The metric server calculates the current load on a cluster host by examining the state of each of its processes. The server appropriately increments a counter, which is later scaled before being used in the load calculation. Scaling is based on:

• CPU power
• Free job slots
• Administratively set bias
• Available memory

To set a bias value, define the system logical name TCPIP$METRIC_CPU_RATING. This factor represents the system manager's estimate of the relative CPU power. Its value can range from 1, the lowest CPU power, to 100, the highest CPU power. To specify the default value, set the logical value to 0.

6.2.3.2 Metric Server Logical Names

Table 6-2 describes the Metric Server's logical names. Define these logical names with the /SYSTEM qualifier, and restart the metric server to make the changes take effect.

Table 6-2 Metric Server Logical Names

Logical Name	Description
TCPIP$METRIC_CPU_RATING value	Sets a bias value that represents your estimate of the relative CPU power. Valid values range from 1 (lowest CPU power) to 100 (highest CPU power). Use a value of 0 (zero) to specify the default.
TCPIP$METRIC_COMPUTE_INTERVAL value	Specifies how often the metric server computes the rating. Valid value (in seconds) is a number from 1 to 300. The default is 10 seconds.
TCPIP$METRIC_DBG_LEVEL value	Turns on diagnostics logged to the file TCPIP$METRIC_SERVER.LOG located in SYS$SPECIFIC:[TCPIP$METRIC]. Valid values are 1 or 2 (2 provides more detailed diagnostics).

Part 3 describes how to set up and manage the network services that, while not required, make the network software more useful.

Chapter 7 describes how to configure the DHCP server so you can centralize the configuration and maintenance of the IP address space.

Chapter 8 describes how to configure the BOOTP server and TFTP so your host can answer bootstrap requests from diskless workstations and other network devices.

Chapter 9 describes how to configure the portmapper service, a service that registers server programs written using RPCs (remote procedure calls). You must run the portmapper service if you intend to run NFS or any customer-developed RPC programs.

Chapter 10 describes how to configure and manage NTP (Network Time Protocol), allowing your host to synchronize its time with that of other internet hosts also running NTP.

Chapter 11 describes how to configure your host so it can answer SNMP (Simple Network Management Protocol) requests from remote SNMP management stations.

Dynamic Host Configuration Protocol (DHCP), a superset of the Bootstrap Protocol (BOOTP), provides a centralized approach to the configuration and maintenance of IP address space. It allows the system manager to configure various clients on the network from a single location.

DHCP allocates temporary or permanent IP addresses from an address pool to client hosts on the network. DHCP can also configure client parameters such as default gateway parameters, domain name server parameters, and subnet masks for each host running a DHCP client.

This chapter reviews key DHCP and BOOTP concepts, explains how to configure and customize the DHCP server and how to configure the DHCP client information.

7.1 Reviewing Key Concepts

With DHCP, system managers can centralize TCP/IP network configurations and management tasks involved with network connections. DHCP makes network administration easier by allowing:

• Consistent application of network parameters, such as subnet masks and default routers, to all hosts on a network
• Support for both DHCP and BOOTP clients
• Static (permanent) mapping of hardware addresses to IP addresses
• Dynamic (temporary) mapping of hardware addresses to IP addresses where the client leases the IP address for a defined length of time

In addition, the TCP/IP Services implementation of DHCP includes support for DHCP server failover in a cluster environment.

The DHCP protocol is a superset of BOOTP. In addition to the BOOTP functionality, DHCP offers robust configuration services including IP addresses, subnet masks, default gateways, and more.

Based on the BOOTP functionality, DHCP is built on the client-server model:

• The DHCP server is a host that provides initialization parameters.
• The DHCP client is a host that requests initialization parameters from a DHCP server. A router cannot be a DHCP client.

DHCP consists of two components:

• A mechanism for allocating network addresses to clients
• A set of rules for delivering client-specific configuration parameters from a DHCP server to a client

DHCP operates as follows:

• When a DHCP client boots, it broadcasts a DHCP request asking for any DHCP server on the network to provide it with an IP address and configuration parameters.
• A DHCP server on the network authorized to configure this client sends the client a reply that offers an IP address.
• When the client receives the offer, it can decide to accept it or wait for another offer from other servers on the network.
• Once the client accepts an offer, it sends an acceptance message to the server.
• When the server receives the acceptance message, it sends an acknowledgment with the offered IP address and any other configuration parameters that the client requested. (The server only responds to specific client requests; it does not impose any parameters on the client.)
• If the dynamic address allocation method is used, the IP address offered to the client has a specific lease time that determines how long the IP address is valid.
  During the lifetime of the lease, the client repeatedly asks the server to renew. If the client chooses not to renew, the lease expires.
  Once the lease expires, the IP address can be recycled and given to another client. When the client reboots, it can be given the old address if available or assigned a new address.

For more information about how DHCP operates, see RFC 2131 and RFC 1534.

7.1.2 How You Configure the DHCP Server and Client Information

TCP/IP Services provides an easy-to-use DHCP graphical user interface (GUI) for you to do the following:

• Configure dynamic and static IP addressing for all clients ( Section 7.2.4)
• Configure the client information appropriate for your client base ( Section 7.3)
• Set DHCP parameters to customize the DHCP server ( Section 8.3)

To use the GUI, you need the following privileges:

• BYPASS
• SYSNAM
• PRMMBX

Instead of using the GUI, you can manually edit the DHCP client configuration files and modify the databases using command-line utilities (see Section 7.4).

7.1.3 How DHCP Allocates IP Addresses

With TCP/IP Services, DHCP uses the dynamic and static IP address mapping methods outlined in Table 7-1 to service DHCP and BOOTP-only client requests.

Table 7-1 DHCP IP Address Allocation Methods

Method	Applicable Client	Description
Dynamic	DHCP and BOOTP	The DHCP server assigns an IP address from an address pool to a client for a limited amount of time (or until the client explicitly relinquishes the address). Addresses no longer needed by clients can be reused. Use dynamic allocation when: Clients plan to be connected to the network only temporarily. You have a limited pool of IP addresses that must be shared among clients that do not need permanent IP addresses. IP addresses are scarce, and you need to reclaim retired addresses so you can assign them the new clients being permanently connected to the network. For BOOTP clients: DHCP assigns dynamic IP addresses from the address pool and stores the addresses in the lease database by assigning each lease a time of infinity.
Static	DHCP and BOOTP	The system manager manually assigns (in the DHCPCAP. file) an IP address to a client and uses DHCP to pass the assigned address to the client. Use static allocation in an error-prone environment where it is desirable to manage IP address assignment outside of the DHCP functionality.
Finite	BOOTP-only	Enhancement provided by the JOIN software product. The DHCP server assigns an IP address from the pool to the BOOTP client and defines a lease time based on certain parameters you define in the SERVER.PCY file. When the lease expires, the DHCP server pings the IP address. If the server receives a reply, it extends the lease and does not offer the address to a new client. If not, the address is free and can be assigned to a new client.

Section 7.3 explains how to configure the different types of addressing for clients on your network.

The typical network uses a combination of static and dynamic DHCP addressing. As the local system manager or network administrator, you can apply any of the IP addressing methods as appropriate for your specific policies and environment.

7.1.4 Understanding the Relationship Between DHCP and BOOTP

From the client's perspective, DHCP is an extension of the BOOTP functionality. DHCP allows existing BOOTP clients to operate with DHCP servers without the need to change the clients' initialization software.

Based on the format of BOOTP messages, the DHCP message format:

• Captures the BOOTP relay agents and eliminates the need to have a DHCP server on each physical network segment.
• Allows existing BOOTP clients to operate with DHCP servers.
  Messages that include a DHCP message-type option are assumed to have been sent by a DHCP client. Messages without the DHCP message-type option are assumed to have been sent by a BOOTP client.

However, DHCP improves the BOOTP-only functionality as follows:

• DHCP allows the serial reassignment of network addresses to different clients by assigning a network address for a finite lease period.
• DHCP allows clients to acquire all of the IP configuration parameters they need to operate.

With BOOTP, a client is identified by its unique Media Access Control (MAC) address that is associated with the network adapter card.

DHCP uses a Client Identifier (ID) to uniquely identify the client and associate it with a lease: The client creates the Client ID from one of the following types of addresses:

• The MAC address.
• A variation of the MAC address. For example, Windows 95 and Windows NT create the Client ID by prepending the hardware type to the hardware address.

If the client does not include a Client ID in the request, the server uses the client's MAC address.

7.2 Configuring the DHCP Server

To configure the DHCP server on an OpenVMS host, you need to do the following tasks:

Task	Go to...
Enable DHCP on your system and set up DHCP files and databases.	Section 7.2.1
Set up the NETMASKS. file, if appropriate.	Section 7.2.3
Define IP addressing.	Section 7.2.4

In addition, you can:

• Define additional client information ( Section 7.3)
• Set up a cluster environment for DHCP server failover ( Section 7.5)
• Define DHCP server parameters ( Section 8.3)

To enable DHCP initially, run TCPIP$CONFIG and choose DHCP from the Server Components menu by entering the following command:

$ SYS$STARTUP:@TCPIP$CONFIG

The configuration procedure asks if you want to convert your existing BOOTP entries to your DHCP database:

Do you want to rollover old-style BOOTP entries into the DHCP database? [Y]

If you answer yes: The TCPIP$DHCP_BOOTPTODHCP.COM procedure tries to locate the existing BOOTP database. Once it locates a file, the configuration procedure asks you to confirm its selection or make a new selection:

Name of file to use for old-style BOOTP: SYS$SYSTEM:TCPIP$BOOTP.DAT Press return or enter new file name:

The configuration procedure takes the file and does the following:

• Converts any existing BOOTP information to the appropriate DHCP format in the new DHCPCAP. configuration file.
• Sets up the DHCP server to support BOOTP clients.
• Sets up permanent leases for existing BOOTP clients.

During TCPIP$CONFIG, all DHCP-related files are placed in the SYS$SYSDEVICE:[TCPIP$DHCP] directory unless you define the logical name TCPIP$DHCP_CONFIG (see Table 7-4).

If you answer no: The new DHCP configuration file DHCPCAP. remains empty, and your BOOTP clients will not be served.

TCPIP$CONFIG invokes the command procedure SYS$MANAGER:TCPIP$DHCP_SETUPCOMMANDS.COM, which defines the GUI Server Management Console and DHCP utilities as OpenVMS foreign commands.

One of the DHCP utilities that is defined in TCPIP$DHCP_SETUPCOMMANDS.COM is the TCPIP$DHCP_SIGNAL utility, which provides interprocess signaling in a manner similar to the UNIX kill signal delivery utility.

The commands available for the TCPIP$DHCP_SIGNAL utility are as follows. PRMMBX and SYSNAM privileges are required to run TCPIP$DHCP_SIGNAL. (This is the image behind the DHCPSIGTERM, DHCPSIGHUP, and DHCPSIGUSR1 commands.)

Command	Description
dhcpsighup	Causes the ASCII configuration files to be read again, then flushes the binary databases.
dhcpsigterm	Causes an orderly shutdown of DHCP.
dhcpusr1	Causes a dump of the ASCII configuration files, then flushes the binary databases.

7.2.2 Returning to the BOOTP-Only Configuration

You can return to a BOOTP-only configuration at any time. Further, you can use the previous TCPIP$BOOTP.DAT database file and the client entries it contains. If you deleted the TCPIP$BOOTP.DAT file, you can create a new one and populate it with entries (see Chapter 8).

To enable BOOTP after you have configured your host for DHCP, run TCPIP$CONFIG and enable the BOOTP component from the Server Components menu. Your existing DHCP files will remain for future use, as appropriate.

7.2.3 Defining Network Masks

DHCP requires that you define network masks correctly. The standard netmask for each network class is as follows:

Network Class	Netmask
A	255.0.0.0
B	255.255.0.0
C	255.255.255.0

If your network is subnetted in a format that is not consistent with the standard A, B, or C netmask address, make sure you edit the SYS$SYSDEVICE:[TCPIP$DHCP]NETMASKS. file and include an entry for each network. Each entry in the file must include two fields: the network address and the netmask address. For example:

# Network Netmask 192.1.1.0 255.255.255.224 192.1.1.96 255.255.255.240

7.2.4 Using the GUI to Define IP Addressing

After you convert your existing BOOTP file to the new DHCPCAP. file as described in Section 7.2.1, you are ready to begin serving your existing BOOTP clients without any further changes.

This section explains how to use the GUI to configure dynamic, static, or finite IP addressing for any DHCP/BOOTP clients you add in the future, as appropriate.

7.2.4.1 Defining Dynamic IP Addressing

To use dynamic addressing, you define a pool of IP addresses for each subnet or network that you want the DHCP server to service. Dynamic addresses have an associated valid lease time that you specify. The lease time is specific to your network and your clients.

Use a short lease time if you have:

• Client systems that start up and shut down frequently
• A limited number of addresses for a large number of clients

A shorter lease time allows you to use your limited address pool more efficiently.

Use a longer lease time if you have:

• Client systems that seldom go down
• A large address pool for a few clients

To define dynamic IP addressing, do the following:

1. Invoke the GUI. Enter the utility program command as follows:

   $ dhcpgui

2. Click the Server/Security tab.
3. Choose IP Ranges, then choose [New IP Range].
4. Enter the subnet address.
5. Enter the DHCP server address.
6. Enter the range of IP addresses.

To define static IP addressing, specify a specific IP address for a specific MAC address as follows:

1. Start the GUI. Enter:

   $ dhcpgui

2. Click the Nodes tab.
3. Choose [New Record].
4. Enter the host name (Name).
5. Enter the MAC/hardware address. For example, 08:00:20:3f:12:4b.
6. Choose Hardware Type from Key Parameters. Enter the type of network on which the node resides. Enter the hardware type using the symbol or the type number as shown in Table 7-2.
7. Choose [Host IP Address].
8. Enter the Host IP address of the host computer for this node.

Table 7-2 Hardware Type Symbols and Numbers

Symbol	Number	Network Type
ethernet or ether	1	10 Mb Ethernet
ethernet3 or ether3	2	3 Mb experimental
ax.25	3	AX.25 Amateur Radio
protnet	4	Protnet proNET Token Ring
chaos	5	Chaos
token-ring,tr,ieee802	6	IEEE802
arcnet	7	ARCNET

7.2.4.3 Defining Static IP Addressing for DHCP Clients

Select static addressing if you want to assign a specific IP address with a permanent lease time to a DHCP client, and you do not want the client to be able to release this IP address. Also, select static addressing if you need to select an IP address that is not part of any IP address pool.

Selecting an IP address from outside an IP address pool allows the server to specify a permanent mapping between a DHCP client's MAC address and the desired IP address. A client can reuse and release any address within an IP pool.

To configure a specific, permanent address for a DHCP client, do the following:

1. Start the GUI. Enter:

   $ dhcpgui

2. Click the Server/Security tab.
3. Choose Active IP Snapshot, then choose [New Record].
4. Enter the MAC address.
5. Enter the MAC type.
6. Enter the MAC address length.
7. Enter an IP address that does not belong to any IP address pools.
8. Enter -1 (infinite lease) for the lease expiration.
9. Enter the Server IP Address.
10. If you want a name associated with the client, specify the client's host name and domain name.

If you set the "Use MAC addr as Client ID" parameter to True, the server uses the MAC address to uniquely identify the clients. The MAC address field might not be the actual MAC address of the client's network adapter. Clients that modify the structure of the MAC address before sending it to the server include:

• Windows 95, Windows NT, and Windows for Workgroups 3.11 with Microsoft TCP/IP
  On these platforms, the MAC Address is prefixed with the hardware type. The MAC type is 0 and the length is 7 (instead of 6). For example, if your Ethernet address is 11:22:33:44:55:66, you need to specify the following for the static IP mapping:
  • MAC Address: 01:11:22:33:44:55:66
  • MAC Type: 0
  • MAC length: 7
• FTP Software's OnNet client
  On this platform, the string "cid-" prefixes the MAC Address. The MAC type is 0 and the length is 16. For example, if your Ethernet address is 11:22:33:44:55:66, you need to specify the following for the static IP mapping:
  • MAC Address: cid-112233445566
  • MAC Type: 0
  • MAC length: 16