[TASK] add no-respondd

config_example.toml

@@ -55,6 +55,12 @@ save_interval = "5s"
 # Set node to offline if not seen within this period
 offline_after = "10m"
 
+## Verify if node is really down by ping last seen address of node
+# send x pings to verify if node is offline (for disable set count < 1)
+ping_count = 3
+# timeout of sending ping to a node
+ping_timeout = "1s"
+
 
 ## [[nodes.output.example]]
 # Each output format has its own config block and needs to be enabled by adding:

database/graphite/global.go

@@ -36,6 +36,7 @@ func (c *Connection) InsertGlobals(stats *runtime.GlobalStats, time time.Time, s
 func GlobalStatsFields(name string, stats *runtime.GlobalStats) []graphigo.Metric {
 	return []graphigo.Metric{
 		{Name: name + ".nodes", Value: stats.Nodes},
+		{Name: name + ".nodes.no_respondd", Value: stats.NodesNoRespondd},
 		{Name: name + ".gateways", Value: stats.Gateways},
 		{Name: name + ".clients.total", Value: stats.Clients},
 		{Name: name + ".clients.wifi", Value: stats.ClientsWifi},

database/influxdb/global.go

@@ -42,12 +42,13 @@ func (conn *Connection) InsertGlobals(stats *runtime.GlobalStats, time time.Time
 // GlobalStatsFields returns fields for InfluxDB
 func GlobalStatsFields(stats *runtime.GlobalStats) map[string]interface{} {
 	return map[string]interface{}{
-		"nodes":          stats.Nodes,
-		"gateways":       stats.Gateways,
-		"clients.total":  stats.Clients,
-		"clients.wifi":   stats.ClientsWifi,
-		"clients.wifi24": stats.ClientsWifi24,
-		"clients.wifi5":  stats.ClientsWifi5,
+		"nodes":             stats.Nodes,
+		"nodes.no_respondd": stats.NodesNoRespondd,
+		"gateways":          stats.Gateways,
+		"clients.total":     stats.Clients,
+		"clients.wifi":      stats.ClientsWifi,
+		"clients.wifi24":    stats.ClientsWifi24,
+		"clients.wifi5":     stats.ClientsWifi5,
 	}
 }
 

database/logging/file.go

@@ -50,7 +50,7 @@ func (conn *Connection) InsertLink(link *runtime.Link, time time.Time) {
 }
 
 func (conn *Connection) InsertGlobals(stats *runtime.GlobalStats, time time.Time, site string, domain string) {
-	conn.log("InsertGlobals: [", time.String(), "] site: ", site, " domain: ", domain, ", nodes: ", stats.Nodes, ", clients: ", stats.Clients, " models: ", len(stats.Models))
+	conn.log("InsertGlobals: [", time.String(), "] site: ", site, " domain: ", domain, ", nodes: ", stats.Nodes, " (no respondd: ", stats.NodesNoRespondd, "), clients: ", stats.Clients, " models: ", len(stats.Models))
 }
 
 func (conn *Connection) PruneNodes(deleteAfter time.Duration) {

docs/docs_configuration.md

@@ -203,6 +203,8 @@ state_path     = "/var/lib/yanic/state.json"
 prune_after    = "7d"
 save_interval  = "5s"
 offline_after  = "10m"
+ping_count = 3
+ping_timeout = "1s"
 ```
 {% endmethod %}
 
@@ -246,6 +248,26 @@ offline_after = "10m"
 ```
 {% endmethod %}
 
+### ping_count
+{% method %}
+Verify if node is really down by ping last seen address of node
+send x pings to verify if node is offline (for disable set count < 1)
+{% sample lang="toml" %}
+```toml
+ping_count = 3
+```
+{% endmethod %}
+
+
+### ping_timeout
+{% method %}
+Timeout of sending ping to a node
+{% sample lang="toml" %}
+```toml
+ping_timeout = "1s"
+```
+{% endmethod %}
+
 
 ## [[nodes.output.example]]
 {% method %}

runtime/node.go

@@ -13,6 +13,7 @@ type Node struct {
 	Firstseen  jsontime.Time    `json:"firstseen"`
 	Lastseen   jsontime.Time    `json:"lastseen"`
 	Online     bool             `json:"online"`
+	NoRespondd bool             `json:"-"`
 	Statistics *data.Statistics `json:"statistics"`
 	Nodeinfo   *data.NodeInfo   `json:"nodeinfo"`
 	Neighbours *data.Neighbours `json:"-"`

runtime/nodes.go

@@ -176,7 +176,16 @@ func (nodes *Nodes) expire() {
 			delete(nodes.List, id)
 		} else if node.Lastseen.Before(offlineAfter) {
 			// set to offline
-			node.Online = false
+			if nodes.config.PingCount > 0 && nodes.ping(node) {
+				node.Online = true
+				node.NoRespondd = true
+
+				node.Statistics = nil
+				node.Neighbours = nil
+			} else {
+				node.Online = false
+				node.NoRespondd = false
+			}
 		}
 	}
 }

runtime/nodes_config.go

@@ -7,5 +7,7 @@ type NodesConfig struct {
 	SaveInterval duration.Duration `toml:"save_interval"` // Save nodes periodically
 	OfflineAfter duration.Duration `toml:"offline_after"` // Set node to offline if not seen within this period
 	PruneAfter   duration.Duration `toml:"prune_after"`   // Remove nodes after n days of inactivity
+	PingCount    int               `toml:"ping_count"`    // send x pings to verify if node is offline (for disable count < 1)
+	PingTimeout  duration.Duration `toml:"ping_timeout"`  // timeout of sending ping to a node
 	Output       map[string]interface{}
 }

runtime/nodes_ping.go

@@ -0,0 +1,38 @@
+package runtime
+
+import (
+	"github.com/bdlm/log"
+	"github.com/sparrc/go-ping"
+)
+
+func (nodes *Nodes) ping(node *Node) bool {
+	logNode := log.WithField("node_id", "unknown")
+	if node.Nodeinfo != nil {
+		logNode = logNode.WithField("node_id", node.Nodeinfo.NodeID)
+	}
+	if node.Address == nil {
+		logNode.Debug("error no address found")
+		return false
+	}
+	addr := node.Address.IP.String()
+	if node.Address.IP.IsLinkLocalUnicast() {
+		addr += "%" + node.Address.Zone
+	}
+
+	logAddr := logNode.WithField("addr", addr)
+
+	pinger, err := ping.NewPinger(addr)
+	if err != nil {
+		logAddr.Debugf("error during ping: %s", err)
+		return false
+	}
+	//pinger.SetPrivileged(true)
+	pinger.Count = nodes.config.PingCount
+	pinger.Timeout = nodes.config.PingTimeout.Duration
+	pinger.Run() // blocks until finished
+	stats := pinger.Statistics()
+	logAddr.WithFields(map[string]interface{}{
+		"pkg_lost": stats.PacketLoss,
+	}).Debug("pong")
+	return stats.PacketLoss < 100
+}

runtime/nodes_ping_test.go

@@ -0,0 +1,38 @@
+package runtime
+
+import (
+	"net"
+	"testing"
+	"time"
+
+	"github.com/bdlm/log"
+	"github.com/stretchr/testify/assert"
+
+	"github.com/FreifunkBremen/yanic/data"
+)
+
+func TestPing(t *testing.T) {
+	log.SetLevel(log.DebugLevel)
+
+	assert := assert.New(t)
+	config := &NodesConfig{
+		PingCount: 1,
+	}
+	config.OfflineAfter.Duration = time.Minute * 10
+	// to get default (100%) path of testing
+	// config.PruneAfter.Duration = time.Hour * 24 * 6
+	nodes := &Nodes{
+		config:        config,
+		List:          make(map[string]*Node),
+		ifaceToNodeID: make(map[string]string),
+	}
+
+	node := nodes.Update("expire", &data.ResponseData{NodeInfo: &data.NodeInfo{NodeID: "nodeID-Lola"}})
+	node.Address = &net.UDPAddr{Zone: "bat0"}
+	// error during ping
+	assert.False(nodes.ping(node))
+
+	node.Address.IP = net.ParseIP("fe80::1")
+	// error during ping
+	assert.False(nodes.ping(node))
+}

runtime/stats.go

@@ -11,12 +11,13 @@ type CounterMap map[string]uint32
 
 // GlobalStats struct
 type GlobalStats struct {
-	Clients       uint32
-	ClientsWifi   uint32
-	ClientsWifi24 uint32
-	ClientsWifi5  uint32
-	Gateways      uint32
-	Nodes         uint32
+	Clients         uint32
+	ClientsWifi     uint32
+	ClientsWifi24   uint32
+	ClientsWifi5    uint32
+	Gateways        uint32
+	Nodes           uint32
+	NodesNoRespondd uint32
 
 	Firmwares   CounterMap
 	Models      CounterMap
@@ -81,6 +82,9 @@ func (s *GlobalStats) Add(node *Node) {
 		s.ClientsWifi5 += stats.Clients.Wifi5
 		s.ClientsWifi += stats.Clients.Wifi
 	}
+	if node.NoRespondd {
+		s.NodesNoRespondd++
+	}
 	if node.IsGateway() {
 		s.Gateways++
 	}

runtime/stats_test.go

@@ -22,6 +22,7 @@ func TestGlobalStats(t *testing.T) {
 	//check GLOBAL_SITE stats
 	assert.EqualValues(1, stats[GLOBAL_SITE][GLOBAL_DOMAIN].Gateways)
 	assert.EqualValues(3, stats[GLOBAL_SITE][GLOBAL_DOMAIN].Nodes)
+	assert.EqualValues(1, stats[GLOBAL_SITE][GLOBAL_DOMAIN].NodesNoRespondd)
 	assert.EqualValues(25, stats[GLOBAL_SITE][GLOBAL_DOMAIN].Clients)
 
 	// check models
@@ -98,7 +99,8 @@ func createTestNodes() *Nodes {
 	nodes.AddNode(nodeData)
 
 	nodes.AddNode(&Node{
-		Online: true,
+		Online:     true,
+		NoRespondd: true,
 		Statistics: &data.Statistics{
 			Clients: data.Clients{
 				Total: 2,