Проект OpenNet: MAN bootpgw (8) Команды системного администрирования (FreeBSD и Linux)

Интерактивная система просмотра системных руководств (man-ов)

bootpgw (8)

>> bootpgw (8) ( FreeBSD man: Команды системного администрирования )

BSD mandoc

NAME



bootpd , bootpgw

 - Internet Boot Protocol server/gateway

SYNOPSIS

[-i | s ] [-c chdir-path ] [-d level ] [-h hostname ] [-t timeout ] [bootptab [dumpfile ] ]
bootpgw [-i | s ] [-d level ] [-h hostname ] [-t timeout ] server

DESCRIPTION

The utility implements an Internet Bootstrap Protocol (BOOTP) server as defined in RFC951, RFC1532, and RFC1533. The bootpgw utility implements a simple BOOTP gateway which can be used to forward requests and responses between clients on one subnet and a BOOTP server (i.e. ) on another subnet. While either or bootpgw will forward BOOTREPLY packets, only bootpgw will forward BOOTREQUEST packets.

One host on each network segment is normally configured to run either or bootpgw from inetd(8) by including one of the following lines in the file /etc/inetd.conf

bootps dgram udp wait root /usr/libexec/bootpd bootpd /etc/bootptab

bootps dgram udp wait root /usr/libexec/bootpgw bootpgw server

This mode of operation is referred to as "inetd mode" and causes (or bootpgw to be started only when a boot request arrives. If it does not receive another packet within fifteen minutes of the last one it received, it will exit to conserve system resources. The -t option controls this timeout (see OPTIONS).

It is also possible to run (or bootpgw in "standalone mode" (without inetd(8)) by simply invoking it from a shell like any other regular command. Standalone mode is particularly useful when is used with a large configuration database, where the start up delay might otherwise prevent timely response to client requests. (Automatic start up in standalone mode can be done by invoking from within /etc/rc.local for example.) Standalone mode is less useful for bootpgw which has very little start up delay because it does not read a configuration file.

Either program automatically detects whether it was invoked from inetd or from a shell and automatically selects the appropriate mode. The -s or -i option may be used to force standalone or inetd mode respectively (see OPTIONS).

OPTIONS

The following options are available:

-t timeout: Specify the timeout value (in minutes) that a or bootpgw process will wait for a BOOTP packet before exiting. If no packets are received for timeout minutes, then the program will exit. A timeout value of zero means "run forever". In standalone mode, this option is forced to zero.
-d debug-level: Set the debug-level variable that controls the amount of debugging messages generated. For example, -d 4 or -d 4 will set the debugging level to 4. For compatibility with older versions of , omitting the numeric parameter (i.e., just -d will simply increment the debug level by one.
-c chdir-path: Set the current directory used by while checking the existence and size of client boot files. This is useful when client boot files are specified as relative pathnames, and needs to use the same current directory as the TFTP server (typically /tftpboot ) This option is not recognized by bootpgw
-h hostname: Specify the hostname corresponding to the IP address to listen on. By default, listens on the IP address corresponding to the machine's hostname, as returned by gethostname(3).
-i: Force inetd mode. This option is obsolete, but remains for compatibility with older versions of .
-s: Force standalone mode. This option is obsolete, but remains for compatibility with older versions of .
bootptab: Specify the name of the configuration file from which loads its database of known clients and client options ( only).
dumpfile: Specify the name of the file that will dump its internal database into when it receives a SIGUSR1 signal ( only). This option is only recognized if was compiled with the -DDEBUG flag.
server: Specify the name of a BOOTP server to which bootpgw will forward all BOOTREQUEST packets it receives ( bootpgw only).

OPERATION

Both and bootpgw operate similarly in that both listen for any packets sent to the bootps port, and both simply forward any BOOTREPLY packets. They differ in their handling of BOOTREQUEST packets.

When bootpgw is started, it determines the address of a BOOTP server whose name is provided as a command line parameter. When bootpgw receives a BOOTREQUEST packet, it sets the "gateway address" and "hop count" fields in the packet and forwards the packet to the BOOTP server at the address determined earlier. Requests are forwarded only if they indicate that the client has been waiting for at least three seconds.

When is started it reads a configuration file, (normally /etc/bootptab that initializes the internal database of known clients and client options. This internal database is reloaded from the configuration file when receives a hangup signal (SIGHUP) or when it discovers that the configuration file has changed.

When receives a BOOTREQUEST packet, it looks for a database entry matching the client request. If the client is known, composes a BOOTREPLY packet using the database entry found above, and sends the reply to the client (possibly using a gateway). If the client is unknown, the request is discarded (with a notice if debug > 0).

If is compiled with the -DDEBUG option, receipt of a SIGUSR1 signal causes it to dump its internal database to the file /tmp/bootpd.dump or the dumpfile specified as a command line parameter.

During initialization, both programs determine the UDP port numbers to be used by calling getservbyname(3) (which normally uses /etc/services ) Two service names (and port numbers) are used: