![[Go to CFHT Home Page]](/images/cfht-icon.gif){kind=icon}
Man Pages 
Back to Software Index 
Manpage Top Level
The bootparams file contains a list of client entries that diskless clients use for booting. Diskless booting clients retrieve this information by issuing requests to a server running the rpc.bootparamd(1M) program. The bootparams file may be used in conjunction with or in place of other sources for the bootparams information. See nsswitch.conf(4) .
For each client the file contains an entry with the client’s name and a list of boot parameter values for that client. Each entry should have the form:
clientname identifier-specifier ...
The first item of each entry is the host name of the diskless client. The asterisk (’*’) character may be used as a "wildcard" in place of the client name in a single entry. That entry will apply to all clients for whom there is not an entry that specifically names them.
This is followed by one or more whitespace characters and a series of identifier-specifiers separated by whitespace characters.
Each identifier-specifier has the form:
identifier=server:pathname
or
identifier=domain-name
The first form is used for file-specific identifiers. A file-specific identifier is a key that is used by diskless clients to identify a file or filesystem. server is the name of the server that will provide the file or filesystem to the diskless client, and pathname is the path to the exported file or filesystem on the specified server. The equal sign (’=’) and colon (’:’) characters are used in the indicated positions. There should not be any whitespace within an identifier-specifier.
Non-file-specific identifiers use the second form of identifier-specifier. One non-file-specific value for identifier is supported: the assignment of the client’s domain name. In this case, the value used for identifier is domain. domain-name must be the client’s domain name. The algorithm for determining a client’s domain name is to first check for a domain identifier in the client-specific entry and then in "wildcard" entry. If none is found, the server’s domain name is used.
An entry may be split across multiple lines of the file. The backslash (’\’) character should be used as the last character of a line to signify that the entry continues on the next line. The line may only be split in places where whitespace is allowed in the entry.
A variation of the first form (identifier=server:pathname) is used for the ns key which forces sysidtool(1M) to use a specific name service. By default, sysidtool uses NIS+ in preference to NIS if it can find a NIS+ server for the system’s domain on the subnet. This key may be necessary if you are trying to set up a hands-off installation, or if the name server is on a different subnet, which is common with NIS+ .
If this key is not used, sysidtool uses broadcast to attempt to bind to either a NIS+ or NIS server; if a name server is not on the local subnet, which is possible for NIS+ , the bind will fail, automatic configuration of the name service will fail, and an interactive screen is displayed, prompting the user to specify the name service.
The ns entry has the form:
ns=[server] : [nameservice] [(netmask)]
where:
The ns keyword can be set in add_install_client or by Host Manager.
Here is an example of an entry in the bootparams file:
client1 root=server1:/export/client1/root \ swap=server1:/export/client1/swap \ domain=bldg1.workco.com root=server2:/export/client2/root ns=:nis root=server2:/export/client2/root ns=watson: root=server2:/export/client2/root ns=mach:nisplus(255.255.255.0)
Solaris diskless clients use the identifiers "root", "swap", and "dump" to look up the pathnames for the root filesystem, a swap area, and a dump area, respectively. These are the only identifiers meaningful for SPARC diskless booting clients.
For x86 booting clients, the additional keyword identifiers "numbootfiles," "bootfile," and "bootaddr" are used (see rpld(1M) ).