The Z Common Data Provider Data Streamer streams operational
data to configured subscribers in the appropriate format. It receives the data from the System Data
Engine, splits it into individual messages when necessary, and sends the data to the
subscriber.
Before you begin
The Data Streamer can stream data to both on-platform and off-platform subscribers. To reduce general CPU usage and costs, you can run the Data Streamer on System z® Integrated Information Processors (zIIPs).
About this task
To configure the Data Streamer, you must create the Data Streamer started task by copying the
sample procedure HBODSPRO
in the hlq.SHBOSAMP
library, and updating the copy.
If you want to run the Data Streamer as a job rather than a procedure, use the sample job
HBODS001
in the hlq.SHBOSAMP
library rather
than procedure HBODSPRO.
The user ID that is associated with the Data Streamer started task must have the appropriate
authority to access the Z Common Data Provider
program files, which include the installation files and the policy file. It must also have
read/execute permissions to the Java™ libraries in the UNIX System Services file system.
Procedure
To create the started task, complete the following steps:
-
Copy the procedure
HBODSPRO
in the
hlq.SHBOSAMP
library to a user procedure library.
Tip: You can rename this procedure according to your installation conventions. When the
name HBODSPRO
is used in the Z Common Data Provider documentation, including in the
messages, it means the Data Streamer started task.
-
In your copy of the procedure
HBODSPRO
, customize the following parameter
values for your environment:
- /usr/lpp/IBM/zcdp/v5r1m0/DS/LIB
- Replace this value with the directory where the Data Streamer is installed in your environment.
This directory contains the
startup.sh
script for the Data Streamer.
- /etc/cdpConfig/myPolicy.policy
- Replace this value with the policy file path and name for your environment.
nnnnn
- Replace this value with the port number on which
the Data Streamer listens for data from the data gatherers. The default port on which the Data
Streamer listens for data is 51401.
Important: All data gatherers must send data to the
Data Streamer through this port. If you update this port in the Data Streamer configuration, you
must also update it in the configuration for all data gatherers. For more information, see
Data Streamer port definition.
start=w
and start=c
- This parameter is optional. The letter
w
stands for warm, and the letter c
stands
for cold. When stopping, the Data Streamer saves data that has not been sent to the subscriber in a
file buffer so that the data can be sent when the Data Streamer is started again. If you want the
Data Streamer to load the data in the file buffer on startup and then send it to the subscriber, use
start=w
. If you want the Data Streamer to ignore the buffered data at startup, use
start=c
. If this parameter is absent, the default value start=w
takes effect.
statint=numberh/d
- This parameter is optional. This parameter specifies whether to enable the metrics capture
function or not. If it doe not exist, or the value of number is 0, then the
metrics capture function is disabled. If the value of number is a nonzero value,
then the metrics capture function will be enabled and the statistic interval is
number h/d. The
h/d
is the interval time unit, the
h
means hour, and the d
means day.
host=ip_address
- This parameter is optional. If specified, the Data Streamer will bind to the specified IP
address and can accept connections on that IP address only. The value of the host can be a host
name, an IPv4 address, or an IPv6 address. The specified IP address, or the resolved IP address of
the specified host name, must be a valid IP address that is active on any of the TCP/IP stacks on
the LPAR, or a dynamic virtual IP address (DVIPA) which can be activated by any of the TCP/IP stacks
on the LPAR. If this parameter is omitted, or is 0.0.0.0, the Data Stream does a generic bind() to
accept connection requests from any IP address associated with any TCP/IP stacks in the
LPAR.
Note: The Data Streamer can only accept connections from the same LPAR where the Data Streamer
runs. You must run the System Data Engine in the same LPAR as the Data Streamer.
-
In your copy of the procedure
HBODSPRO
, set the following environment
variables for your environment:
- JAVA_HOME
- Specify the Java installation directory.
- CDP_HOME
- Specify the location of the Data Streamer working directory. The Data Streamer working directory
contains files that are created and used during the operation of the Data Streamer, including the
Data Streamer truststore and file buffers.
- Guidelines for the working directory
- Use the following guidelines to help you decide which directory to use as the working directory:
- The directory must be readable and writable by the user ID that runs the Data Streamer.
- To avoid possible conflicts, do not use a directory that is defined as the Configuration Tool
working directory.
Important: Do not update, delete, or move the files in the
CDP_HOME directory. The value of CDP_HOME must not be a
symlink.
- TZ
- Specify the time zone offset for the Data Streamer. For more information, see the information
about the format of the TZ environment variable in the z/OS® product
documentation.
Important: If the value of the
TZ environment variable is incorrect, the
time interval that is set in the
Time Filter transform is directly affected.
For more information about the
Time Filter transform, see
webui_transform_time.html.
- RESOLVER_CONFIG
- If a TCP/IP resolver must be explicitly provided, uncomment the
RESOLVER_CONFIG environment variable, and specify the correct TCP/IP resolver.
The Data Streamer must have access to a TCP/IP resolver. For more information, see resolver_search_order_verify.html#resolver_search_order_verify.
- _BPXK_SETIBMOPT_TRANSPORT
- If you want the Data Streamer to have affinity to a certain TCP/IP stack, uncomment the
_BPXK_SETIBMOPT_TRANSPORT environment variable, and specify that TCP/IP
stack.
Important: If the LPAR has multiple TCP/IP stacks, you must specify which stack
you want the Data Streamer to use and specify the same TCP/IP stack for the System Data Engine (as
instructed in
Creating the System Data Engine started task for streaming SMF data). Otherwise, the System Data Engine might be unable to
connect to the Data Streamer.
- DEFAULT_HEAP
- The heap value that is used by the Data Streamer Java application by default.
DEFAULT_HEAP=4g
means that the heap size is 4 GB by default. If you want to change
the default heap size, uncomment the parameter DEFAULT_HEAP and set a new size.
The default value is 4 GB.
- MAXIMUM_HEAP
- The maximum heap value that is available to the Data Streamer Java application.
MAXIMUM_HEAP=4g
means that the maximum heap size is 4 GB.
If you want to change the maximum heap size, uncomment the parameter
MAXIMUM_HEAP and set a new size. This value must be no less than the value of
DEFAULT_HEAP. The default value is 4 GB. For more information about the maximum
heap size, see The Data Streamer maximum heap size.Important:
- In most cases, a heap size of 4 GB is sufficient.
- If you are running out of memory, you might increase the heap size.
- If your heap is never filled, you can reduce the heap size to free up storage.
- For best performance, set the same value for both DEFAULT_HEAP and
MAXIMUM_HEAP.
-
Update your security software, such as the Resource Access Control Facility (RACF®), to permit the Data Streamer started task to run in your
environment.
- Set
HBODSPRO
as SYSSTC or other high priority WLM service
classes.