NETCONFIG(8N) COMMAND REFERENCE NETCONFIG(8N) NAME netconfig - configure workstation for network operation SYNOPSIS /etc/netconfig [ -q ] [ -h hostname ] [ -i interfacename -a address -b broadcast address -m subnet mask ] [ -e net ] [ -d net ] [ -P ] DESCRIPTION netconfig is used to configure a workstation for operation on a local area network. netconfig allows the user to change the following network attributes on the workstation: Hostname Host ID Standard Network Utilities (enabled or disabled) Internet address, broadcast address and subnet mask for each network interface. After changing any of these attributes the workstation must be rebooted. This is so the nameserver(8n) daemon is restarted with the proper new hostname and addresses; otherwise the network utilities will not work correctly. netconfig writes these changes into the network.conf(5n) file. When rebooted, rc.net(8n) is executed which invokes netconfig with the -q switch. When netconfig is invoked with this switch, it looks at the network.conf(5n) file to set up the host ID and hostname. rc.net(8n) then to decide which of the daemons to start up. Typically netconfig is invoked with no options. The program will then prompt the user with questions to set the attributes mentioned above. When it asks for a hostname, the user should enter no more than 31 characters taken from the following set: a-z, A-Z, 0-9, -, _ The first character must not be a number. Remember that the name assigned to your workstation should be unique throughout the network. The first time netconfig is run on the workstation, the Internet address will not be set (actually it is set to the invalid address 0.0.0.0). Accordingly, the user must supply a valid address so that the workstation may be used on the local network. Printed 4/6/89 1
NETCONFIG(8N) COMMAND REFERENCE NETCONFIG(8N) The Internet address is a 32-bit number (expressed as four 8-bit bytes) which is comprised of a network component and local host component. Further, the Internet address may be in one of three classes, namely, A, B, or C. The distinction between the classes is in how many of the 32 bits are used to express the network and local host components, respectively. In a Class A Internet address, one byte is used for the network component; in Class B, two bytes; and in Class C, three bytes. Consequently, in a Class A network with three bytes (24 bits) remaining for the local host component, there are up to 16,777,216 separate node addresses available (Class B - 65,536, and Class C - 256). When netconfig prompts for a network number the user selects which class to use by entering the data as follows: User enters: Class assumed: Where x, y, and z = x A 0 < x <= 127 x.y B 128 <= x <= 191, 0 <= y <= 255 x.y.z C 192 <= x <= 223, 0 <= y <= 255, 0 <= z <= 255 In the above, x, y, and z are decimal integers. NOTE: The class A network number 127 is reserved for the local loopback interface and should not be assigned for other purposes. Once the network number has been entered (and assuming that an Internet address has not already been assigned to this workstation) netconfig will suggest an Internet address. This suggested address is based on the network number already supplied plus a host number derived from the LAN interface's physical address. Since the physical address is guaranteed to be unique, and the host number of the Internet address must be unique on the local network, by basing the Internet address on the physical address we increase the likelihood that the suggested address is unique. If you have selected class A or B addressing, and have only 6130, 4132, or 4300 series workstations on your network, then the user can feel confident that the suggested address is unique. If other vendors' equipment also appears on the network, then before using the suggested Internet address, verify that no other equipment uses that address. If the suggested address is not appropriate, enter the host number component of the Internet address as follows: Printed 4/6/89 2
NETCONFIG(8N) COMMAND REFERENCE NETCONFIG(8N) Class User enters: Where x, y, and z = A x.y.z 0 <= x,y,z <= 255 B x.y 0 <= x,y <= 255 C x 0 < x <= 254 In the above, x, y, and z are decimal integers, and at least one of them is non-zero and is less than 255 (i.e., the aggregate host number must not be 0 nor may all of the bits of the host number be 1). After the internet address has been determined, netconfig will ask for the subnet mask and broadcast address. If subnetting is not being used on the network being connected to, the subnet mask should be set to have one's exactly corresponding to the net portion of the internet address. That is Class User enters A 255.0.0.0 B 255.255.0.0 C 255.255.255.0 If subnetting is being used, the user should consult with the local network administrator as to the proper subnet mask. The broadcast address is the address used to send internet broadcast messages. Originally, the broadcast address was determined by using the same net portion of the internet address and setting the host portion to zeros. This is the broadcast address used by previous releases of UTek and version of Unix based of Berkeley 4.2 BSD release. Newer versions of the internet protocols specify that the broadcast addreses should use ones instaed of zeros. The user must determine whether the network being connected to is using zeros or ones in the broadcast addresses and set the broadcast address appropriately. The broadcast address should be specified in the standard "dot" notation used for internet addresses, e.g., if the internet address is 241.45.1.23, and ones are being used for broadcast addresses, the user should enter 241.45.255.255. If subnetting is being used, the choice of broadcast address can become more complicated. The user should consult with the local network administrator for the proper choice of broadcast address. Next netconfig asks whether to enable the regular network daemons. These daemons include those that handle remote logins (rlogind(8n)) and remote command execution (rshd(8n)). See the file /etc/rc.net for what daemons will Printed 4/6/89 3
NETCONFIG(8N) COMMAND REFERENCE NETCONFIG(8N) be started. Also see inetd(8n). OPTIONS Typically the user would invoke netconfig without any switches or with the -P switch. The full list of capabilities follows. -a address Specifies the internet address when setting the internet address parameters from the command line. This option must be used in conjunction with the -i , -b and -m options. All three address parameters (internet address, broadcast address and subnet mask) must be specified along with the interface name. If one or more of the paramters are not specified, netconfig will prompt for all the parameters. -b broadcast address Specifies the broadcast address when setting the internet address parameters from the command line. This option must be used in conjunction with the -i , -a and -m options. All three address parameters (internet address, broadcast address and subnet mask) must be specified along with the interface name. If one or more of the paramters are not specified, netconfig will prompt for all the parameters. -d net netconfig writes the net_disabled string into the network.conf(5n) file. No prompting occurs. -e net This option causes the net_enabled string to be written into the network.conf(5n) file. Prompting may occur if the Internet address is not set for the interface(s). -h hostname Set the hostname; no prompting is provided. -i interfacename Indicates the interface name to act upon using the -a option. -m subnet mask Specifies the subnet mask when setting the internet address parameters from the command line. This option must be used in conjunction with the -i , -a and -b options. All three address parameters (internet address, broadcast address and subnet mask) must be specified along with the interface name. If one or more of the paramters are not specified, netconfig will prompt for all the parameters. Printed 4/6/89 4
NETCONFIG(8N) COMMAND REFERENCE NETCONFIG(8N) -q This option is used when netconfig is invoked from rc.net(8n) at boot time. It causes netconfig to prompt only for attributes for which we have no known previous value. -P Print out the Internet and physical address for the interface on the workstation. RETURN VALUE [NO_ERRS] Command completed without error. [USAGE] Incorrect command line syntax. Execution terminated. [P_ERR] A system error occurred. Execution terminated. See intro(2) for more information on system errors. [NP_WARN] An error warranting a warning message occurred. Execution continues. [0] Indicates to rc.net(8n) not to enable networking. [1] Indicates to rc.net(8n) to enable networking. CAVEATS Input validation for command line invocation is minimal. SEE ALSO hostid(1n), hostname(1n), gethostname(2), inet(3n), hosts(5n), and network.conf(5n). Printed 4/6/89 5
%%index%% na:384,107; sy:491,700; de:1191,2848;4495,2901;7852,3155;11463,53; op:11516,3816;15788,559; rv:16347,741; ca:17088,141; se:17229,305; %%index%%000000000167