listing for C
caa_start - Starts resources that have been registered with Cluster
Application Availability (CAA).
/usr/sbin/caa_start resource_name [...] [-c cluster_member] [-q] [-f]
/usr/sbin/caa_start -all [-q]
/usr/sbin/caa_start [USR_attribute_name=value] [...] resource_name [-c
Name(s) (as specified in a resource profile) of one or more resources
to be started. The resource must be registered with CAA.
Starts each indicated resource on the specified member if the cluster
member is allowed by the placement policy and resource dependencies.
If the cluster member specified is not allowed by the placement policy
and resource dependiencies, the caa_start command fails. If the
specified member is not available, the command fails. If one of the
resources specified fails to start, the command will still attempt to
start other resources listed on the command line.
Starts all registered CAA application resources on active cluster
members, according to their placement policies and required resource
-q Runs the caa_start command in quiet mode; no messages are displayed on
-f Forces starting of a specified application resource if all required
resources for that resource are available or can be started. Any
application resources in the specified resource profile's required
resource list are started if not currently running or are relocated if
they are running on another cluster member. Use this option with only
one specified application resource at a time.
The caa_start command sets an application resource target state to ONLINE
and attempts to start specified registered application resources. The
target state is the state that CAA attempts to achieve.
The caa_start command also can be used to reset the target state of a non-
application resource to "ONLINE" if it has been set to "OFFLINE" due to the
failure count exceeding the failure threshold. Any application resource
that is dependent upon this resource that has been stopped due to this
resource failing is restarted as well.
If one or more user-defined attributes have been defined for application
resources, you can specify values for these attributes when starting an
application with caa_start. The specified value is passed to the action
script as an environment variable with the attribute name, see caa(4).
If the action script start entry point fails to execute successfully, the
stop entry point is run. If the stop entry point fails to execute
successfully, the application resource state is marked as "UNKNOWN" and
relocation attempts are aborted. If the stop entry point succeeds, the
state is set to "OFFLINE". The target state remains "ONLINE" however, so
subsequent cluster member failures or restarts may attempt to restart the
application. If you have not specified one member to start the application
on and there are available cluster members that satisfy the placement
criteria, CAA will attempt to start the application on one of the available
You must stop a resource in the state "UNKNOWN" before restarting it (see
caa_stop(8)). You should also investigate why the action script is failing.
The caa_start command will start dependent resources as defined in resource
profile REQUIRED_RESOURCE fields.
You must have root privilege to use this command.
After a registered resource is started, you must use the caa_stop command
to end its execution.
The caa_start command displays status messages on the screen. If there are
troubles starting an application, feedback will be displayed.
If network, tape, or media changer resources are defined as required
resources for an application resource that is started, messages saying that
CAA is attempting to start these non-application resources will appear.
CAA is not actually attempting to start these resources but is testing the
state of the devices. These messages can be ignored.
An application resource has the working directory set to
Standard error and standard output from a resource action script invoked by
caa_start are redirected to the standard error and standard output for
caa_start. Note that if the CAA daemon starts an application, standard
error and standard output of the action script is lost. Within an action
script, you can check for user invocation of the action script using reason
codes. For more information on reason codes see caa(4).
1. The following example starts an application:
Attempting to start `clock` on member `polishham`
Start of `clock` on member `polishham` succeeded.
2. The following example starts an application on a specific cluster
/usr/sbin/caa_start -c provolone clock
Attempting to start `clock` on member `provolone`
Start of `clock` on member `provolone` succeeded.
3. If a user-defined attribute USR_DEBUG has been defined, see caa(4),
the following example runs the start entry point of the action script
with the USR_DEBUG environment variable set to true, overriding any
value set in the resource profile.
In the corresponding action script, if you add the following line to
the appropriate section of the action script you can view the value:
Then execute the command:
# caa_start USR_DEBUG=true database
Attempting to start `database` on member `provolone`
Start of `database` on member `provolone` succeeded.
The following exit values are returned:
0 Successful completion.
A failure has occurred.
A corresponding text message will be given for the following
No root privilege.
The CAA daemon is not running.
The application is running.
A start action script is not found.
The cluster member was specified and it was not available.
A timeout was reached before the startup was unsuccessful.
Commands: caa_stat(1), caa_profile(8), caa_register(8), caa_unregister(8),
TruCluster Server Cluster Administration
listing for C