To create a message queue polling server definition
in the SAS Metadata Repository, perform the following steps:
-
In SAS Management Console,
select the Server Manager and then select
ActionsNew Server. The New Server Wizard appears.
-
Select
Message
Queue Polling Server and then click
Next.
-
Specify a name and an
optional description. Click
Next.
-
Specify your configuration
settings for the following fields:
specifies a command
that is used to invoke SAS and process messages. You can modify the
command to include invocation options.
In your SAS command
or in the script that you use to invoke SAS, you must specify a SAS
program file by using the –SYSIN option. The SAS program that
you specify should contain messaging code to read messages from the
queue and process the message contents.
select the credentials
that are used to start SAS server sessions. The credentials that you
specify must have permissions to access the resources, such as data
libraries, that your SAS program will access.
If you select
(None),
then the object spawner's credentials are used to start the session.
specifies the machine
where the polling server runs. The polling server must run on the
same machine as an object spawner that it is associated with.
specifies the queue
(or destination) that the polling server monitors for messages.
-
For JMS polling servers,
specify the following options on the
Polling tab:
specifies the maximum
number of concurrent server sessions.
specifies the number
of server sessions that are kept running. If you specify 0, then one
server session is running.
Note: If the value of
Minimum
sessions is greater than the value of
Maximum
sessions, then the value of
Maximum sessions is
used.
Queue polling process timeout
specifies the time
(in seconds) to wait for the server sessions to end when the object
spawner is shutting down. If you specify a value that is greater than
zero, and any server sessions are still running after the time has
elapsed, then the spawner terminates the sessions. If you specify
0, then there is no time limit for the server sessions to end. The
default value is 0.
It is recommended that
you use the default value of 0. Make sure that the code that is run
by your server sessions checks for stop messages. See
Checking for Stop Messages.
Note: If you specify a value that
is greater than zero, then the spawner log might contain an error
message, "Failed to locate the server indicated in the kill request"
for each server session that ended normally. These messages do not
indicate a problem.
Note: If you specify a value that
is greater than zero, then the spawner always waits the full time-out
period when shutting down. For example, if you specify 30 seconds
as the time-out value, then the spawner always waits 30 seconds to
shut down, even if all of the server sessions end before 30 seconds.
If a spawner manages multiple polling servers, then the polling servers
are shut down sequentially. The time delay for shutting down the spawner
is cumulative.
For WebSphere MQ polling
servers, specify the following options on the
Polling tab:
specifies the maximum
ratio of messages to server sessions. If the message threshold is
exceeded, then the object spawner creates a new server session. The
default value is 10 (10 messages t o 1 server session).
For example, a polling
server is configured with a message threshold value of 10. The message
queue contains 21 messages, and two server sessions are running. Because
the ratio of messages per server session (10.5) is greater than the
threshold value (10), the object spawner creates a new server session.
specifies the interval
(in seconds) at which the server checks the depth of the message queue.
The default value is 10.
specifies the maximum
number of concurrent server sessions. If you specify 0, then an unlimited
number of server sessions can be created. The default value is 1.
specifies the minimum
number of server sessions that are running. The default value is 0.
Note: If you specify 0, then one
server session is created when the object spawner is started. Also,
the object spawner maintains at least one server session if there
are any messages in the queue.
Queue polling process timeout
specifies the time
(in seconds) to wait for the server sessions to end when the object
spawner is shutting down. If you specify a value that is greater than
zero, and any server sessions are still running after the time has
elapsed, then the spawner terminates the sessions. If you specify
0, then there is no time limit for the server sessions to end. The
default value is 0.
It is recommended that
you use the default value of 0. Make sure that the code that is run
by your server sessions checks for stop messages. See
Checking for Stop Messages.
Note: If you specify a value that
is greater than zero, then the spawner log might contain an error
message, "Failed to locate the server indicated in the kill request"
for each server session that ended normally. These messages do not
indicate a problem.
Note: If you specify a value that
is greater than zero, then the spawner always waits the full time-out
period when shutting down. For example, if you specify 30 seconds
as the time-out value, then the spawner always waits 30 seconds to
shut down, even if all of the server sessions end before 30 seconds.
If a spawner manages multiple polling servers, then the polling servers
are shut down sequentially. The time delay for shutting down the spawner
is cumulative.
-
For WebSphere MQ polling
servers, select the
WebSphere Options tab
and specify whether the MQ Server interface is used for monitoring
the queue depth. If you choose to use the MQ Server interface, then
the object spawner and the queue manager must be on the same machine.
-
Click
OK to
return to the New Server Wizard, and then click
Next.
-
Review your server settings,
and then click
Finish to create the server
definition.