OSPFv3 Component

Enabling the OSPFv3 Component

The OSPFv3 configuration is held under the top level ospfv3 object in /etc/flockd/flockd.json. If the ospfv3 object exists OSPFv3 will be enabled.

OSPFv3 (RFC 5340) is the IPv6 version of OSPF. The configuration structure mirrors OSPFv2.

With this configuration file:

• OSPFv3 will be enabled

• The OSPFv3 router will advertise a router id of 10.0.100.1

• All interfaces with names starting with r301 will be placed in OSPF area 0.

• The loopback interface will also be placed in area 0.

    "ospfv3": {
        "vrfs": {
            "default": {
                "router_id": "10.0.100.1",
                "areas": {
                    "0.0.0.0": {
                        "intfs": {
                            "lo": {},
                            "^r301": {
                                "hello_interval": 1
                            }
                        }
                    }
                }
            }
        }
    }
  

Redistribution of routes into OSPFv3

You may wish to redistribute routes from the RIB into OSPFv3. Use the redistribute JSON array. The origin field specifies the protocol that programmed the routes into the RIB: static, kernel-connected or kernel-static, as for OSPFv2.

"ospfv3": {
    "vrfs": {
        "default": {
            "router_id": "10.0.100.1",
            "redistribute": [
                {
                    "origin": "static",
                    "metric_type": 2,
                    "metric": 1000
                }
            ],
            "areas": {
                "0.0.0.0": {
                    "intfs": {
                        "lo": {},
                        "^r301": {}
                    }
                }
            }
        }
    }
}

Implicit and Explicit Router Id

As with OSPFv2, the router id can be set explicitly via the router_id field, or left unset to use the highest IPv4 address.

Interface Configuration

OSPFv3 interfaces support the same configuration options as OSPFv2:

"ospfv3": {
    "vrfs": {
        "<vrf-name>": {
            "areas": {
                "<area-id>": {
                    "intfs": {
                        "<interface-name>": {
                            # Interface type (default: broadcast)
                            "ospf_intf_type": ["broadcast" | "point-to-point" | "nbma" | "point-to-multipoint"],
                            # Interface output cost (default: 10)
                            "cost": ( 1..65535 ),
                            # Router Priority (default: 1)
                            "priority": ( 0..255 ),
                            # HelloInterval in seconds (default: 10)
                            "hello_interval": ( 1..65535 ),
                            # RouterDeadInterval in seconds (default: 40)
                            "dead_interval": ( 1..65535 ),
                            # Retransmit interval in milliseconds (default: 5000)
                            "rxmt_interval": ( 1..65535 ),
                            # Ignore MTU mismatch with neighbor
                            "mtu_ignore": [ true | false ],
                            # BFD failure detection
                            "bfd": {
                                "multiplier": 3,
                                "min_tx": 1000
                            }
                        }
                    }
                }
            }
        }
    }
}

Area Configuration

Area summary ranges

Area summary ranges allow aggregation of prefixes at area boundaries. When a summary range is configured, individual prefixes within that range are suppressed and replaced with a single summary prefix.

"ospfv3": {
    "vrfs": {
        "default": {
            "areas": {
                "0.0.0.1": {
                    "summary": {
                        "fc00:1::/32": {
                            "advertise": true
                        }
                    },
                    "intfs": {
                        "^eth": {}
                    }
                }
            }
        }
    }
}

Global Configuration Options

Stub router

Configure the router to advertise maximum metric in its Router LSA, causing other routers to avoid using it as a transit node.

"ospfv3": {
    "vrfs": {
        "default": {
            "stub_router": {
                "mode": "MaxMetric",
                "always": true
            }
        }
    }
}

Default route origination

"ospfv3": {
    "vrfs": {
        "default": {
            "default_originate": {
                "always": true
            }
        }
    }
}

Administrative distance

"ospfv3": {
    "vrfs": {
        "default": {
            "admin_distance": 115
        }
    }
}

LSA arrival timer

The timer_lsa_arrival sets the minimum interval (in milliseconds) between accepting the same LSA from a neighbor. This helps dampen rapid LSA flooding.

"ospfv3": {
    "vrfs": {
        "default": {
            "timer_lsa_arrival": 1000
        }
    }
}

LSA mode

lsa_mode selects which RFC 8362 LSA behaviour OSPFv3 uses (there is no on-the-wire negotiation, so it is operator config):

• legacy (default): classic RFC 5340 LSAs only.

• sparse: compute SPF from the classic LSAs but also originate extended LSAs to carry Prefix SIDs (RFC 8362 section 6.2) — interoperates with a classic OSPFv3 router while still advertising Segment Routing.

• full-extended: originate and compute SPF from the RFC 8362 extended LSA set (section 6.1), for an all-flock area.

  "ospfv3": { "vrfs": { "default": { "lsa_mode": "full-extended" } } }

lsa_mode is set per OSPFv3 instance and applies to all of its areas; there is no per-area override.

The mode an instance is running in is reported as lsa_mode by flockc ospfv3 instances.

lsa_mode may be changed on a running instance. A change between legacy and sparse is applied in place, only adding or removing the extended LSA that carries Prefix SIDs. A change to or from full-extended resets the OSPFv3 instance so it re-originates every LSA in the new mode, which briefly drops the instance's adjacencies; make such a change during a maintenance window. An area running full-extended must run it on every router: a legacy or sparse router ignores the extended LSAs a full-extended router originates, and a full-extended router likewise ignores the classic LSAs the extended ones replace. A router left on the other mode is therefore not routed to at all.

Migrating to full-extended

There is no node-by-node migration into full-extended. A router originates either the classic or the extended form of each topology LSA and never both, so it leaves the topology of every router still in the other mode as soon as it changes, and they leave its own.

There is no per-area migration either. lsa_mode is instance-wide, so setting full-extended converts every area the instance serves at once; an ABR cannot migrate one of its areas and leave the other. The E-AS-External-LSA also floods the whole routing domain rather than a single area, and a legacy or sparse router anywhere in that domain ignores it. Plan the change as one maintenance window covering every OSPFv3 router in the domain.

A router left behind does not degrade, it disappears from the topology. Its adjacencies still reach Full, because a router stores and floods every LSA form whatever mode it is in, so the neighbour table looks healthy while the route table has no route to the router or through it. A Full neighbour that no route uses is the signature of a mode mismatch; check lsa_mode on every router after migrating.

Segment Routing and Flexible Algorithm

OSPFv3 supports Segment Routing with Prefix SIDs and Flexible Algorithms (RFC 9350).

Prefix SIDs are carried in the extended Intra-Area-Prefix LSA, which only sparse and full-extended originate. With lsa_mode left at its legacy default a configured prefix_sid_map is not advertised, so set the mode as well as the SID map.

Prefix SID Map

Prefix SIDs assign MPLS labels to IPv6 prefixes, enabling Segment Routing forwarding.

"ospfv3": {
    "vrfs": {
        "default": {
            "router_id": "10.0.100.53",
            "prefix_sid_map": {
                "fc00::0161/128": {
                    "prefix_sid": 353,
                    "flex_algo_sids": {
                        "201": {
                            "sid": 90569
                        }
                    }
                }
            }
        }
    }
}

Flexible Algorithm

Flexible Algorithms allow defining custom constraint-based paths through the network.

"ospfv3": {
    "vrfs": {
        "default": {
            "flex_algos": {
                "201": {
                    "advertise_definition": true
                },
                "202": {
                    "advertise_definition": true
                }
            }
        }
    }
}

Operational State Overview

Check OSPFv3 is enabled

Check OSPFv3 is listed in the enabled_protocols field.

flock@r301:~$ flockc sys overview
{...,"enabled_protocols":["OSPFv3","Static"],...}

Show OSPFv3 Overview

flock@r301:~$ flockc ospfv3 inst default ospfv3_default show
{"my_router":{"id":"10.0.100.1","class":"ASBR"},"vrf_name":"default", ...,"instance_name":"ospfv3_default", ...,"version":"V3","lsa_mode":"full-extended", ...,"admin_distance":110, ...,"stats":{"ls_db_counters":{...},"route_count":11,"neigh_state_count":{"down":0,"attempt":0,"init":0,"two_way":0,"ex_start":0,"exchange":0,"loading":0,"full":4}}}

An OSPFv2 instance has no RFC 8362 mode and reports "lsa_mode":null.

Show all neighbors on an interface in Area 0

flock@r301:~$ flockc ospfv3 inst default ospfv3_default area 0 intf r301-eth0 neighbors

Show Area Link State Database

flock@r301:~$ flockc ospfv3 inst default ospfv3_default area 0 lsdb

Filter LSDB by LSA type

--lsa-type takes the OSPFv3 LS Type as a decimal number, including the U bit and flooding scope bits it carries in its top three bits. The classic RFC 5340 types and the RFC 8362 extended types that replace them are:

Router Information LSAs (40972, 0xa00c) carry the router's hostname and its Segment Routing and Flexible Algorithm capabilities. They are TLV based and have no classic form, so they are originated in every mode.

A type number outside this set is not rejected. It matches no LSA, so the filter returns an empty list rather than an error.

flock@r301:~$ flockc ospfv3 inst default ospfv3_default area 0 lsdb --lsa-type 40993

Filter LSDB by originating router ID

flock@r301:~$ flockc ospfv3 inst default ospfv3_default area 0 lsdb --router-id <router-id>

Show an interface's link-local LSDB

The link-local-scope LSDB (OSPFv3 Link-LSAs) for one interface:

flock@r301:~$ flockc ospfv3 inst default ospfv3_default area 0 intf r301-eth0 lsdb

The same --lsa-type, --router-id, and --lsa-id filters apply as for the area LSDB.

Show network route table prefixes

flock@r301:~$ flockc ospfv3 inst default ospfv3_default rib lookup <ipv6-network>
flock@r301:~$ flockc ospfv3 inst default ospfv3_default rib walk

Show router route table prefixes

flock@r301:~$ flockc ospfv3 inst default ospfv3_default router-rib lookup <router-id>

Show event buffer

flock@r301:~$ flockc ospfv3 event-log

Show redistributed RIB

flock@r301:~$ flockc ospfv3 inst default ospfv3_default redist-rib walk

Show Flexible Algorithm routes

flock@r301:~$ flockc ospfv3 inst default ospfv3_default fa-rib <algo-id> walk

OSPFv3 Operation Commands Reference

Help

flockc ospfv3 -h

List OSPFv3 instances

flockc ospfv3 instances

Overview of an instance (default VRF, default instance name ospfv3_default)

flockc ospfv3 inst default ospfv3_default show

List areas

flockc ospfv3 inst default ospfv3_default areas

Area overview

flockc ospfv3 inst default ospfv3_default area <area-id> show

All interfaces in an area

flockc ospfv3 inst default ospfv3_default area <area-id> interfaces

All neighbors on an interface

flockc ospfv3 inst default ospfv3_default area <area-id> intf <intf-name> neighbors

Specific neighbor detail

flockc ospfv3 inst default ospfv3_default area <area-id> intf <intf-name> neigh <router-id>

Link State Database

flockc ospfv3 inst default ospfv3_default lsdb

Area Link State Database

flockc ospfv3 inst default ospfv3_default area <area-id> lsdb

LSDB filtered by LSA type

flockc ospfv3 inst default ospfv3_default area <area-id> lsdb --lsa-type <type>

LSDB filtered by originating router ID

flockc ospfv3 inst default ospfv3_default area <area-id> lsdb --router-id <router-id>

Interface link-local LSDB (OSPFv3 Link-LSAs)

flockc ospfv3 inst default ospfv3_default area <area-id> intf <intf-name> lsdb [--lsa-type <type>] [--router-id <router-id>] [--lsa-id <id>]

Network route table prefixes

flockc ospfv3 inst default ospfv3_default rib lookup <ipv6-network>
flockc ospfv3 inst default ospfv3_default rib walk

Router route table prefixes

flockc ospfv3 inst default ospfv3_default router-rib lookup <router-id>
flockc ospfv3 inst default ospfv3_default router-rib walk

Redistributed RIB lookup / walk

flockc ospfv3 inst default ospfv3_default redist-rib lookup <ipv6-network>
flockc ospfv3 inst default ospfv3_default redist-rib walk

Event buffer

flockc ospfv3 event-log

Flexible Algorithm routes

flockc ospfv3 inst default ospfv3_default fa-rib <algo-id> lookup <ipv6-network>
flockc ospfv3 inst default ospfv3_default fa-rib <algo-id> walk