VOL-3049 patch for alarm sync and audit
Change-Id: I400eddba48732d9b3ef5b6212fbda03b5c1f7ef1
diff --git a/internal/pkg/onuadaptercore/alarm_manager.go b/internal/pkg/onuadaptercore/alarm_manager.go
index fdb17b3..6665bd4 100644
--- a/internal/pkg/onuadaptercore/alarm_manager.go
+++ b/internal/pkg/onuadaptercore/alarm_manager.go
@@ -21,14 +21,15 @@
"context"
"errors"
"fmt"
- "sync"
- "time"
-
+ "github.com/looplab/fsm"
"github.com/opencord/omci-lib-go"
me "github.com/opencord/omci-lib-go/generated"
"github.com/opencord/voltha-lib-go/v4/pkg/events/eventif"
"github.com/opencord/voltha-lib-go/v4/pkg/log"
"github.com/opencord/voltha-protos/v4/go/voltha"
+ "reflect"
+ "sync"
+ "time"
)
const (
@@ -36,8 +37,33 @@
physicalPathTerminationPointEthernetUniClassID = me.PhysicalPathTerminationPointEthernetUniClassID
onuGClassID = me.OnuGClassID
aniGClassID = me.AniGClassID
+ auditInterval = 300
+ defaultTimeoutDelay = 10
+ alarmBitMapSizeBytes = 28
)
+const (
+ // events of alarm sync FSM
+ asEvStart = "asEvStart"
+ asEvStop = "asEvStop"
+ asEvAudit = "asEvAudit"
+ asEvSync = "asEvSync"
+ asEvSuccess = "asEvSuccess"
+ asEvFailure = "asEvFailure"
+ asEvResync = "asEvResync"
+)
+const (
+ // states of alarm sync FSM
+ asStStarting = "asStStarting"
+ asStDisabled = "asStDisabled"
+ asStInSync = "asStInSync"
+ asStAuditing = "asStAuditing"
+ asStResynchronizing = "asStResynchronizing"
+ asStIdle = "asStIdle"
+)
+
+//const cAsFsmIdleState = asStIdle not using idle state currently
+
type alarmInfo struct {
classID me.ClassID
instanceID uint16
@@ -46,6 +72,12 @@
type alarms map[alarmInfo]struct{}
+type meAlarmKey struct {
+ classID me.ClassID
+ instanceID uint16
+}
+
+type alarmBitMapDB map[meAlarmKey][alarmBitMapSizeBytes]byte
type onuAlarmManager struct {
pDeviceHandler *deviceHandler
eventProxy eventif.EventProxy
@@ -54,7 +86,16 @@
onuAlarmManagerLock sync.RWMutex
processMessage bool
activeAlarms alarms
+ alarmBitMapDB alarmBitMapDB
onuEventsList map[onuDevice]onuDeviceEvent
+ lastAlarmSequence uint8
+ alarmSyncFsm *AdapterFsm
+ oltDbCopy alarmBitMapDB
+ onuDBCopy alarmBitMapDB
+ bufferedNotifications []*omci.AlarmNotificationMsg
+ alarmUploadSeqNo uint16
+ alarmUploadNoOfCmds uint16
+ stopAlarmAuditTimer chan struct{}
}
type onuDevice struct {
@@ -70,13 +111,15 @@
func newAlarmManager(ctx context.Context, dh *deviceHandler) *onuAlarmManager {
var alarmManager onuAlarmManager
- logger.Debugw(ctx, "init-alarmManager", log.Fields{"device-id": dh.deviceID})
+ logger.Debugw(ctx, "init-alarm-manager", log.Fields{"device-id": dh.deviceID})
alarmManager.pDeviceHandler = dh
alarmManager.eventProxy = dh.EventProxy // Or event proxy should be on cluster address ??
alarmManager.eventChannel = make(chan Message)
alarmManager.stopProcessingOmciMessages = make(chan bool)
alarmManager.processMessage = false
alarmManager.activeAlarms = make(map[alarmInfo]struct{})
+ alarmManager.alarmBitMapDB = make(map[meAlarmKey][alarmBitMapSizeBytes]byte)
+ alarmManager.stopAlarmAuditTimer = make(chan struct{})
alarmManager.onuEventsList = map[onuDevice]onuDeviceEvent{
{classID: circuitPackClassID, alarmno: 0}: {EventName: "ONU_EQUIPMENT",
EventCategory: voltha.EventCategory_EQUIPMENT, EventSubCategory: voltha.EventSubCategory_ONU, EventDescription: "onu equipment"},
@@ -115,15 +158,338 @@
{classID: aniGClassID, alarmno: 6}: {EventName: "ONU_LASER_BIAS_CURRENT",
EventCategory: voltha.EventCategory_EQUIPMENT, EventSubCategory: voltha.EventSubCategory_ONU, EventDescription: "onu laser bias current"},
}
+ alarmManager.alarmSyncFsm = NewAdapterFsm("AlarmSync", dh.deviceID, alarmManager.eventChannel)
+ alarmManager.alarmSyncFsm.pFsm = fsm.NewFSM(
+ asStDisabled,
+ fsm.Events{
+ {Name: asEvStart, Src: []string{asStDisabled}, Dst: asStStarting},
+ {Name: asEvAudit, Src: []string{asStStarting, asStInSync}, Dst: asStAuditing},
+ {Name: asEvSync, Src: []string{asStStarting}, Dst: asStInSync},
+ {Name: asEvSuccess, Src: []string{asStAuditing, asStResynchronizing}, Dst: asStInSync},
+ {Name: asEvFailure, Src: []string{asStAuditing, asStResynchronizing}, Dst: asStAuditing},
+ {Name: asEvResync, Src: []string{asStAuditing}, Dst: asStResynchronizing},
+ {Name: asEvStop, Src: []string{asStDisabled, asStStarting, asStAuditing, asStInSync, asStIdle, asStResynchronizing}, Dst: asStDisabled},
+ },
+ fsm.Callbacks{
+ "enter_state": func(e *fsm.Event) { alarmManager.alarmSyncFsm.logFsmStateChange(ctx, e) },
+ "enter_" + asStStarting: func(e *fsm.Event) { alarmManager.asFsmStarting(ctx, e) },
+ "enter_" + asStAuditing: func(e *fsm.Event) { alarmManager.asFsmAuditing(ctx, e) },
+ "enter_" + asStInSync: func(e *fsm.Event) { alarmManager.asFsmInSync(ctx, e) },
+ "enter_" + asStResynchronizing: func(e *fsm.Event) { alarmManager.asFsmResynchronizing(ctx, e) },
+ },
+ )
return &alarmManager
}
-// getDeviceEventData returns the event data for a device
-func (am *onuAlarmManager) getDeviceEventData(ctx context.Context, classID me.ClassID, alarmNo uint8) (onuDeviceEvent, error) {
- if onuEventDetails, ok := am.onuEventsList[onuDevice{classID: classID, alarmno: alarmNo}]; ok {
- return onuEventDetails, nil
+func (am *onuAlarmManager) asFsmStarting(ctx context.Context, e *fsm.Event) {
+ logger.Debugw(ctx, "alarm-sync-fsm-start-processing-msgs-in-state", log.Fields{"state": e.FSM.Current(), "device-id": am.pDeviceHandler.deviceID})
+ go am.processAlarmSyncMessages(ctx)
+ // Start the first audit, if audit interval configured, else reach the sync state
+ if auditInterval > 0 {
+ select {
+ //Transition into auditing state, using a very shorter timeout delay here, hence it is the first audit
+ case <-time.After(defaultTimeoutDelay * time.Second):
+ go func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvAudit); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-auditing", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }()
+ case <-am.stopAlarmAuditTimer:
+ logger.Infow(ctx, "stopping-alarm-timer", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ return
+ }
+ } else {
+ // Transition into sync state directly.
+ go func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvSync); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-sync", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }()
}
- return onuDeviceEvent{}, errors.New("onu Event Detail not found")
+}
+
+func (am *onuAlarmManager) asFsmAuditing(ctx context.Context, e *fsm.Event) {
+ logger.Debugw(ctx, "alarm-sync-fsm-start-auditing", log.Fields{"state": e.FSM.Current(), "device-id": am.pDeviceHandler.deviceID})
+ // Always reset the buffered notifications and db copies before starting the audit
+ am.onuAlarmManagerLock.Lock()
+ am.bufferedNotifications = make([]*omci.AlarmNotificationMsg, 0)
+ am.oltDbCopy = make(map[meAlarmKey][alarmBitMapSizeBytes]byte)
+ am.onuDBCopy = make(map[meAlarmKey][alarmBitMapSizeBytes]byte)
+ am.onuAlarmManagerLock.Unlock()
+ failureTransition := func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvFailure); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-failure", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }
+ if err := am.pDeviceHandler.pOnuOmciDevice.PDevOmciCC.sendGetAllAlarm(log.WithSpanFromContext(context.TODO(), ctx), 0,
+ am.pDeviceHandler.pOpenOnuAc.omciTimeout, true); err != nil {
+ // Transition to failure so that alarm sync can be restarted again
+ go failureTransition()
+ }
+}
+
+func (am *onuAlarmManager) asFsmResynchronizing(ctx context.Context, e *fsm.Event) {
+ logger.Debugw(ctx, "alarm-sync-fsm", log.Fields{"state": e.FSM.Current(), "device-id": am.pDeviceHandler.deviceID})
+ failureTransition := func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvFailure); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-failure", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }
+ // See if there is any onu only diff, meaning the class and entity is only in onu DB
+ for alarm := range am.onuDBCopy {
+ if _, exists := am.oltDbCopy[meAlarmKey{
+ classID: alarm.classID,
+ instanceID: alarm.instanceID,
+ }]; !exists {
+ // We need to raise all such alarms as OLT wont have received notification for these alarms
+ omciAlarmMessage := &omci.AlarmNotificationMsg{
+ MeBasePacket: omci.MeBasePacket{
+ EntityClass: alarm.classID,
+ EntityInstance: alarm.instanceID,
+ },
+ AlarmBitmap: am.onuDBCopy[alarm],
+ }
+ if err := am.processAlarmData(ctx, omciAlarmMessage); err != nil {
+ logger.Errorw(ctx, "unable-to-process-alarm-notification", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ // Transition to failure.
+ go failureTransition()
+ return
+ }
+ }
+ }
+ // See if there is any olt only diff, meaning the class and entity is only in olt DB
+ for alarm := range am.oltDbCopy {
+ if _, exists := am.onuDBCopy[meAlarmKey{
+ classID: alarm.classID,
+ instanceID: alarm.instanceID,
+ }]; !exists {
+ // We need to clear all such alarms as OLT might have stale data and the alarms are already cleared.
+ omciAlarmMessage := &omci.AlarmNotificationMsg{
+ MeBasePacket: omci.MeBasePacket{
+ EntityClass: alarm.classID,
+ EntityInstance: alarm.instanceID,
+ },
+ AlarmBitmap: am.oltDbCopy[alarm],
+ }
+ if err := am.processAlarmData(ctx, omciAlarmMessage); err != nil {
+ logger.Errorw(ctx, "unable-to-process-alarm-notification", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ // Transition to failure
+ go failureTransition()
+ return
+ }
+ }
+ }
+ // See if there is any attribute difference
+ for alarm := range am.onuDBCopy {
+ if _, exists := am.oltDbCopy[alarm]; exists {
+ if am.onuDBCopy[alarm] != am.oltDbCopy[alarm] {
+ omciAlarmMessage := &omci.AlarmNotificationMsg{
+ MeBasePacket: omci.MeBasePacket{
+ EntityClass: alarm.classID,
+ EntityInstance: alarm.instanceID,
+ },
+ AlarmBitmap: am.onuDBCopy[alarm],
+ }
+ // We will assume that onudb is correct always in this case and process the changed bitmap.
+ if err := am.processAlarmData(ctx, omciAlarmMessage); err != nil {
+ logger.Errorw(ctx, "unable-to-process-alarm-notification", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ // Transition to failure
+ go failureTransition()
+ return
+ }
+ }
+ }
+ }
+ // Send the buffered notifications if no failure.
+ for _, notif := range am.bufferedNotifications {
+ logger.Debugw(ctx, "processing-buffered-alarm-notification", log.Fields{"device-id": am.pDeviceHandler.deviceID,
+ "notification": notif})
+ if err := am.processAlarmData(ctx, notif); err != nil {
+ logger.Errorw(ctx, "unable-to-process-alarm-notification", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ go failureTransition()
+ }
+ }
+ go func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvSuccess); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-sync", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }()
+}
+
+func (am *onuAlarmManager) asFsmInSync(ctx context.Context, e *fsm.Event) {
+ logger.Debugw(ctx, "alarm-sync-fsm", log.Fields{"state": e.FSM.Current(), "device-id": am.pDeviceHandler.deviceID})
+ if auditInterval > 0 {
+ select {
+ case <-time.After(auditInterval * time.Second):
+ go func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvAudit); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-auditing", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }()
+ case <-am.stopAlarmAuditTimer:
+ logger.Infow(ctx, "stopping-alarm-timer", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ return
+ }
+ }
+}
+
+func (am *onuAlarmManager) processAlarmSyncMessages(ctx context.Context) {
+ logger.Debugw(ctx, "start-routine-to-process-omci-messages-for-alarm-sync", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ am.flushAlarmSyncChannels(ctx)
+loop:
+ for {
+ message, ok := <-am.eventChannel
+ if !ok {
+ logger.Info(ctx, "alarm-sync-omci-message-could-not-be-read-from-channel", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ break loop
+ }
+ logger.Debugw(ctx, "alarm-sync-omci-message-received", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+
+ switch message.Type {
+ case OMCI:
+ msg, _ := message.Data.(OmciMessage)
+ am.handleOmciMessage(ctx, msg)
+ default:
+ logger.Warn(ctx, "alarm-sync-unknown-message-type-received", log.Fields{"device-id": am.pDeviceHandler.deviceID, "message.Type": message.Type})
+ }
+ }
+ logger.Info(ctx, "alarm-sync-stopped-handling-of-alarm-sync-omci-message", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ _ = am.alarmSyncFsm.pFsm.Event(asEvStop)
+}
+
+func (am *onuAlarmManager) handleOmciMessage(ctx context.Context, msg OmciMessage) {
+ logger.Debugw(ctx, "alarm-sync-omci-message-received", log.Fields{"device-id": am.pDeviceHandler.deviceID,
+ "msg-type": msg.OmciMsg.MessageType, "msg": msg})
+ switch msg.OmciMsg.MessageType {
+ case omci.GetAllAlarmsResponseType:
+ am.handleOmciGetAllAlarmsResponseMessage(ctx, msg)
+ case omci.GetAllAlarmsNextResponseType:
+ am.handleOmciGetAllAlarmNextResponseMessage(ctx, msg)
+ default:
+ logger.Warnw(ctx, "unknown-message-type", log.Fields{"msg-type": msg.OmciMsg.MessageType})
+
+ }
+}
+
+func (am *onuAlarmManager) handleOmciGetAllAlarmsResponseMessage(ctx context.Context, msg OmciMessage) {
+ msgLayer := (*msg.OmciPacket).Layer(omci.LayerTypeGetAllAlarmsResponse)
+ if msgLayer == nil {
+ logger.Errorw(ctx, "omci-msg-layer-could-not-be-detected", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ return
+ }
+ msgObj, msgOk := msgLayer.(*omci.GetAllAlarmsResponse)
+ if !msgOk {
+ logger.Errorw(ctx, "omci-msg-layer-could-not-be-assigned", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ return
+ }
+ logger.Debugw(ctx, "get-all-alarm-response-data", log.Fields{"device-id": am.pDeviceHandler.deviceID, "data-fields": msgObj})
+ if am.alarmSyncFsm.pFsm.Is(asStDisabled) {
+ logger.Debugw(ctx, "alarm-sync-fsm-is-disabled-ignoring-response-message", log.Fields{"device-id": am.pDeviceHandler.deviceID, "data-fields": msgObj})
+ return
+ }
+ am.alarmUploadNoOfCmds = msgObj.NumberOfCommands
+ failureTransition := func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvFailure); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-failure", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }
+ if am.alarmUploadSeqNo < am.alarmUploadNoOfCmds {
+ // Reset Onu Alarm Sequence
+ am.onuAlarmManagerLock.Lock()
+ am.resetAlarmSequence()
+ // Get a copy of the alarm bit map db.
+ for alarms, bitmap := range am.alarmBitMapDB {
+ am.oltDbCopy[alarms] = bitmap
+ }
+ am.onuAlarmManagerLock.Unlock()
+ if err := am.pDeviceHandler.pOnuOmciDevice.PDevOmciCC.sendGetAllAlarmNext(
+ log.WithSpanFromContext(context.TODO(), ctx), am.pDeviceHandler.pOpenOnuAc.omciTimeout, true); err != nil {
+ // Transition to failure
+ go failureTransition()
+ }
+ } else if am.alarmUploadNoOfCmds == 0 {
+ // Reset Onu Alarm Sequence
+ am.onuAlarmManagerLock.Lock()
+ am.resetAlarmSequence()
+ // Get a copy of the alarm bit map db.
+ for alarms, bitmap := range am.alarmBitMapDB {
+ am.oltDbCopy[alarms] = bitmap
+ }
+ am.onuAlarmManagerLock.Unlock()
+ if am.isAlarmDBDiffPresent(ctx) {
+ // transition to resync state
+ go func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvResync); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-resynchronizing", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }()
+ } else {
+ // Transition to sync state
+ go func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvSuccess); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-sync", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }()
+ }
+ } else {
+ logger.Errorw(ctx, "invalid-number-of-commands-received", log.Fields{"device-id": am.pDeviceHandler.deviceID,
+ "upload-no-of-cmds": am.alarmUploadNoOfCmds, "upload-seq-no": am.alarmUploadSeqNo})
+ go failureTransition()
+ }
+}
+
+func (am *onuAlarmManager) handleOmciGetAllAlarmNextResponseMessage(ctx context.Context, msg OmciMessage) {
+ msgLayer := (*msg.OmciPacket).Layer(omci.LayerTypeGetAllAlarmsNextResponse)
+
+ if msgLayer == nil {
+ logger.Errorw(ctx, "omci-msg-layer-could-not-be-detected", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ return
+ }
+ msgObj, msgOk := msgLayer.(*omci.GetAllAlarmsNextResponse)
+ if !msgOk {
+ logger.Errorw(ctx, "omci-msg-layer-could-not-be-assigned", log.Fields{"device-id": am.pDeviceHandler.deviceID})
+ return
+ }
+ logger.Debugw(ctx, "get-all-alarms-next-response-data",
+ log.Fields{"device-id": am.pDeviceHandler.deviceID, "data-fields": msgObj})
+ meClassID := msgObj.AlarmEntityClass
+ meEntityID := msgObj.AlarmEntityInstance
+ meAlarmBitMap := msgObj.AlarmBitMap
+
+ am.onuAlarmManagerLock.Lock()
+ am.onuDBCopy[meAlarmKey{
+ classID: meClassID,
+ instanceID: meEntityID,
+ }] = meAlarmBitMap
+ am.onuAlarmManagerLock.Unlock()
+ failureTransition := func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvFailure); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-failure", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }
+ if am.alarmUploadSeqNo < am.alarmUploadNoOfCmds {
+ if err := am.pDeviceHandler.pOnuOmciDevice.PDevOmciCC.sendGetAllAlarmNext(
+ log.WithSpanFromContext(context.TODO(), ctx), am.pDeviceHandler.pOpenOnuAc.omciTimeout, true); err != nil {
+ // Transition to failure
+ go failureTransition()
+ } //TODO: needs to handle timeouts
+ } else {
+ if am.isAlarmDBDiffPresent(ctx) {
+ // transition to resync state
+ go func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvResync); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-resynchronizing", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }()
+ } else {
+ // Transition to sync state
+ go func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvSuccess); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-sync", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }()
+ }
+ }
}
func (am *onuAlarmManager) startOMCIAlarmMessageProcessing(ctx context.Context) {
@@ -132,37 +498,55 @@
if am.activeAlarms == nil {
am.activeAlarms = make(map[alarmInfo]struct{})
}
+ am.alarmBitMapDB = make(map[meAlarmKey][alarmBitMapSizeBytes]byte)
am.onuAlarmManagerLock.Unlock()
+ if am.alarmSyncFsm.pFsm.Is(asStDisabled) {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvStart); err != nil {
+ logger.Errorw(ctx, "alarm-sync-fsm-can-not-go-to-state-starting", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ return
+ }
+ } else {
+ logger.Errorw(ctx, "wrong-state-of-alarm-sync-fsm-want-disabled", log.Fields{"state": string(am.alarmSyncFsm.pFsm.Current()),
+ "device-id": am.pDeviceHandler.deviceID})
+ return
+ }
+ logger.Debugw(ctx, "alarm-sync-fsm-started", log.Fields{"state": string(am.alarmSyncFsm.pFsm.Current())})
+
if stop := <-am.stopProcessingOmciMessages; stop {
am.onuAlarmManagerLock.Lock()
am.processMessage = false
am.activeAlarms = nil
+ am.alarmBitMapDB = nil
+ am.alarmUploadNoOfCmds = 0
+ am.alarmUploadSeqNo = 0
am.onuAlarmManagerLock.Unlock()
+
}
}
func (am *onuAlarmManager) handleOmciAlarmNotificationMessage(ctx context.Context, msg OmciMessage) {
- logger.Debugw(ctx, "OMCI Alarm Notification Msg", log.Fields{"device-id": am.pDeviceHandler.deviceID,
- "msgType": msg.OmciMsg.MessageType})
- am.onuAlarmManagerLock.Lock()
- defer am.onuAlarmManagerLock.Unlock()
+ logger.Debugw(ctx, "omci-alarm-notification-msg", log.Fields{"device-id": am.pDeviceHandler.deviceID,
+ "msg-type": msg.OmciMsg.MessageType})
+ am.onuAlarmManagerLock.RLock()
+ processMessage := am.processMessage
+ am.onuAlarmManagerLock.RUnlock()
- if am.processMessage {
+ if processMessage {
msgLayer := (*msg.OmciPacket).Layer(omci.LayerTypeAlarmNotification)
if msgLayer == nil {
- logger.Errorw(ctx, "Omci Msg layer could not be detected for Alarm Notification",
+ logger.Errorw(ctx, "omci-msg-layer-could-not-be-detected-for-alarm-notification",
log.Fields{"device-id": am.pDeviceHandler.deviceID})
return
}
msgObj, msgOk := msgLayer.(*omci.AlarmNotificationMsg)
if !msgOk {
- logger.Errorw(ctx, "Omci Msg layer could not be assigned for Alarm Notification",
+ logger.Errorw(ctx, "omci-msg-layer-could-not-be-assigned-for-alarm-notification",
log.Fields{"device-id": am.pDeviceHandler.deviceID})
return
}
//Alarm Notification decoding at omci lib validates that the me class ID supports the
// alarm notifications.
- logger.Debugw(ctx, "Alarm Notification Data", log.Fields{"device-id": am.pDeviceHandler.deviceID, "data-fields": msgObj})
+ logger.Debugw(ctx, "alarm-notification-data", log.Fields{"device-id": am.pDeviceHandler.deviceID, "data-fields": msgObj})
if err := am.processAlarmData(ctx, msgObj); err != nil {
logger.Errorw(ctx, "unable-to-process-alarm-notification", log.Fields{"device-id": am.pDeviceHandler.deviceID})
}
@@ -184,12 +568,25 @@
logger.Errorw(ctx, "me-class-instance-not-present", log.Fields{"class-id": classID, "instance-id": meInstance})
return fmt.Errorf("me-class-%d-instance-%d-not-present", classID, meInstance)
}
- /*
- if sequenceNo > 0 {
- // TODO Need Auditing if sequence no does not matches, after incrementing the last sequence no by 1
+ am.onuAlarmManagerLock.Lock()
+ defer am.onuAlarmManagerLock.Unlock()
+ if sequenceNo > 0 {
+ if am.alarmSyncFsm.pFsm.Is(asStAuditing) || am.alarmSyncFsm.pFsm.Is(asStResynchronizing) {
+ am.bufferedNotifications = append(am.bufferedNotifications, msg)
+ logger.Debugw(ctx, "adding-notification-to-buffered-notification-list", log.Fields{"device-id": am.pDeviceHandler.deviceID,
+ "notification": msg})
+ return nil
}
- */
-
+ am.incrementAlarmSequence()
+ if sequenceNo != am.lastAlarmSequence && auditInterval > 0 {
+ // signal early audit, if no match(if we are reaching here it means that audit is not going on currently)
+ go func() {
+ if err := am.alarmSyncFsm.pFsm.Event(asEvAudit); err != nil {
+ logger.Debugw(ctx, "alarm-sync-fsm-cannot-go-to-state-auditing", log.Fields{"device-id": am.pDeviceHandler.deviceID, "err": err})
+ }
+ }()
+ }
+ }
entity, omciErr := me.LoadManagedEntityDefinition(classID,
me.ParamData{EntityID: meInstance})
if omciErr.StatusCode() != me.Success {
@@ -202,6 +599,11 @@
logger.Error(ctx, "unable-to-get-managed-entity-alarm-map", log.Fields{"class-id": classID, "instance-id": meInstance})
return fmt.Errorf("unable-to-get-managed-entity-alarm-map-%d-instance-%d", classID, meInstance)
}
+
+ am.alarmBitMapDB[meAlarmKey{
+ classID: classID,
+ instanceID: meInstance,
+ }] = alarmBitmap
// Loop over the supported alarm list for this me
for alarmNo := range meAlarmMap {
// Check if alarmNo was previously active in the alarms, if yes clear it and remove it from active alarms
@@ -245,6 +647,7 @@
instanceID: instanceID,
alarmNo: alarm,
}] = struct{}{}
+
go am.sendAlarm(ctx, classID, instanceID, alarm, true)
}
@@ -255,6 +658,13 @@
instanceID: instanceID,
alarmNo: alarm,
})
+ key := meAlarmKey{
+ classID: classID,
+ instanceID: instanceID,
+ }
+ if am.alarmBitMapDB[key] == [alarmBitMapSizeBytes]byte{0} {
+ delete(am.alarmBitMapDB, key)
+ }
}
func (am *onuAlarmManager) getIntfIDAlarm(ctx context.Context, classID me.ClassID, instanceID uint16) *uint32 {
@@ -308,3 +718,43 @@
_ = am.eventProxy.SendDeviceEvent(ctx, deviceEvent, eventDetails.EventCategory, eventDetails.EventSubCategory,
raisedTimestamp)
}
+
+func (am *onuAlarmManager) isAlarmDBDiffPresent(ctx context.Context) bool {
+ return !reflect.DeepEqual(am.onuDBCopy, am.oltDbCopy)
+}
+
+func (am *onuAlarmManager) incrementAlarmSequence() {
+ //alarm sequence number wraps from 255 to 1.
+ if am.lastAlarmSequence == 255 {
+ am.lastAlarmSequence = 1
+ } else {
+ am.lastAlarmSequence++
+ }
+}
+
+func (am *onuAlarmManager) resetAlarmSequence() {
+ am.lastAlarmSequence = 0
+}
+
+// flushAlarmSyncChannels flushes all alarm sync channels to discard any previous response
+func (am *onuAlarmManager) flushAlarmSyncChannels(ctx context.Context) {
+ // flush alarm sync channel
+ select {
+ case <-am.eventChannel:
+ logger.Debug(ctx, "flushed-alarm-sync-channel")
+ default:
+ }
+ select {
+ case <-am.stopAlarmAuditTimer:
+ logger.Debug(ctx, "flushed-alarm-audit-timer-channel")
+ default:
+ }
+}
+
+// getDeviceEventData returns the event data for a device
+func (am *onuAlarmManager) getDeviceEventData(ctx context.Context, classID me.ClassID, alarmNo uint8) (onuDeviceEvent, error) {
+ if onuEventDetails, ok := am.onuEventsList[onuDevice{classID: classID, alarmno: alarmNo}]; ok {
+ return onuEventDetails, nil
+ }
+ return onuDeviceEvent{}, errors.New("onu Event Detail not found")
+}
diff --git a/internal/pkg/onuadaptercore/device_handler.go b/internal/pkg/onuadaptercore/device_handler.go
index 3422c43..59cd8d0 100644
--- a/internal/pkg/onuadaptercore/device_handler.go
+++ b/internal/pkg/onuadaptercore/device_handler.go
@@ -3163,8 +3163,12 @@
dh.setAlarmManagerIsRunning(true)
if stop := <-dh.stopAlarmManager; stop {
logger.Debugw(ctx, "stopping-collector-for-onu", log.Fields{"device-id": dh.device.Id})
- dh.pAlarmMgr.stopProcessingOmciMessages <- true // Stop the OMCI routines if any
dh.setAlarmManagerIsRunning(false)
+ go func() {
+ _ = dh.pAlarmMgr.alarmSyncFsm.pFsm.Event(asEvStop)
+ }()
+ dh.pAlarmMgr.stopAlarmAuditTimer <- struct{}{}
+ dh.pAlarmMgr.stopProcessingOmciMessages <- true // Stop the OMCI routines if any(This will stop the fsms also)
}
}
diff --git a/internal/pkg/onuadaptercore/omci_cc.go b/internal/pkg/onuadaptercore/omci_cc.go
index 3a46f45..1f32be0 100644
--- a/internal/pkg/onuadaptercore/omci_cc.go
+++ b/internal/pkg/onuadaptercore/omci_cc.go
@@ -156,7 +156,6 @@
omciCC.hpTid = 0x8000
omciCC.uploadSequNo = 0
omciCC.uploadNoOfCmds = 0
-
omciCC.txQueue = list.New()
omciCC.rxSchedulerMap = make(map[uint16]callbackPairEntry)
@@ -779,6 +778,58 @@
return oo.send(ctx, pkt, timeout, 0, highPrio, omciRxCallbackPair)
}
+func (oo *omciCC) sendGetAllAlarm(ctx context.Context, alarmRetreivalMode uint8, timeout int, highPrio bool) error {
+ logger.Debugw(ctx, "send GetAllAlarms-msg to:", log.Fields{"device-id": oo.deviceID})
+ request := &omci.GetAllAlarmsRequest{
+ MeBasePacket: omci.MeBasePacket{
+ EntityClass: me.OnuDataClassID,
+ },
+ AlarmRetrievalMode: byte(alarmRetreivalMode),
+ }
+ tid := oo.getNextTid(highPrio)
+ pkt, err := serialize(ctx, omci.GetAllAlarmsRequestType, request, tid)
+ if err != nil {
+ logger.Errorw(ctx, "Cannot serialize GetAllAlarmsRequest", log.Fields{
+ "Err": err, "device-id": oo.deviceID})
+ return err
+ }
+ oo.pBaseDeviceHandler.pAlarmMgr.alarmUploadSeqNo = 0
+ oo.pBaseDeviceHandler.pAlarmMgr.alarmUploadNoOfCmds = 0
+
+ omciRxCallbackPair := callbackPair{
+ cbKey: tid,
+ cbEntry: callbackPairEntry{(*oo.pBaseDeviceHandler.pAlarmMgr).eventChannel,
+ oo.receiveOmciResponse, true},
+ }
+ return oo.send(ctx, pkt, timeout, 0, highPrio, omciRxCallbackPair)
+}
+
+func (oo *omciCC) sendGetAllAlarmNext(ctx context.Context, timeout int, highPrio bool) error {
+ alarmUploadSeqNo := oo.pBaseDeviceHandler.pAlarmMgr.alarmUploadSeqNo
+ logger.Debugw(ctx, "send sendGetAllAlarmNext-msg to:", log.Fields{"device-id": oo.deviceID,
+ "alarmUploadSeqNo": alarmUploadSeqNo})
+ request := &omci.GetAllAlarmsNextRequest{
+ MeBasePacket: omci.MeBasePacket{
+ EntityClass: me.OnuDataClassID,
+ },
+ CommandSequenceNumber: alarmUploadSeqNo,
+ }
+ tid := oo.getNextTid(highPrio)
+ pkt, err := serialize(ctx, omci.GetAllAlarmsNextRequestType, request, tid)
+ if err != nil {
+ logger.Errorw(ctx, "Cannot serialize GetAllAlarmsNextRequest", log.Fields{
+ "Err": err, "device-id": oo.deviceID})
+ return err
+ }
+ oo.pBaseDeviceHandler.pAlarmMgr.alarmUploadSeqNo++
+
+ omciRxCallbackPair := callbackPair{
+ cbKey: tid,
+ cbEntry: callbackPairEntry{(*oo.pBaseDeviceHandler.pAlarmMgr).eventChannel, oo.receiveOmciResponse, true},
+ }
+ return oo.send(ctx, pkt, timeout, 0, highPrio, omciRxCallbackPair)
+}
+
func (oo *omciCC) sendCreateGalEthernetProfile(ctx context.Context, timeout int, highPrio bool) *me.ManagedEntity {
tid := oo.getNextTid(highPrio)
logger.Debugw(ctx, "send GalEnetProfile-Create-msg:", log.Fields{"device-id": oo.deviceID,