2. INSTALLATION › 2.3 Activating Accounting for Data Sources › 2.3.4 External Journal File in Accounting-Only Unit › 2.3.4.1 Linux Process Accounting Journal File
2.3.4.1 Linux Process Accounting Journal File
The ACTJLX journal file is used to account for process
activity captured as log data using the Linux ACCT command.
The Linux Option Specification Member (MWF;4;2;3;6;8) is
used to specify the ddnames, dataset names, and other
information about the Linux data to be processed by an
accounting-only unit. Accounting-only unit external files
are processed like standard CA MICS journal files, except
the associated external journal file is created by a
generated job instead of by the CA MICS DAILY job.
When you select the ACTJLX journal file and then select the
Linux Option Specification Member (MWF;4;2;3;6;8), you enter
an ISPF EDIT session for dataset prefix.MICS.PARMS(LINUXP).
This is where you describe the Linux data to be processed by
the unit.
prefix.MICS.PARMS(LINUXP) contains comments that explain the
rules for updating the member. Those rules are as follows:
o Lines that begin with an asterisk (*), are ignored.
o Two line types are supported:
- GLOBAL Statements
- Dataset Description Statement
These statements are described in detail below.
GLOBAL Statement
-----------------------
There are two types of GLOBAL statements supported: GLOBAL
PLATFORM and GLOBAL TIMESHIFT.
o GLOBAL PLATFORM
Linux systems running on Intel platforms use a method of
numeric data value representation known as "little
endian." Those running on mainframe platforms, such as
z/VM, represent numeric data using "big endian." You must
identify the platform so that CA MICS knows how to input
and interpret the raw measurement data.
The Dataset Description statements allow you to specify
the platform associated with each ddname, but to save time
and coding, you can simply enter a GLOBAL PLATFORM
statement ahead of the Dataset Description statements that
describe data generated on a specific platform.
The GLOBAL PLATFORM statement is coded as follows:
GLOBAL PLATFORM=p
where p can be I (INTEL) or M (Mainframe)
o GLOBAL TIMESHIFT
Linux systems record process level accounting data time
and date fields in Greenwich Mean Time (GMT). If you want
CA MICS to convert GMT to local time, you must supply the
timeshift (number of hours) that you need to add or
subtract from GTM to convert to your local time.
The Dataset Description statements allow you to specify
the timeshift associated with each ddname, but to save
time and coding, you can simply enter a GLOBAL TIMESHIFT
statement ahead of the Dataset Description statements that
describe data generated in a particular timezone.
GLOBAL TIMESHIFT=+/- nn
where + means "add"
- means "subtract"
nn hours convert GMT to local time
GLOBAL statements may appear anywhere in the input stream.
When encountered, a GLOBAL statement will override any
previous GLOBAL statement for the same function (PLATFORM or
TIMESHIFT).
To minimize usage of the P= and T= Dataset Description
statement parameters, discussed below, it is useful to group
DDs and timezones for the same platform together and precede
the statement groups with the appropriate GLOBAL statements.
Dataset Description Statement
-----------------------------
A Dataset Description statement line supports five
parameters as follows:
DDNAME DSN SYSID P=I/M T=+/-nn
where the following are required:
DDNAME - 1-8 characters
DSN - 1-44 characters
SYSID - 1-4 characters
Note: If you define a new SYSID, you must add the new SYSID
in prefix.MICS.PARMS(sysid) and submit
prefix.MICS.CNTL(BASPGEN).
A fourth parameter is required if GLOBAL PLATFORM is not
specified, where:
P=I means INTEL
P=M means Mainframe
If coded, it overrides GLOBAL PLATFORM.
The fifth parameter is optional, where:
T=+5 means add 5 hours to GMT value
T=-3 means subtract 3 hours from GMT value
If coded, it overrides GLOBAL TIMESHIFT. If neither GLOBAL
TIMESHIFT nor T= is specified, the data will reflect GMT.
Example 1
---------
GLOBAL TIMESHIFT=4
GLOBAL PLATFORM=M
DDNAME1 ABC.LINUX.FB.DATA.INTEL.JULY10
DDNAME2 ABC.LINUX.DATA.PACCT.USILGPAX P=I
Example 1 results in JCL generation for two datasets:
//DDNAME1 DD DSN=ABC.LINUX.FB.DATA.INTEL.JULY10 etc.
//DDNAME2 DD DSN=ABC.LINUX.DATA.PACCT.USILGPAX etc.
CA MICS will add 4 hours to date/time metrics for each of
these datasets due to the GLOBAL TIMESHIFT=4 statement.
CA MICS will treat the numeric data from DDNAME1 as "big
endian" due to the GLOBAL PLATFORM=M statement.
CA MICS will treat the numeric data from DDNAME2 as "little
endian" due to the P=I parameter on the Dataset Description
statement for DDNAME2, which overrides the GLOBAL PLATFORM
statement.
Example 2
---------
DDNAME1 ABC.LINUX.FB.DATA.INTEL.JULY10 T=-6 P=I
DDNAME2 ABC.LINUX.DATA.PACCT.USILGPAX T=3 P=M
DDNAME3 ABC.LINUX.DATA.PACCT.INF3TTYY T=3 P=M
Example 2 results in JCL generation for three datasets:
//DDNAME1 DD DSN=ABC.LINUX.FB.DATA.INTEL.JULY10 etc.
//DDNAME2 DD DSN=ABC.LINUX.DATA.PACCT.USILGPAX etc.
//DDNAME3 DD DSN=ABC.LINUX.DATA.PACCT.INF3TTYY etc.
CA MICS will subtract 6 hours from date/time metrics for the
DDNAME1 dataset due to the T=-6 parameter on the Dataset
Description statement.
CA MICS will treat the numeric data from DDNAME1 as "little
endian" due to the P=I parameter on the Dataset Description
statement.
CA MICS will add 3 hours to date/time metrics for the
DDNAME2 and DDNAME3 datasets due to the T=3 parameter on the
Dataset Description statements.
CA MICS will treat the numeric data from the DDNAME2 and
DDNAME3 datasets as "big endian" due to the P=M parameter on
the Dataset Description statements.
GLOBAL statements could have been used in Example 2 to
achieve the same results for DDNAME2 and DDNAME3 as follows:
DDNAME1 ABC.LINUX.FB.DATA.INTEL.JULY10 T=-6 P=I
GLOBAL TIMESHIFT=3
GLOBAL PLATFORM=M
DDNAME2 ABC.LINUX.DATA.PACCT.USILGPAX
DDNAME3 ABC.LINUX.DATA.PACCT.INF3TTYY
Note that P= and T= parameters are not required for DDNAME2
and DDNAME3 because the GLOBAL statement provides the
platform and timeshift information.
