Copyright (c) 2003, Ryan McBride. All rights reserved.
Copyright (c) 2011, Gleb Smirnoff <glebius@FreeBSD.org>
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
1. Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
SUCH DAMAGE.
$FreeBSD: releng/10.2/share/man/man4/carp.4 267100 2014-06-05 11:48:36Z marck $
.Dd February 21, 2013 .Dt CARP 4 .Os .Sh NAME .Nm carp .Nd Common Address Redundancy Protocol .Sh SYNOPSIS .Cd "device carp" .Sh DESCRIPTION The CARP allows multiple hosts on the same local network to share a set of IPv4 and/or IPv6 addresses. Its primary purpose is to ensure that these addresses are always available.
p To use .Nm , the administrator needs to configure at a minimum a common virtual host ID (vhid), and attach at least one IP address to this vhid on each machine which is to take part in the virtual group. Additional parameters can also be set on a per-vhid basis: .Cm advbase and .Cm advskew , which are used to control how frequently the host sends advertisements when it is the master for a virtual host, and .Cm pass which is used to authenticate .Nm advertisements. The .Cm advbase parameter stands for .Dq "advertisement base" . It is measured in seconds and specifies the base of the advertisement interval. The .Cm advskew parameter stands for .Dq "advertisement skew" . It is measured in 1/256 of seconds. It is added to the base advertisement interval to make one host advertise a bit slower that the other does. Both .Cm advbase and .Cm advskew are put inside CARP advertisements. These values can be configured using .Xr ifconfig 8 , or through the .Dv SIOCSVH .Xr ioctl 2 .
p CARP virtual hosts can be configured on multicast-capable interfaces: Ethernet, layer 2 VLAN, FDDI and Token Ring. An arbitrary number of virtual host IDs can be configured on an interface. An arbitrary number of IPv4 or IPv6 addresses can be attached to a particular vhid. It is important that all hosts participating in a vhid have the same list of prefixes configured on the vhid, since all prefixes are included in the cryptographic checksum supplied in each advertisement. Multiple vhids running on one interface participate in master/backup elections independently.
p
Additionally, there are a number of global parameters which can be set using
.Xr sysctl 8 :
l -tag -width ".Va net.inet.carp.ifdown_demotion_factor" t Va net.inet.carp.allow Accept incoming
.Nm
packets.
Enabled by default.
t Va net.inet.carp.preempt Allow virtual hosts to preempt each other.
When enabled, a vhid in a backup state would preempt a master that
is announcing itself with a lower advskew.
Disabled by default.
t Va net.inet.carp.log Determines what events relating to
.Nm
vhids are logged.
A value of 0 disables any logging.
A value of 1 enables logging state changes of
.Nm
vhids.
Values above 1 enable logging of bad
.Nm
packets.
The default value is 1.
t Va net.inet.carp.demotion This value shows current level of CARP demotion.
The value is added to the actual advskew sent in announcements for
all vhids.
At normal system operation the demotion factor is zero.
However, problematic conditions raise its level: when
.Nm
experiences problem with sending announcements, when an interface
running a vhid goes down, or while the
.Xr pfsync 4
interface is not synchronized.
The demotion factor can be adjusted writing to the sysctl oid.
The signed value supplied to the
.Xr sysctl 8
command is added to current demotion factor.
This allows to control
.Nm
behaviour depending on some external conditions, for example on the status
of some daemon utility.
t Va net.inet.carp.ifdown_demotion_factor This value is added to
.Va net.inet.carp.demotion
when an interface running a vhid goes down.
The default value is 240 (the maximum advskew value).
t Va net.inet.carp.senderr_demotion_factor This value is added to
.Va net.inet.carp.demotion
when
.Nm
experiences errors sending its announcements.
The default value is 240 (the maximum advskew value).
.El
.Sh ARP level load balancing
A
.Nm
interface has limited abilities for load balancing incoming connections
between hosts in an Ethernet network.
For load-balancing operation, one needs several CARP interfaces that
are configured to the same IP address, but to a different vhids.
Once an ARP request is received, the CARP protocol will use a hashing
function against the source IP address in the ARP request to determine
which vhid the request will be assigned to.
If the corresponding CARP interface is the current
master interface, a reply will
be sent to the ARP request;
otherwise it will be ignored.
See the
.Sx EXAMPLES
section for a practical example of load balancing.
.Pp
The ARP load balancing implemented in
.Nm
has some limitations.
First, ARP balancing only works on the local network segment.
It cannot balance traffic that crosses a router, because the
router itself will always be balanced to the same virtual host.
Second, ARP load balancing can lead to asymmetric routing
of incoming and outgoing traffic, and thus combining it with
.Xr pfsync 4
is dangerous, because this creates a race condition between
balanced routers and a host they are serving.
Imagine an incoming packet creating state on the first router, being
forwarded to its destination, and the destination replying faster
than the state information is packed and synced with the second router.
If the reply would be load balanced to second router, it will be
dropped since the second router has not yet received information about
the connection state.
.Sh STATE CHANGE NOTIFICATIONS
Sometimes it is useful to get notified about
.Nm
status change events.
This can be accomplished by using
.Xr devd 8
hooks.
Master/slave events are signalled under system
.Dv CARP .
The subsystem specifies the vhid and name of the interface where
the master/slave event occurred.
The type of the message displays the new state of the vhid.
Please see
.Xr devd.conf 5
and the
.Sx EXAMPLES
section for more information.
.Sh EXAMPLES
For firewalls and routers with multiple interfaces, it is desirable to
failover all of the addresses running
.Nm
together, when one of the physical interfaces goes down.
This is achieved by the use of the preempt option.
Enable it on both hosts A and B:
p .Dl sysctl net.inet.carp.preempt=1
p Assume that host A is the preferred master and we are running the 192.168.1.0/24 prefix on em0 and 192.168.2.0/24 on em1. This is the setup for host A (advskew is above 0 so it could be overwritten in the emergency situation from the other host): d -literal -offset indent ifconfig em0 vhid 1 advskew 100 pass mekmitasdigoat 192.168.1.1/24 ifconfig em1 vhid 2 advskew 100 pass mekmitasdigoat 192.168.2.1/24 .Ed
p The setup for host B is identical, but it has a higher .Cm advskew : d -literal -offset indent ifconfig em0 vhid 1 advskew 200 pass mekmitasdigoat 192.168.1.1/24 ifconfig em1 vhid 2 advskew 200 pass mekmitasdigoat 192.168.2.1/24 .Ed
p
When one of the physical interfaces of host A fails,
.Cm advskew
is demoted to a configured value on all its
.Nm
vhids.
Due to the preempt option, host B would start announcing itself, and thus
preempt host A on both interfaces instead of just the failed one.
.Pp
In order to set up an ARP balanced virtual host, it is necessary to configure
one virtual host for each physical host which would respond to ARP requests
and thus handle the traffic.
In the following example, two virtual hosts are configured on two hosts to
provide balancing and failover for the IP address 192.168.1.10.
.Pp
First the
.Nm
interfaces on host A are configured.
The
.Cm advskew
of 100 on the second virtual host means that its advertisements will be sent
out slightly less frequently.
.Bd -literal -offset indent
ifconfig carp0 create
ifconfig carp0 vhid 1 advskew 100 pass mekmitasdigoat 192.168.1.10/24
ifconfig carp1 create
ifconfig carp1 vhid 2 advskew 200 pass mekmitasdigoat 192.168.1.10/24
.Ed
.Pp
The configuration for host B is identical, except the
.Cm advskew
is on virtual host 1 rather than virtual host 2.
.Bd -literal -offset indent
ifconfig carp0 create
ifconfig carp0 vhid 1 advskew 200 pass mekmitasdigoat 192.168.1.10/24
ifconfig carp1 create
ifconfig carp1 vhid 2 advskew 100 pass mekmitasdigoat 192.168.1.10/24
.Ed
.Pp
Finally, the ARP balancing feature must be enabled on both hosts:
.Pp
.Dl sysctl net.inet.carp.arpbalance=1
.Pp
When the hosts receive an ARP request for 192.168.1.10, the source IP address
of the request is used to compute which virtual host should answer the request.
The host which is master of the selected virtual host will reply to the
request, the other(s) will ignore it.
.Pp
This way, locally connected systems will receive different ARP replies and
subsequent IP traffic will be balanced among the hosts.
If one of the hosts fails, the other will take over the virtual MAC address,
and begin answering ARP requests on its behalf.
p Processing of .Nm status change events can be set up by using the following devd.conf rule: d -literal -offset indent notify 0 { match "system" "CARP"; match "subsystem" "[0-9]+@[0-9a-z]+"; match "type" "(MASTER|BACKUP)"; action "/root/carpcontrol.sh $subsystem $type"; }; .Ed
p To see .Nm packets decoded in .Xr tcpdump 8 output, one needs to specify .Fl T Ar carp option, otherwise .Xr tcpdump 8 tries to interpret them as VRRP packets: d -literal -offset indent tcpdump -npi vlan0 -T carp .Ed .Sh SEE ALSO .Xr inet 4 , .Xr pfsync 4 , .Xr rc.conf 5 , .Xr devd.conf 5 , .Xr ifconfig 8 , .Xr sysctl 8 .Xr tcpdump 8 .Sh HISTORY The .Nm device first appeared in .Ox 3.5 . The .Nm device was imported into .Fx 5.4 . In .Fx 10.0 , .Nm was significantly rewritten, and is no longer a pseudo-interface.