24,9 +24,7 @@ import (
 	"github.com/ethereum/go-ethereum/metrics"
 	"github.com/ethereum/go-ethereum/swarm/log"
 	bv "github.com/ethereum/go-ethereum/swarm/network/bitvector"
-	"github.com/ethereum/go-ethereum/swarm/spancontext"
 	"github.com/ethereum/go-ethereum/swarm/storage"
-	"github.com/opentracing/opentracing-go"
 )
 
 var syncBatchTimeout = 30 * time.Second
@@ -201,12 +199,6 @@ func (m OfferedHashesMsg) String() string {
 func (p *Peer) handleOfferedHashesMsg(ctx context.Context, req *OfferedHashesMsg) error {
 	metrics.GetOrRegisterCounter("peer.handleofferedhashes", nil).Inc(1)
 
-	var sp opentracing.Span
-	ctx, sp = spancontext.StartSpan(
-		ctx,
-		"handle.offered.hashes")
-	defer sp.Finish()
-
 	c, _, err := p.getOrSetClient(req.Stream, req.From, req.To)
 	if err != nil {
 		return err
@@ -297,21 +289,21 @@ func (p *Peer) handleOfferedHashesMsg(ctx context.Context, req *OfferedHashesMsg
 		From:   from,
 		To:     to,
 	}
-	go func() {
+
 	log.Trace("sending want batch", "peer", p.ID(), "stream", msg.Stream, "from", msg.From, "to", msg.To)
 	select {
 	case err := <-c.next:
 		if err != nil {
 			log.Warn("c.next error dropping peer", "err", err)
 			p.Drop()
-				return
+			return err
 		}
 	case <-c.quit:
 		log.Debug("client.handleOfferedHashesMsg() quit")
-			return
+		return nil
 	case <-ctx.Done():
 		log.Debug("client.handleOfferedHashesMsg() context done", "ctx.Err()", ctx.Err())
-			return
+		return nil
 	}
 	log.Trace("sending want batch", "peer", p.ID(), "stream", msg.Stream, "from", msg.From, "to", msg.To)
 
@@ -320,11 +312,11 @@ func (p *Peer) handleOfferedHashesMsg(ctx context.Context, req *OfferedHashesMsg
 		metrics.GetOrRegisterResettingTimer("handleoffered.wantdelay", nil).UpdateSince(wantDelay)
 	}
 
-		err := p.SendPriority(ctx, msg, c.priority)
+	err = p.SendPriority(ctx, msg, c.priority)
 	if err != nil {
 		log.Warn("SendPriority error", "err", err)
 	}
-	}()
+
 	return nil
 }
 

swarm/storage/localstore/localstore.go

@@ -73,7 +73,7 @@ type DB struct {
 	pullTriggers   map[uint8][]chan struct{}
 	pullTriggersMu sync.RWMutex
 
-	// binIDs stores the latest chunk serial ID for very
+	// binIDs stores the latest chunk serial ID for every
 	// proximity order bin
 	binIDs shed.Uint64Vector
 

swarm/storage/localstore/mode_get.go

@@ -47,7 +47,7 @@ func (db *DB) Get(ctx context.Context, mode chunk.ModeGet, addr chunk.Address) (
 
 	defer func() {
 		if err != nil {
-			metrics.GetOrRegisterCounter(fmt.Sprintf(metricName+".error", mode), nil).Inc(1)
+			metrics.GetOrRegisterCounter(metricName+".error", nil).Inc(1)
 		}
 	}()
 

swarm/storage/localstore/subscription_pull.go

@@ -36,7 +36,7 @@ import (
 // Pull syncing index can be only subscribed to a particular proximity order bin. If since
 // is not 0, the iteration will start from the first item stored after that id. If until is not 0,
 // only chunks stored up to this id will be sent to the channel, and the returned channel will be
-// closed. The since-until interval is closed on the both sides [since,until]. Returned stop
+// closed. The since-until interval is open on since side, and closed on until side: (since,until] <=> [since+1,until]. Returned stop
 // function will terminate current and further iterations without errors, and also close the returned channel.
 // Make sure that you check the second returned parameter from the channel to stop iteration when its value
 // is false.