Overview

l2tpns is half of a complete L2TP implementation. It supports only the LNS side of the connection.

L2TP (Layer 2 Tunneling Protocol) is designed to allow any layer 2 protocol (e.g. Ethernet, PPP) to be tunneled over an IP connection. l2tpns implements PPP over L2TP only.

There are a couple of other L2TP implementations, of which l2tpd is probably the most popular. l2tpd also will handle being either end of a tunnel, and is a lot more configurable than l2tpns. However, due to the way it works, it is nowhere near as scalable.

l2tpns uses the TUN/TAP interface provided by the Linux kernel to receive and send packets. Using some packet manipulation it doesn't require a single interface per connection, as l2tpd does.

This allows it to scale extremely well to very high loads and very high numbers of connections.

It also has a plugin architecture which allows custom code to be run during processing. An example of this is in the walled garden module included.

Installation

Requirements

• Linux kernel version 2.4 or above, with the Tun/Tap interface either compiled in, or as a module.

• libcli 1.8.5 or greater. You can get this from SourceForge

Compiling

You can generally get away with just running make from the source directory. This will compile the daemon, associated tools and any modules shipped with the distribution.

Installing

After you have successfully compiled everything, run make install to install it. By default, the binaries are installed into /usr/sbin, the configuration into /etc/l2tpns, and the modules into /usr/lib/l2tpns.

You will definately need to edit the configuration files before you start. See Configuration for more information.

Running

You only need to run /usr/sbin/l2tpns as root to start it. It does not normally detach to a daemon process (see the -d option), so you should perhaps run it from init.

By default there is no log destination set, so all log messages will go to stdout.

Configuration

All configuration of the software is done from the files installed into /etc/l2tpns.

startup-config

This is the main configuration file for l2tpns. The format of the file is a list of commands that can be run through the command-line interface. This file can also be written directly by the l2tpns process if a user runs the write memory command, so any comments will be lost. However if your policy is not to write the config by the program, then feel free to comment the file with a # or ! at the beginning of the line.

A list of the possible configuration directives follows. Each of these should be set by a line like: set configstring "value" set ipaddress 192.168.1.1 set boolean true

debug (int)

Sets the level of messages that will be written to the log file. The value should be between 0 and 5, with 0 being no debugging, and 5 being the highest. A rough description of the levels is:

0: Critical Errors

Things are probably broken

1: Errors

Things might have gone wrong, but probably will recover

2: Warnings

Just in case you care what is not quite perfect

3: Information

Parameters of control packets

4: Calls

For tracing the execution of the code

5: Packets

Everything, including a hex dump of all packets processed... probably twice

Note that the higher you set the debugging level, the slower the program will run. Also, at level 5 a lot of information will be logged. This should only ever be used for working out why it doesn't work at all.

log_file (string)

This will be where all logging and debugging information is written to. This may be either a filename, such as /var/log/l2tpns, or the special magic string syslog:facility, where facility is any one of the syslog logging facilities, such as local5.

pid_file (string)

If set, the process id will be written to the specified file. The value must be an absolute path.

random_device (string)

Path to random data source (default /dev/urandom). Use "" to use the rand() library function.

l2tp_secret (string)

The secret used by l2tpns for authenticating tunnel request. Must be the same as the LAC, or authentication will fail. Only actually be used if the LAC requests authentication.

l2tp_mtu (int)

MTU of interface for L2TP traffic (default: 1500). Used to set link MRU and adjust TCP MSS.

mp_mrru (int)

MRRU for MP traffic (default: 1614). Can be set to 0 to disable MP negociation.

ppp_restart_time (int); ppp_max_configure (int); ppp_max_failure (int)

PPP counter and timer values, as described in §4.1 of RFC1661.

lcp_renegotiation (string)

By default (always), we renegotiate LCP even if the LAC already did with the client.

We can avoid the LCP renegotiation (proxy LCP negotiation), unless what was already negotiated is not fine for us (on-mismatch). To be noted: we will accept not using our preferred RADIUS authentication method. This is notably useful when the LAC cannot pass LCP configuration through, and thus we have to try to accept what was already negotiated by the LAC with the client.

primary_dns (ip address); econdary_dns (ip address)

Whenever a PPP connection is established, DNS servers will be sent to the user, both a primary and a secondary. If either is set to 0.0.0.0, then that one will not be sent.

primary_radius (ip address); secondary_radius (ip address)

Sets the RADIUS servers used for both authentication and accounting. If the primary server does not respond, then the secondary RADIUS server will be tried.

In addition to the source IP address and identifier, the RADIUS server must include the source port when detecting duplicates to suppress (in order to cope with a large number of sessions coming on-line simultaneously l2tpns uses a set of udp sockets, each with a separate identifier).

primary_radius_port (short); secondary_radius_port (short)

Sets the authentication ports for the primary and secondary RADIUS servers. The accounting port is one more than the authentication port. If no RADIUS ports are given, the authentication port defaults to 1812, and the accounting port to 1813.

radius_accounting (boolean)

If set to true, then RADIUS accounting packets will be sent. This means that a Start record will be sent when the session is successfully authenticated, and a Stop record will be sent when the session is closed.

radius_interim (int)

If radius_accounting is on, defines the interval between sending of RADIUS interim accounting records (in seconds).

radius_secret (string)

This secret will be used in all RADIUS queries. If this is not set then RADIUS queries will fail.

radius_require_message_authenticator (string)

If set to true, RADIUS answers to AccessRequests will have to contain a valid MessageAuthenticator. If set to auto (default), if the first RADIUS answer to AccessRequests contains a valid MessageAuthenticator, subsequent answers will have to contain one. If set to no (not recommended), RADIUS answers to AccessRequests do not have to contain a valid MessageAuthenticator. It is advised to set this to true after checking that your RADIUS server does send MessageAuthenticator.

radius_authtypes (string)

A comma separated list of supported RADIUS authentication methods (pap or chap), in order of preference (default pap).

radius_bind_min (short); radius_bind_max (short)

Define a port range in which to bind sockets used to send and receive RADIUS packets. Must be at least RADIUS_FDS (64) wide. Simplifies firewalling of RADIUS ports (default: dynamically assigned).

radius_dae_port (short)

Port for DAE RADIUS (Packet of Death/Disconnect, Change of Authorization) requests (default: 3799).

allow_duplicate_users (boolean)

Allow multiple logins with the same username. If false (the default), any prior session with the same username will be dropped when a new session is established.

guest_account (string)

Allow multiple logins matching this specific username.

bind_address (ip address)

When the tun interface is created, it is assigned the address specified here. If no address is given, 1.1.1.1 is used. Packets containing user traffic should be routed via this address if given, otherwise the primary address of the machine.

peer_address (ip address)

Address to send to clients as the default gateway.

route_protocol (short)

Protocol number to record when adding a route (see /usr/share/iproute2/rt_protos). This allows to separate routes added by several l2tpns instances. The default is 42.

route_metric (int)

Metric of the route. Routes with lower values are preferred. This should not be changed during execution, since l2tpns would then not recognize the routes it added itself. The default is 1.

ipv6_prefix (ipv6 address)

Enable negotiation of IPv6. This forms the the first 64 bits of the client allocated address. The remaining 64 come from the allocated IPv4 address and 4 bytes of 0s.

send_garp (boolean)

Determines whether or not to send a gratuitous ARP for the bind_address when the server is ready to handle traffic (default: true). This value is ignored if BGP is configured.

throttle_speed (int)

Sets the default speed (in kbits/s) which sessions will be limited to. If this is set to 0, then throttling will not be used at all. Note: You can set this by the CLI, but changes will not affect currently connected users.

throttle_buckets (int)

Number of token buckets to allocate for throttling. Each throttled session requires two buckets (in and out).

accounting_dir (string)

If set to a directory, then every 5 minutes the current usage for every connected use will be dumped to a file in this directory. Each file dumped begins with a header, where each line is prefixed by #. Following the header is a single line for every connected user, fields separated by a space.

The fields are username, ip, qos, uptxoctets, downrxoctets. The qos field is 1 if a standard user, and 2 if the user is throttled.

dump_speed (boolean)

If set to true, then the current bandwidth utilization will be logged every second. Even if this is disabled, you can see this information by running the uptime command on the CLI.

multi_read_count (int)

Number of packets to read off each of the UDP and TUN fds when returned as readable by select (default: 10). Avoids incurring the unnecessary system call overhead of select on busy servers.

scheduler_fifo (boolean)

Sets the scheduling policy for the l2tpns process to SCHED_FIFO. This causes the kernel to immediately preempt any currently running SCHED_OTHER (normal) process in favour of l2tpns when it becomes runnable. Ignored on uniprocessor systems.

lock_pages (boolean)

Keep all pages mapped by the l2tpns process in memory.

icmp_rate (int)

Maximum number of host unreachable ICMP packets to send per second.

packet_limit (int)

Maximum number of packets of downstream traffic to be handled each tenth of a second per session. If zero, no limit is applied (default: 0). Intended as a DoS prevention mechanism and not a general throttling control (packets are dropped, not queued).

cluster_address (ip address)

Multicast cluster address (default: 239.192.13.13). See Clustering for more information.

cluster_port (udp port)

UDP cluster port (default: 32792). See Clustering for more information.

cluster_interface (string)

Interface for cluster packets (default: none). An IPv4 address can also be specified, e.g. to run several l2tpns on the same host (but beware of enabling kernel acceleration only on one instance, to set a different bind_portremotelns, and to set a different route_metric. It is also useful to set a different route_protocol number)

cluster_mcast_ttl (int)

TTL for multicast packets (default: 1).

cluster_hb_interval (int)

Interval in tenths of a second between cluster heartbeat/pings.

cluster_hb_timeout (int)

Cluster heartbeat timeout in tenths of a second. A new master will be elected when this interval has been passed without seeing a heartbeat from the master.

cluster_master_min_adv (int)

Determines the minimum number of up to date slaves required before the master will drop routes (default: 1).

BGP

BGP routing configuration is entered by the command: router bgp as where as specifies the local AS number.

Subsequent lines prefixed with neighbour peer define the attributes of BGP neighhbours. Valid commands are: neighbour peer remote-as as neighbour peer timers keepalive hold

Where peer specifies the BGP neighbour as either a hostname or IP address, as is the remote AS number and keepalive, hold are the timer values in seconds.

Access Lists

Named access-lists are configured using one of the commands: ip access-list standard name ip access-list extended name

Subsequent lines prefixed with permit or deny define the body of the access-list. Standard access-list syntax:

{permit|deny} {host|source source-wildcard|any} [{host|destination destination-wildcard|any}]

Extended access-lists:

{permit|deny} ip {host|source source-wildcard|any} {host|destination destination-wildcard|any} [fragments]

{permit|deny} udp {host|source source-wildcard|any} [{eq|neq|gt|lt} port|range from to] {host|destination destination-wildcard|any} [{eq|neq|gt|lt} port|range from to] [fragments]

{permit|deny} tcp {host|source source-wildcard|any} [{eq|neq|gt|lt} port|range from to] {host|destination destination-wildcard|any} [{eq|neq|gt|lt} port|range from to] [{established|{match-any|match-all} {+|-}{fin|syn|rst|psh|ack|urg} ...|fragments]

users

Usernames and passwords for the command-line interface are stored in this file. The format is username:password where password may either by plain text, an MD5 digest (prefixed by $1salt$) or a DES password, distinguished from plain text by the prefix {crypt}.

The username enable has a special meaning and is used to set the enable password.

ip_pool

This file is used to configure the IP address pool which user addresses are assigned from. This file should contain either an IP address or a CIDR network per line. e.g.:

192.168.1.1
192.168.1.2
192.168.1.3
192.168.4.0/24
172.16.0.0/16
10.0.0.0/8

Keep in mind that l2tpns can only handle 65535 connections per process, so don't put more than 65535 IP addresses in the configuration file. They will be wasted.

build-garden

The garden plugin on startup creates a NAT table called "garden" then sources the build-garden script to populate that table. All packets from gardened users will be sent through this table. Example:

iptables -t nat -A garden -p tcp -m tcp --dport 25 -j DNAT --to 192.168.1.1
iptables -t nat -A garden -p udp -m udp --dport 53 -j DNAT --to 192.168.1.1
iptables -t nat -A garden -p tcp -m tcp --dport 53 -j DNAT --to 192.168.1.1
iptables -t nat -A garden -p tcp -m tcp --dport 80 -j DNAT --to 192.168.1.1
iptables -t nat -A garden -p tcp -m tcp --dport 110 -j DNAT --to 192.168.1.1
iptables -t nat -A garden -p tcp -m tcp --dport 443 -j DNAT --to 192.168.1.1
iptables -t nat -A garden -p icmp -m icmp --icmp-type echo-request -j DNAT --to 192.168.1.1
iptables -t nat -A garden -p icmp -j ACCEPT
iptables -t nat -A garden -j DROP

Operation

A running l2tpns process can be controlled in a number of ways. The primary method of control is by the Command-Line Interface (CLI).

You can also remotely send commands to modules via the nsctl client provided.

There are also a number of signals that l2tpns understands and takes action when it receives them.

Command-Line Interface

You can access the command line interface by telneting to port 23. There is no IP address restriction, so it's a good idea to firewall this port off from anyone who doesn't need access to it. See for information on restricting access based on a username and password.

The CLI gives you real-time control over almost everything in the process. The interface is designed to look like a Cisco device, and supports things like command history, line editing and context sensitive help. This is provided by linking with the libcli library. Some general documentation of the interface is here.

After you have connected to the telnet port (and perhaps logged in), you will be presented with a hostname> prompt.

Enter help to get a list of possible commands, or press ? for context-specific help.

A brief overview of the more important commands follows:

show session [ID]

: Detailed information for a specific session is presented if you specify a session ID argument.

If no ID is given, a summary of all connected sessions is produced.
Note that this summary list can be around 185 columns wide, so you
should probably use a wide terminal.

The columns listed in the summary are:

  -------------- -------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------
  `SID`          Session ID                                                                                                     
  `TID`          Tunnel ID                                                                                                      See also the [show tunnel](#operation-cli-show-tunnel) CLI command.
  `Username`     The username given in the PPP authentication.                                                                  If this is \*, then LCP authentication has not completed.
  `IP`           The IP address given to the session.                                                                           If this is 0.0.0.0, IPCP negotiation has not completed
  `I`            Intercept                                                                                                      Y or N: indicates whether the session is being snooped. See also the [snoop](#operation-cli-snoop) CLI command.
  `T`            Throttled                                                                                                      Y or N: indicates whether the session is currently throttled. See also the [throttle](#operation-cli-throttle) CLI command.
  `G`            Walled Garden                                                                                                  Y or N: indicates whether the user is trapped in the walled garden. This field is present even if the garden module is not loaded.
  `6`            IPv6                                                                                                           Y or N: indicates whether the session has IPv6 active (IPV6CP open)
  `opened`       The number of seconds since the session started                                                                
  `downloaded`   Number of bytes downloaded by the user                                                                         
  `uploaded`     Number of bytes uploaded by the user                                                                           
  `idle`         The number of seconds since traffic was detected on the session                                                
  `LAC`          The IP address of the LAC the session is connected to.                                                         
  `CLI`          The Calling-Line-Identification field provided during the session setup. This field is generated by the LAC.   
  -------------- -------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------

show users; show user username

: With no arguments, display a list of currently connected users. If an argument is given, the session details for the given username are displayed.

show tunnel [ID]

Produce a summary list of all open tunnels, or detail on a specific tunnel ID.

The columns listed in the summary are:

TID	Tunnel ID
Hostname	The hostname for the tunnel as provided by the LAC. This has no relation to DNS, it is just a text field.
IP	The IP address of the LAC
State	Tunnel state: Free, Open, Dieing, Opening
Sessions	The number of open sessions on the tunnel

show pool

Displays the current IP address pool allocation. This will only display addresses that are in use, or are reserved for re-allocation to a disconnected user.

If an address is not currently in use, but has been used, then in the User column the username will be shown in square brackets, followed by the time since the address was used:

IP Address      Used  Session User
192.168.100.6     N           [joe.user] 1548s

show radius

Show a summary of the in-use RADIUS sessions. This list should not be very long, as RADIUS sessions should be cleaned up as soon as they are used. The columns listed are:

Radius	The ID of the RADIUS request. This is sent in the packet to the RADIUS server for identification
State	The state of the request: WAIT, CHAP, AUTH, IPCP, START, STOP or NULL
Session	The session ID that this RADIUS request is associated with
Retry	If a response does not appear to the request, it will retry at this time. This is a Unix timestamp
Try	Retry count. The RADIUS request is discarded after 3 retries

show running-config

This will list the current running configuration. This is in a format that can either be pasted into the configuration file, or run directly at the command line.

show counters

Internally, counters are kept of key values, such as bytes and packets transferred, as well as function call counters. This function displays all these counters, and is probably only useful for debugging.

You can reset these counters by running clear counters.

show cluster

Show cluster status. Shows the cluster state for this server (Master/Slave), information about known peers and (for slaves) the master IP address, last packet seen and up-to-date status. See Clustering for more information.

write memory

This will write the current running configuration to the config file startup-config, which will be run on a restart.

snoop user IP port

: You must specify a username, IP address and port. All packets for the current session for that username will be forwarded to the given host/port. Specify no snoop username to disable interception for the session.

If you want interception to be permanent, you will have to modify
the RADIUS response for the user. See [Interception](#interception).

throttle user [in|out] rate

: You must specify a username, which will be throttled for the current session to rate Kbps. Prefix rate with in or out to set different upstream and downstream rates.

Specify `no throttle
        username` to disable throttling for the current session.

If you want throttling to be permanent, you will have to modify the
RADIUS response for the user. See [Throttling](#throttling).

drop session

This will cleanly disconnect the session specified by session ID.

drop tunnel

This will cleanly disconnect the tunnel specified by tunnel ID, as well as all sessions on that tunnel.

uptime

This will show how long the l2tpns process has been running, and the current bandwidth utilization:

17:10:35 up 8 days, 2212 users, load average: 0.21, 0.17, 0.16
Bandwidth: UDP-ETH:6/6  ETH-UDP:13/13  TOTAL:37.6   IN:3033 OUT:2569

The bandwidth line contains 4 sets of values:

UDP-ETH	The current bandwidth going from the LAC to the ethernet (user uploads), in mbits/sec.
ETH-UDP	The current bandwidth going from ethernet to the LAC (user downloads).
TOTAL	The total aggregate bandwidth in mbits/s.
IN and OUT	Packets/per-second going between UDP-ETH and ETH-UDP.

These counters are updated every second.

configure terminal

Enter configuration mode. Use exit or ^Z to exit this mode.

The following commands are valid in this mode:

load plugin name

: Load a plugin. You must specify the plugin name, and it will search in /usr/lib/l2tpns for name.so. You can unload a loaded plugin with remove plugin name.

set ...

Set a configuration variable. You must specify the variable name, and the value. If the value contains any spaces, you should quote the value with double (") or single (') quotes.

You can set any configuration value in this way, although some may require a restart to take effect. See .

router bgp ...

Configure BGP. See BGP.

ip access-list ...

Configure a named access list. See Access Lists.

nsctl

nsctl sends messages to a running l2tpns instance to be control plugins.

Arguments are command and optional args. See nsctl(8).

Built-in command are load_plugin, unload_plugin and help. Any other commands are passed to plugins for processing by the plugin_control function.

Signals

While the process is running, you can send it a few different signals, using the kill command.

killall -HUP l2tpns

The signals understood are:

SIGHUP

Reload the config from disk and re-open log file.

SIGTERM; SIGINT

Stop process. Tunnels and sessions are not terminated. This signal should be used to stop l2tpns on a cluster node where there are other machines to continue handling traffic. See Clustering

SIGQUIT

Shut down tunnels and sessions, exit process when complete.

Throttling

l2tpns contains support for slowing down user sessions to whatever speed you desire. The global setting throttle_speed defines the default throttle rate.

To throttle a sesion permanently, add a Cisco-AVPair RADIUS attribute. The autothrotle module interprets the following attributes:

You can also enable and disable throttling an active session using the throttle CLI command.

Interception

You may have to deal with legal requirements to be able to intercept a user's traffic at any time. l2tpns allows you to begin and end interception on the fly, as well as at authentication time.

When a user is being intercepted, a copy of every packet they send and receive will be sent wrapped in a UDP packet to a specified host.

The UDP packet contains just the raw IP frame, with no extra headers. The script scripts/l2tpns-capture may be used as the end-point for such intercepts, writing the data in PCAP format (suitable for inspection with tcpdump).

To enable or disable interception of a connected user, use the snoop and no snoop CLI commands. These will enable interception immediately.

If you wish the user to be intercepted whenever they reconnect, you will need to modify the RADIUS response to include the Vendor-Specific value Cisco-AVPair="intercept=ip:port". For this feature to be enabled, you need to have the autosnoop module loaded.

Plugins

So as to make l2tpns as flexible as possible, a plugin API is include which you can use to hook into certain events.

There are a some standard modules included which may be used as examples: autosnoop, autothrottle, garden, sessionctl, setrxspeed, snoopctl, stripdomain and throttlectl.

When an event occurs that has a hook, l2tpns looks for a predefined function name in every loaded module, and runs them in the order the modules were loaded.

The function should return PLUGIN_RET_OK if it is all OK. If it returns PLUGIN_RET_STOP, then it is assumed to have worked, but that no further modules should be run for this event.

A return of PLUGIN_RET_ERROR means that this module failed, and no further processing should be done for this event.

Most event functions take a specific structure named param_event, which varies in content with each event. The function name for each event will be plugin_event, so for the event timer, the function declaration should look like:

int plugin_timer(struct param_timer *data);

A list of the available events follows, with a list of all the fields in the supplied structure:

 int

char

Walled Garden

A "Walled Garden" is implemented so that you can provide perhaps limited service to sessions that incorrectly authenticate.

Whenever a session provides incorrect authentication, and the RADIUS server responds with Auth-Reject, the walled garden module (if loaded) will force authentication to succeed, but set the walled_garden flag in the session structure, and adds an iptables rule to the garden_users chain to cause all packets for the session to traverse the garden chain.

This doesn't just work. To set this all up, you will to setup the garden nat table with the build-garden script with rules to limit user's traffic.

For example, to force all traffic except DNS to be forwarded to 192.168.1.1, add these entries to your build-garden script:

iptables -t nat -A garden -p tcp --dport ! 53 -j DNAT --to 192.168.1.1
iptables -t nat -A garden -p udp --dport ! 53 -j DNAT --to 192.168.1.1

l2tpns will add entries to the garden_users chain as appropriate.

You can check the amount of traffic being captured using the following command:

iptables -t nat -L garden -nvx

Filtering

Sessions may be filtered by specifying Filter-Id attributes in the RADIUS reply. filter.in specifies that the named access-list filter should be applied to traffic from the customer, filter.out specifies a list for traffic to the customer.

Clustering

An l2tpns cluster consists of one* or more servers configured with the same configuration, notably the multicast cluster_address and the cluster_port

*A stand-alone server is simply a degraded cluster.

Initially servers come up as cluster slaves, and periodically (every cluster_hb_interval/10 seconds) send out ping packets containing the start time of the process to the multicast cluster_address on cluster_port.

A cluster master sends heartbeat rather than ping packets, which contain those session and tunnel changes since the last heartbeat.

When a slave has not seen a heartbeat within cluster_hb_timeout/10 seconds it "elects" a new master by examining the list of peers it has seen pings from and determines which of these and itself is the "best" candidate to be master. "Best" in this context means the server with the highest uptime (the highest IP address is used as a tie-breaker in the case of equal uptimes).

After discovering a master, and determining that it is up-to-date (has seen an update for all in-use sessions and tunnels from heartbeat packets) will raise a route (see Routing) for the bind_address and for all addresses/networks in ip_pool.

Any packets received by the slave which would alter the session state, as well as packets for throttled or gardened sessions are forwarded to the master for handling. In addition, byte counters for session traffic are periodically forwarded.

The heartbeat exchange use timing, so the master and slaves need to have their time synchronized, e.g. with ntp.

Since all the masters and slaves should be able to receive all packets, bind_address needs to be used to make all of them agree on the IP to be used. If l2tpns is to behave as LAC, bind_address_remotelns also needs to be set.

To balance the load, bind_multi_address can be used to make them support several IPs, and one can then prioritize one or the other on each server, and make e.g. radius pick up one of them randomly for Tunnel-Server-Endpoint. There is currently no support for such load balancing for LAC support.

*Configurable with cluster_master_min_adv

Multiple clusters can be run on the same network by just using different multicast cluster_address. However, for a given host to be part of multiple clusters without mixing the clusters, cluster_port must be different for each cluster.

Routing

If you are running a single instance, you may simply statically route the IP pools to the bind_address (l2tpns will send a gratuitous arp).

For a cluster, configure the members as BGP neighbours on your router and configure multi-path load-balancing. Cisco uses maximum-paths ibgp for IBGP. If this is not supported by your IOS revision, you can use maximum-paths (which works for EBGP) and set as_number to a private value such as 64512.