routed(1Mtcp)


routed -- network unicast routing daemon

Synopsis

in.routed [-sqdghmAt] [-T tracefile] [-F net[/mask][,metric]] [-P parameter[=value]...]

Description

routed is a daemon invoked at boot time to manage the network routing tables. It uses Routing Information Protocol, RIPv1 (RFC 1058), RIPv2 (RFC 1723), and Internet Router Discovery Protocol (RFC 1256) to maintain the kernel routing table. The RIPv1 protocol is based on the reference 4.3BSD daemon.

routed listens on the UDP socket of the route service (see services(4tcp)) for Routing Information Protocol packets, and also sends and receives multicast Router Discovery ICMP messages. If the host is a router, routed periodically supplies copies of its routing tables to any directly connected hosts and networks. It also advertises or solicits default routes using Router Discovery ICMP messages.

When routed starts (or when a network interface is later turned on), it finds out which of the system's directly connected interfaces are marked ``up''. routed then adds necessary routes for the interfaces to the kernel routing table. Soon after it first starts, and provided there is at least one interface on which RIP has not been disabled, routed deletes all pre-existing non-static routes from the kernel routing table. Static routes in the kernel routing table are preserved and included in RIP responses if they have a valid RIP metric (see route(7tcp)).

If more than one interface is present (not counting the loopback interface), routed checks if the host should forward packets between the connected networks. After transmitting a RIP request and Router Discovery advertisements or solicitations over a new interface, routed enters a loop, listening for RIP request and response packets and Router Discovery packets from other hosts.

When a request packet is received, routed formulates a reply based on the information maintained in its internal tables. The response packet generated contains a list of known routes, each marked with a ``hop count'' metric (a count of 16 or greater is considered infinite). Advertised metrics reflect the metric associated with an interface (see ifconfig(1Mtcp)), so setting the metric on an interface is an effective way to steer traffic.

Responses do not contain routes that define a first hop on the requesting network. This partly implements a ``split-horizon''. Requests from query programs such as rtquery(1Mtcp) are answered with the complete table.

The routing table maintained by routed includes space for several gateways for each destination to speed recovery if a router fails. Received RIP response packets are used to update the routing tables provided they come from one of the several currently recognized gateways or they advertise a better metric than at least one of the existing gateways.

When an update is applied, routed records the change in its own tables and updates the kernel routing table if the best route to the destination changes. The change in the kernel routing table is reflected in the next batch of response packets sent. If the next response is not scheduled for a while, routed sends a ``flash update'' response containing only recently changed routes.

In addition to processing incoming packets, routed also periodically checks the routing table entries. If an entry has not been updated for 3 minutes, the entry's metric is set to infinity and marked for deletion. Deletions are delayed until the route has been advertised with an infinite metric to insure the invalidation is propagated throughout the local internet. This is a form of ``poison reverse'' .

Routes in the kernel table that are added or changed as a result of ICMP Redirect messages are deleted after a while to minimize ``black-holes'' . When a TCP connection suffers a timeout, the kernel tells routed, which deletes all redirected routes through the gateway involved, advances the age of all RIP routes through the gateway to allow an alternate to be chosen, and advances the age of any relevant Router Discovery Protocol default routes.

Hosts acting as internetwork routers gratuitously supply their routing tables every 30 seconds to all directly connected hosts and networks. routed sends RIP responses to the broadcast address on networks that support broadcasting, to the destination address on point-to-point links, and to the router's own address on other networks. If RIPv2 is enabled, routed sends multicast packets on interfaces that support multicasting.

If no response is received on a remote interface, if there are errors while sending responses, or if there are more errors than input or output (see netstat(1Mtcp)), then the cable or some other part of the interface is assumed to be disconnected or broken, and routes are adjusted appropriately.

Internet Router Discovery

The Internet Router Discovery Protocol is handled similarly to RIP. When the daemon is supplying RIP routes, it also listens for Router Discovery Solicitations and sends Advertisements. When it is quiet and only listening to other RIP routers, it sends Solicitations and listens for Advertisements. If it receives a good Advertisement, it stops listening for broadcast or multicast RIP responses. It tracks several advertising routers to speed recovery when the currently chosen router dies. If all discovered routers disappear, routed resumes listening to RIP responses.

While using Router Discovery (which happens by default when the system has a single network interface and a Router Discovery Advertisement is received), there is a single default route and a variable number of redirected host routes in the kernel table.

The Router Discovery standard requires that advertisements have a default ``lifetime'' of 30 minutes. That means should something happen, a client can be without a good route for up to 30 minutes. It is a good idea to reduce the default to 45 seconds by specifying -P rdisc_interval=45 on the command line or rdisc_interval=45 in the /etc/inet/gateways file (see ``Parameters'').

On a host with more than one network interface, the default route will be via only one of the interfaces. Thus, multi-homed hosts running with -q might need to specify no_rdisc (described in ``Parameters'').

See the description of pm_rdisc in ``Parameters'' for details of how to support legacy systems that do not handle RIPv2 or Router Discovery.

By default, Router Discovery advertisements and solicitations are not sent over point-to-point links (such as PPP). routed uses the netmask associated with point-to-point links (such as SLIP and PPP links which set the IFF_POINTOPOINT flag) to infer the netmask used by the remote system when RIPv1 is used.

Options

routed understands the following options:

-s
Force routed to supply routing information. This is the default if multiple network interfaces are present on which RIP or Router Discovery have not been disabled, and if the kernel parameters ipforwarding and ipsendredirects have the value 1.

-q
Do not supply routing information (the opposite of the -s option). This is the default when only one interface is present.

-d
Do not run in the background. This option is meant for interactive use.

-g
Used on internetwork routers to offer a route to the ``default'' destination. This is equivalent to setting net[/mask][,metric] to ``0/0,1'' and is present mostly for historical reasons. It is preferable to specify -P pm_rdisc on the command line or pm_rdisc in the /etc/inet/gateways file. This will use a larger metric, reducing the spread of the potentially dangerous default route.

This option is typically used on a gateway to the Internet, or on a gateway that uses another routing protocol whose routes are not reported to other local routers.


WARNING: This feature is dangerous because it uses a metric of 1. It may accidentally wreak havoc by creating a routing loop.


-h
Do not advertise host or point-to-point routes, provided there is a network route going the same direction. This provides a limited kind of aggregation which may be useful on gateways to ethernets that have other gateway machines connected with point-to-point links such as SLIP.

-m
Advertise a host or point-to-point route to its primary interface. It is useful on multi-homed machines such as NFS servers. This option should not be used except when the cost of the host routes it generates is justified by the popularity of the server. It is effective only when the machine is supplying routing information, because there is more than one interface. The -m option overrides the -q option to the limited extent of advertising the host route.

-A
Do not ignore RIPv2 authentication if we do not care about RIPv2 authentication. This option is required for conformance with RFC 1723. However, it makes no sense and breaks using RIP as a discovery protocol to ignore all RIPv2 packets that carry authentication when this machine does not care about authentication.

-T tracefile
Increase the debugging level to at least 1 and causes debugging information to be appended to the trace file. Note that because of security concerns, it is wisest to not run routed routinely with tracing directed to a file.

-t
Increase the debugging level, which causes more information to be logged on the standard output or to the tracefile specified with the -T option. The debugging level can be increased or decreased by using kill(1) to send the SIGUSR1 or SIGUSR2 signals to the routed process or using the rtquery command.

The debug levels are:


0
no tracing

1
trace actions only

2
trace actions and packets

3
trace actions and history of packets and contents after change

4
trace all actions, packets and contents

-F net[/mask][,metric]
Minimize routes in transmissions via interfaces with addresses that match the first mask (0-32) bits in the IP address specified by net, and synthesizes a default route to this machine with the specified metric. The default value for mask is 32.

The intention is to reduce RIP traffic on slow, point-to-point links such as PPP links by replacing many large UDP packets of RIP information with a single, small packet containing a ``fake'' default route. If metric is absent, a value of 14 is assumed to limit the spread of the fake default route.


WARNING: This is a dangerous feature that when used carelessly can cause routing loops. Note also that more than one interface can match the specified network number and mask.

See also the description of the -g option.


-P parameter[=value]...
Equivalent to specifying the parameter(s) on a line in the /etc/inet/gateways file. See ``Parameters'' for more details.

Any other argument supplied on the command line is interpreted as the name of a trace file in which the actions of routed should be logged. It is preferable to use the -T option to specify the name of a trace file.

Passive, external, and active gateways

routed supports the notion of ``distant'' passive or active gateways. When routed is started, it reads the file /etc/inet/gateways to find out which local and non-local (distant) gateways are defined as passive, and to obtain other configuration parameters. Gateways should be marked passive if they are not expected to exchange routing information, or they are distant gateways which cannot be located using information from a routing socket. Routes through passive gateways are installed in the kernel's routing tables once upon startup and are not included in transmitted RIP responses.

Gateways marked extern (external) are also passive, but are not placed in the kernel routing table nor are they included in routing updates. The function of external entries is to indicate that another routing process will install such a route if necessary, and that alternate routes to that destination should not be installed by routed. Such entries are only required when both routers may learn of routes to the same destination.

Gateways marked active should be willing to exchange RIP packets.

Distant active gateways are treated like network interfaces. RIP responses are sent to the distant active gateway. If no responses are received, the associated route is deleted from the kernel table and RIP responses advertised via other interfaces. If the distant gateway resumes sending RIP responses, the associated route is restored. Such gateways can be useful on media, such as some ATM networks, that do not support broadcasts or multicasts but otherwise act like classic shared media like Ethernets. You can list all RIP routers reachable on an ATM network by defining a series of host lines in the /etc/inet/gateways file.

Format of the /etc/inet/gateways file

The /etc/inet/gateways file is comprised of a series of lines. Each line must define a route to a network or a host, or one of the parameters described in ``Parameters''.

A route to a network is defined using a line with the following format:

net Nname[/mask] gateway Gname metric value passive|extern|active

A route to a host is defined using a line with the following format:

host Hname gateway Gname metric value passive|extern|active

Nname or Hname is the name of the destination network or host. It may be a symbolic network name or an Internet address specified in ``dot'' notation (see inet(3N)). (If it is a name, then it must either be defined in /etc/networks or /etc/hosts , or named(1Mtcp) must have been started before routed.)

mask is an optional number between 1 and 32 indicating the length of the netmask associated with Nname. (Note that host Hname is equivalent to net Nname/32.)

Gname is the name or address of the gateway to which RIP responses should be forwarded.

value is the hop count to the destination host or network.

One of the keywords passive, active or extern must be present to indicate whether the gateway should be treated as passive or active (as described in ``Passive, external, and active gateways''), or whether the gateway is external to the scope of the RIP protocol.

Parameters

Lines in /etc/inet/gateways that do not start with net or host must consist of one or more of the following parameter settings, separated by commas or blanks:

bcast_rdisc
Specifies that Router Discovery packets should be broadcast instead of multicast.

fake_default=metric
Has an identical effect to net[/mask][,metric] with the network and mask coming from the specified interface.

if=ifname
Indicates that the other parameters on the line apply to the interface name ifname .

md5_passwd=XXX1|KeyID [start|stop]
Specifies a RIPv2 MD5 password. The KeyID must be unique. If present, start and stop are timestamps in the form:
year/month/day@hour:minute
They specify when the password is valid. The valid password with the most future is used on output packets, unless all passwords have expired, in which case the password that expired most recently is used, or unless no passwords are valid yet, in which case no password is output. Incoming packets can carry any password that is valid, will be valid within 24 hours, or that was valid within 24 hours.

To protect the secrets, this parameter setting is valid only in the /etc/inet/gateways file and only when that file is readable only by UID 0.


no_ag
Turns off aggregation of subnets in RIPv1 and RIPv2 responses.

no_rdisc
Disables the Internet Router Discovery Protocol.

no_rdisc_adv
Disables the transmission of Router Discovery Advertisements

no_rip
Disables all RIP processing on the specified interface. If no interfaces are allowed to process RIP packets, routed acts purely as a router discovery daemon.

Note that using rdisc_adv or the -s option to turn off RIP without explicitly turning on router discovery advertisements causes routed to act as a client router discovery daemon without advertising.


no_ripv1_in
Causes RIPv1 received responses to be ignored.

no_ripv2_in
Causes RIPv2 received responses to be ignored.

no_solicit
Disables the transmission of Router Discovery Solicitations.

no_super_ag
Turns off aggregation of networks into supernets in RIPv2 responses.

passive
Marks the interface to not be advertised in updates sent via other interfaces, and turns off all RIP and router discovery through the interface.

passwd=XXX1[|KeyID[start|stop]]
Specifies a RIPv2 cleartext password that will be included on all RIPv2 responses sent, and checked on all RIPv2 responses received. Any blanks, tab characters, commas, or ``#'', ``|'', or NULL characters in the password must be escaped with a backslash (\). The common escape sequences ``\n'', ``\r'', ``\t'', ``\b'', and ``\xxx'' have their usual meanings. The KeyID must be unique but is ignored for cleartext passwords. If present, start and stop are timestamps in the form:
year/month/day@hour:minute
They specify when the password is valid. The valid password with the most future is used on output packets, unless all passwords have expired, in which case the password that expired most recently is used, or unless no passwords are valid yet, in which case no password is output. Incoming packets can carry any password that is valid, will be valid within 24 hours, or that was valid within 24 hours.

pm_rdisc
Similar to fake_default . When RIPv2 routes are multicast, so that RIPv1 listeners cannot receive them, this feature causes a RIPv1 default route to be broadcast to RIPv1 listeners. Unless modified with fake_default, the default route is broadcast with a metric of 14. This serves as a ``poor man's router discovery'' protocol.

rdisc_adv
Specifies that Router Discovery Advertisements should be sent, even on point-to-point links, which by default only listen to Router Discovery messages

rdisc_pref=N
Sets the preference in Router Discovery Advertisements to the integer N.

rdisc_interval=N
Sets the nominal interval with which Router Discovery Advertisements are transmitted to N seconds and their lifetime to 3*N.

redirect_ok
Causes RIP to allow ICMP Redirect messages when the system is acting as a router and forwarding packets. Otherwise, ICMP Redirect messages are are overridden.

ripv2
Equivalent to both no_ripv1_in and ripv2_out being specified.

ripv2_out
Turns off RIPv1 output and causes RIPv2 advertisements to be multicast when possible.

send_solicit
Specifies that Router Discovery solicitations should be sent, even on point-to-point links, which by default only listen to Router Discovery messages.

subnet=Nname[/mask][,metric]
Advertises a route to network Nname with the optionally specified mask and metric (default 1). This is useful for filling ``holes'' in CIDR allocations. This parameter must appear by itself on a line.


WARNING: Do not use this feature unless necessary. It can cause routing loops if used incorrectly.


trust_gateway=rname
Causes RIP packets from that router and other routers named in other trust_gateway keywords to be accepted, and packets from other routers to be ignored.

verbose=boolean
If set to 1, routed will log illegal routes contained in RIP advertisements, such as those to loopback interfaces (127.*). By default (set to 0), routed will ignore illegal routes.

Files


/etc/inet/gateways
defines distant gateways

References

gated(1Mtcp), htable(1Mtcp), icmp(7tcp), netstat(1Mtcp), ripquery(1Mtcp), route(1Mtcp), rtquery(1Mtcp), udp(7tcp)

RFC 1058, RFC 1256, RFC 1321, RFC 1723

Notices

routed does not always detect unidirectional failures in network interfaces (for example, when the output side fails).
© 2004 The SCO Group, Inc. All rights reserved.
UnixWare 7 Release 7.1.4 - 25 April 2004