Main Server (main)
|
Identifies the cluster with the main server for a failover pair.
|
Backup Server (backup)
|
Identifies the cluster that contains the backup server for a failover pair.
|
Scope Template (scopetemplate)
|
Associates a scope template with a specified failover pair.
|
Failover Settings
|
failover
|
Enables failover configuration. If you disable this attribute, you turn off failover on attached subnets without changing
configuration fundamentals.
|
mclt
|
Sets the maximum client lead time in seconds. This attribute controls how far ahead of the backup server that you can make
the client lease expiration. You must define this value on both the main and backup servers, and make sure the value is identical
on both servers.
|
backup-pct
|
Controls the percentage of available addresses that the main server sends to the backup server. Set this value on the main
server. If it is set on a backup server, it is ignored (to enable copying of configurations). Unless you explicitly set this
value on a scope and you disable load balancing, the value set here becomes the default value.
|
dynamic-bootp-backup-pct
|
Determines the percentage of available addresses that the main server sends to the backup server for scopes on which dynamic
BOOTP is enabled. If defined, it must be defined on the main server. If it is defined in a backup server, it is ignored (to
enable copying of configurations). If it is not defined at all or the value is 0, the backup-pct is used instead. This parameter is separate from "backup-pct" because if dynamic BOOTP is enabled on a scope, a server will
never, even in PARTNER-DOWN state, grant leases on addresses that are available to the other server because they can never
safely be assumed to be available again.
The MCLT has no meaning for dynamic BOOTP leases.
|
load-balancing
|
Determines whether load balancing (RFC 3074) is enabled on a failover pair. The default is disabled. When enabled, the backup-pct is ignored and the main and backup server evenly split the client load and available leases for all scopes in the failover
relationship (that is, as if backup-pct were configured at 50%).
|
rebind-limit
|
Sets the limit on T2 (rebinding time) beyond T1 (renewal time) that the failover partner that would not normally respond
to a REQUEST, RENEW, or REBIND request uses when responding on behalf of the partner.
When set to a non-zero value, this expedites returning the client to being serviced by the failover partner fairly quickly
once the client starts renewing and failover is in NORMAL state.
As clients will typically re-transmit a renewal request after a few seconds and retry using an exponential background algorithm,
the reasonable values for rebind-limit are likely in the range of 60 to 600 seconds.
Note: For DHCPv6, this can also cause a server to drop a RENEW request it would have serviced based on the server-id option (if the failover partner is operational and should service the client).
|
safe-period
|
Controls the safe period, in seconds. It does not have to be the same on both main and backup servers. It only has meaning
if use-safe-period is enabled.
|
use-safe-period
|
Controls whether a server can enter PARTNER-DOWN state without an operator command. If disabled, a server never enters PARTNER-DOWN
without an operator command.
|
Relay Health Check
|
relay-health-check
|
Sets whether health checking is enabled and for which protocols when failover communication is in the NORMAL state. It can
be set to disabled (default), v4-only, v6-only, or both.
|
rhc-monitor-first-relay
|
Determines if DHCPv6 traffic should be monitored based on innermost (first) relay or outermost relay (from where DHCP server
receives the message). It is recommended to monitor traffic based on outermost relay, as it will reduce the number of relays
to be monitored and hence reduce memory usage.
|
rhc-unresponsive-time
|
Sets the health check unresponsive time (in seconds). This is the minimum amount of time over which this server must fail
to receive a (DHCPv4) DHCPREQUEST or (DHCPv6) Request packet addressed to another server before it starts responding to requests
on behalf of its partner as it assumes there are communication issues between the relay and that partner.
|
rhc-request-count
|
Sets the health check partner request count. This is the number of (DHCPv4) DHCPDISCOVER or (DHCPv6) Solicit requests the
partner could have responded to before this server starts responding to client requests that it would normally not respond
to.
|
rhc-restart-time
|
Sets the health check restart time (in seconds). When no (DHCPv4) DHCPDISCOVER or (DHCPv6) Solicit requests to which the partner
would respond are received for this interval of time, the times and counts for the monitored relay will be restarted. This
is to assure that the health check is based on very recent data.
|
rhc-warning-interval
|
Sets the minimum interval between logging of warning messages when the server is responding for its partner because of likely
communication issues between its partner and a relay agent.
|
rhc-elapsed-time
|
Sets the minimum time a client must report in the DHCPv4 secs field or DHCPv6 elapsed time (8) option before the server will
respond for its partner when the communication between the partner and relay is determined to be down.
If set to 0 (not recommended), the server will respond to all requests.
|
rhc-server-preference
|
Sets the DHCPv6 server preference to use when responding to a Solicit on behalf of the failover partner. Only used if a server
preference option has been configured and the rhc-server-preference value is lower.
|
rhc-response-time
|
Sets the health check response time (in seconds). This is the amount of time this server will respond to (DHCPv4) DHCPDISCOVER
or (DHCPv6) Solicit requests on behalf of its partner as it assumes there are communication issues between the relay and that
partner. This is to assure that this server does not respond forever if clients keep accepting offers from this server.
If set to 0, this server will continue to respond to clients until it receives a (DHCPv4) DHCPREQUEST or (DHCPv6) Request
packet addressed to another server.
|
Failover Server Addresses
|
main-server
|
Controls the IPv4 address used for the failover protocol on the main server. If this value is unset, the address specified
for the main cluster is used. Cisco recommends setting this value only if the server is configured with different interfaces
for configuration management and clients requests.
This value may be set to 0.0.0.0 to disable use of IPv4 for failover communication.
If both IPv4 and IPv6 addresses are available, the servers will try both transports for the TCP connection and use whichever
comes up first.
|
backup-server
|
Controls the IPv4 address used for the failover protocol on the backup server. If this value is unset, the address specified
for the backup cluster is used. Cisco recommends setting this value only if the server is configured with different interfaces
for configuration management and clients requests.
This value may be set to 0.0.0.0 to disable use of IPv4 for failover communication.
If both IPv4 and IPv6 addresses are available, the servers will try both transports for the TCP connection and use whichever
comes up first.
|
main-ip6address
|
Controls the IPv6 address used for the failover protocol on the main server. If this value is unset, the address specified
for the main cluster is used. Cisco recommends setting this value only if the server is configured with different addresses
for configuration management and clients requests.
This value may be set to 0::0 to disable use of IPv6 for failover communication.
If both IPv4 and IPv6 addresses are available, the servers will try both transports for the TCP connection and use whichever
comes up first.
|
backup-ip6address
|
Controls the IPv6 address used for the failover protocol on the backup server. If this value is unset, the address specified
for the backup cluster is used. Cisco recommends setting this value only if the server is configured with different addresses
for configuration management and clients requests.
This value may be set to 0::0 to disable use of IPv6 for failover communication.
If both IPv4 and IPv6 addresses are available, the servers will try both transports for the TCP connection and use whichever
comes up first.
|