admfailoverip(1M) DG/UX R4.11MU05 admfailoverip(1M)
NAME
admfailoverip - manage IP Takeover
SYNOPSIS
admfailoverip -o add -h hostname -l local-interface-name -r remote-
interface-name -m failover-method [ -bsS ] ip-services-
name
admfailoverip -o delete -h hostname [ -S ] [ ip-services-name ... ]
admfailoverip -o modify -h hostname [ -n new-ip-services-name ] [ -l
new-local-interface-name ] [ -r new-remote-interface-
name ] -m failover-method [ -bsS ] ip-services-name
admfailoverip -o list [ -qv ] [ -h hostname ] [ ip-services-name ...
]
admfailoverip -o give -h hostname [ ip-services-name ... ]
admfailoverip -o take -h hostname [ -T ] [ ip-services-name ... ]
admfailoverip -o start [ -h hostname ] [ ip-services-name ... ]
admfailoverip -o stop [ -h hostname ] [ ip-services-name ... ]
admfailoverip -o sync -h hostname
DESCRIPTION
The admfailoverip command is used to manage the
/etc/failover/failoverip database, and to control the usage of the IP
address used to supply NFS services to clients. Servers in a dual-
initiator configuration can now provide a level of NFS and login
services failover by using IP Takeover. This is achieved by
providing the NFS and login services through an IP address that can
be cooperatively managed by all of the hosts in the configuration.
The IP Takeover functionality depends on the presence of a dual-
initiator configuration and the user having already set up physical
disk failover.
The admfailoverip command provides functions to maintain the
database, as well as functions to control the ownership of the IP
address. It is important to note that not all applications may
benefit from IP/Takeover. IP/Takeover was designed to meet the IP
failover requirements of a very small set of applications: Telnet,
rlogin, and NFS. Central to qualifying an application for use with
IP/Takeover is the concept of statelessness:
"An application is stateless with respect to IP/Takeover if and
only if, neither the client nor the server side of an application
employing the operation ever makes use, directly or indirectly, of
information that is specific to just one of the IP/Takeover servers"
For example, a client instance of an application that holds a pointer
into a server's virtual memory space is almost certainly not
stateless: if IP/Takeover occurred after the client had obtained such
a pointer, and the client subsequently attempted to present that
pointer to the backup server, it would almost certainly not point to
backup-server memory analogous to the primary-server memory to which
it originally referred.
The ip-services-name is the name of an IP address that is to be
associated with the physical disks to be failed over, or to be used
as the name that users will rlogin or telnet to. The ip-services-name
should not be the primary address of either of the systems in the
dual-initiator configuration. The ip-services-name should also be
limited to nine characters in length. When the address is started,
it is catenated with _<local-interface-name> to form the actual
interface name passed to the kernel. The kernel requires that this
name does not exceed 15 characters in length.
The IP/Takeover functionality described here is only supported for
the following network interfaces:
cien
dgen
hken
inen
pefn
Operations
add Add a failoverip database entry. This operation will add
the specified failoverip database entry, and can optionally
start the use of the IP address at this time.
delete Delete a failoverip database entry. This operation will
cause the specified IP address to be stopped (if it was
running), and the database entry to be deleted.
modify Modify a failoverip database entry. This operation is used
to modify the database entry and control restarting the IP
address with any new information.
list List failoverip database entries. The list operation
reports the following failoverip information to stdout:
the name of host to share floating IP address with
the floating IP address name to share
the local network interface name
the remote network interface name
flag indicating whether or not the floating address is to be
restarted when the system is rebooted
status flag indicating whether or not the floating IP address
is currently active on this host
the failover method used for the address
With the "verbose" format (-v), information is printed in
aligned columns with headers. With the "quiet" format (-q)
headers are suppressed and each ip database entry is
printed on a separate line. If both -q and -v are
specified, the output will be in "quiet" format.
give Give a floating ip address to another host. This operation
should be used in conjunction with the give operation of
the admfailoverdisk(1M) command. When a configuration is
using Machine Initiated Failover, the following command
should be added to the regain-pulse script that is to be
executed by the failovermon(1M) monitor to return control
of the floating IP addresses associated with the physical
disks that will be returned:
admfailoverip -o give -h <hostname>
take Take a floating ip address from another host. This
operation should be used in conjunction with the take
operation of the admfailoverdisk(1M) command. When a
configuration is using Machine Initiated Failover, the
following command should be added to the lost-pulse script
that is to be executed by the failovermon(1M) monitor to
take control of the floating IP addresses associated with
the physical disks that will be taken:
admfailoverip -o take -h <hostname> -T
start Start the floating ip address on this host. This operation
is performed after the entry has been added to the
failoverip database. It is also performed when a system is
booting. If the operation is performed during a system
reboot, the address will be "ping"ed to see if it is
already in use by another host in the configuration.
stop Stop the floating ip address on this host. This operation
is performed when a system is shutting down, or if you want
to stop the floating ip address.
sync Synchronize failoverip address databases. This operation
will copy the contents of the failoverip database to the
specified host.
Options
The following option is required for the add, delete, modify, give,
take, and sync operations, and is optional on the list, start, and
stop operations:
-h hostname
This is the primary name of the host that is the target of
the operation. In a dual-initiator configuration in which
the IP/Takeover will be used for NFS or rlogin/telnet
access, the hostname should be the name of one of the
systems in the configuration. This name should not be
confused with the ip-services-name, which is what will be
used to remotely mount file systems, and users will
rlogin/telnet to.
The following option can be used with the add, delete, or modify
operations:
-S Synchronize failover databases. This option will cause the
failover databases to be synchronized upon successful
completion of the current operation.
The following option can be used with the add or modify operations:
-l local-interface-name
This is the name of the network interface on the current
host that will be used to send and receive requests for the
specified IP address.
-r remote-interface-name
This is the name of the network interface on the remote
host that will be used to send and receive requests for the
specified IP address.
-m failover-method
Failover-method specifies the failover method used for the
Floating IP address. The argument can be either arp or
multicast. When no method is specified on an add
operation, the default is arp.
-s Start. This option instructs the admfailoverip command to
start the floating IP address after the failoverip database
entry has been added or modified.
-b Start on reboot. This option instructs the admfailoverip
command to start the floating IP address after the system
has been rebooted. If the system is rebooted and the
address is detected as active, it will not be started at
this time. If this option is used on a modify operation, it
will toggle the current value of the start on reboot flag
(i.e., change YES to NO, or NO to YES).
The following option can be used with the modify operation:
-n new-ip-services-name
This option is used to change the name of the current ip-
services-name of an failoverip database entry.
The following options can be used with the list operation:
-q Quiet. Produce an unformatted listing with no headers,
fields delimited by a single space.
-v Verbose. Produce a formatted listing with headers and
aligned columns. This option is the default.
The following option can be used with the take operation:
-T Trespass. Unconditionally take control of the floating IP
address. This option should be used if the host you are
taking the IP address from has failed.
EXAMPLE
To add an entry to the failover ip database on hosta for an IP
address services_a_b that can be managed between hosta and hostb, for
network controllers dgen0 (hosta) and dgen1 (hostb) you should
execute the following command line:
admfailoverip -o add -h hostb -l dgen0 -r dgen1 servicesab
To synchronize the databases of hosta and hostb, you should execute
the following command line:
admfailoverip -o sync -h hostb
To give control of the IP address services_a_b to hostb while logged
into hosta, you should execute the following command line:
admfailoverip -o give -h hostb servicesab
FILES
/etc/failover/failoverip.
DIAGNOSTICS
Warnings
None.
Errors
- An attempt was made to add an IP entry that already exists in
the failoverip database.
- An attempt was made to modify, delete, give, or take, an IP
entry that does not exist in the failoverip database.
- An attempt was made to take an IP address that is already in
use on this system.
- An attempt was made to add a failoverip entry on a system that
does not have physical disk failover set up.
- An attempt was made to specify a network interface that does
not exist, or is not supported for IP/Takeover.
Exit Codes
0 The operation was successful.
1 The operation was unsuccessful.
2 The operation failed due to access restrictions.
3 There was an error in the command line.
SEE ALSO
admfailoverdisk(1M), failovermon(1M), failover(4M), failoverd(1M),
appropriateprivilege(5), capdefaults(5).
NOTES
You must have appropriate privilege to perform all operations except
list. For systems supporting the DG/UX Capability Option,
appropriate privilege is defined as having one or more specific
capabilities enabled in the effective capability set of the user.
See capdefaults(5) for the default capabilities for this command.
On systems without the DG/UX Capability Option, appropriate privilege
means that your process has an effective UID of root. See the
appropriateprivilege(5) man page for more information.
When using sysadm and the admfailoverip command directly, certain
operations may take a while to complete. Operations that involve
determining the status of the IP address (add, delete, modify, list,
start, and stop) will attempt to ping the address to see if it is
active. This is done to determine which system is using the address,
and to prevent multiple systems from broadcasting the address.
Licensed material--property of copyright holder(s)