 |
Index for
Section 8 |
|
 |
Alphabetical
listing for C |
|
 |
Bottom of
page |
|
cluamgr(8)
NAME
cluamgr - Specify and manage cluster aliases
SYNOPSIS
/usr/sbin/cluamgr -a alias=alias,option[=value][,option]...
/usr/sbin/cluamgr {-f | -F file}
/usr/sbin/cluamgr -h
/usr/sbin/cluamgr -r option[,option]...
/usr/sbin/cluamgr -s [alias | all ]
OPTIONS
Specify only one of the primary flags (-a, -f, -F, -h, -r, -s) each time
you run cluamgr. You cannot combine these flags.
-a alias=alias,option[=value],...
The -a flag precedes a list of comma-separated attributes that
you use to specify, join, modify, and leave aliases. (Attributes
are separated only by commas; do not use spaces to separate
attributes within a list.)
You must specify a cluster alias on each cluster member that you
want to join that alias. For example, for three cluster members
to belong to an alias, you must run the cluamgr command on each
member.
Alias attribute options are member-specific. You must run the
cluamgr command on the cluster member whose alias attributes you
want to modify.
The following are valid options for the -a flag:
alias=alias
The name of a cluster alias, either in dot notation
(n.n.n.n), or as a hostname, to which the other
specified option[=value] attributes apply. When using
a hostname, make sure that there is an entry
associating the hostname with an IP address in
/etc/hosts before running cluamgr.
If alias=DEFAULTALIAS, the alias in question is the
default cluster alias: the only alias whose hostname is
the name of the cluster. (The default cluster alias is
specified during cluster installation. There is only
one default cluster alias per cluster, and all cluster
members are automatically members of this alias.)
The following command line specifies, but does not join
an alias. This means that the cluster member can route
for the alias, but will not be the end recipient of any
packets addressed to that alias:
# cluamgr -a alias=testalias
join This cluster member joins the specified alias. In
order to join an alias, you must specify it. For
example, to join an alias named testalias:
# cluamgr -a alias=testalias,join
leave This cluster member leaves the specified alias. This is
the opposite of joining an alias.
If the member is an alias router, it will not receive
any packets as a final destination, but it will still
forward packets to current members of the alias (if
this is the only member available, it will still accept
packets even when selw=0).
If this member runs gated under aliasd control, setting
rpri=0 and restarting alias routing (cluamgr -r start)
will remove routing entries for this member from its
/etc/gated.conf.membern file.
Note that leaving a cluster alias does not delete that
alias from the system.
mask=n.n.n.n
Optional subnet mask used when advertising net routes
for virtual subnets.
When a cluster alias IP address resides in a virtual
subnet, the cluster alias daemon advertises both a host
route to the alias address and, if virtual=t, a net
route to the virtual subnet containing the alias
address.
If a subnet mask is not specified, the alias daemon
uses the network mask of the first interface through
which the virtual subnet will be advertised. Therefore,
if all the network interfaces on the system use the
same network mask, and if you want the virtual subnet
to use this mask, you do not have to specify a
mask=n.n.n.n flag when defining an alias address in a
virtual subnet.
There are two reasons for specifying a subnet mask:
· The network interfaces on the system do not use
the same mask.
· The network interfaces on the system use the same
mask, but you want to use a different mask for the
virtual subnet.
rpri=n An integer value from 0 to 100 inclusive, representing
this alias member's router priority (rpri). When a
cluster alias IP address is configured in a common
subnet, router priorities from 1 to 100 establish a
hierarchy within the members of that alias. The member
with the highest rpri value for a given alias will act
as proxy ARP master for that alias and will respond to
ARP requests for the alias address.
Only one member at a time can act as an alias's proxy
ARP master. If more than one member shares the highest
rpri value, the cluster alias software selects one of
these members to act as proxy ARP master. Should the
member, or members, with the highest rpri value become
unavailable, the member with the next highest rpri
value becomes the proxy ARP master for the alias.
When a cluster has more than one cluster alias, you can
use router priority to spread the routing overhead for
aliases among cluster members.
This option sets the router priority only for the
current member. It does not affect the router priority
for other cluster members of the same alias. The rpri
option does not control which members will broadcast
host or network routes for aliases. The rpri option is
irrelevant for an alias whose address is in a virtual
subnet.
The default router priority is 1.
A value of 0 is a special case. If you set rpri=0 for
an alias on a member system, that system will not route
cluster alias traffic into the cluster unless two
conditions are met: all active members of that alias
have rpri=0, and that system is elected to provide the
proxy ARP response for the alias.
A value of 100 is also a special case. For rpri values
of 1-99, aliasd advertises a host route cost metric of
14. For an rpri value of 100, aliasd advertises a host
route cost metric of 10. When aliasd advertises a host
route with a cost metric of 10, RIP clients will
replace an existing host route to an alias (with a cost
metric of 14) with the new lower-cost route.
selp=n An integer value from 1 to 100 inclusive, representing
this alias member's selection priority. The selection
priority determines the order in which members of an
alias receive new connection requests. The selection
priority establishes a hierarchy within the members of
an alias. Connection requests are distributed among
those members sharing the highest selection priority
value. If an alias has three members, two with selp=10
and one with selp=5, no connection requests or messages
are given to the selp=5 member as long as either of the
selp=10 members is available.
You can use selection priority values to set up a
failover order for members of a particular cluster
alias.
This option sets the selection priority only for the
current member. It does not affect the selection
priority for other cluster members of the same alias.
The default selection priority is 1.
selw=n An integer value from 0 to 100 inclusive, representing
this alias member's selection weight. The selection
weight indicates the number of connections (on average)
this member is given before connections are given to
the next alias member with the same selp value. (The
selp value determines the order in which members are
eligible to receive requests or messages; the selw
value determines how many requests or messages a member
gets once it is eligible.)
Selection weight applies only to applications that are
registered as multi-instance applications, either
through an entry in /etc/clua_services or as the result
of a call to clua_registerservice().
As an example, assume that there are two members in an
alias with the same selp value: the first has selw=3
and the second has selw=2. The first will receive three
connections, then the second will receive two
connections, then the first will receive three
connections, and so on.
If all alias members have the same selection weight,
incoming connections to this alias are distributed
equally among all alias members.
If selw=0, no incoming connections are given to this
member. This is useful when the member will handle
only outbound connections. For example, in a mixed
cluster, a system without the compute power of the
other cluster members might have selw=0 for the default
cluster alias; the system is not burdened with incoming
requests but mail sent from it is labeled as coming
from the cluster.
Another reason for setting selw=0 is when an alias
member is doing compute-intensive work and you do not
want it to spend time handling connection requests.
This option sets the selection weight only for the
current member. It does not affect the selection weight
for other cluster members of the same alias.
The default selection weight is 1.
virtual[={t|f}]
This option enables or disables advertising a network
route for the virtual subnet in which the alias address
resides.
If virtual or virtual=t, this cluster member will
advertise a network route to the virtual subnet that
contains the alias's IP address. The following two
commands both set the virtual flag for an alias named
virtualalias:
# cluamgr -a alias=virtualalias,join,virtual=t
# cluamgr -a alias=virtualalias,join,virtual
The virtual option indicates that this cluster member
should use gated to advertise a network route for this
alias's virtual subnet.
Do not enable virtual for an alias on a common subnet.
Doing so can result in the cluster acting as a conduit
for traffic between subnets, instead of functioning
correctly as a router.
If you do not want a cluster member to act as a router
for a virtual subnet, it is perfectly legal to not
specify virtual for an alias that is in a virtual
subnet. For example, to join but not route for an alias
in a virtual subnet:
# cluamgr -a alias=virtualalias,join
Once you specify the virtual=t option, you do not have
to specify it again when you modify another option that
is associated with the alias. For example, you can
modify an alias's selp, selw, or rpri values without
affecting that alias's virtual setting.
The virtual=f option disables network route advertising
for the virtual subnet that contains this alias IP
address.
-f Load (or reload) the /etc/clua_services configuration file on
this member. If you modify the clua_services file, run cluamgr -f
on each cluster member in order for the modifications to take
effect the next time a service starts. Note that reloading the
clua_services file does not affect currently running services.
For example, the telnet service is started by inetd from
/etc/inetd.conf. If you modified the service attributes for
telnet in /etc/clua_services, you would have to run cluamgr -f
and then stop and restart inetd in order for the changes to take
effect. Otherwise the changes will take effect at the next
reboot.
-F [file] Load (or reload) the specified configuration file on this member.
The F flag gives you the option of loading a configuration file
whose name is not /etc/clua_services.
-h Display help text.
-r option[,option]...
The -r flag precedes a list of comma-separated options that you
can use to start and stop routing for cluster aliases, or to
modify the behavior of the alias daemon, aliasd. The options do
not start or kill the alias daemon.
Note
Under normal circumstances you should not use these
cluamgr -r options to modify the daemon's behavior. The
daemon is designed to operate correctly from its
startup file, /sbin/init.d/clu_alias. The daemon gets
information from the cluster alias subsystem and
automatically updates routes as needed. These cluamgr
options are provided in case you need to modify daemon
behavior; for example, to help diagnose a problem with
alias routing on a cluster member.
Each pair of -r options acts like toggle switch to turn a
specific aliasd behavior on or off; for example: start, stop;
gated, nogated; enable=interface, disable=interface; resvok,
noresv; and vmac, novmac. Once an option is specified, it is in
effect for the life of the daemon or until its opposite is
specified. (At boot time, the internal defaults for the daemon
correspond to the following cluamgr -r options: start, gated,
enable, noresv, and novmac.)
Options are passed to the alias daemon in the order they are
passed to cluamgr. You can specify behaviors before starting
routing so they are in effect from the time routing starts.
The following are valid options for the -r flag:
[start | stop]
If stop is specified, stop alias routing operations on
this cluster member. If gated is the routing daemon on
this member, the alias daemon stops gated, and then
restarts gated with /etc/gated.conf as its
configuration file. If this member is also the proxy
ARP router for an alias (this member has the highest
rpri value in the cluster for the alias), the cluster
member with the next highest rpri value becomes the
proxy ARP router for the alias. You cannot combine the
stop option with any other -r options. For example,
cluamgr -r stop,start will return an error message.
When start is specified, if routing was stopped by a
cluamgr -r stop command, resume routing. If routing is
already started, check with the kernel and update
routing information. (If you specify additional aliases
after starting routing, you must reissue the cluamgr -r
start command in order to route for these new aliases.)
Once routing is started, you do not have to specify the
start option again in order to specify additional -r
options. For example:
# cluamgr -r start
# cluamgr -r resvok
In the previous example, routing is started, followed
by the resvok option. In this case, until the aliasd
daemon receives the resvok option, it automatically
uses the default for that behavior, noresvok. If you
want a behavior in effect from the moment routing is
started with a cluamgr -r start command, do one of the
following:
Issue the other routing option first. For example:
# cluamgr -r resvok
# cluamgr -r start
Put the option before the start option in a list of
routing options (options are processed in order). For
example:
# cluamgr -r resvok,start
Specify the option as a value for the
CLUAMGR_ROUTE_ARGS configuration variable either in a
cluster member's /etc/rc.config file, or in the shared
/etc/rc.config.common file (if all cluster members
should use the option). For example, to set
CLUAMGR_ROUTE_ARGS to resvok in /etc/rc.config.common:
# rcmgr -c set CLUAMGR_ROUTE_ARGS resvok
This is the preferred approach for options that you
want to invoke each time a system boots.
[gated | nogated]
These options affect how the gated daemon is used with
regard to cluster aliases.
If gated is specified, the aliasd daemon uses gated to
advertise host routes for all alias addresses, and
network routes for virtual subnets. The aliasd daemon
creates a /etc/gated.conf.membern file for this cluster
member, stops the gated daemon, and restarts gated
using /etc/gated.conf.membern as gated's configuration
file. This is the default behavior when you run cluamgr
-r start (and at boot time). The only time you would
explicitly specify gated is if you had previously
specified nogated and now want to return aliasd to its
default behavior.
If nogated is specified, the aliasd daemon modifies the
member's /etc/gated.conf.membern file to add route
entries as though the gated option were specified, but
it does not restart the gated daemon using
/etc/gated.conf.membern. This lets an experienced
network administrator rename and customize the
configuration file (for example, to use OSPF instead of
RIP, or to use static routing), and restart the gated
daemon manually.
Note that in both cases (gated and nogated), aliasd
does not modify a member's /etc/gated.conf file. The
difference between the gated and nogated options is
whether aliasd stops the gated daemon and restarts it
using /etc/gated.conf.membern as gated's configuration
file. The default behavior at boot time is for aliasd
to stop gated and restart it using
/etc/gated.conf.membern.
The following information is supplied only for
experienced network administrators. The aliasd daemon
supports only the Routing Information Protocol (RIP).
If you have a customized routing environment, need use
a routing protocol such as OSPF with the cluster alias,
or use static routing, you can start alias routing on a
cluster member with the nogated option. Then use the
version of /etc/gated.conf.membern created by cluamgr
-r nogated,start as the basis for edits to a customized
gated configuration file. You are responsible for
correctly merging the cluster alias information from
the /etc/gated.conf.membern file into your customized
configuration file. Set the CLUAMGR_ROUTE_ARGS
variable to nogated in /etc/rc.config.common:
# rcmgr -c set CLUAMGR_ROUTE_ARGS nogated
This ensures that aliasd will not point gated at
/etc/gated.conf.membern when a cluster member reboots.
(You can override this clusterwide nogated setting for
an individual cluster member by setting
CLUAMGR_ROUTE_ARGS to gated in that member's rc.config
file.)
When the nogated option is in effect, aliasd does not
advertise routes to cluster aliases on virtual subnets.
These aliases will not be reachable unless configured
as static routes on clients.
[enable=interface,...] | [disable=interface,...]
By default, alias IP addresses are published over all
broadcast capable network interfaces, with the
exception of the cluster interconnect and the loopback
interfaces. The default behavior is to enable the
advertising of alias addresses through these broadcast
capable interfaces.
If the disable option is specified, aliases are not
advertised through the specified interface or
interfaces. Use the disable option to explicitly
disable the advertising of alias addresses on a
particular interface. The following command disables
the advertising of cluster aliases on two interfaces:
# cluamgr -r disable=tu0,disable=tu1
Issuing the enable option in later cluamgr commands
flushes the list of restricted interfaces and allows
the advertising of alias addresses through these
interfaces.
You cannot combine the enable and disable options. For
example, cluamgr -r enable=tu0,disable=tu1 will return
an error message.
[resvok | noresv]
By default, aliasd operates as though the noresv (no
reserved) option is specified. The daemon does not
advertise alias IP addresses that are configured in any
of the reserved private address spaces defined in RFC
1918 (such as 10.0.0.0, 172.16.0.0, and 198.168.0.0).
If the resvok (reserved OK) option is specified, aliasd
advertises alias IP addresses that reside in reserved
subnets, and also advertises alias IP addresses on
network interfaces configured within these reserved
subnets. This is useful when a company uses a reserved
subnet for something other than cluster interconnect
addresses; for example, the entire company intranet
resides in a reserved subnet. In this case, unless
resvok is specified, aliasd will not advertise net or
host routes to alias addresses.
[vmac | novmac]
By default, aliasd operates as though the novmac (no
virtual media access control (vMAC)) addressing option
is specified.
If vMAC support is enabled, when a cluster member
becomes the proxy ARP master for a cluster alias, it
creates a virtual MAC address for use with that cluster
alias. A virtual MAC address consists of a prefix (the
default is AA:01) followed by the IP address of the
alias in hexadecimal format. For example, the default
vMAC address for an alias whose IP address is
16.140.112.209 is AA:01:10:8C:70:D1:
Default vMAC prefix: AA:01
Cluster Alias IP Address: 16.140.112.209
IP address in hex. format: 10.8C.70:D1
vMAC for this alias: AA:01:10:8C:70:D1
When another cluster member becomes the proxy ARP
master for this alias, the virtual MAC address moves
with the alias so that a consistent MAC address is
presented within the common subnet for each cluster
alias.
When configuring vMAC support, configure all cluster
members identically. For this reason, set vMAC
configuration variables in /etc/rc.config.common.
To enable vMAC support, run the following rcmgr command
on any cluster member to put an entry in
/etc/rc.config.common:
# rcmgr -c set VMAC_ENABLED yes
This ensures that vMAC support is automatically enabled
at boot time. However, because setting this variable
only affects a member when it reboots, the currently
running cluster does not have vMAC support enabled.
To manually enable vMAC support for the currently
running cluster, enter the following command on all
cluster members:
# cluamgr -r vmac
You do not have to add the cluamgr -r vmac command to
each cluster member's /etc/clu_alias.config file.
Running the cluamgr -r vmac command manually on each
member enables vMAC support now; setting VMAC_ENABLED
to yes in the shared /etc/rc.config.common file
automatically enables vMAC support at boot time.
To disable vMAC support once it is enabled, run the
following rcmgr command on any cluster member to put an
entry in /etc/rc.config.common:
# rcmgr -c set VMAC_ENABLED no
Then run the following cluamgr command on all cluster
members:
# cluamgr -r novmac
To change the default AA:01 vMAC prefix, run the
following rcmgr command to put an entry in
/etc/rc.config.common:
# rcmgr -c set VMAC_PREFACE xx:xx
Where xx:xx is the new prefix.
-s [alias | all]
Print status of alias to standard output. To print the status of
the default cluster alias, you can specify the alias by its host
name or use the keyword DEFAULTALIAS. For example:
# cluamgr -s DEFAULTALIAS
The command cluamgr -s all prints status for all aliases that
have active members. If no members have joined an alias, no
status is printed for that alias.
DESCRIPTION
The cluamgr utility configures and initializes the in-memory cluster alias
database for a cluster member. Each cluster member that intends to
participate in an alias must run the cluamgr utility to specify and join
that alias.
Cluster alias definitions are volatile. Aliases must be configured using
cluamgr each time a system boots. At boot time, each member's
/etc/clu_alias.config file is called from the /sbin/init/clu_alias script.
The /etc/clu_alias.config file is a CDSL that points to a member-specific
clu_alias.config file. Each member's clu_alias.config file contains the
cluamgr commands that configure aliases, including the default cluster
alias, for that member. Although you can manually edit this file, the
preferred interface is the through the SysMan Menu.
When a cluster member invokes the cluamgr utility, the utility registers
its currently enabled port database based on the contents of the
/etc/clua_services file. This involves sending a list of all the ports that
can receive connections using the alias address to all alias router nodes
for the alias.
How a member participates in advertising the alias address depends upon
whether the address is in a common subnet or a virtual subnet as follows:
common subnet
To configure alias IP addresses in an existing subnet, use the
cluamgr utility with the -a option, specifying the -r option to
start and stop routing. Using a common subnet for cluster aliases
works well when the cluster is connected to only a single local
area network, and that network is managed as a single IP address
domain.
Cluster alias routing in a common subnet is based on proxy ARP
support. The node elected to advertise the alias configures the
alias's IP address to be advertised using proxy ARP over a
network interface in the same subnet. Other cluster members that
join the alias and are configured to be routing members become
capable of taking over the proxy ARP function if necessary. The
rpri attribute determine the order in which cluster members that
have specified an alias can act as the router for that alias.
When a designated alias router or network interface fails, other
potential routers are notified. They then elect a new router for
the alias.
virtual subnet
A cluster alias resides in a virtual subnet if its address is in
a subnet that is not associated with a physical interface. A
virtual subnet must not have any real systems in it.
Cluster alias routing in a virtual subnet uses host routes. The
alias configuration process modifies the gated.conf.membern
configuration file to advertise the alias address.
EXAMPLES
In the following examples, remember that cluamgr commands affect only the
cluster member on which they are run. For example, setting the selection
weight for an alias on this system does not affect the selection weight
assigned to the alias by another cluster member.
To specify and join alias testalias, setting the selection weight, but
taking the default values for selection priority and routing priority:
# cluamgr -a alias=testalias,join,selw=4
To change the selection weight for the default cluster alias (without
changing any of its other attributes):
# cluamgr -a alias=DEFAULTALIAS,selw=2
To specify, but not join, alias virtualalias, whose IP address is in a
virtual subnet; explicitly define the subnet mask used when advertising a
net route to this virtual subnet:
# cluamgr -a alias=virtualalias,virtual,mask=255.255.255.0
To start routing for the previously specified aliases:
# cluamgr -r start
To specify and join another alias, specifying non-default values for
selection weight and selection priority. Set the router priority so this
node will be the router for this alias:
# cluamgr -a alias=16.140.112.209,join,selw=2,selp=10,rpri=100
To start routing again in order to add alias 16.140.112.209 to the list of
aliases for which this node will route. (Note that you do not have to stop
and restart routing.)
# cluamgr -r start
FILES
/etc/clua_services
Static configuration file where services using well-known ports
are registered for cluster alias connections.
/sbin/init.d/clu_alias
Startup script for the default cluster alias.
/etc/clu_alias.config
Startup commands for cluster aliases.
SEE ALSO
Commands: clua_active(8), rcmgr(8)
Daemons: aliasd(8)
Functions: clua_error(3), clua_getaliasaddress(3), clua_getaliasinfo(3),
clua_getdefaultalias(3), clua_isalias(3), clua_registerservice(3),
clusvc_getcommport(3)
Files: clua_services(4), clu_alias.config(4), exports.aliases(4)
Miscellaneous: sys_attrs_clua(5)
|
Index for
Section 8 |
|
|
Alphabetical
listing for C |
|
 |
Top of
page |
|