7 :abbr:`BGP` stands for Border Gateway Protocol. The latest BGP version is 4.
8 BGP-4 is one of the Exterior Gateway Protocols and the de facto standard
9 interdomain routing protocol. BGP-4 is described in :rfc:`1771` and updated by
10 :rfc:`4271`. :rfc:`2858` adds multiprotocol support to BGP-4.
17 The default configuration file of *bgpd* is :file:`bgpd.conf`. *bgpd* searches
18 the current directory first, followed by |INSTALL_PREFIX_ETC|/bgpd.conf. All of
19 *bgpd*'s commands must be configured in :file:`bgpd.conf` when the integrated
20 config is not being used.
22 *bgpd* specific invocation options are described below. Common options may also
23 be specified (:ref:`common-invocation-options`).
27 .. option:: -p, --bgp_port <port>
29 Set the bgp protocol's port number. When port number is 0, that means do not
32 .. option:: -l, --listenon
34 Specify specific IP addresses for bgpd to listen on, rather than its default
35 of ``0.0.0.0`` / ``::``. This can be useful to constrain bgpd to an internal
36 address, or to run multiple bgpd processes on one host. Multiple addresses
39 In the following example, bgpd is started listening for connections on the
40 addresses 100.0.1.2 and fd00::2:2. The options -d (runs in daemon mode) and
41 -f (uses specific configuration file) are also used in this example as we
42 are likely to run multiple bgpd instances, each one with different
43 configurations, when using -l option.
45 Note that this option implies the --no_kernel option, and no learned routes will be installed into the linux kernel.
49 # /usr/lib/frr/bgpd -d -f /some-folder/bgpd.conf -l 100.0.1.2 -l fd00::2:2
51 .. option:: -n, --no_kernel
53 Do not install learned routes into the linux kernel. This option is useful
54 for a route-reflector environment or if you are running multiple bgp
55 processes in the same namespace. This option is different than the --no_zebra
56 option in that a ZAPI connection is made.
58 This option can also be toggled during runtime by using the
59 ``[no] bgp no-rib`` commands in VTY shell.
61 Note that this option will persist after saving the configuration during
62 runtime, unless unset by the ``no bgp no-rib`` command in VTY shell prior to
63 a configuration write operation.
65 .. option:: -S, --skip_runas
67 Skip the normal process of checking capabilities and changing user and group
70 .. option:: -e, --ecmp
72 Run BGP with a limited ecmp capability, that is different than what BGP
73 was compiled with. The value specified must be greater than 0 and less
74 than or equal to the MULTIPATH_NUM specified on compilation.
76 .. option:: -Z, --no_zebra
78 Do not communicate with zebra at all. This is different than the --no_kernel
79 option in that we do not even open a ZAPI connection to the zebra process.
81 .. option:: -s, --socket_size
83 When opening tcp connections to our peers, set the socket send buffer
84 size that the kernel will use for the peers socket. This option
85 is only really useful at a very large scale. Experimentation should
86 be done to see if this is helping or not at the scale you are running
92 .. option:: -I, --int_num
94 Set zclient id. This is required when using Zebra label manager in proxy mode.
96 .. _bgp-basic-concepts:
101 .. _bgp-autonomous-systems:
108 An AS is a connected group of one or more IP prefixes run by one or more
109 network operators which has a SINGLE and CLEARLY DEFINED routing policy.
111 Each AS has an identifying number associated with it called an :abbr:`ASN
112 (Autonomous System Number)`. This is a two octet value ranging in value from 1
113 to 65535. The AS numbers 64512 through 65535 are defined as private AS numbers.
114 Private AS numbers must not be advertised on the global Internet.
116 The :abbr:`ASN (Autonomous System Number)` is one of the essential elements of
117 BGP. BGP is a distance vector routing protocol, and the AS-Path framework
118 provides distance vector metric and loop detection to BGP.
120 .. seealso:: :rfc:`1930`
122 .. _bgp-address-families:
127 Multiprotocol extensions enable BGP to carry routing information for multiple
128 network layer protocols. BGP supports an Address Family Identifier (AFI) for
129 IPv4 and IPv6. Support is also provided for multiple sets of per-AFI
130 information via the BGP Subsequent Address Family Identifier (SAFI). FRR
131 supports SAFIs for unicast information, labeled information (:rfc:`3107` and
132 :rfc:`8277`), and Layer 3 VPN information (:rfc:`4364` and :rfc:`4659`).
134 .. _bgp-route-selection:
139 The route selection process used by FRR's BGP implementation uses the following
140 decision criterion, starting at the top of the list and going towards the
141 bottom until one of the factors can be used.
145 Prefer higher local weight routes to lower routes.
147 2. **Local preference check**
149 Prefer higher local preference routes to lower.
151 3. **Local route check**
153 Prefer local routes (statics, aggregates, redistributed) to received routes.
155 4. **AS path length check**
157 Prefer shortest hop-count AS_PATHs.
161 Prefer the lowest origin type route. That is, prefer IGP origin routes to
162 EGP, to Incomplete routes.
166 Where routes with a MED were received from the same AS, prefer the route
167 with the lowest MED. :ref:`bgp-med`.
169 7. **External check**
171 Prefer the route received from an external, eBGP peer over routes received
172 from other types of peers.
174 8. **IGP cost check**
176 Prefer the route with the lower IGP cost.
178 9. **Multi-path check**
180 If multi-pathing is enabled, then check whether the routes not yet
181 distinguished in preference may be considered equal. If
182 :clicmd:`bgp bestpath as-path multipath-relax` is set, all such routes are
183 considered equal, otherwise routes received via iBGP with identical AS_PATHs
184 or routes received from eBGP neighbours in the same AS are considered equal.
186 10. **Already-selected external check**
188 Where both routes were received from eBGP peers, then prefer the route
189 which is already selected. Note that this check is not applied if
190 :clicmd:`bgp bestpath compare-routerid` is configured. This check can
191 prevent some cases of oscillation.
193 11. **Router-ID check**
195 Prefer the route with the lowest `router-ID`. If the route has an
196 `ORIGINATOR_ID` attribute, through iBGP reflection, then that router ID is
197 used, otherwise the `router-ID` of the peer the route was received from is
200 12. **Cluster-List length check**
202 The route with the shortest cluster-list length is used. The cluster-list
203 reflects the iBGP reflection path the route has taken.
207 Prefer the route received from the peer with the higher transport layer
208 address, as a last-resort tie-breaker.
210 .. _bgp-capability-negotiation:
212 Capability Negotiation
213 ----------------------
215 When adding IPv6 routing information exchange feature to BGP. There were some
216 proposals. :abbr:`IETF (Internet Engineering Task Force)`
217 :abbr:`IDR (Inter Domain Routing)` adopted a proposal called Multiprotocol
218 Extension for BGP. The specification is described in :rfc:`2283`. The protocol
219 does not define new protocols. It defines new attributes to existing BGP. When
220 it is used exchanging IPv6 routing information it is called BGP-4+. When it is
221 used for exchanging multicast routing information it is called MBGP.
223 *bgpd* supports Multiprotocol Extension for BGP. So if a remote peer supports
224 the protocol, *bgpd* can exchange IPv6 and/or multicast routing information.
226 Traditional BGP did not have the feature to detect a remote peer's
227 capabilities, e.g. whether it can handle prefix types other than IPv4 unicast
228 routes. This was a big problem using Multiprotocol Extension for BGP in an
229 operational network. :rfc:`2842` adopted a feature called Capability
230 Negotiation. *bgpd* use this Capability Negotiation to detect the remote peer's
231 capabilities. If a peer is only configured as an IPv4 unicast neighbor, *bgpd*
232 does not send these Capability Negotiation packets (at least not unless other
233 optional BGP features require capability negotiation).
235 By default, FRR will bring up peering with minimal common capability for the
236 both sides. For example, if the local router has unicast and multicast
237 capabilities and the remote router only has unicast capability the local router
238 will establish the connection with unicast only capability. When there are no
239 common capabilities, FRR sends Unsupported Capability error and then resets the
242 .. _bgp-router-configuration:
244 BGP Router Configuration
245 ========================
250 First of all you must configure BGP router with the :clicmd:`router bgp ASN`
251 command. The AS number is an identifier for the autonomous system. The BGP
252 protocol uses the AS number for detecting whether the BGP connection is
253 internal or external.
255 .. clicmd:: router bgp ASN
257 Enable a BGP protocol process with the specified ASN. After
258 this statement you can input any `BGP Commands`.
260 .. clicmd:: bgp router-id A.B.C.D
262 This command specifies the router-ID. If *bgpd* connects to *zebra* it gets
263 interface and address information. In that case default router ID value is
264 selected as the largest IP Address of the interfaces. When `router zebra` is
265 not enabled *bgpd* can't get interface information so `router-id` is set to
266 0.0.0.0. So please set router-id by hand.
269 .. _bgp-multiple-autonomous-systems:
271 Multiple Autonomous Systems
272 ---------------------------
274 FRR's BGP implementation is capable of running multiple autonomous systems at
275 once. Each configured AS corresponds to a :ref:`zebra-vrf`. In the past, to get
276 the same functionality the network administrator had to run a new *bgpd*
277 process; using VRFs allows multiple autonomous systems to be handled in a
280 When using multiple autonomous systems, all router config blocks after the
281 first one must specify a VRF to be the target of BGP's route selection. This
282 VRF must be unique within respect to all other VRFs being used for the same
283 purpose, i.e. two different autonomous systems cannot use the same VRF.
284 However, the same AS can be used with different VRFs.
288 The separated nature of VRFs makes it possible to peer a single *bgpd*
289 process to itself, on one machine. Note that this can be done fully within
290 BGP without a corresponding VRF in the kernel or Zebra, which enables some
291 practical use cases such as :ref:`route reflectors <bgp-route-reflector>`
294 Configuration of additional autonomous systems, or of a router that targets a
295 specific VRF, is accomplished with the following command:
297 .. clicmd:: router bgp ASN vrf VRFNAME
299 ``VRFNAME`` is matched against VRFs configured in the kernel. When ``vrf
300 VRFNAME`` is not specified, the BGP protocol process belongs to the default
303 An example configuration with multiple autonomous systems might look like this:
308 neighbor 10.0.0.1 remote-as 20
309 neighbor 10.0.0.2 remote-as 30
311 router bgp 2 vrf blue
312 neighbor 10.0.0.3 remote-as 40
313 neighbor 10.0.0.4 remote-as 50
316 neighbor 10.0.0.5 remote-as 60
317 neighbor 10.0.0.6 remote-as 70
320 .. seealso:: :ref:`bgp-vrf-route-leaking`
321 .. seealso:: :ref:`zebra-vrf`
329 In addition to supporting multiple autonomous systems, FRR's BGP implementation
330 also supports *views*.
332 BGP views are almost the same as normal BGP processes, except that routes
333 selected by BGP are not installed into the kernel routing table. Each BGP view
334 provides an independent set of routing information which is only distributed
335 via BGP. Multiple views can be supported, and BGP view information is always
336 independent from other routing protocols and Zebra/kernel routes. BGP views use
337 the core instance (i.e., default VRF) for communication with peers.
339 .. clicmd:: router bgp AS-NUMBER view NAME
341 Make a new BGP view. You can use an arbitrary word for the ``NAME``. Routes
342 selected by the view are not installed into the kernel routing table.
344 With this command, you can setup Route Server like below.
350 neighbor 10.0.0.1 remote-as 2
351 neighbor 10.0.0.2 remote-as 3
354 neighbor 10.0.0.3 remote-as 4
355 neighbor 10.0.0.4 remote-as 5
357 .. clicmd:: show [ip] bgp view NAME
359 Display the routing table of BGP view ``NAME``.
365 .. clicmd:: bgp bestpath as-path confed
367 This command specifies that the length of confederation path sets and
368 sequences should should be taken into account during the BGP best path
371 .. clicmd:: bgp bestpath as-path multipath-relax
373 This command specifies that BGP decision process should consider paths
374 of equal AS_PATH length candidates for multipath computation. Without
375 the knob, the entire AS_PATH must match for multipath computation.
377 .. clicmd:: bgp bestpath compare-routerid
379 Ensure that when comparing routes where both are equal on most metrics,
380 including local-pref, AS_PATH length, IGP cost, MED, that the tie is broken
383 If this option is enabled, then the already-selected check, where
384 already selected eBGP routes are preferred, is skipped.
386 If a route has an `ORIGINATOR_ID` attribute because it has been reflected,
387 that `ORIGINATOR_ID` will be used. Otherwise, the router-ID of the peer the
388 route was received from will be used.
390 The advantage of this is that the route-selection (at this point) will be
391 more deterministic. The disadvantage is that a few or even one lowest-ID
392 router may attract all traffic to otherwise-equal paths because of this
393 check. It may increase the possibility of MED or IGP oscillation, unless
394 other measures were taken to avoid these. The exact behaviour will be
395 sensitive to the iBGP and reflection topology.
399 Administrative Distance Metrics
400 -------------------------------
402 .. clicmd:: distance bgp (1-255) (1-255) (1-255)
404 This command change distance value of BGP. The arguments are the distance
405 values for for external routes, internal routes and local routes
408 .. clicmd:: distance (1-255) A.B.C.D/M
410 .. clicmd:: distance (1-255) A.B.C.D/M WORD
412 Sets the administrative distance for a particular route.
414 .. _bgp-requires-policy:
416 Require policy on EBGP
417 -------------------------------
419 .. clicmd:: bgp ebgp-requires-policy
421 This command requires incoming and outgoing filters to be applied
422 for eBGP sessions as part of RFC-8212 compliance. Without the incoming
423 filter, no routes will be accepted. Without the outgoing filter, no
424 routes will be announced.
426 This is enabled by default for the traditional configuration and
427 turned off by default for datacenter configuration.
429 When you enable/disable this option you MUST clear the session.
431 When the incoming or outgoing filter is missing you will see
432 "(Policy)" sign under ``show bgp summary``:
436 exit1# show bgp summary
438 IPv4 Unicast Summary:
439 BGP router identifier 10.10.10.1, local AS number 65001 vrf-id 0
441 RIB entries 7, using 1344 bytes of memory
442 Peers 2, using 43 KiB of memory
444 Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd PfxSnt
445 192.168.0.2 4 65002 8 10 0 0 0 00:03:09 5 (Policy)
446 fe80:1::2222 4 65002 9 11 0 0 0 00:03:09 (Policy) (Policy)
448 Additionally a `show bgp neighbor` command would indicate in the `For address family:`
453 exit1# show bgp neighbor
455 For address family: IPv4 Unicast
456 Update group 1, subgroup 1
457 Packet Queue length 0
458 Inbound soft reconfiguration allowed
459 Community attribute sent to this neighbor(all)
460 Inbound updates discarded due to missing policy
461 Outbound updates discarded due to missing policy
464 Reject routes with AS_SET or AS_CONFED_SET types
465 ------------------------------------------------
467 .. clicmd:: bgp reject-as-sets
469 This command enables rejection of incoming and outgoing routes having AS_SET or AS_CONFED_SET type.
471 Suppress duplicate updates
472 --------------------------
474 .. clicmd:: bgp suppress-duplicates
476 For example, BGP routers can generate multiple identical announcements with
477 empty community attributes if stripped at egress. This is an undesired behavior.
478 Suppress duplicate updates if the route actually not changed.
481 Disable checking if nexthop is connected on EBGP sessions
482 ---------------------------------------------------------
484 .. clicmd:: bgp disable-ebgp-connected-route-check
486 This command is used to disable the connection verification process for EBGP peering sessions
487 that are reachable by a single hop but are configured on a loopback interface or otherwise
488 configured with a non-directly connected IP address.
490 .. _bgp-route-flap-dampening:
495 .. clicmd:: bgp dampening [(1-45) [(1-20000) (1-20000) (1-255)]]
497 This command enables (with optionally specified dampening parameters) or
498 disables route-flap dampening for all routes of a BGP instance.
500 .. clicmd:: neighbor PEER dampening [(1-45) [(1-20000) (1-20000) (1-255)]]
502 This command enables (with optionally specified dampening parameters) or
503 disables route-flap dampening for all routes learned from a BGP peer.
505 .. clicmd:: neighbor GROUP dampening [(1-45) [(1-20000) (1-20000) (1-255)]]
507 This command enables (with optionally specified dampening parameters) or
508 disables route-flap dampening for all routes learned from peers of a peer
512 Half-life time for the penalty in minutes (default value: 15).
515 Value to start reusing a route (default value: 750).
518 Value to start suppressing a route (default value: 2000).
521 Maximum duration to suppress a stable route in minutes (default value:
524 The route-flap damping algorithm is compatible with :rfc:`2439`. The use of
525 these commands is not recommended nowadays.
527 At the moment, route-flap dampening is not working per VRF and is working only
528 for IPv4 unicast and multicast.
530 With different parameter sets configurable for BGP instances, peer groups and
531 peers, the active dampening profile for a route is chosen on the fly,
532 allowing for various changes in configuration (i.e. peer group memberships)
533 during runtime. The parameter sets are taking precedence in the following
540 The negating commands do not allow to exclude a peer/peer group from a peer
541 group/BGP instances configuration.
544 https://www.ripe.net/publications/docs/ripe-378
548 Multi-Exit Discriminator
549 ------------------------
551 The BGP :abbr:`MED (Multi-Exit Discriminator)` attribute has properties which
552 can cause subtle convergence problems in BGP. These properties and problems
553 have proven to be hard to understand, at least historically, and may still not
554 be widely understood. The following attempts to collect together and present
555 what is known about MED, to help operators and FRR users in designing and
556 configuring their networks.
558 The BGP :abbr:`MED` attribute is intended to allow one AS to indicate its
559 preferences for its ingress points to another AS. The MED attribute will not be
560 propagated on to another AS by the receiving AS - it is 'non-transitive' in the
563 E.g., if AS X and AS Y have 2 different BGP peering points, then AS X might set
564 a MED of 100 on routes advertised at one and a MED of 200 at the other. When AS
565 Y selects between otherwise equal routes to or via AS X, AS Y should prefer to
566 take the path via the lower MED peering of 100 with AS X. Setting the MED
567 allows an AS to influence the routing taken to it within another, neighbouring
570 In this use of MED it is not really meaningful to compare the MED value on
571 routes where the next AS on the paths differs. E.g., if AS Y also had a route
572 for some destination via AS Z in addition to the routes from AS X, and AS Z had
573 also set a MED, it wouldn't make sense for AS Y to compare AS Z's MED values to
574 those of AS X. The MED values have been set by different administrators, with
575 different frames of reference.
577 The default behaviour of BGP therefore is to not compare MED values across
578 routes received from different neighbouring ASes. In FRR this is done by
579 comparing the neighbouring, left-most AS in the received AS_PATHs of the routes
580 and only comparing MED if those are the same.
582 Unfortunately, this behaviour of MED, of sometimes being compared across routes
583 and sometimes not, depending on the properties of those other routes, means MED
584 can cause the order of preference over all the routes to be undefined. That is,
585 given routes A, B, and C, if A is preferred to B, and B is preferred to C, then
586 a well-defined order should mean the preference is transitive (in the sense of
587 orders [#med-transitivity-rant]_) and that A would be preferred to C.
589 However, when MED is involved this need not be the case. With MED it is
590 possible that C is actually preferred over A. So A is preferred to B, B is
591 preferred to C, but C is preferred to A. This can be true even where BGP
592 defines a deterministic 'most preferred' route out of the full set of A,B,C.
593 With MED, for any given set of routes there may be a deterministically
594 preferred route, but there need not be any way to arrange them into any order
595 of preference. With unmodified MED, the order of preference of routes literally
598 That MED can induce non-transitive preferences over routes can cause issues.
599 Firstly, it may be perceived to cause routing table churn locally at speakers;
600 secondly, and more seriously, it may cause routing instability in iBGP
601 topologies, where sets of speakers continually oscillate between different
604 The first issue arises from how speakers often implement routing decisions.
605 Though BGP defines a selection process that will deterministically select the
606 same route as best at any given speaker, even with MED, that process requires
607 evaluating all routes together. For performance and ease of implementation
608 reasons, many implementations evaluate route preferences in a pair-wise fashion
609 instead. Given there is no well-defined order when MED is involved, the best
610 route that will be chosen becomes subject to implementation details, such as
611 the order the routes are stored in. That may be (locally) non-deterministic,
612 e.g.: it may be the order the routes were received in.
614 This indeterminism may be considered undesirable, though it need not cause
615 problems. It may mean additional routing churn is perceived, as sometimes more
616 updates may be produced than at other times in reaction to some event .
618 This first issue can be fixed with a more deterministic route selection that
619 ensures routes are ordered by the neighbouring AS during selection.
620 :clicmd:`bgp deterministic-med`. This may reduce the number of updates as routes
621 are received, and may in some cases reduce routing churn. Though, it could
622 equally deterministically produce the largest possible set of updates in
623 response to the most common sequence of received updates.
625 A deterministic order of evaluation tends to imply an additional overhead of
626 sorting over any set of n routes to a destination. The implementation of
627 deterministic MED in FRR scales significantly worse than most sorting
628 algorithms at present, with the number of paths to a given destination. That
629 number is often low enough to not cause any issues, but where there are many
630 paths, the deterministic comparison may quickly become increasingly expensive
633 Deterministic local evaluation can *not* fix the second, more major, issue of
634 MED however. Which is that the non-transitive preference of routes MED can
635 cause may lead to routing instability or oscillation across multiple speakers
636 in iBGP topologies. This can occur with full-mesh iBGP, but is particularly
637 problematic in non-full-mesh iBGP topologies that further reduce the routing
638 information known to each speaker. This has primarily been documented with iBGP
639 :ref:`route-reflection <bgp-route-reflector>` topologies. However, any
640 route-hiding technologies potentially could also exacerbate oscillation with MED.
642 This second issue occurs where speakers each have only a subset of routes, and
643 there are cycles in the preferences between different combinations of routes -
644 as the undefined order of preference of MED allows - and the routes are
645 distributed in a way that causes the BGP speakers to 'chase' those cycles. This
646 can occur even if all speakers use a deterministic order of evaluation in route
649 E.g., speaker 4 in AS A might receive a route from speaker 2 in AS X, and from
650 speaker 3 in AS Y; while speaker 5 in AS A might receive that route from
651 speaker 1 in AS Y. AS Y might set a MED of 200 at speaker 1, and 100 at speaker
652 3. I.e, using ASN:ID:MED to label the speakers:
658 X:2------|--A:4-------A:5--|-Y:1:200
664 Assuming all other metrics are equal (AS_PATH, ORIGIN, 0 IGP costs), then based
665 on the RFC4271 decision process speaker 4 will choose X:2 over Y:3:100, based
666 on the lower ID of 2. Speaker 4 advertises X:2 to speaker 5. Speaker 5 will
667 continue to prefer Y:1:200 based on the ID, and advertise this to speaker 4.
668 Speaker 4 will now have the full set of routes, and the Y:1:200 it receives
669 from 5 will beat X:2, but when speaker 4 compares Y:1:200 to Y:3:100 the MED
670 check now becomes active as the ASes match, and now Y:3:100 is preferred.
671 Speaker 4 therefore now advertises Y:3:100 to 5, which will also agrees that
672 Y:3:100 is preferred to Y:1:200, and so withdraws the latter route from 4.
673 Speaker 4 now has only X:2 and Y:3:100, and X:2 beats Y:3:100, and so speaker 4
674 implicitly updates its route to speaker 5 to X:2. Speaker 5 sees that Y:1:200
675 beats X:2 based on the ID, and advertises Y:1:200 to speaker 4, and the cycle
678 The root cause is the lack of a clear order of preference caused by how MED
679 sometimes is and sometimes is not compared, leading to this cycle in the
680 preferences between the routes:
685 /---> X:2 ---beats---> Y:3:100 --\\
688 \\---beats--- Y:1:200 <---beats---/
692 This particular type of oscillation in full-mesh iBGP topologies can be
693 avoided by speakers preferring already selected, external routes rather than
694 choosing to update to new a route based on a post-MED metric (e.g. router-ID),
695 at the cost of a non-deterministic selection process. FRR implements this, as
696 do many other implementations, so long as it is not overridden by setting
697 :clicmd:`bgp bestpath compare-routerid`, and see also
698 :ref:`bgp-route-selection`.
700 However, more complex and insidious cycles of oscillation are possible with
701 iBGP route-reflection, which are not so easily avoided. These have been
702 documented in various places. See, e.g.:
704 - [bgp-route-osci-cond]_
705 - [stable-flexible-ibgp]_
706 - [ibgp-correctness]_
708 for concrete examples and further references.
710 There is as of this writing *no* known way to use MED for its original purpose;
711 *and* reduce routing information in iBGP topologies; *and* be sure to avoid the
712 instability problems of MED due the non-transitive routing preferences it can
713 induce; in general on arbitrary networks.
715 There may be iBGP topology specific ways to reduce the instability risks, even
716 while using MED, e.g.: by constraining the reflection topology and by tuning
717 IGP costs between route-reflector clusters, see :rfc:`3345` for details. In the
718 near future, the Add-Path extension to BGP may also solve MED oscillation while
719 still allowing MED to be used as intended, by distributing "best-paths per
720 neighbour AS". This would be at the cost of distributing at least as many
721 routes to all speakers as a full-mesh iBGP would, if not more, while also
722 imposing similar CPU overheads as the "Deterministic MED" feature at each
725 More generally, the instability problems that MED can introduce on more
726 complex, non-full-mesh, iBGP topologies may be avoided either by:
728 - Setting :clicmd:`bgp always-compare-med`, however this allows MED to be compared
729 across values set by different neighbour ASes, which may not produce
730 coherent desirable results, of itself.
731 - Effectively ignoring MED by setting MED to the same value (e.g.: 0) using
732 :clicmd:`set metric METRIC` on all received routes, in combination with
733 setting :clicmd:`bgp always-compare-med` on all speakers. This is the simplest
734 and most performant way to avoid MED oscillation issues, where an AS is happy
735 not to allow neighbours to inject this problematic metric.
737 As MED is evaluated after the AS_PATH length check, another possible use for
738 MED is for intra-AS steering of routes with equal AS_PATH length, as an
739 extension of the last case above. As MED is evaluated before IGP metric, this
740 can allow cold-potato routing to be implemented to send traffic to preferred
741 hand-offs with neighbours, rather than the closest hand-off according to the
744 Note that even if action is taken to address the MED non-transitivity issues,
745 other oscillations may still be possible. E.g., on IGP cost if iBGP and IGP
746 topologies are at cross-purposes with each other - see the Flavel and Roughan
747 paper above for an example. Hence the guideline that the iBGP topology should
748 follow the IGP topology.
750 .. clicmd:: bgp deterministic-med
752 Carry out route-selection in way that produces deterministic answers
753 locally, even in the face of MED and the lack of a well-defined order of
754 preference it can induce on routes. Without this option the preferred route
755 with MED may be determined largely by the order that routes were received
758 Setting this option will have a performance cost that may be noticeable when
759 there are many routes for each destination. Currently in FRR it is
760 implemented in a way that scales poorly as the number of routes per
761 destination increases.
763 The default is that this option is not set.
765 Note that there are other sources of indeterminism in the route selection
766 process, specifically, the preference for older and already selected routes
767 from eBGP peers, :ref:`bgp-route-selection`.
769 .. clicmd:: bgp always-compare-med
771 Always compare the MED on routes, even when they were received from
772 different neighbouring ASes. Setting this option makes the order of
773 preference of routes more defined, and should eliminate MED induced
776 If using this option, it may also be desirable to use
777 :clicmd:`set metric METRIC` to set MED to 0 on routes received from external
780 This option can be used, together with :clicmd:`set metric METRIC` to use
781 MED as an intra-AS metric to steer equal-length AS_PATH routes to, e.g.,
785 .. _bgp-graceful-restart:
790 BGP graceful restart functionality as defined in
791 `RFC-4724 <https://tools.ietf.org/html/rfc4724/>`_ defines the mechanisms that
792 allows BGP speaker to continue to forward data packets along known routes
793 while the routing protocol information is being restored.
796 Usually, when BGP on a router restarts, all the BGP peers detect that the
797 session went down and then came up. This "down/up" transition results in a
798 "routing flap" and causes BGP route re-computation, generation of BGP routing
799 updates, and unnecessary churn to the forwarding tables.
801 The following functionality is provided by graceful restart:
803 1. The feature allows the restarting router to indicate to the helping peer the
804 routes it can preserve in its forwarding plane during control plane restart
805 by sending graceful restart capability in the OPEN message sent during
806 session establishment.
807 2. The feature allows helping router to advertise to all other peers the routes
808 received from the restarting router which are preserved in the forwarding
809 plane of the restarting router during control plane restart.
816 (R1)-----------------------------------------------------------------(R2)
818 1. BGP Graceful Restart Capability exchanged between R1 & R2.
820 <--------------------------------------------------------------------->
822 2. Kill BGP Process at R1.
824 ---------------------------------------------------------------------->
826 3. R2 Detects the above BGP Restart & verifies BGP Restarting
829 4. Start BGP Process at R1.
831 5. Re-establish the BGP session between R1 & R2.
833 <--------------------------------------------------------------------->
835 6. R2 Send initial route updates, followed by End-Of-Rib.
837 <----------------------------------------------------------------------
839 7. R1 was waiting for End-Of-Rib from R2 & which has been received
842 8. R1 now runs BGP Best-Path algorithm. Send Initial BGP Update,
843 followed by End-Of Rib
845 <--------------------------------------------------------------------->
848 .. _bgp-GR-preserve-forwarding-state:
850 BGP-GR Preserve-Forwarding State
851 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
853 BGP OPEN message carrying optional capabilities for Graceful Restart has
854 8 bit “Flags for Address Family” for given AFI and SAFI. This field contains
855 bit flags relating to routes that were advertised with the given AFI and SAFI.
864 The most significant bit is defined as the Forwarding State (F) bit, which
865 can be used to indicate whether the forwarding state for routes that were
866 advertised with the given AFI and SAFI has indeed been preserved during the
867 previous BGP restart. When set (value 1), the bit indicates that the
868 forwarding state has been preserved.
869 The remaining bits are reserved and MUST be set to zero by the sender and
870 ignored by the receiver.
872 .. clicmd:: bgp graceful-restart preserve-fw-state
874 FRR gives us the option to enable/disable the "F" flag using this specific
875 vty command. However, it doesn't have the option to enable/disable
876 this flag only for specific AFI/SAFI i.e. when this command is used, it
877 applied to all the supported AFI/SAFI combinations for this peer.
879 .. _bgp-end-of-rib-message:
881 End-of-RIB (EOR) message
882 ^^^^^^^^^^^^^^^^^^^^^^^^
884 An UPDATE message with no reachable Network Layer Reachability Information
885 (NLRI) and empty withdrawn NLRI is specified as the End-of-RIB marker that can
886 be used by a BGP speaker to indicate to its peer the completion of the initial
887 routing update after the session is established.
889 For the IPv4 unicast address family, the End-of-RIB marker is an UPDATE message
890 with the minimum length. For any other address family, it is an UPDATE message
891 that contains only the MP_UNREACH_NLRI attribute with no withdrawn routes for
894 Although the End-of-RIB marker is specified for the purpose of BGP graceful
895 restart, it is noted that the generation of such a marker upon completion of
896 the initial update would be useful for routing convergence in general, and thus
897 the practice is recommended.
899 .. _bgp-route-selection-deferral-timer:
901 Route Selection Deferral Timer
902 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
904 Specifies the time the restarting router defers the route selection process
907 Restarting Router : The usage of route election deferral timer is specified
908 in https://tools.ietf.org/html/rfc4724#section-4.1
910 Once the session between the Restarting Speaker and the Receiving Speaker is
911 re-established, the Restarting Speaker will receive and process BGP messages
914 However, it MUST defer route selection for an address family until it either.
916 1. Receives the End-of-RIB marker from all its peers (excluding the ones with
917 the "Restart State" bit set in the received capability and excluding the ones
918 that do not advertise the graceful restart capability).
919 2. The Selection_Deferral_Timer timeout.
921 .. clicmd:: bgp graceful-restart select-defer-time (0-3600)
923 This is command, will set deferral time to value specified.
926 .. clicmd:: bgp graceful-restart rib-stale-time (1-3600)
928 This is command, will set the time for which stale routes are kept in RIB.
930 .. clicmd:: bgp graceful-restart stalepath-time (1-4095)
932 This is command, will set the max time (in seconds) to hold onto
933 restarting peer's stale paths.
935 It also controls Enhanced Route-Refresh timer.
937 If this command is configured and the router does not receive a Route-Refresh EoRR
938 message, the router removes the stale routes from the BGP table after the timer
939 expires. The stale path timer is started when the router receives a Route-Refresh
942 .. _bgp-per-peer-graceful-restart:
944 BGP Per Peer Graceful Restart
945 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
947 Ability to enable and disable graceful restart, helper and no GR at all mode
948 functionality at peer level.
950 So bgp graceful restart can be enabled at modes global BGP level or at per
951 peer level. There are two FSM, one for BGP GR global mode and other for peer
954 Default global mode is helper and default peer per mode is inherit from global.
955 If per peer mode is configured, the GR mode of this particular peer will
956 override the global mode.
958 .. _bgp-GR-global-mode-cmd:
960 BGP GR Global Mode Commands
961 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
963 .. clicmd:: bgp graceful-restart
965 This command will enable BGP graceful restart ifunctionality at the global
968 .. clicmd:: bgp graceful-restart disable
970 This command will disable both the functionality graceful restart and helper
974 .. _bgp-GR-peer-mode-cmd:
976 BGP GR Peer Mode Commands
977 ^^^^^^^^^^^^^^^^^^^^^^^^^
979 .. clicmd:: neighbor A.B.C.D graceful-restart
981 This command will enable BGP graceful restart ifunctionality at the peer
984 .. clicmd:: neighbor A.B.C.D graceful-restart-helper
986 This command will enable BGP graceful restart helper only functionality
989 .. clicmd:: neighbor A.B.C.D graceful-restart-disable
991 This command will disable the entire BGP graceful restart functionality
997 Administrative Shutdown
998 -----------------------
1000 .. clicmd:: bgp shutdown [message MSG...]
1002 Administrative shutdown of all peers of a bgp instance. Drop all BGP peers,
1003 but preserve their configurations. The peers are notified in accordance with
1004 `RFC 8203 <https://tools.ietf.org/html/rfc8203/>`_ by sending a
1005 ``NOTIFICATION`` message with error code ``Cease`` and subcode
1006 ``Administrative Shutdown`` prior to terminating connections. This global
1007 shutdown is independent of the neighbor shutdown, meaning that individually
1008 shut down peers will not be affected by lifting it.
1010 An optional shutdown message `MSG` can be specified.
1018 .. clicmd:: network A.B.C.D/M
1020 This command adds the announcement network.
1025 address-family ipv4 unicast
1029 This configuration example says that network 10.0.0.0/8 will be
1030 announced to all neighbors. Some vendors' routers don't advertise
1031 routes if they aren't present in their IGP routing tables; `bgpd`
1032 doesn't care about IGP routes when announcing its routes.
1035 .. clicmd:: bgp network import-check
1037 This configuration modifies the behavior of the network statement.
1038 If you have this configured the underlying network must exist in
1039 the rib. If you have the [no] form configured then BGP will not
1040 check for the networks existence in the rib. For versions 7.3 and
1041 before frr defaults for datacenter were the network must exist,
1042 traditional did not check for existence. For versions 7.4 and beyond
1043 both traditional and datacenter the network must exist.
1045 .. _bgp-ipv6-support:
1050 .. clicmd:: neighbor A.B.C.D activate
1052 This configuration modifies whether to enable an address family for a
1053 specific neighbor. By default only the IPv4 unicast address family is
1059 address-family ipv6 unicast
1060 neighbor 2001:0DB8::1 activate
1061 network 2001:0DB8:5009::/64
1064 This configuration example says that network 2001:0DB8:5009::/64 will be
1065 announced and enables the neighbor 2001:0DB8::1 to receive this announcement.
1067 By default, only the IPv4 unicast address family is announced to all
1068 neighbors. Using the 'no bgp default ipv4-unicast' configuration overrides
1069 this default so that all address families need to be enabled explicitly.
1074 no bgp default ipv4-unicast
1075 neighbor 10.10.10.1 remote-as 2
1076 neighbor 2001:0DB8::1 remote-as 3
1077 address-family ipv4 unicast
1078 neighbor 10.10.10.1 activate
1079 network 192.168.1.0/24
1081 address-family ipv6 unicast
1082 neighbor 2001:0DB8::1 activate
1083 network 2001:0DB8:5009::/64
1086 This configuration demonstrates how the 'no bgp default ipv4-unicast' might
1087 be used in a setup with two upstreams where each of the upstreams should only
1088 receive either IPv4 or IPv6 annocuments.
1091 .. _bgp-route-aggregation:
1096 .. _bgp-route-aggregation-ipv4:
1098 Route Aggregation-IPv4 Address Family
1099 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1101 .. clicmd:: aggregate-address A.B.C.D/M
1103 This command specifies an aggregate address.
1105 .. clicmd:: aggregate-address A.B.C.D/M route-map NAME
1107 Apply a route-map for an aggregated prefix.
1109 .. clicmd:: aggregate-address A.B.C.D/M origin <egp|igp|incomplete>
1111 Override ORIGIN for an aggregated prefix.
1113 .. clicmd:: aggregate-address A.B.C.D/M as-set
1115 This command specifies an aggregate address. Resulting routes include
1118 .. clicmd:: aggregate-address A.B.C.D/M summary-only
1120 This command specifies an aggregate address. Aggregated routes will
1123 .. clicmd:: aggregate-address A.B.C.D/M matching-MED-only
1125 Configure the aggregated address to only be created when the routes MED
1126 match, otherwise no aggregated route will be created.
1128 .. clicmd:: aggregate-address A.B.C.D/M suppress-map NAME
1130 Similar to `summary-only`, but will only suppress more specific routes that
1131 are matched by the selected route-map.
1134 This configuration example sets up an ``aggregate-address`` under the ipv4
1140 address-family ipv4 unicast
1141 aggregate-address 10.0.0.0/8
1142 aggregate-address 20.0.0.0/8 as-set
1143 aggregate-address 40.0.0.0/8 summary-only
1144 aggregate-address 50.0.0.0/8 route-map aggr-rmap
1148 .. _bgp-route-aggregation-ipv6:
1150 Route Aggregation-IPv6 Address Family
1151 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1153 .. clicmd:: aggregate-address X:X::X:X/M
1155 This command specifies an aggregate address.
1157 .. clicmd:: aggregate-address X:X::X:X/M route-map NAME
1159 Apply a route-map for an aggregated prefix.
1161 .. clicmd:: aggregate-address X:X::X:X/M origin <egp|igp|incomplete>
1163 Override ORIGIN for an aggregated prefix.
1165 .. clicmd:: aggregate-address X:X::X:X/M as-set
1167 This command specifies an aggregate address. Resulting routes include
1170 .. clicmd:: aggregate-address X:X::X:X/M summary-only
1172 This command specifies an aggregate address. Aggregated routes will
1175 .. clicmd:: aggregate-address X:X::X:X/M matching-MED-only
1177 Configure the aggregated address to only be created when the routes MED
1178 match, otherwise no aggregated route will be created.
1180 .. clicmd:: aggregate-address X:X::X:X/M suppress-map NAME
1182 Similar to `summary-only`, but will only suppress more specific routes that
1183 are matched by the selected route-map.
1186 This configuration example sets up an ``aggregate-address`` under the ipv6
1192 address-family ipv6 unicast
1193 aggregate-address 10::0/64
1194 aggregate-address 20::0/64 as-set
1195 aggregate-address 40::0/64 summary-only
1196 aggregate-address 50::0/64 route-map aggr-rmap
1200 .. _bgp-redistribute-to-bgp:
1205 Redistribution configuration should be placed under the ``address-family``
1206 section for the specific AF to redistribute into. Protocol availability for
1207 redistribution is determined by BGP AF; for example, you cannot redistribute
1208 OSPFv3 into ``address-family ipv4 unicast`` as OSPFv3 supports IPv6.
1210 .. clicmd:: redistribute <babel|connected|eigrp|isis|kernel|openfabric|ospf|ospf6|rip|ripng|sharp|static|table> [metric (0-4294967295)] [route-map WORD]
1212 Redistribute routes from other protocols into BGP.
1214 .. clicmd:: redistribute vnc-direct
1216 Redistribute VNC direct (not via zebra) routes to BGP process.
1218 .. clicmd:: bgp update-delay MAX-DELAY
1220 .. clicmd:: bgp update-delay MAX-DELAY ESTABLISH-WAIT
1222 This feature is used to enable read-only mode on BGP process restart or when
1223 a BGP process is cleared using 'clear ip bgp \*'. Note that this command is
1224 configured at the global level and applies to all bgp instances/vrfs. It
1225 cannot be used at the same time as the "update-delay" command described below,
1226 which is entered in each bgp instance/vrf desired to delay update installation
1227 and advertisements. The global and per-vrf approaches to defining update-delay
1228 are mutually exclusive.
1230 When applicable, read-only mode would begin as soon as the first peer reaches
1231 Established status and a timer for max-delay seconds is started. During this
1232 mode BGP doesn't run any best-path or generate any updates to its peers. This
1233 mode continues until:
1235 1. All the configured peers, except the shutdown peers, have sent explicit EOR
1236 (End-Of-RIB) or an implicit-EOR. The first keep-alive after BGP has reached
1237 Established is considered an implicit-EOR.
1238 If the establish-wait optional value is given, then BGP will wait for
1239 peers to reach established from the beginning of the update-delay till the
1240 establish-wait period is over, i.e. the minimum set of established peers for
1241 which EOR is expected would be peers established during the establish-wait
1242 window, not necessarily all the configured neighbors.
1243 2. max-delay period is over.
1245 On hitting any of the above two conditions, BGP resumes the decision process
1246 and generates updates to its peers.
1248 Default max-delay is 0, i.e. the feature is off by default.
1251 .. clicmd:: update-delay MAX-DELAY
1253 .. clicmd:: update-delay MAX-DELAY ESTABLISH-WAIT
1255 This feature is used to enable read-only mode on BGP process restart or when
1256 a BGP process is cleared using 'clear ip bgp \*'. Note that this command is
1257 configured under the specific bgp instance/vrf that the feaure is enabled for.
1258 It cannot be used at the same time as the global "bgp update-delay" described
1259 above, which is entered at the global level and applies to all bgp instances.
1260 The global and per-vrf approaches to defining update-delay are mutually
1263 When applicable, read-only mode would begin as soon as the first peer reaches
1264 Established status and a timer for max-delay seconds is started. During this
1265 mode BGP doesn't run any best-path or generate any updates to its peers. This
1266 mode continues until:
1268 1. All the configured peers, except the shutdown peers, have sent explicit EOR
1269 (End-Of-RIB) or an implicit-EOR. The first keep-alive after BGP has reached
1270 Established is considered an implicit-EOR.
1271 If the establish-wait optional value is given, then BGP will wait for
1272 peers to reach established from the beginning of the update-delay till the
1273 establish-wait period is over, i.e. the minimum set of established peers for
1274 which EOR is expected would be peers established during the establish-wait
1275 window, not necessarily all the configured neighbors.
1276 2. max-delay period is over.
1278 On hitting any of the above two conditions, BGP resumes the decision process
1279 and generates updates to its peers.
1281 Default max-delay is 0, i.e. the feature is off by default.
1283 .. clicmd:: table-map ROUTE-MAP-NAME
1285 This feature is used to apply a route-map on route updates from BGP to
1286 Zebra. All the applicable match operations are allowed, such as match on
1287 prefix, next-hop, communities, etc. Set operations for this attach-point are
1288 limited to metric and next-hop only. Any operation of this feature does not
1289 affect BGPs internal RIB.
1291 Supported for ipv4 and ipv6 address families. It works on multi-paths as
1292 well, however, metric setting is based on the best-path only.
1299 .. _bgp-defining-peers:
1304 .. clicmd:: neighbor PEER remote-as ASN
1306 Creates a new neighbor whose remote-as is ASN. PEER can be an IPv4 address
1307 or an IPv6 address or an interface to use for the connection.
1312 neighbor 10.0.0.1 remote-as 2
1314 In this case my router, in AS-1, is trying to peer with AS-2 at 10.0.0.1.
1316 This command must be the first command used when configuring a neighbor. If
1317 the remote-as is not specified, *bgpd* will complain like this: ::
1319 can't find neighbor 10.0.0.1
1321 .. clicmd:: neighbor PEER remote-as internal
1323 Create a peer as you would when you specify an ASN, except that if the
1324 peers ASN is different than mine as specified under the :clicmd:`router bgp ASN`
1325 command the connection will be denied.
1327 .. clicmd:: neighbor PEER remote-as external
1329 Create a peer as you would when you specify an ASN, except that if the
1330 peers ASN is the same as mine as specified under the :clicmd:`router bgp ASN`
1331 command the connection will be denied.
1333 .. clicmd:: bgp listen range <A.B.C.D/M|X:X::X:X/M> peer-group PGNAME
1335 Accept connections from any peers in the specified prefix. Configuration
1336 from the specified peer-group is used to configure these peers.
1340 When using BGP listen ranges, if the associated peer group has TCP MD5
1341 authentication configured, your kernel must support this on prefixes. On
1342 Linux, this support was added in kernel version 4.14. If your kernel does
1343 not support this feature you will get a warning in the log file, and the
1344 listen range will only accept connections from peers without MD5 configured.
1346 Additionally, we have observed that when using this option at scale (several
1347 hundred peers) the kernel may hit its option memory limit. In this situation
1348 you will see error messages like:
1350 ``bgpd: sockopt_tcp_signature: setsockopt(23): Cannot allocate memory``
1352 In this case you need to increase the value of the sysctl
1353 ``net.core.optmem_max`` to allow the kernel to allocate the necessary option
1356 .. clicmd:: coalesce-time (0-4294967295)
1358 The time in milliseconds that BGP will delay before deciding what peers
1359 can be put into an update-group together in order to generate a single
1360 update for them. The default time is 1000.
1362 .. _bgp-configuring-peers:
1367 .. clicmd:: neighbor PEER shutdown [message MSG...] [rtt (1-65535) [count (1-255)]]
1369 Shutdown the peer. We can delete the neighbor's configuration by
1370 ``no neighbor PEER remote-as ASN`` but all configuration of the neighbor
1371 will be deleted. When you want to preserve the configuration, but want to
1372 drop the BGP peer, use this syntax.
1374 Optionally you can specify a shutdown message `MSG`.
1376 Also, you can specify optionally ``rtt`` in milliseconds to automatically
1377 shutdown the peer if round-trip-time becomes higher than defined.
1379 Additional ``count`` parameter is the number of keepalive messages to count
1380 before shutdown the peer if round-trip-time becomes higher than defined.
1382 .. clicmd:: neighbor PEER disable-connected-check
1384 Allow peerings between directly connected eBGP peers using loopback
1387 .. clicmd:: neighbor PEER ebgp-multihop
1389 Specifying ``ebgp-multihop`` allows sessions with eBGP neighbors to
1390 establish when they are multiple hops away. When the neighbor is not
1391 directly connected and this knob is not enabled, the session will not
1394 If the peer's IP address is not in the RIB and is reachable via the
1395 default route, then you have to enable ``ip nht resolve-via-default``.
1397 .. clicmd:: neighbor PEER description ...
1399 Set description of the peer.
1401 .. clicmd:: neighbor PEER version VERSION
1403 Set up the neighbor's BGP version. `version` can be `4`, `4+` or `4-`. BGP
1404 version `4` is the default value used for BGP peering. BGP version `4+`
1405 means that the neighbor supports Multiprotocol Extensions for BGP-4. BGP
1406 version `4-` is similar but the neighbor speaks the old Internet-Draft
1407 revision 00's Multiprotocol Extensions for BGP-4. Some routing software is
1408 still using this version.
1410 .. clicmd:: neighbor PEER interface IFNAME
1412 When you connect to a BGP peer over an IPv6 link-local address, you have to
1413 specify the IFNAME of the interface used for the connection. To specify
1414 IPv4 session addresses, see the ``neighbor PEER update-source`` command
1417 This command is deprecated and may be removed in a future release. Its use
1420 .. clicmd:: neighbor PEER next-hop-self [all]
1422 This command specifies an announced route's nexthop as being equivalent to
1423 the address of the bgp router if it is learned via eBGP. If the optional
1424 keyword `all` is specified the modification is done also for routes learned
1427 .. clicmd:: neighbor PEER attribute-unchanged [{as-path|next-hop|med}]
1429 This command specifies attributes to be left unchanged for advertisements
1430 sent to a peer. Use this to leave the next-hop unchanged in ipv6
1431 configurations, as the route-map directive to leave the next-hop unchanged
1432 is only available for ipv4.
1434 .. clicmd:: neighbor PEER update-source <IFNAME|ADDRESS>
1436 Specify the IPv4 source address to use for the :abbr:`BGP` session to this
1437 neighbour, may be specified as either an IPv4 address directly or as an
1438 interface name (in which case the *zebra* daemon MUST be running in order
1439 for *bgpd* to be able to retrieve interface state).
1444 neighbor foo update-source 192.168.0.1
1445 neighbor bar update-source lo0
1448 .. clicmd:: neighbor PEER default-originate
1450 *bgpd*'s default is to not announce the default route (0.0.0.0/0) even if it
1451 is in routing table. When you want to announce default routes to the peer,
1454 .. clicmd:: neighbor PEER port PORT
1456 .. clicmd:: neighbor PEER password PASSWORD
1458 Set a MD5 password to be used with the tcp socket that is being used
1459 to connect to the remote peer. Please note if you are using this
1460 command with a large number of peers on linux you should consider
1461 modifying the `net.core.optmem_max` sysctl to a larger value to
1462 avoid out of memory errors from the linux kernel.
1464 .. clicmd:: neighbor PEER send-community
1466 .. clicmd:: neighbor PEER weight WEIGHT
1468 This command specifies a default `weight` value for the neighbor's routes.
1470 .. clicmd:: neighbor PEER maximum-prefix NUMBER [force]
1472 Sets a maximum number of prefixes we can receive from a given peer. If this
1473 number is exceeded, the BGP session will be destroyed.
1475 In practice, it is generally preferable to use a prefix-list to limit what
1476 prefixes are received from the peer instead of using this knob. Tearing down
1477 the BGP session when a limit is exceeded is far more destructive than merely
1478 rejecting undesired prefixes. The prefix-list method is also much more
1479 granular and offers much smarter matching criterion than number of received
1480 prefixes, making it more suited to implementing policy.
1482 If ``force`` is set, then ALL prefixes are counted for maximum instead of
1483 accepted only. This is useful for cases where an inbound filter is applied,
1484 but you want maximum-prefix to act on ALL (including filtered) prefixes. This
1485 option requires `soft-reconfiguration inbound` to be enabled for the peer.
1487 .. clicmd:: neighbor PEER maximum-prefix-out NUMBER
1489 Sets a maximum number of prefixes we can send to a given peer.
1491 Since sent prefix count is managed by update-groups, this option
1492 creates a separate update-group for outgoing updates.
1494 .. clicmd:: neighbor PEER local-as AS-NUMBER [no-prepend] [replace-as]
1496 Specify an alternate AS for this BGP process when interacting with the
1497 specified peer. With no modifiers, the specified local-as is prepended to
1498 the received AS_PATH when receiving routing updates from the peer, and
1499 prepended to the outgoing AS_PATH (after the process local AS) when
1500 transmitting local routes to the peer.
1502 If the no-prepend attribute is specified, then the supplied local-as is not
1503 prepended to the received AS_PATH.
1505 If the replace-as attribute is specified, then only the supplied local-as is
1506 prepended to the AS_PATH when transmitting local-route updates to this peer.
1508 Note that replace-as can only be specified if no-prepend is.
1510 This command is only allowed for eBGP peers.
1512 .. clicmd:: neighbor <A.B.C.D|X:X::X:X|WORD> as-override
1514 Override AS number of the originating router with the local AS number.
1516 Usually this configuration is used in PEs (Provider Edge) to replace
1517 the incoming customer AS number so the connected CE (Customer Edge)
1518 can use the same AS number as the other customer sites. This allows
1519 customers of the provider network to use the same AS number across
1522 This command is only allowed for eBGP peers.
1524 .. clicmd:: neighbor <A.B.C.D|X:X::X:X|WORD> allowas-in [<(1-10)|origin>]
1526 Accept incoming routes with AS path containing AS number with the same value
1527 as the current system AS.
1529 This is used when you want to use the same AS number in your sites, but you
1530 can't connect them directly. This is an alternative to
1531 `neighbor WORD as-override`.
1533 The parameter `(1-10)` configures the amount of accepted occurences of the
1534 system AS number in AS path.
1536 The parameter `origin` configures BGP to only accept routes originated with
1537 the same AS number as the system.
1539 This command is only allowed for eBGP peers.
1541 .. clicmd:: neighbor <A.B.C.D|X:X::X:X|WORD> addpath-tx-all-paths
1543 Configure BGP to send all known paths to neighbor in order to preserve multi
1544 path capabilities inside a network.
1546 .. clicmd:: neighbor <A.B.C.D|X:X::X:X|WORD> addpath-tx-bestpath-per-AS
1548 Configure BGP to send best known paths to neighbor in order to preserve multi
1549 path capabilities inside a network.
1551 .. clicmd:: neighbor PEER ttl-security hops NUMBER
1553 This command enforces Generalized TTL Security Mechanism (GTSM), as
1554 specified in RFC 5082. With this command, only neighbors that are the
1555 specified number of hops away will be allowed to become neighbors. This
1556 command is mutually exclusive with *ebgp-multihop*.
1558 .. clicmd:: neighbor PEER capability extended-nexthop
1560 Allow bgp to negotiate the extended-nexthop capability with it's peer.
1561 If you are peering over a v6 LL address then this capability is turned
1562 on automatically. If you are peering over a v6 Global Address then
1563 turning on this command will allow BGP to install v4 routes with
1564 v6 nexthops if you do not have v4 configured on interfaces.
1566 .. clicmd:: bgp fast-external-failover
1568 This command causes bgp to not take down ebgp peers immediately
1569 when a link flaps. `bgp fast-external-failover` is the default
1570 and will not be displayed as part of a `show run`. The no form
1571 of the command turns off this ability.
1573 .. clicmd:: bgp default ipv4-unicast
1575 This command allows the user to specify that v4 peering is turned
1576 on by default or not. This command defaults to on and is not displayed.
1577 The `no bgp default ipv4-unicast` form of the command is displayed.
1579 .. clicmd:: bgp default show-hostname
1581 This command shows the hostname of the peer in certain BGP commands
1582 outputs. It's easier to troubleshoot if you have a number of BGP peers.
1584 .. clicmd:: bgp default show-nexthop-hostname
1586 This command shows the hostname of the next-hop in certain BGP commands
1587 outputs. It's easier to troubleshoot if you have a number of BGP peers
1588 and a number of routes to check.
1590 .. clicmd:: neighbor PEER advertisement-interval (0-600)
1592 Setup the minimum route advertisement interval(mrai) for the
1593 peer in question. This number is between 0 and 600 seconds,
1594 with the default advertisement interval being 0.
1596 .. clicmd:: neighbor PEER timers delayopen (1-240)
1598 This command allows the user enable the
1599 `RFC 4271 <https://tools.ietf.org/html/rfc4271/>` DelayOpenTimer with the
1600 specified interval or disable it with the negating command for the peer. By
1601 default, the DelayOpenTimer is disabled. The timer interval may be set to a
1602 duration of 1 to 240 seconds.
1604 Displaying Information about Peers
1605 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1607 .. clicmd:: show bgp <afi> <safi> neighbors WORD bestpath-routes [json] [wide]
1609 For the given neighbor, WORD, that is specified list the routes selected
1610 by BGP as having the best path.
1612 .. _bgp-peer-filtering:
1617 .. clicmd:: neighbor PEER distribute-list NAME [in|out]
1619 This command specifies a distribute-list for the peer. `direct` is
1622 .. clicmd:: neighbor PEER prefix-list NAME [in|out]
1624 .. clicmd:: neighbor PEER filter-list NAME [in|out]
1626 .. clicmd:: neighbor PEER route-map NAME [in|out]
1628 Apply a route-map on the neighbor. `direct` must be `in` or `out`.
1630 .. clicmd:: bgp route-reflector allow-outbound-policy
1632 By default, attribute modification via route-map policy out is not reflected
1633 on reflected routes. This option allows the modifications to be reflected as
1634 well. Once enabled, it affects all reflected routes.
1636 .. clicmd:: neighbor PEER sender-as-path-loop-detection
1638 Enable the detection of sender side AS path loops and filter the
1639 bad routes before they are sent.
1641 This setting is disabled by default.
1648 Peer groups are used to help improve scaling by generating the same
1649 update information to all members of a peer group. Note that this means
1650 that the routes generated by a member of a peer group will be sent back
1651 to that originating peer with the originator identifier attribute set to
1652 indicated the originating peer. All peers not associated with a
1653 specific peer group are treated as belonging to a default peer group,
1654 and will share updates.
1656 .. clicmd:: neighbor WORD peer-group
1658 This command defines a new peer group.
1660 .. clicmd:: neighbor PEER peer-group PGNAME
1662 This command bind specific peer to peer group WORD.
1664 .. clicmd:: neighbor PEER solo
1666 This command is used to indicate that routes advertised by the peer
1667 should not be reflected back to the peer. This command only is only
1668 meaningful when there is a single peer defined in the peer-group.
1670 Capability Negotiation
1671 ^^^^^^^^^^^^^^^^^^^^^^
1673 .. clicmd:: neighbor PEER strict-capability-match
1676 Strictly compares remote capabilities and local capabilities. If
1677 capabilities are different, send Unsupported Capability error then reset
1680 You may want to disable sending Capability Negotiation OPEN message optional
1681 parameter to the peer when remote peer does not implement Capability
1682 Negotiation. Please use *dont-capability-negotiate* command to disable the
1685 .. clicmd:: neighbor PEER dont-capability-negotiate
1687 Suppress sending Capability Negotiation as OPEN message optional parameter
1688 to the peer. This command only affects the peer is configured other than
1689 IPv4 unicast configuration.
1691 When remote peer does not have capability negotiation feature, remote peer
1692 will not send any capabilities at all. In that case, bgp configures the peer
1693 with configured capabilities.
1695 You may prefer locally configured capabilities more than the negotiated
1696 capabilities even though remote peer sends capabilities. If the peer is
1697 configured by *override-capability*, *bgpd* ignores received capabilities
1698 then override negotiated capabilities with configured values.
1700 Additionally the operator should be reminded that this feature fundamentally
1701 disables the ability to use widely deployed BGP features. BGP unnumbered,
1702 hostname support, AS4, Addpath, Route Refresh, ORF, Dynamic Capabilities,
1703 and graceful restart.
1705 .. clicmd:: neighbor PEER override-capability
1708 Override the result of Capability Negotiation with local configuration.
1709 Ignore remote peer's capability value.
1711 .. _bgp-as-path-access-lists:
1713 AS Path Access Lists
1714 --------------------
1716 AS path access list is user defined AS path.
1718 .. clicmd:: bgp as-path access-list WORD permit|deny LINE
1720 This command defines a new AS path access list.
1724 .. _bgp-bogon-filter-example:
1726 Bogon ASN filter policy configuration example
1727 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1731 bgp as-path access-list 99 permit _0_
1732 bgp as-path access-list 99 permit _23456_
1733 bgp as-path access-list 99 permit _1310[0-6][0-9]_|_13107[0-1]_
1735 .. _bgp-using-as-path-in-route-map:
1737 Using AS Path in Route Map
1738 --------------------------
1740 .. clicmd:: match as-path WORD
1742 For a given as-path, WORD, match it on the BGP as-path given for the prefix
1743 and if it matches do normal route-map actions. The no form of the command
1744 removes this match from the route-map.
1746 .. clicmd:: set as-path prepend AS-PATH
1748 Prepend the given string of AS numbers to the AS_PATH of the BGP path's NLRI.
1749 The no form of this command removes this set operation from the route-map.
1751 .. clicmd:: set as-path prepend last-as NUM
1753 Prepend the existing last AS number (the leftmost ASN) to the AS_PATH.
1754 The no form of this command removes this set operation from the route-map.
1756 .. _bgp-communities-attribute:
1758 Communities Attribute
1759 ---------------------
1761 The BGP communities attribute is widely used for implementing policy routing.
1762 Network operators can manipulate BGP communities attribute based on their
1763 network policy. BGP communities attribute is defined in :rfc:`1997` and
1764 :rfc:`1998`. It is an optional transitive attribute, therefore local policy can
1765 travel through different autonomous system.
1767 The communities attribute is a set of communities values. Each community value
1768 is 4 octet long. The following format is used to define the community value.
1771 This format represents 4 octet communities value. ``AS`` is high order 2
1772 octet in digit format. ``VAL`` is low order 2 octet in digit format. This
1773 format is useful to define AS oriented policy value. For example,
1774 ``7675:80`` can be used when AS 7675 wants to pass local policy value 80 to
1778 ``internet`` represents well-known communities value 0.
1780 ``graceful-shutdown``
1781 ``graceful-shutdown`` represents well-known communities value
1782 ``GRACEFUL_SHUTDOWN`` ``0xFFFF0000`` ``65535:0``. :rfc:`8326` implements
1783 the purpose Graceful BGP Session Shutdown to reduce the amount of
1784 lost traffic when taking BGP sessions down for maintenance. The use
1785 of the community needs to be supported from your peers side to
1786 actually have any effect.
1789 ``accept-own`` represents well-known communities value ``ACCEPT_OWN``
1790 ``0xFFFF0001`` ``65535:1``. :rfc:`7611` implements a way to signal
1791 to a router to accept routes with a local nexthop address. This
1792 can be the case when doing policing and having traffic having a
1793 nexthop located in another VRF but still local interface to the
1794 router. It is recommended to read the RFC for full details.
1796 ``route-filter-translated-v4``
1797 ``route-filter-translated-v4`` represents well-known communities value
1798 ``ROUTE_FILTER_TRANSLATED_v4`` ``0xFFFF0002`` ``65535:2``.
1801 ``route-filter-v4`` represents well-known communities value
1802 ``ROUTE_FILTER_v4`` ``0xFFFF0003`` ``65535:3``.
1804 ``route-filter-translated-v6``
1805 ``route-filter-translated-v6`` represents well-known communities value
1806 ``ROUTE_FILTER_TRANSLATED_v6`` ``0xFFFF0004`` ``65535:4``.
1809 ``route-filter-v6`` represents well-known communities value
1810 ``ROUTE_FILTER_v6`` ``0xFFFF0005`` ``65535:5``.
1813 ``llgr-stale`` represents well-known communities value ``LLGR_STALE``
1814 ``0xFFFF0006`` ``65535:6``.
1815 Assigned and intended only for use with routers supporting the
1816 Long-lived Graceful Restart Capability as described in
1817 [Draft-IETF-uttaro-idr-bgp-persistence]_.
1818 Routers receiving routes with this community may (depending on
1819 implementation) choose allow to reject or modify routes on the
1820 presence or absence of this community.
1823 ``no-llgr`` represents well-known communities value ``NO_LLGR``
1824 ``0xFFFF0007`` ``65535:7``.
1825 Assigned and intended only for use with routers supporting the
1826 Long-lived Graceful Restart Capability as described in
1827 [Draft-IETF-uttaro-idr-bgp-persistence]_.
1828 Routers receiving routes with this community may (depending on
1829 implementation) choose allow to reject or modify routes on the
1830 presence or absence of this community.
1832 ``accept-own-nexthop``
1833 ``accept-own-nexthop`` represents well-known communities value
1834 ``accept-own-nexthop`` ``0xFFFF0008`` ``65535:8``.
1835 [Draft-IETF-agrewal-idr-accept-own-nexthop]_ describes
1836 how to tag and label VPN routes to be able to send traffic between VRFs
1837 via an internal layer 2 domain on the same PE device. Refer to
1838 [Draft-IETF-agrewal-idr-accept-own-nexthop]_ for full details.
1841 ``blackhole`` represents well-known communities value ``BLACKHOLE``
1842 ``0xFFFF029A`` ``65535:666``. :rfc:`7999` documents sending prefixes to
1843 EBGP peers and upstream for the purpose of blackholing traffic.
1844 Prefixes tagged with the this community should normally not be
1845 re-advertised from neighbors of the originating network. Upon receiving
1846 ``BLACKHOLE`` community from a BGP speaker, ``NO_ADVERTISE`` community
1847 is added automatically.
1850 ``no-export`` represents well-known communities value ``NO_EXPORT``
1851 ``0xFFFFFF01``. All routes carry this value must not be advertised to
1852 outside a BGP confederation boundary. If neighboring BGP peer is part of BGP
1853 confederation, the peer is considered as inside a BGP confederation
1854 boundary, so the route will be announced to the peer.
1857 ``no-advertise`` represents well-known communities value ``NO_ADVERTISE``
1858 ``0xFFFFFF02``. All routes carry this value must not be advertise to other
1862 ``local-AS`` represents well-known communities value ``NO_EXPORT_SUBCONFED``
1863 ``0xFFFFFF03``. All routes carry this value must not be advertised to
1864 external BGP peers. Even if the neighboring router is part of confederation,
1865 it is considered as external BGP peer, so the route will not be announced to
1869 ``no-peer`` represents well-known communities value ``NOPEER``
1870 ``0xFFFFFF04`` ``65535:65284``. :rfc:`3765` is used to communicate to
1871 another network how the originating network want the prefix propagated.
1873 When the communities attribute is received duplicate community values in the
1874 attribute are ignored and value is sorted in numerical order.
1876 .. [Draft-IETF-uttaro-idr-bgp-persistence] <https://tools.ietf.org/id/draft-uttaro-idr-bgp-persistence-04.txt>
1877 .. [Draft-IETF-agrewal-idr-accept-own-nexthop] <https://tools.ietf.org/id/draft-agrewal-idr-accept-own-nexthop-00.txt>
1879 .. _bgp-community-lists:
1883 Community lists are user defined lists of community attribute values. These
1884 lists can be used for matching or manipulating the communities attribute in
1887 There are two types of community list:
1890 This type accepts an explicit value for the attribute.
1893 This type accepts a regular expression. Because the regex must be
1894 interpreted on each use expanded community lists are slower than standard
1897 .. clicmd:: bgp community-list standard NAME permit|deny COMMUNITY
1899 This command defines a new standard community list. ``COMMUNITY`` is
1900 communities value. The ``COMMUNITY`` is compiled into community structure.
1901 We can define multiple community list under same name. In that case match
1902 will happen user defined order. Once the community list matches to
1903 communities attribute in BGP updates it return permit or deny by the
1904 community list definition. When there is no matched entry, deny will be
1905 returned. When ``COMMUNITY`` is empty it matches to any routes.
1907 .. clicmd:: bgp community-list expanded NAME permit|deny COMMUNITY
1909 This command defines a new expanded community list. ``COMMUNITY`` is a
1910 string expression of communities attribute. ``COMMUNITY`` can be a regular
1911 expression (:ref:`bgp-regular-expressions`) to match the communities
1912 attribute in BGP updates. The expanded community is only used to filter,
1916 It is recommended to use the more explicit versions of this command.
1918 .. clicmd:: bgp community-list NAME permit|deny COMMUNITY
1920 When the community list type is not specified, the community list type is
1921 automatically detected. If ``COMMUNITY`` can be compiled into communities
1922 attribute, the community list is defined as a standard community list.
1923 Otherwise it is defined as an expanded community list. This feature is left
1924 for backward compatibility. Use of this feature is not recommended.
1926 Note that all community lists share the same namespace, so it's not
1927 necessary to specify ``standard`` or ``expanded``; these modifiers are
1930 .. clicmd:: show bgp community-list [NAME detail]
1932 Displays community list information. When ``NAME`` is specified the
1933 specified community list's information is shown.
1937 # show bgp community-list
1938 Named Community standard list CLIST
1939 permit 7675:80 7675:100 no-export
1941 Named Community expanded list EXPAND
1944 # show bgp community-list CLIST detail
1945 Named Community standard list CLIST
1946 permit 7675:80 7675:100 no-export
1950 .. _bgp-numbered-community-lists:
1952 Numbered Community Lists
1953 ^^^^^^^^^^^^^^^^^^^^^^^^
1955 When number is used for BGP community list name, the number has
1956 special meanings. Community list number in the range from 1 and 99 is
1957 standard community list. Community list number in the range from 100
1958 to 199 is expanded community list. These community lists are called
1959 as numbered community lists. On the other hand normal community lists
1960 is called as named community lists.
1962 .. clicmd:: bgp community-list (1-99) permit|deny COMMUNITY
1964 This command defines a new community list. The argument to (1-99) defines
1965 the list identifier.
1967 .. clicmd:: bgp community-list (100-199) permit|deny COMMUNITY
1969 This command defines a new expanded community list. The argument to
1970 (100-199) defines the list identifier.
1972 .. _bgp-using-communities-in-route-map:
1974 Using Communities in Route Maps
1975 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1977 In :ref:`route-map` we can match on or set the BGP communities attribute. Using
1978 this feature network operator can implement their network policy based on BGP
1979 communities attribute.
1981 The following commands can be used in route maps:
1983 .. clicmd:: match community WORD exact-match [exact-match]
1985 This command perform match to BGP updates using community list WORD. When
1986 the one of BGP communities value match to the one of communities value in
1987 community list, it is match. When `exact-match` keyword is specified, match
1988 happen only when BGP updates have completely same communities value
1989 specified in the community list.
1991 .. clicmd:: set community <none|COMMUNITY> additive
1993 This command sets the community value in BGP updates. If the attribute is
1994 already configured, the newly provided value replaces the old one unless the
1995 ``additive`` keyword is specified, in which case the new value is appended
1996 to the existing value.
1998 If ``none`` is specified as the community value, the communities attribute
2001 It is not possible to set an expanded community list.
2003 .. clicmd:: set comm-list WORD delete
2005 This command remove communities value from BGP communities attribute. The
2006 ``word`` is community list name. When BGP route's communities value matches
2007 to the community list ``word``, the communities value is removed. When all
2008 of communities value is removed eventually, the BGP update's communities
2009 attribute is completely removed.
2011 .. _bgp-communities-example:
2013 Example Configuration
2014 ^^^^^^^^^^^^^^^^^^^^^
2016 The following configuration is exemplary of the most typical usage of BGP
2017 communities attribute. In the example, AS 7675 provides an upstream Internet
2018 connection to AS 100. When the following configuration exists in AS 7675, the
2019 network operator of AS 100 can set local preference in AS 7675 network by
2020 setting BGP communities attribute to the updates.
2025 neighbor 192.168.0.1 remote-as 100
2026 address-family ipv4 unicast
2027 neighbor 192.168.0.1 route-map RMAP in
2030 bgp community-list 70 permit 7675:70
2031 bgp community-list 70 deny
2032 bgp community-list 80 permit 7675:80
2033 bgp community-list 80 deny
2034 bgp community-list 90 permit 7675:90
2035 bgp community-list 90 deny
2037 route-map RMAP permit 10
2039 set local-preference 70
2041 route-map RMAP permit 20
2043 set local-preference 80
2045 route-map RMAP permit 30
2047 set local-preference 90
2050 The following configuration announces ``10.0.0.0/8`` from AS 100 to AS 7675.
2051 The route has communities value ``7675:80`` so when above configuration exists
2052 in AS 7675, the announced routes' local preference value will be set to 80.
2058 neighbor 192.168.0.2 remote-as 7675
2059 address-family ipv4 unicast
2060 neighbor 192.168.0.2 route-map RMAP out
2063 ip prefix-list PLIST permit 10.0.0.0/8
2065 route-map RMAP permit 10
2066 match ip address prefix-list PLIST
2067 set community 7675:80
2070 The following configuration is an example of BGP route filtering using
2071 communities attribute. This configuration only permit BGP routes which has BGP
2072 communities value ``0:80`` or ``0:90``. The network operator can set special
2073 internal communities value at BGP border router, then limit the BGP route
2074 announcements into the internal network.
2079 neighbor 192.168.0.1 remote-as 100
2080 address-family ipv4 unicast
2081 neighbor 192.168.0.1 route-map RMAP in
2084 bgp community-list 1 permit 0:80 0:90
2086 route-map RMAP permit in
2090 The following example filters BGP routes which have a community value of
2091 ``1:1``. When there is no match community-list returns ``deny``. To avoid
2092 filtering all routes, a ``permit`` line is set at the end of the
2098 neighbor 192.168.0.1 remote-as 100
2099 address-family ipv4 unicast
2100 neighbor 192.168.0.1 route-map RMAP in
2103 bgp community-list standard FILTER deny 1:1
2104 bgp community-list standard FILTER permit
2106 route-map RMAP permit 10
2107 match community FILTER
2110 The communities value keyword ``internet`` has special meanings in standard
2111 community lists. In the below example ``internet`` matches all BGP routes even
2112 if the route does not have communities attribute at all. So community list
2113 ``INTERNET`` is the same as ``FILTER`` in the previous example.
2117 bgp community-list standard INTERNET deny 1:1
2118 bgp community-list standard INTERNET permit internet
2121 The following configuration is an example of communities value deletion. With
2122 this configuration the community values ``100:1`` and ``100:2`` are removed
2123 from BGP updates. For communities value deletion, only ``permit``
2124 community-list is used. ``deny`` community-list is ignored.
2129 neighbor 192.168.0.1 remote-as 100
2130 address-family ipv4 unicast
2131 neighbor 192.168.0.1 route-map RMAP in
2134 bgp community-list standard DEL permit 100:1 100:2
2136 route-map RMAP permit 10
2137 set comm-list DEL delete
2140 .. _bgp-extended-communities-attribute:
2142 Extended Communities Attribute
2143 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2145 BGP extended communities attribute is introduced with MPLS VPN/BGP technology.
2146 MPLS VPN/BGP expands capability of network infrastructure to provide VPN
2147 functionality. At the same time it requires a new framework for policy routing.
2148 With BGP Extended Communities Attribute we can use Route Target or Site of
2149 Origin for implementing network policy for MPLS VPN/BGP.
2151 BGP Extended Communities Attribute is similar to BGP Communities Attribute. It
2152 is an optional transitive attribute. BGP Extended Communities Attribute can
2153 carry multiple Extended Community value. Each Extended Community value is
2156 BGP Extended Communities Attribute provides an extended range compared with BGP
2157 Communities Attribute. Adding to that there is a type field in each value to
2158 provides community space structure.
2160 There are two format to define Extended Community value. One is AS based format
2161 the other is IP address based format.
2164 This is a format to define AS based Extended Community value. ``AS`` part
2165 is 2 octets Global Administrator subfield in Extended Community value.
2166 ``VAL`` part is 4 octets Local Administrator subfield. ``7675:100``
2167 represents AS 7675 policy value 100.
2170 This is a format to define IP address based Extended Community value.
2171 ``IP-Address`` part is 4 octets Global Administrator subfield. ``VAL`` part
2172 is 2 octets Local Administrator subfield.
2174 .. _bgp-extended-community-lists:
2176 Extended Community Lists
2177 ^^^^^^^^^^^^^^^^^^^^^^^^
2179 .. clicmd:: bgp extcommunity-list standard NAME permit|deny EXTCOMMUNITY
2181 This command defines a new standard extcommunity-list. `extcommunity` is
2182 extended communities value. The `extcommunity` is compiled into extended
2183 community structure. We can define multiple extcommunity-list under same
2184 name. In that case match will happen user defined order. Once the
2185 extcommunity-list matches to extended communities attribute in BGP updates
2186 it return permit or deny based upon the extcommunity-list definition. When
2187 there is no matched entry, deny will be returned. When `extcommunity` is
2188 empty it matches to any routes.
2190 .. clicmd:: bgp extcommunity-list expanded NAME permit|deny LINE
2192 This command defines a new expanded extcommunity-list. `line` is a string
2193 expression of extended communities attribute. `line` can be a regular
2194 expression (:ref:`bgp-regular-expressions`) to match an extended communities
2195 attribute in BGP updates.
2197 Note that all extended community lists shares a single name space, so it's
2198 not necessary to specify their type when creating or destroying them.
2200 .. clicmd:: show bgp extcommunity-list [NAME detail]
2202 This command displays current extcommunity-list information. When `name` is
2203 specified the community list's information is shown.
2206 .. _bgp-extended-communities-in-route-map:
2208 BGP Extended Communities in Route Map
2209 """""""""""""""""""""""""""""""""""""
2211 .. clicmd:: match extcommunity WORD
2213 .. clicmd:: set extcommunity rt EXTCOMMUNITY
2215 This command set Route Target value.
2217 .. clicmd:: set extcommunity soo EXTCOMMUNITY
2219 This command set Site of Origin value.
2221 .. clicmd:: set extcommunity bandwidth <(1-25600) | cumulative | num-multipaths> [non-transitive]
2223 This command sets the BGP link-bandwidth extended community for the prefix
2224 (best path) for which it is applied. The link-bandwidth can be specified as
2225 an ``explicit value`` (specified in Mbps), or the router can be told to use
2226 the ``cumulative bandwidth`` of all multipaths for the prefix or to compute
2227 it based on the ``number of multipaths``. The link bandwidth extended
2228 community is encoded as ``transitive`` unless the set command explicitly
2229 configures it as ``non-transitive``.
2231 .. seealso:: :ref:`wecmp_linkbw`
2233 Note that the extended expanded community is only used for `match` rule, not for
2236 .. _bgp-large-communities-attribute:
2238 Large Communities Attribute
2239 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
2241 The BGP Large Communities attribute was introduced in Feb 2017 with
2244 The BGP Large Communities Attribute is similar to the BGP Communities Attribute
2245 except that it has 3 components instead of two and each of which are 4 octets
2246 in length. Large Communities bring additional functionality and convenience
2247 over traditional communities, specifically the fact that the ``GLOBAL`` part
2248 below is now 4 octets wide allowing seamless use in networks using 4-byte ASNs.
2250 ``GLOBAL:LOCAL1:LOCAL2``
2251 This is the format to define Large Community values. Referencing :rfc:`8195`
2252 the values are commonly referred to as follows:
2254 - The ``GLOBAL`` part is a 4 octet Global Administrator field, commonly used
2255 as the operators AS number.
2256 - The ``LOCAL1`` part is a 4 octet Local Data Part 1 subfield referred to as
2258 - The ``LOCAL2`` part is a 4 octet Local Data Part 2 field and referred to
2259 as the parameter subfield.
2261 As an example, ``65551:1:10`` represents AS 65551 function 1 and parameter
2262 10. The referenced RFC above gives some guidelines on recommended usage.
2264 .. _bgp-large-community-lists:
2266 Large Community Lists
2267 """""""""""""""""""""
2269 Two types of large community lists are supported, namely `standard` and
2272 .. clicmd:: bgp large-community-list standard NAME permit|deny LARGE-COMMUNITY
2274 This command defines a new standard large-community-list. `large-community`
2275 is the Large Community value. We can add multiple large communities under
2276 same name. In that case the match will happen in the user defined order.
2277 Once the large-community-list matches the Large Communities attribute in BGP
2278 updates it will return permit or deny based upon the large-community-list
2279 definition. When there is no matched entry, a deny will be returned. When
2280 `large-community` is empty it matches any routes.
2282 .. clicmd:: bgp large-community-list expanded NAME permit|deny LINE
2284 This command defines a new expanded large-community-list. Where `line` is a
2285 string matching expression, it will be compared to the entire Large
2286 Communities attribute as a string, with each large-community in order from
2287 lowest to highest. `line` can also be a regular expression which matches
2288 this Large Community attribute.
2290 Note that all community lists share the same namespace, so it's not
2291 necessary to specify ``standard`` or ``expanded``; these modifiers are
2294 .. clicmd:: show bgp large-community-list
2296 .. clicmd:: show bgp large-community-list NAME detail
2298 This command display current large-community-list information. When
2299 `name` is specified the community list information is shown.
2301 .. clicmd:: show ip bgp large-community-info
2303 This command displays the current large communities in use.
2305 .. _bgp-large-communities-in-route-map:
2307 Large Communities in Route Map
2308 """"""""""""""""""""""""""""""
2310 .. clicmd:: match large-community LINE [exact-match]
2312 Where `line` can be a simple string to match, or a regular expression. It
2313 is very important to note that this match occurs on the entire
2314 large-community string as a whole, where each large-community is ordered
2315 from lowest to highest. When `exact-match` keyword is specified, match
2316 happen only when BGP updates have completely same large communities value
2317 specified in the large community list.
2319 .. clicmd:: set large-community LARGE-COMMUNITY
2321 .. clicmd:: set large-community LARGE-COMMUNITY LARGE-COMMUNITY
2323 .. clicmd:: set large-community LARGE-COMMUNITY additive
2325 These commands are used for setting large-community values. The first
2326 command will overwrite any large-communities currently present.
2327 The second specifies two large-communities, which overwrites the current
2328 large-community list. The third will add a large-community value without
2329 overwriting other values. Multiple large-community values can be specified.
2331 Note that the large expanded community is only used for `match` rule, not for
2339 *bgpd* supports :abbr:`L3VPN (Layer 3 Virtual Private Networks)` :abbr:`VRFs
2340 (Virtual Routing and Forwarding)` for IPv4 :rfc:`4364` and IPv6 :rfc:`4659`.
2341 L3VPN routes, and their associated VRF MPLS labels, can be distributed to VPN
2342 SAFI neighbors in the *default*, i.e., non VRF, BGP instance. VRF MPLS labels
2343 are reached using *core* MPLS labels which are distributed using LDP or BGP
2344 labeled unicast. *bgpd* also supports inter-VRF route leaking.
2347 .. _bgp-vrf-route-leaking:
2352 BGP routes may be leaked (i.e. copied) between a unicast VRF RIB and the VPN
2353 SAFI RIB of the default VRF for use in MPLS-based L3VPNs. Unicast routes may
2354 also be leaked between any VRFs (including the unicast RIB of the default BGP
2355 instanced). A shortcut syntax is also available for specifying leaking from one
2356 VRF to another VRF using the default instance's VPN RIB as the intemediary. A
2357 common application of the VRF-VRF feature is to connect a customer's private
2358 routing domain to a provider's VPN service. Leaking is configured from the
2359 point of view of an individual VRF: ``import`` refers to routes leaked from VPN
2360 to a unicast VRF, whereas ``export`` refers to routes leaked from a unicast VRF
2366 Routes exported from a unicast VRF to the VPN RIB must be augmented by two
2369 - an :abbr:`RD (Route Distinguisher)`
2370 - an :abbr:`RTLIST (Route-target List)`
2372 Configuration for these exported routes must, at a minimum, specify these two
2375 Routes imported from the VPN RIB to a unicast VRF are selected according to
2376 their RTLISTs. Routes whose RTLIST contains at least one route-target in
2377 common with the configured import RTLIST are leaked. Configuration for these
2378 imported routes must specify an RTLIST to be matched.
2380 The RD, which carries no semantic value, is intended to make the route unique
2381 in the VPN RIB among all routes of its prefix that originate from all the
2382 customers and sites that are attached to the provider's VPN service.
2383 Accordingly, each site of each customer is typically assigned an RD that is
2384 unique across the entire provider network.
2386 The RTLIST is a set of route-target extended community values whose purpose is
2387 to specify route-leaking policy. Typically, a customer is assigned a single
2388 route-target value for import and export to be used at all customer sites. This
2389 configuration specifies a simple topology wherein a customer has a single
2390 routing domain which is shared across all its sites. More complex routing
2391 topologies are possible through use of additional route-targets to augment the
2392 leaking of sets of routes in various ways.
2394 When using the shortcut syntax for vrf-to-vrf leaking, the RD and RT are
2397 General configuration
2398 ^^^^^^^^^^^^^^^^^^^^^
2400 Configuration of route leaking between a unicast VRF RIB and the VPN SAFI RIB
2401 of the default VRF is accomplished via commands in the context of a VRF
2404 .. clicmd:: rd vpn export AS:NN|IP:nn
2406 Specifies the route distinguisher to be added to a route exported from the
2407 current unicast VRF to VPN.
2409 .. clicmd:: rt vpn import|export|both RTLIST...
2411 Specifies the route-target list to be attached to a route (export) or the
2412 route-target list to match against (import) when exporting/importing between
2413 the current unicast VRF and VPN.
2415 The RTLIST is a space-separated list of route-targets, which are BGP
2416 extended community values as described in
2417 :ref:`bgp-extended-communities-attribute`.
2419 .. clicmd:: label vpn export (0..1048575)|auto
2421 Enables an MPLS label to be attached to a route exported from the current
2422 unicast VRF to VPN. If the value specified is ``auto``, the label value is
2423 automatically assigned from a pool maintained by the Zebra daemon. If Zebra
2424 is not running, or if this command is not configured, automatic label
2425 assignment will not complete, which will block corresponding route export.
2427 .. clicmd:: nexthop vpn export A.B.C.D|X:X::X:X
2429 Specifies an optional nexthop value to be assigned to a route exported from
2430 the current unicast VRF to VPN. If left unspecified, the nexthop will be set
2431 to 0.0.0.0 or 0:0::0:0 (self).
2433 .. clicmd:: route-map vpn import|export MAP
2435 Specifies an optional route-map to be applied to routes imported or exported
2436 between the current unicast VRF and VPN.
2438 .. clicmd:: import|export vpn
2440 Enables import or export of routes between the current unicast VRF and VPN.
2442 .. clicmd:: import vrf VRFNAME
2444 Shortcut syntax for specifying automatic leaking from vrf VRFNAME to
2445 the current VRF using the VPN RIB as intermediary. The RD and RT
2446 are auto derived and should not be specified explicitly for either the
2447 source or destination VRF's.
2449 This shortcut syntax mode is not compatible with the explicit
2450 `import vpn` and `export vpn` statements for the two VRF's involved.
2451 The CLI will disallow attempts to configure incompatible leaking
2457 Ethernet Virtual Network - EVPN
2458 -------------------------------
2460 .. _bgp-evpn-advertise-pip:
2465 In a EVPN symmetric routing MLAG deployment, all EVPN routes advertised
2466 with anycast-IP as next-hop IP and anycast MAC as the Router MAC (RMAC - in
2467 BGP EVPN Extended-Community).
2468 EVPN picks up the next-hop IP from the VxLAN interface's local tunnel IP and
2469 the RMAC is obtained from the MAC of the L3VNI's SVI interface.
2470 Note: Next-hop IP is used for EVPN routes whether symmetric routing is
2471 deployed or not but the RMAC is only relevant for symmetric routing scenario.
2473 Current behavior is not ideal for Prefix (type-5) and self (type-2)
2474 routes. This is because the traffic from remote VTEPs routed sub optimally
2475 if they land on the system where the route does not belong.
2477 The advertise-pip feature advertises Prefix (type-5) and self (type-2)
2478 routes with system's individual (primary) IP as the next-hop and individual
2479 (system) MAC as Router-MAC (RMAC), while leaving the behavior unchanged for
2482 To support this feature there needs to have ability to co-exist a
2483 (system-MAC, system-IP) pair with a (anycast-MAC, anycast-IP) pair with the
2484 ability to terminate VxLAN-encapsulated packets received for either pair on
2485 the same L3VNI (i.e associated VLAN). This capability is need per tenant
2488 To derive the system-MAC and the anycast MAC, there needs to have a
2489 separate/additional MAC-VLAN interface corresponding to L3VNI’s SVI.
2490 The SVI interface’s MAC address can be interpreted as system-MAC
2491 and MAC-VLAN interface's MAC as anycast MAC.
2493 To derive system-IP and anycast-IP, the default BGP instance's router-id is used
2494 as system-IP and the VxLAN interface’s local tunnel IP as the anycast-IP.
2496 User has an option to configure the system-IP and/or system-MAC value if the
2497 auto derived value is not preferred.
2499 Note: By default, advertise-pip feature is enabled and user has an option to
2500 disable the feature via configuration CLI. Once the feature is disable under
2501 bgp vrf instance or MAC-VLAN interface is not configured, all the routes follow
2502 the same behavior of using same next-hop and RMAC values.
2504 .. clicmd:: advertise-pip [ip <addr> [mac <addr>]]
2506 Enables or disables advertise-pip feature, specifiy system-IP and/or system-MAC
2512 All-Active Multihoming is used for redundancy and load sharing. Servers
2513 are attached to two or more PEs and the links are bonded (link-aggregation).
2514 This group of server links is referred to as an Ethernet Segment.
2518 An Ethernet Segment can be configured by specifying a system-MAC and a
2519 local discriminatior against the bond interface on the PE (via zebra) -
2521 .. clicmd:: evpn mh es-id (1-16777215)
2523 .. clicmd:: evpn mh es-sys-mac X:X:X:X:X:X
2525 The sys-mac and local discriminator are used for generating a 10-byte,
2526 Type-3 Ethernet Segment ID.
2528 Type-1 (EAS-per-ES and EAD-per-EVI) routes are used to advertise the locally
2529 attached ESs and to learn off remote ESs in the network. Local Type-2/MAC-IP
2530 routes are also advertised with a destination ESI allowing for MAC-IP syncing
2531 between Ethernet Segment peers.
2532 Reference: RFC 7432, RFC 8365
2534 EVPN-MH is intended as a replacement for MLAG or Anycast VTEPs. In
2535 multihoming each PE has an unique VTEP address which requires the introduction
2536 of a new dataplane construct, MAC-ECMP. Here a MAC/FDB entry can point to a
2537 list of remote PEs/VTEPs.
2541 Type-4 (ESR) routes are used for Designated Forwarder (DF) election. DFs
2542 forward BUM traffic received via the overlay network. This implementation
2543 uses a preference based DF election specified by draft-ietf-bess-evpn-pref-df.
2544 The DF preference is configurable per-ES (via zebra) -
2546 .. clicmd:: evpn mh es-df-pref (1-16777215)
2548 BUM traffic is rxed via the overlay by all PEs attached to a server but
2549 only the DF can forward the de-capsulated traffic to the access port. To
2550 accomodate that non-DF filters are installed in the dataplane to drop
2553 Similarly traffic received from ES peers via the overlay cannot be forwarded
2554 to the server. This is split-horizon-filtering with local bias.
2558 Some vendors do not send EAD-per-EVI routes. To interop with them we
2559 need to relax the dependency on EAD-per-EVI routes and activate a remote
2560 ES-PE based on just the EAD-per-ES route.
2562 Note that by default we advertise and expect EAD-per-EVI routes.
2564 .. clicmd:: disable-ead-evi-rx
2566 .. clicmd:: disable-ead-evi-tx
2570 As the primary purpose of EVPN-MH is redundancy keeping the failover efficient
2571 is a recurring theme in the implementation. Following sub-features have
2572 been introduced for the express purpose of efficient ES failovers.
2574 - Layer-2 Nexthop Groups and MAC-ECMP via L2NHG.
2576 - Host routes (for symmetric IRB) via L3NHG.
2577 On dataplanes that support layer3 nexthop groups the feature can be turned
2578 on via the following BGP config -
2580 .. clicmd:: use-es-l3nhg
2582 - Local ES (MAC/Neigh) failover via ES-redirect.
2583 On dataplanes that do not have support for ES-redirect the feature can be
2584 turned off via the following zebra config -
2586 .. clicmd:: evpn mh redirect-off
2588 Uplink/Core tracking
2589 """"""""""""""""""""
2590 When all the underlay links go down the PE no longer has access to the VxLAN
2591 +overlay. To prevent blackholing of traffic the server/ES links are
2592 protodowned on the PE. A link can be setup for uplink tracking via the
2593 following zebra configuration -
2595 .. clicmd:: evpn mh uplink
2597 Proxy advertisements
2598 """"""""""""""""""""
2599 To handle hitless upgrades support for proxy advertisement has been added
2600 as specified by draft-rbickhart-evpn-ip-mac-proxy-adv. This allows a PE
2601 (say PE1) to proxy advertise a MAC-IP rxed from an ES peer (say PE2). When
2602 the ES peer (PE2) goes down PE1 continues to advertise hosts learnt from PE2
2603 for a holdtime during which it attempts to establish local reachability of
2604 the host. This holdtime is configurable via the following zebra commands -
2606 .. clicmd:: evpn mh neigh-holdtime (0-86400)
2608 .. clicmd:: evpn mh mac-holdtime (0-86400)
2612 When a switch is rebooted we wait for a brief period to allow the underlay
2613 and EVPN network to converge before enabling the ESs. For this duration the
2614 ES bonds are held protodown. The startup delay is configurable via the
2615 following zebra command -
2617 .. clicmd:: evpn mh startup-delay (0-3600)
2619 +Support with VRF network namespace backend
2620 +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2621 It is possible to separate overlay networks contained in VXLAN interfaces from
2622 underlay networks by using VRFs. VRF-lite and VRF-netns backends can be used for
2623 that. In the latter case, it is necessary to set both bridge and vxlan interface
2624 in the same network namespace, as below example illustrates:
2626 .. code-block:: shell
2630 ip link add name vxlan101 type vxlan id 101 dstport 4789 dev eth0 local 10.1.1.1
2631 ip link set dev vxlan101 netns vrf1
2632 ip netns exec vrf1 ip link set dev lo up
2633 ip netns exec vrf1 brctl addbr bridge101
2634 ip netns exec vrf1 brctl addif bridge101 vxlan101
2636 This makes it possible to separate not only layer 3 networks like VRF-lite networks.
2637 Also, VRF netns based make possible to separate layer 2 networks on separate VRF
2640 .. _bgp-conditional-advertisement:
2642 BGP Conditional Advertisement
2643 -----------------------------
2644 The BGP conditional advertisement feature uses the ``non-exist-map`` or the
2645 ``exist-map`` and the ``advertise-map`` keywords of the neighbor advertise-map
2646 command in order to track routes by the route prefix.
2649 1. If a route prefix is not present in the output of non-exist-map command,
2650 then advertise the route specified by the advertise-map command.
2652 2. If a route prefix is present in the output of non-exist-map command,
2653 then do not advertise the route specified by the addvertise-map command.
2656 1. If a route prefix is present in the output of exist-map command,
2657 then advertise the route specified by the advertise-map command.
2659 2. If a route prefix is not present in the output of exist-map command,
2660 then do not advertise the route specified by the advertise-map command.
2662 This feature is useful when some prefixes are advertised to one of its peers
2663 only if the information from the other peer is not present (due to failure in
2664 peering session or partial reachability etc).
2666 The conditional BGP announcements are sent in addition to the normal
2667 announcements that a BGP router sends to its peer.
2669 The conditional advertisement process is triggered by the BGP scanner process,
2670 which runs every 60 seconds. This means that the maximum time for the conditional
2671 advertisement to take effect is 60 seconds. The conditional advertisement can take
2672 effect depending on when the tracked route is removed from the BGP table and
2673 when the next instance of the BGP scanner occurs.
2675 .. clicmd:: neighbor A.B.C.D advertise-map NAME [exist-map|non-exist-map] NAME
2677 This command enables BGP scanner process to monitor routes specified by
2678 exist-map or non-exist-map command in BGP table and conditionally advertises
2679 the routes specified by advertise-map command.
2681 Sample Configuration
2682 ^^^^^^^^^^^^^^^^^^^^^
2686 ip address 10.10.10.2/24
2689 ip address 10.10.20.2/24
2692 ip address 203.0.113.1/32
2695 bgp log-neighbor-changes
2696 no bgp ebgp-requires-policy
2697 neighbor 10.10.10.1 remote-as 1
2698 neighbor 10.10.20.3 remote-as 3
2700 address-family ipv4 unicast
2701 neighbor 10.10.10.1 soft-reconfiguration inbound
2702 neighbor 10.10.20.3 soft-reconfiguration inbound
2703 neighbor 10.10.20.3 advertise-map ADV-MAP non-exist-map EXIST-MAP
2706 ip prefix-list DEFAULT seq 5 permit 192.0.2.5/32
2707 ip prefix-list DEFAULT seq 10 permit 192.0.2.1/32
2708 ip prefix-list EXIST seq 5 permit 10.10.10.10/32
2709 ip prefix-list DEFAULT-ROUTE seq 5 permit 0.0.0.0/0
2710 ip prefix-list IP1 seq 5 permit 10.139.224.0/20
2712 bgp community-list standard DC-ROUTES seq 5 permit 64952:3008
2713 bgp community-list standard DC-ROUTES seq 10 permit 64671:501
2714 bgp community-list standard DC-ROUTES seq 15 permit 64950:3009
2715 bgp community-list standard DEFAULT-ROUTE seq 5 permit 65013:200
2717 route-map ADV-MAP permit 10
2718 match ip address prefix-list IP1
2720 route-map ADV-MAP permit 20
2721 match community DC-ROUTES
2723 route-map EXIST-MAP permit 10
2724 match community DEFAULT-ROUTE
2725 match ip address prefix-list DEFAULT-ROUTE
2731 When default route is present in R2'2 BGP table, 10.139.224.0/20 and 192.0.2.1/32 are not advertised to R3.
2735 Router2# show ip bgp
2736 BGP table version is 20, local router ID is 203.0.113.1, vrf id 0
2737 Default local pref 100, local AS 2
2738 Status codes: s suppressed, d damped, h history, * valid, > best, = multipath,
2739 i internal, r RIB-failure, S Stale, R Removed
2740 Nexthop codes: @NNN nexthop's vrf id, < announce-nh-self
2741 Origin codes: i - IGP, e - EGP, ? - incomplete
2743 Network Next Hop Metric LocPrf Weight Path
2744 *> 0.0.0.0/0 10.10.10.1 0 0 1 i
2745 *> 10.139.224.0/20 10.10.10.1 0 0 1 ?
2746 *> 192.0.2.1/32 10.10.10.1 0 0 1 i
2747 *> 192.0.2.5/32 10.10.10.1 0 0 1 i
2749 Displayed 4 routes and 4 total paths
2750 Router2# show ip bgp neighbors 10.10.20.3
2752 !--- Output suppressed.
2754 For address family: IPv4 Unicast
2755 Update group 7, subgroup 7
2756 Packet Queue length 0
2757 Inbound soft reconfiguration allowed
2758 Community attribute sent to this neighbor(all)
2759 Condition NON_EXIST, Condition-map *EXIST-MAP, Advertise-map *ADV-MAP, status: Withdraw
2762 !--- Output suppressed.
2764 Router2# show ip bgp neighbors 10.10.20.3 advertised-routes
2765 BGP table version is 20, local router ID is 203.0.113.1, vrf id 0
2766 Default local pref 100, local AS 2
2767 Status codes: s suppressed, d damped, h history, * valid, > best, = multipath,
2768 i internal, r RIB-failure, S Stale, R Removed
2769 Nexthop codes: @NNN nexthop's vrf id, < announce-nh-self
2770 Origin codes: i - IGP, e - EGP, ? - incomplete
2772 Network Next Hop Metric LocPrf Weight Path
2773 *> 0.0.0.0/0 0.0.0.0 0 1 i
2774 *> 192.0.2.5/32 0.0.0.0 0 1 i
2776 Total number of prefixes 2
2778 When default route is not present in R2'2 BGP table, 10.139.224.0/20 and 192.0.2.1/32 are advertised to R3.
2782 Router2# show ip bgp
2783 BGP table version is 21, local router ID is 203.0.113.1, vrf id 0
2784 Default local pref 100, local AS 2
2785 Status codes: s suppressed, d damped, h history, * valid, > best, = multipath,
2786 i internal, r RIB-failure, S Stale, R Removed
2787 Nexthop codes: @NNN nexthop's vrf id, < announce-nh-self
2788 Origin codes: i - IGP, e - EGP, ? - incomplete
2790 Network Next Hop Metric LocPrf Weight Path
2791 *> 10.139.224.0/20 10.10.10.1 0 0 1 ?
2792 *> 192.0.2.1/32 10.10.10.1 0 0 1 i
2793 *> 192.0.2.5/32 10.10.10.1 0 0 1 i
2795 Displayed 3 routes and 3 total paths
2797 Router2# show ip bgp neighbors 10.10.20.3
2799 !--- Output suppressed.
2801 For address family: IPv4 Unicast
2802 Update group 7, subgroup 7
2803 Packet Queue length 0
2804 Inbound soft reconfiguration allowed
2805 Community attribute sent to this neighbor(all)
2806 Condition NON_EXIST, Condition-map *EXIST-MAP, Advertise-map *ADV-MAP, status: Advertise
2809 !--- Output suppressed.
2811 Router2# show ip bgp neighbors 10.10.20.3 advertised-routes
2812 BGP table version is 21, local router ID is 203.0.113.1, vrf id 0
2813 Default local pref 100, local AS 2
2814 Status codes: s suppressed, d damped, h history, * valid, > best, = multipath,
2815 i internal, r RIB-failure, S Stale, R Removed
2816 Nexthop codes: @NNN nexthop's vrf id, < announce-nh-self
2817 Origin codes: i - IGP, e - EGP, ? - incomplete
2819 Network Next Hop Metric LocPrf Weight Path
2820 *> 10.139.224.0/20 0.0.0.0 0 1 ?
2821 *> 192.0.2.1/32 0.0.0.0 0 1 i
2822 *> 192.0.2.5/32 0.0.0.0 0 1 i
2824 Total number of prefixes 3
2832 .. clicmd:: show debug
2834 Show all enabled debugs.
2836 .. clicmd:: show bgp listeners
2838 Display Listen sockets and the vrf that created them. Useful for debugging of when
2839 listen is not working and this is considered a developer debug statement.
2841 .. clicmd:: debug bgp neighbor-events
2843 Enable or disable debugging for neighbor events. This provides general
2844 information on BGP events such as peer connection / disconnection, session
2845 establishment / teardown, and capability negotiation.
2847 .. clicmd:: debug bgp updates
2849 Enable or disable debugging for BGP updates. This provides information on
2850 BGP UPDATE messages transmitted and received between local and remote
2853 .. clicmd:: debug bgp keepalives
2855 Enable or disable debugging for BGP keepalives. This provides information on
2856 BGP KEEPALIVE messages transmitted and received between local and remote
2859 .. clicmd:: debug bgp bestpath <A.B.C.D/M|X:X::X:X/M>
2861 Enable or disable debugging for bestpath selection on the specified prefix.
2863 .. clicmd:: debug bgp nht
2865 Enable or disable debugging of BGP nexthop tracking.
2867 .. clicmd:: debug bgp update-groups
2869 Enable or disable debugging of dynamic update groups. This provides general
2870 information on group creation, deletion, join and prune events.
2872 .. clicmd:: debug bgp zebra
2874 Enable or disable debugging of communications between *bgpd* and *zebra*.
2876 Dumping Messages and Routing Tables
2877 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2879 .. clicmd:: dump bgp all PATH [INTERVAL]
2881 .. clicmd:: dump bgp all-et PATH [INTERVAL]
2884 Dump all BGP packet and events to `path` file.
2885 If `interval` is set, a new file will be created for echo `interval` of
2886 seconds. The path `path` can be set with date and time formatting
2887 (strftime). The type ‘all-et’ enables support for Extended Timestamp Header
2888 (:ref:`packet-binary-dump-format`).
2890 .. clicmd:: dump bgp updates PATH [INTERVAL]
2892 .. clicmd:: dump bgp updates-et PATH [INTERVAL]
2895 Dump only BGP updates messages to `path` file.
2896 If `interval` is set, a new file will be created for echo `interval` of
2897 seconds. The path `path` can be set with date and time formatting
2898 (strftime). The type ‘updates-et’ enables support for Extended Timestamp
2899 Header (:ref:`packet-binary-dump-format`).
2901 .. clicmd:: dump bgp routes-mrt PATH
2903 .. clicmd:: dump bgp routes-mrt PATH INTERVAL
2906 Dump whole BGP routing table to `path`. This is heavy process. The path
2907 `path` can be set with date and time formatting (strftime). If `interval` is
2908 set, a new file will be created for echo `interval` of seconds.
2910 Note: the interval variable can also be set using hours and minutes: 04h20m00.
2913 .. _bgp-other-commands:
2918 The following are available in the top level *enable* mode:
2920 .. clicmd:: clear bgp \*
2924 .. clicmd:: clear bgp ipv4|ipv6 \*
2926 Clear all peers with this address-family activated.
2928 .. clicmd:: clear bgp ipv4|ipv6 unicast \*
2930 Clear all peers with this address-family and sub-address-family activated.
2932 .. clicmd:: clear bgp ipv4|ipv6 PEER
2934 Clear peers with address of X.X.X.X and this address-family activated.
2936 .. clicmd:: clear bgp ipv4|ipv6 unicast PEER
2938 Clear peer with address of X.X.X.X and this address-family and sub-address-family activated.
2940 .. clicmd:: clear bgp ipv4|ipv6 PEER soft|in|out
2942 Clear peer using soft reconfiguration in this address-family.
2944 .. clicmd:: clear bgp ipv4|ipv6 unicast PEER soft|in|out
2946 Clear peer using soft reconfiguration in this address-family and sub-address-family.
2948 The following are available in the ``router bgp`` mode:
2950 .. clicmd:: write-quanta (1-64)
2952 BGP message Tx I/O is vectored. This means that multiple packets are written
2953 to the peer socket at the same time each I/O cycle, in order to minimize
2954 system call overhead. This value controls how many are written at a time.
2955 Under certain load conditions, reducing this value could make peer traffic
2956 less 'bursty'. In practice, leave this settings on the default (64) unless
2957 you truly know what you are doing.
2959 .. clicmd:: read-quanta (1-10)
2961 Unlike Tx, BGP Rx traffic is not vectored. Packets are read off the wire one
2962 at a time in a loop. This setting controls how many iterations the loop runs
2963 for. As with write-quanta, it is best to leave this setting on the default.
2965 The following command is available in ``config`` mode as well as in the
2966 ``router bgp`` mode:
2968 .. clicmd:: bgp graceful-shutdown
2970 The purpose of this command is to initiate BGP Graceful Shutdown which
2971 is described in :rfc:`8326`. The use case for this is to minimize or
2972 eliminate the amount of traffic loss in a network when a planned
2973 maintenance activity such as software upgrade or hardware replacement
2974 is to be performed on a router. The feature works by re-announcing
2975 routes to eBGP peers with the GRACEFUL_SHUTDOWN community included.
2976 Peers are then expected to treat such paths with the lowest preference.
2977 This happens automatically on a receiver running FRR; with other
2978 routing protocol stacks, an inbound policy may have to be configured.
2979 In FRR, triggering graceful shutdown also results in announcing a
2980 LOCAL_PREF of 0 to iBGP peers.
2982 Graceful shutdown can be configured per BGP instance or globally for
2983 all of BGP. These two options are mutually exclusive. The no form of
2984 the command causes graceful shutdown to be stopped, and routes will
2985 be re-announced without the GRACEFUL_SHUTDOWN community and/or with
2986 the usual LOCAL_PREF value. Note that if this option is saved to
2987 the startup configuration, graceful shutdown will remain in effect
2988 across restarts of *bgpd* and will need to be explicitly disabled.
2990 .. _bgp-displaying-bgp-information:
2992 Displaying BGP Information
2993 ==========================
2995 The following four commands display the IPv6 and IPv4 routing tables, depending
2996 on whether or not the ``ip`` keyword is used.
2997 Actually, :clicmd:`show ip bgp` command was used on older `Quagga` routing
2998 daemon project, while :clicmd:`show bgp` command is the new format. The choice
2999 has been done to keep old format with IPv4 routing table, while new format
3000 displays IPv6 routing table.
3002 .. clicmd:: show ip bgp [all] [wide|json]
3004 .. clicmd:: show ip bgp A.B.C.D [json]
3006 .. clicmd:: show bgp [all] [wide|json]
3008 .. clicmd:: show bgp X:X::X:X [json]
3010 These commands display BGP routes. When no route is specified, the default
3011 is to display all BGP routes.
3015 BGP table version is 0, local router ID is 10.1.1.1
3016 Status codes: s suppressed, d damped, h history, * valid, > best, i - internal
3017 Origin codes: i - IGP, e - EGP, ? - incomplete
3019 Network Next Hop Metric LocPrf Weight Path
3020 \*> 1.1.1.1/32 0.0.0.0 0 32768 i
3022 Total number of prefixes 1
3024 If ``wide`` option is specified, then the prefix table's width is increased
3025 to fully display the prefix and the nexthop.
3027 This is especially handy dealing with IPv6 prefixes and
3028 if :clicmd:`[no] bgp default show-nexthop-hostname` is enabled.
3030 If ``all`` option is specified, ``ip`` keyword is ignored, show bgp all and
3031 show ip bgp all commands display routes for all AFIs and SAFIs.
3033 If ``json`` option is specified, output is displayed in JSON format.
3035 Some other commands provide additional options for filtering the output.
3037 .. clicmd:: show [ip] bgp regexp LINE
3039 This command displays BGP routes using AS path regular expression
3040 (:ref:`bgp-regular-expressions`).
3042 .. clicmd:: show [ip] bgp [all] summary [wide] [json]
3044 Show a bgp peer summary for the specified address family.
3046 The old command structure :clicmd:`show ip bgp` may be removed in the future
3047 and should no longer be used. In order to reach the other BGP routing tables
3048 other than the IPv6 routing table given by :clicmd:`show bgp`, the new command
3049 structure is extended with :clicmd:`show bgp [afi] [safi]`.
3051 ``wide`` option gives more output like ``LocalAS`` and extended ``Desc`` to
3056 exit1# show ip bgp summary wide
3058 IPv4 Unicast Summary:
3059 BGP router identifier 192.168.100.1, local AS number 65534 vrf-id 0
3061 RIB entries 5, using 920 bytes of memory
3062 Peers 1, using 27 KiB of memory
3064 Neighbor V AS LocalAS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd PfxSnt Desc
3065 192.168.0.2 4 65030 123 15 22 0 0 0 00:07:00 0 1 us-east1-rs1.frrouting.org
3067 Total number of neighbors 1
3070 .. clicmd:: show bgp [afi] [safi] [all] [wide|json]
3072 .. clicmd:: show bgp <ipv4|ipv6> <unicast|multicast|vpn|labeled-unicast>
3074 These commands display BGP routes for the specific routing table indicated by
3075 the selected afi and the selected safi. If no afi and no safi value is given,
3076 the command falls back to the default IPv6 routing table.
3077 For EVPN prefixes, you can display the full BGP table for this AFI/SAFI
3078 using the standard `show bgp [afi] [safi]` syntax.
3080 .. clicmd:: show bgp l2vpn evpn route [type <macip|2|multicast|3|es|4|prefix|5>]
3082 Additionally, you can also filter this output by route type.
3084 .. clicmd:: show bgp [afi] [safi] [all] summary [json]
3086 Show a bgp peer summary for the specified address family, and subsequent
3089 .. clicmd:: show bgp [afi] [safi] [all] summary failed [json]
3091 Show a bgp peer summary for peers that are not succesfully exchanging routes
3092 for the specified address family, and subsequent address-family.
3094 .. clicmd:: show bgp [afi] [safi] [all] summary established [json]
3096 Show a bgp peer summary for peers that are succesfully exchanging routes
3097 for the specified address family, and subsequent address-family.
3099 .. clicmd:: show bgp [afi] [safi] neighbor [PEER]
3101 This command shows information on a specific BGP peer of the relevant
3102 afi and safi selected.
3104 .. clicmd:: show bgp [afi] [safi] [all] dampening dampened-paths [wide|json]
3106 Display paths suppressed due to dampening of the selected afi and safi
3109 .. clicmd:: show bgp [afi] [safi] [all] dampening flap-statistics [wide|json]
3111 Display flap statistics of routes of the selected afi and safi selected.
3113 .. clicmd:: show bgp [afi] [safi] statistics
3115 Display statistics of routes of the selected afi and safi.
3117 .. clicmd:: show bgp statistics-all
3119 Display statistics of routes of all the afi and safi.
3121 .. clicmd:: show [ip] bgp [afi] [safi] [all] cidr-only [wide|json]
3123 Display routes with non-natural netmasks.
3125 .. clicmd:: show [ip] bgp [afi] [safi] [all] neighbors A.B.C.D [advertised-routes|received-routes|filtered-routes] [json|wide]
3127 Display the routes advertised to a BGP neighbor or received routes
3128 from neighbor or filtered routes received from neighbor based on the
3131 If ``wide`` option is specified, then the prefix table's width is increased
3132 to fully display the prefix and the nexthop.
3134 This is especially handy dealing with IPv6 prefixes and
3135 if :clicmd:`[no] bgp default show-nexthop-hostname` is enabled.
3137 If ``all`` option is specified, ``ip`` keyword is ignored and,
3138 routes displayed for all AFIs and SAFIs.
3139 if afi is specified, with ``all`` option, routes will be displayed for
3140 each SAFI in the selcted AFI
3142 If ``json`` option is specified, output is displayed in JSON format.
3144 .. _bgp-display-routes-by-community:
3146 Displaying Routes by Community Attribute
3147 ----------------------------------------
3149 The following commands allow displaying routes based on their community
3152 .. clicmd:: show [ip] bgp <ipv4|ipv6> [all] community [wide|json]
3154 .. clicmd:: show [ip] bgp <ipv4|ipv6> [all] community COMMUNITY [wide|json]
3156 .. clicmd:: show [ip] bgp <ipv4|ipv6> [all] community COMMUNITY exact-match [wide|json]
3158 These commands display BGP routes which have the community attribute.
3159 attribute. When ``COMMUNITY`` is specified, BGP routes that match that
3160 community are displayed. When `exact-match` is specified, it display only
3161 routes that have an exact match.
3163 .. clicmd:: show [ip] bgp <ipv4|ipv6> community-list WORD
3165 .. clicmd:: show [ip] bgp <ipv4|ipv6> community-list WORD exact-match
3167 These commands display BGP routes for the address family specified that
3168 match the specified community list. When `exact-match` is specified, it
3169 displays only routes that have an exact match.
3171 If ``wide`` option is specified, then the prefix table's width is increased
3172 to fully display the prefix and the nexthop.
3174 This is especially handy dealing with IPv6 prefixes and
3175 if :clicmd:`[no] bgp default show-nexthop-hostname` is enabled.
3177 If ``all`` option is specified, ``ip`` keyword is ignored and,
3178 routes displayed for all AFIs and SAFIs.
3179 if afi is specified, with ``all`` option, routes will be displayed for
3180 each SAFI in the selcted AFI
3182 If ``json`` option is specified, output is displayed in JSON format.
3184 .. clicmd:: show bgp labelpool <chunks|inuse|ledger|requests|summary> [json]
3186 These commands display information about the BGP labelpool used for
3187 the association of MPLS labels with routes for L3VPN and Labeled Unicast
3189 If ``chunks`` option is specified, output shows the current list of label
3190 chunks granted to BGP by Zebra, indicating the start and end label in
3193 If ``inuse`` option is specified, output shows the current inuse list of
3194 label to prefix mappings
3196 If ``ledger`` option is specified, output shows ledger list of all
3197 label requests made per prefix
3199 If ``requests`` option is specified, output shows current list of label
3200 requests which have not yet been fulfilled by the labelpool
3202 If ``summary`` option is specified, output is a summary of the counts for
3203 the chunks, inuse, ledger and requests list along with the count of
3204 outstanding chunk requests to Zebra and the nummber of zebra reconnects
3207 If ``json`` option is specified, output is displayed in JSON format.
3209 .. _bgp-display-routes-by-lcommunity:
3211 Displaying Routes by Large Community Attribute
3212 ----------------------------------------------
3214 The following commands allow displaying routes based on their
3215 large community attribute.
3217 .. clicmd:: show [ip] bgp <ipv4|ipv6> large-community
3219 .. clicmd:: show [ip] bgp <ipv4|ipv6> large-community LARGE-COMMUNITY
3221 .. clicmd:: show [ip] bgp <ipv4|ipv6> large-community LARGE-COMMUNITY exact-match
3223 .. clicmd:: show [ip] bgp <ipv4|ipv6> large-community LARGE-COMMUNITY json
3225 These commands display BGP routes which have the large community attribute.
3226 attribute. When ``LARGE-COMMUNITY`` is specified, BGP routes that match that
3227 large community are displayed. When `exact-match` is specified, it display
3228 only routes that have an exact match. When `json` is specified, it display
3229 routes in json format.
3231 .. clicmd:: show [ip] bgp <ipv4|ipv6> large-community-list WORD
3233 .. clicmd:: show [ip] bgp <ipv4|ipv6> large-community-list WORD exact-match
3235 .. clicmd:: show [ip] bgp <ipv4|ipv6> large-community-list WORD json
3237 These commands display BGP routes for the address family specified that
3238 match the specified large community list. When `exact-match` is specified,
3239 it displays only routes that have an exact match. When `json` is specified,
3240 it display routes in json format.
3242 .. _bgp-display-routes-by-as-path:
3245 Displaying Routes by AS Path
3246 ----------------------------
3248 .. clicmd:: show bgp ipv4|ipv6 regexp LINE
3250 This commands displays BGP routes that matches a regular
3251 expression `line` (:ref:`bgp-regular-expressions`).
3253 .. clicmd:: show [ip] bgp ipv4 vpn
3255 .. clicmd:: show [ip] bgp ipv6 vpn
3257 Print active IPV4 or IPV6 routes advertised via the VPN SAFI.
3259 .. clicmd:: show bgp ipv4 vpn summary
3261 .. clicmd:: show bgp ipv6 vpn summary
3263 Print a summary of neighbor connections for the specified AFI/SAFI combination.
3265 Displaying Update Group Information
3266 -----------------------------------
3268 .. clicmd:: show bgp update-groups [advertise-queue|advertised-routes|packet-queue]
3270 Display Information about each individual update-group being used.
3271 If SUBGROUP-ID is specified only display about that particular group. If
3272 advertise-queue is specified the list of routes that need to be sent
3273 to the peers in the update-group is displayed, advertised-routes means
3274 the list of routes we have sent to the peers in the update-group and
3275 packet-queue specifies the list of packets in the queue to be sent.
3277 .. clicmd:: show bgp update-groups statistics
3279 Display Information about update-group events in FRR.
3281 .. _bgp-route-reflector:
3286 BGP routers connected inside the same AS through BGP belong to an internal
3287 BGP session, or IBGP. In order to prevent routing table loops, IBGP does not
3288 advertise IBGP-learned routes to other routers in the same session. As such,
3289 IBGP requires a full mesh of all peers. For large networks, this quickly becomes
3290 unscalable. Introducing route reflectors removes the need for the full-mesh.
3292 When route reflectors are configured, these will reflect the routes announced
3293 by the peers configured as clients. A route reflector client is configured
3296 .. clicmd:: neighbor PEER route-reflector-client
3299 To avoid single points of failure, multiple route reflectors can be configured.
3301 A cluster is a collection of route reflectors and their clients, and is used
3302 by route reflectors to avoid looping.
3304 .. clicmd:: bgp cluster-id A.B.C.D
3306 .. clicmd:: bgp no-rib
3308 To set and unset the BGP daemon ``-n`` / ``--no_kernel`` options during runtime
3309 to disable BGP route installation to the RIB (Zebra), the ``[no] bgp no-rib``
3310 commands can be used;
3312 Please note that setting the option during runtime will withdraw all routes in
3313 the daemons RIB from Zebra and unsetting it will announce all routes in the
3314 daemons RIB to Zebra. If the option is passed as a command line argument when
3315 starting the daemon and the configuration gets saved, the option will persist
3316 unless removed from the configuration with the negating command prior to the
3317 configuration write operation.
3319 .. clicmd:: bgp send-extra-data zebra
3321 This Command turns off the ability of BGP to send extra data to zebra.
3322 In this case it's the AS-Path being used for the path. The default behavior
3323 in BGP is to send this data and to turn it off enter the no form of the command.
3324 If extra data was sent to zebra, and this command is turned on there is no
3325 effort to clean up this data in the rib.
3327 .. _bgp-suppress-fib:
3329 Suppressing routes not installed in FIB
3330 =======================================
3332 The FRR implementation of BGP advertises prefixes learnt from a peer to other
3333 peers even if the routes do not get installed in the FIB. There can be
3334 scenarios where the hardware tables in some of the routers (along the path from
3335 the source to destination) is full which will result in all routes not getting
3336 installed in the FIB. If these routes are advertised to the downstream routers
3337 then traffic will start flowing and will be dropped at the intermediate router.
3339 The solution is to provide a configurable option to check for the FIB install
3340 status of the prefixes and advertise to peers if the prefixes are successfully
3341 installed in the FIB. The advertisement of the prefixes are suppressed if it is
3342 not installed in FIB.
3344 The following conditions apply will apply when checking for route installation
3347 1. The advertisement or suppression of routes based on FIB install status
3348 applies only for newly learnt routes from peer (routes which are not in
3350 2. If the route received from peer already exists in BGP local RIB and route
3351 attributes have changed (best path changed), the old path is deleted and
3352 new path is installed in FIB. The FIB install status will not have any
3353 effect. Therefore only when the route is received first time the checks
3355 3. The feature will not apply for routes learnt through other means like
3356 redistribution to bgp from other protocols. This is applicable only to
3358 4. If a route is installed in FIB and then gets deleted from the dataplane,
3359 then routes will not be withdrawn from peers. This will be considered as
3361 5. The feature will slightly increase the time required to advertise the routes
3362 to peers since the route install status needs to be received from the FIB
3363 6. If routes are received by the peer before the configuration is applied, then
3364 the bgp sessions need to be reset for the configuration to take effect.
3365 7. If the route which is already installed in dataplane is removed for some
3366 reason, sending withdraw message to peers is not currently supported.
3368 .. clicmd:: bgp suppress-fib-pending
3370 This command is applicable at the global level and at an individual
3371 bgp level. If applied at the global level all bgp instances will
3372 wait for fib installation before announcing routes and there is no
3373 way to turn it off for a particular bgp vrf.
3380 You can set different routing policy for a peer. For example, you can set
3381 different filter for a peer.
3387 neighbor 10.0.0.1 remote-as 2
3388 address-family ipv4 unicast
3389 neighbor 10.0.0.1 distribute-list 1 in
3393 neighbor 10.0.0.1 remote-as 2
3394 address-family ipv4 unicast
3395 neighbor 10.0.0.1 distribute-list 2 in
3398 This means BGP update from a peer 10.0.0.1 goes to both BGP view 1 and view 2.
3399 When the update is inserted into view 1, distribute-list 1 is applied. On the
3400 other hand, when the update is inserted into view 2, distribute-list 2 is
3404 .. _bgp-regular-expressions:
3406 BGP Regular Expressions
3407 =======================
3409 BGP regular expressions are based on :t:`POSIX 1003.2` regular expressions. The
3410 following description is just a quick subset of the POSIX regular expressions.
3414 Matches any single character.
3417 Matches 0 or more occurrences of pattern.
3420 Matches 1 or more occurrences of pattern.
3423 Match 0 or 1 occurrences of pattern.
3426 Matches the beginning of the line.
3429 Matches the end of the line.
3432 The ``_`` character has special meanings in BGP regular expressions. It
3433 matches to space and comma , and AS set delimiter ``{`` and ``}`` and AS
3434 confederation delimiter ``(`` and ``)``. And it also matches to the
3435 beginning of the line and the end of the line. So ``_`` can be used for AS
3436 value boundaries match. This character technically evaluates to
3440 .. _bgp-configuration-examples:
3442 Miscellaneous Configuration Examples
3443 ====================================
3445 Example of a session to an upstream, advertising only one prefix to it.
3450 bgp router-id 10.236.87.1
3451 neighbor upstream peer-group
3452 neighbor upstream remote-as 64515
3453 neighbor upstream capability dynamic
3454 neighbor 10.1.1.1 peer-group upstream
3455 neighbor 10.1.1.1 description ACME ISP
3457 address-family ipv4 unicast
3458 network 10.236.87.0/24
3459 neighbor upstream prefix-list pl-allowed-adv out
3462 ip prefix-list pl-allowed-adv seq 5 permit 82.195.133.0/25
3463 ip prefix-list pl-allowed-adv seq 10 deny any
3465 A more complex example including upstream, peer and customer sessions
3466 advertising global prefixes and NO_EXPORT prefixes and providing actions for
3467 customer routes based on community values. Extensive use is made of route-maps
3468 and the 'call' feature to support selective advertising of prefixes. This
3469 example is intended as guidance only, it has NOT been tested and almost
3470 certainly contains silly mistakes, if not serious flaws.
3475 bgp router-id 10.236.87.1
3476 neighbor upstream capability dynamic
3477 neighbor cust capability dynamic
3478 neighbor peer capability dynamic
3479 neighbor 10.1.1.1 remote-as 64515
3480 neighbor 10.1.1.1 peer-group upstream
3481 neighbor 10.2.1.1 remote-as 64516
3482 neighbor 10.2.1.1 peer-group upstream
3483 neighbor 10.3.1.1 remote-as 64517
3484 neighbor 10.3.1.1 peer-group cust-default
3485 neighbor 10.3.1.1 description customer1
3486 neighbor 10.4.1.1 remote-as 64518
3487 neighbor 10.4.1.1 peer-group cust
3488 neighbor 10.4.1.1 description customer2
3489 neighbor 10.5.1.1 remote-as 64519
3490 neighbor 10.5.1.1 peer-group peer
3491 neighbor 10.5.1.1 description peer AS 1
3492 neighbor 10.6.1.1 remote-as 64520
3493 neighbor 10.6.1.1 peer-group peer
3494 neighbor 10.6.1.1 description peer AS 2
3496 address-family ipv4 unicast
3497 network 10.123.456.0/24
3498 network 10.123.456.128/25 route-map rm-no-export
3499 neighbor upstream route-map rm-upstream-out out
3500 neighbor cust route-map rm-cust-in in
3501 neighbor cust route-map rm-cust-out out
3502 neighbor cust send-community both
3503 neighbor peer route-map rm-peer-in in
3504 neighbor peer route-map rm-peer-out out
3505 neighbor peer send-community both
3506 neighbor 10.3.1.1 prefix-list pl-cust1-network in
3507 neighbor 10.4.1.1 prefix-list pl-cust2-network in
3508 neighbor 10.5.1.1 prefix-list pl-peer1-network in
3509 neighbor 10.6.1.1 prefix-list pl-peer2-network in
3512 ip prefix-list pl-default permit 0.0.0.0/0
3514 ip prefix-list pl-upstream-peers permit 10.1.1.1/32
3515 ip prefix-list pl-upstream-peers permit 10.2.1.1/32
3517 ip prefix-list pl-cust1-network permit 10.3.1.0/24
3518 ip prefix-list pl-cust1-network permit 10.3.2.0/24
3520 ip prefix-list pl-cust2-network permit 10.4.1.0/24
3522 ip prefix-list pl-peer1-network permit 10.5.1.0/24
3523 ip prefix-list pl-peer1-network permit 10.5.2.0/24
3524 ip prefix-list pl-peer1-network permit 192.168.0.0/24
3526 ip prefix-list pl-peer2-network permit 10.6.1.0/24
3527 ip prefix-list pl-peer2-network permit 10.6.2.0/24
3528 ip prefix-list pl-peer2-network permit 192.168.1.0/24
3529 ip prefix-list pl-peer2-network permit 192.168.2.0/24
3530 ip prefix-list pl-peer2-network permit 172.16.1/24
3532 bgp as-path access-list asp-own-as permit ^$
3533 bgp as-path access-list asp-own-as permit _64512_
3535 ! #################################################################
3536 ! Match communities we provide actions for, on routes receives from
3537 ! customers. Communities values of <our-ASN>:X, with X, have actions:
3539 ! 100 - blackhole the prefix
3540 ! 200 - set no_export
3541 ! 300 - advertise only to other customers
3542 ! 400 - advertise only to upstreams
3543 ! 500 - set no_export when advertising to upstreams
3544 ! 2X00 - set local_preference to X00
3546 ! blackhole the prefix of the route
3547 bgp community-list standard cm-blackhole permit 64512:100
3549 ! set no-export community before advertising
3550 bgp community-list standard cm-set-no-export permit 64512:200
3552 ! advertise only to other customers
3553 bgp community-list standard cm-cust-only permit 64512:300
3555 ! advertise only to upstreams
3556 bgp community-list standard cm-upstream-only permit 64512:400
3558 ! advertise to upstreams with no-export
3559 bgp community-list standard cm-upstream-noexport permit 64512:500
3561 ! set local-pref to least significant 3 digits of the community
3562 bgp community-list standard cm-prefmod-100 permit 64512:2100
3563 bgp community-list standard cm-prefmod-200 permit 64512:2200
3564 bgp community-list standard cm-prefmod-300 permit 64512:2300
3565 bgp community-list standard cm-prefmod-400 permit 64512:2400
3566 bgp community-list expanded cme-prefmod-range permit 64512:2...
3568 ! Informational communities
3570 ! 3000 - learned from upstream
3571 ! 3100 - learned from customer
3572 ! 3200 - learned from peer
3574 bgp community-list standard cm-learnt-upstream permit 64512:3000
3575 bgp community-list standard cm-learnt-cust permit 64512:3100
3576 bgp community-list standard cm-learnt-peer permit 64512:3200
3578 ! ###################################################################
3579 ! Utility route-maps
3581 ! These utility route-maps generally should not used to permit/deny
3582 ! routes, i.e. they do not have meaning as filters, and hence probably
3583 ! should be used with 'on-match next'. These all finish with an empty
3584 ! permit entry so as not interfere with processing in the caller.
3586 route-map rm-no-export permit 10
3587 set community additive no-export
3588 route-map rm-no-export permit 20
3590 route-map rm-blackhole permit 10
3591 description blackhole, up-pref and ensure it cannot escape this AS
3592 set ip next-hop 127.0.0.1
3593 set local-preference 10
3594 set community additive no-export
3595 route-map rm-blackhole permit 20
3597 ! Set local-pref as requested
3598 route-map rm-prefmod permit 10
3599 match community cm-prefmod-100
3600 set local-preference 100
3601 route-map rm-prefmod permit 20
3602 match community cm-prefmod-200
3603 set local-preference 200
3604 route-map rm-prefmod permit 30
3605 match community cm-prefmod-300
3606 set local-preference 300
3607 route-map rm-prefmod permit 40
3608 match community cm-prefmod-400
3609 set local-preference 400
3610 route-map rm-prefmod permit 50
3612 ! Community actions to take on receipt of route.
3613 route-map rm-community-in permit 10
3614 description check for blackholing, no point continuing if it matches.
3615 match community cm-blackhole
3617 route-map rm-community-in permit 20
3618 match community cm-set-no-export
3621 route-map rm-community-in permit 30
3622 match community cme-prefmod-range
3624 route-map rm-community-in permit 40
3626 ! #####################################################################
3627 ! Community actions to take when advertising a route.
3628 ! These are filtering route-maps,
3630 ! Deny customer routes to upstream with cust-only set.
3631 route-map rm-community-filt-to-upstream deny 10
3632 match community cm-learnt-cust
3633 match community cm-cust-only
3634 route-map rm-community-filt-to-upstream permit 20
3636 ! Deny customer routes to other customers with upstream-only set.
3637 route-map rm-community-filt-to-cust deny 10
3638 match community cm-learnt-cust
3639 match community cm-upstream-only
3640 route-map rm-community-filt-to-cust permit 20
3642 ! ###################################################################
3643 ! The top-level route-maps applied to sessions. Further entries could
3644 ! be added obviously..
3647 route-map rm-cust-in permit 10
3648 call rm-community-in
3650 route-map rm-cust-in permit 20
3651 set community additive 64512:3100
3652 route-map rm-cust-in permit 30
3654 route-map rm-cust-out permit 10
3655 call rm-community-filt-to-cust
3657 route-map rm-cust-out permit 20
3659 ! Upstream transit ASes
3660 route-map rm-upstream-out permit 10
3661 description filter customer prefixes which are marked cust-only
3662 call rm-community-filt-to-upstream
3664 route-map rm-upstream-out permit 20
3665 description only customer routes are provided to upstreams/peers
3666 match community cm-learnt-cust
3669 ! outbound policy is same as for upstream
3670 route-map rm-peer-out permit 10
3671 call rm-upstream-out
3673 route-map rm-peer-in permit 10
3674 set community additive 64512:3200
3677 Example of how to set up a 6-Bone connection.
3681 ! bgpd configuration
3682 ! ==================
3684 ! MP-BGP configuration
3687 bgp router-id 10.0.0.1
3688 neighbor 3ffe:1cfa:0:2:2a0:c9ff:fe9e:f56 remote-as `as-number`
3691 network 3ffe:506::/32
3692 neighbor 3ffe:1cfa:0:2:2a0:c9ff:fe9e:f56 activate
3693 neighbor 3ffe:1cfa:0:2:2a0:c9ff:fe9e:f56 route-map set-nexthop out
3694 neighbor 3ffe:1cfa:0:2:2c0:4fff:fe68:a231 remote-as `as-number`
3695 neighbor 3ffe:1cfa:0:2:2c0:4fff:fe68:a231 route-map set-nexthop out
3698 ipv6 access-list all permit any
3700 ! Set output nexthop address.
3702 route-map set-nexthop permit 10
3703 match ipv6 address all
3704 set ipv6 nexthop global 3ffe:1cfa:0:2:2c0:4fff:fe68:a225
3705 set ipv6 nexthop local fe80::2c0:4fff:fe68:a225
3711 .. include:: routeserver.rst
3713 .. include:: rpki.rst
3715 .. include:: wecmp_linkbw.rst
3717 .. include:: flowspec.rst
3719 .. [#med-transitivity-rant] For some set of objects to have an order, there *must* be some binary ordering relation that is defined for *every* combination of those objects, and that relation *must* be transitive. I.e.:, if the relation operator is <, and if a < b and b < c then that relation must carry over and it *must* be that a < c for the objects to have an order. The ordering relation may allow for equality, i.e. a < b and b < a may both be true and imply that a and b are equal in the order and not distinguished by it, in which case the set has a partial order. Otherwise, if there is an order, all the objects have a distinct place in the order and the set has a total order)
3720 .. [bgp-route-osci-cond] McPherson, D. and Gill, V. and Walton, D., "Border Gateway Protocol (BGP) Persistent Route Oscillation Condition", IETF RFC3345
3721 .. [stable-flexible-ibgp] Flavel, A. and M. Roughan, "Stable and flexible iBGP", ACM SIGCOMM 2009
3722 .. [ibgp-correctness] Griffin, T. and G. Wilfong, "On the correctness of IBGP configuration", ACM SIGCOMM 2002