CORD-348 Fabric multicast support - error handling
Automatically failover to backup spine if
- ingress - transit link down
- transit - egress link down
- transit device down
Can recover from fatal error with human involved
- ingress switch down
- egress switch down
- all links to spine down
Scan through McastRouteStore when
- SR activate
- link up
Also include following features
- Use flow objective context in McastHandler
- Update Mcast VLAN config sample
Change-Id: I75007d9efd7646e7c4e57fa6d3fc6943543153cf
diff --git a/apps/segmentrouting/src/main/java/org/onosproject/segmentrouting/SegmentRoutingManager.java b/apps/segmentrouting/src/main/java/org/onosproject/segmentrouting/SegmentRoutingManager.java
index 48bb1a7..5eec207 100644
--- a/apps/segmentrouting/src/main/java/org/onosproject/segmentrouting/SegmentRoutingManager.java
+++ b/apps/segmentrouting/src/main/java/org/onosproject/segmentrouting/SegmentRoutingManager.java
@@ -762,6 +762,8 @@
defaultRoutingHandler.populateRoutingRulesForLinkStatusChange(null);
//log.trace("processLinkAdded: re-starting route population process");
//defaultRoutingHandler.startPopulationProcess();
+
+ mcastHandler.init();
}
private void processLinkRemoved(Link link) {
@@ -775,6 +777,8 @@
defaultRoutingHandler.populateRoutingRulesForLinkStatusChange(link);
//log.trace("processLinkRemoved: re-starting route population process");
//defaultRoutingHandler.startPopulationProcess();
+
+ mcastHandler.processLinkDown(link);
}
private void processDeviceAdded(Device device) {
@@ -784,43 +788,50 @@
+ "processed after config completes.", device.id());
return;
}
+ processDeviceAddedInternal(device.id());
+ }
+
+ private void processDeviceAddedInternal(DeviceId deviceId) {
// Irrespective of whether the local is a MASTER or not for this device,
// we need to create a SR-group-handler instance. This is because in a
// multi-instance setup, any instance can initiate forwarding/next-objectives
// for any switch (even if this instance is a SLAVE or not even connected
// to the switch). To handle this, a default-group-handler instance is necessary
// per switch.
- if (groupHandlerMap.get(device.id()) == null) {
+ log.debug("Current groupHandlerMap devs: {}", groupHandlerMap.keySet());
+ if (groupHandlerMap.get(deviceId) == null) {
DefaultGroupHandler groupHandler;
try {
groupHandler = DefaultGroupHandler.
- createGroupHandler(device.id(),
- appId,
- deviceConfiguration,
- linkService,
- flowObjectiveService,
- this);
+ createGroupHandler(deviceId,
+ appId,
+ deviceConfiguration,
+ linkService,
+ flowObjectiveService,
+ this);
} catch (DeviceConfigNotFoundException e) {
log.warn(e.getMessage() + " Aborting processDeviceAdded.");
return;
}
- groupHandlerMap.put(device.id(), groupHandler);
+ log.debug("updating groupHandlerMap with new config for device: {}",
+ deviceId);
+ groupHandlerMap.put(deviceId, groupHandler);
// Also, in some cases, drivers may need extra
// information to process rules (eg. Router IP/MAC); and so, we send
// port addressing rules to the driver as well irrespective of whether
// this instance is the master or not.
- defaultRoutingHandler.populatePortAddressingRules(device.id());
+ defaultRoutingHandler.populatePortAddressingRules(deviceId);
}
- if (mastershipService.isLocalMaster(device.id())) {
- hostHandler.readInitialHosts(device.id());
- DefaultGroupHandler groupHandler = groupHandlerMap.get(device.id());
+ if (mastershipService.isLocalMaster(deviceId)) {
+ hostHandler.readInitialHosts(deviceId);
+ DefaultGroupHandler groupHandler = groupHandlerMap.get(deviceId);
groupHandler.createGroupsFromSubnetConfig();
- routingRulePopulator.populateSubnetBroadcastRule(device.id());
- groupHandler.createGroupsForXConnect(device.id());
- routingRulePopulator.populateXConnectBroadcastRule(device.id());
+ routingRulePopulator.populateSubnetBroadcastRule(deviceId);
+ groupHandler.createGroupsForXConnect(deviceId);
+ routingRulePopulator.populateXConnectBroadcastRule(deviceId);
}
- netcfgHandler.initVRouters(device.id());
+ netcfgHandler.initVRouters(deviceId);
}
private void processDeviceRemoved(Device device) {
@@ -829,34 +840,29 @@
.forEach(entry -> {
nsNextObjStore.remove(entry.getKey());
});
-
subnetNextObjStore.entrySet().stream()
.filter(entry -> entry.getKey().deviceId().equals(device.id()))
.forEach(entry -> {
subnetNextObjStore.remove(entry.getKey());
});
-
portNextObjStore.entrySet().stream()
.filter(entry -> entry.getKey().deviceId().equals(device.id()))
.forEach(entry -> {
portNextObjStore.remove(entry.getKey());
});
-
xConnectNextObjStore.entrySet().stream()
.filter(entry -> entry.getKey().deviceId().equals(device.id()))
.forEach(entry -> {
xConnectNextObjStore.remove(entry.getKey());
});
-
subnetVidStore.entrySet().stream()
.filter(entry -> entry.getKey().deviceId().equals(device.id()))
.forEach(entry -> {
subnetVidStore.remove(entry.getKey());
});
-
groupHandlerMap.remove(device.id());
-
defaultRoutingHandler.purgeEcmpGraph(device.id());
+ mcastHandler.removeDevice(device.id());
}
private void processPortRemoved(Device device, Port port) {
@@ -900,48 +906,11 @@
tunnelHandler, policyStore);
for (Device device : deviceService.getDevices()) {
- // Irrespective of whether the local is a MASTER or not for this device,
- // we need to create a SR-group-handler instance. This is because in a
- // multi-instance setup, any instance can initiate forwarding/next-objectives
- // for any switch (even if this instance is a SLAVE or not even connected
- // to the switch). To handle this, a default-group-handler instance is necessary
- // per switch.
- log.debug("Current groupHandlerMap devs: {}", groupHandlerMap.keySet());
- if (groupHandlerMap.get(device.id()) == null) {
- DefaultGroupHandler groupHandler;
- try {
- groupHandler = DefaultGroupHandler.
- createGroupHandler(device.id(),
- appId,
- deviceConfiguration,
- linkService,
- flowObjectiveService,
- segmentRoutingManager);
- } catch (DeviceConfigNotFoundException e) {
- log.warn(e.getMessage() + " Aborting configureNetwork.");
- return;
- }
- log.debug("updating groupHandlerMap with new config for "
- + "device: {}", device.id());
- groupHandlerMap.put(device.id(), groupHandler);
-
- // Also, in some cases, drivers may need extra
- // information to process rules (eg. Router IP/MAC); and so, we send
- // port addressing rules to the driver as well, irrespective of whether
- // this instance is the master or not.
- defaultRoutingHandler.populatePortAddressingRules(device.id());
- }
- if (mastershipService.isLocalMaster(device.id())) {
- hostHandler.readInitialHosts(device.id());
- DefaultGroupHandler groupHandler = groupHandlerMap.get(device.id());
- groupHandler.createGroupsFromSubnetConfig();
- routingRulePopulator.populateSubnetBroadcastRule(device.id());
- groupHandler.createGroupsForXConnect(device.id());
- routingRulePopulator.populateXConnectBroadcastRule(device.id());
- }
+ processDeviceAddedInternal(device.id());
}
defaultRoutingHandler.startPopulationProcess();
+ mcastHandler.init();
}
@Override