Confirming the correctness of the installation
The best way to configure capture is to use the Syncsort Capacity Management application System Manager, which is designed for this task. When you have completed the Acquire installation, and if you do not have immediate access to System Manager, you can test it by performing a trial data capture run.
-
Check that enough space is available in the relevant filesystem to write the captured data. 10 MB will be sufficient for a short test.
-
Operation of Acquire is controlled by certain run-time parameters. In the normal use of the product, these are set up using System Manager, but at the moment you will have to set them up manually on the UNIX machine. To do this:
-
Open the file APMconfg for editing. Lines 90 (or thereabouts) to the end of the file contain the runtime parameters. For the purposes of this test it is assumed that you will want to capture just System data.
-
Alter the line CollectSystem=n to CollectSystem=y. Save the APMconfg file.
-
In the scr directory there are a number of specimen crontab files. One of these is named crontab.02. This file instructs Acquire to perform data capture every 2 minutes from 09:00 until 16:58 and to go inactive again at 17:00. Install this crontab by entering the command:
UNIX – crontab < scr/crontab.02
Or:
Linux – crontab scr/crontab.02
After a few seconds, check that the new table is in effect. To do this, enter the command:
crontab -l
Data capture should start at the next time that is a multiple of 2 minutes past the hour.
-
The metron user’s home directory contains a status “flag” file for each type of data capture it is capable of performing. These files are all named APMst.xxxx where xxxx relates to the type of data being captured. Issue the command:
ls -l
- You should see that APMst.system is 7 bytes in length, while all the others are 9 bytes in length. This is because these two files contain the character string active (because you have requested this type of data to be captured), and the other status files contain the character string inactive.
-
Verify that data capture is proceeding correctly by inspecting the contents of the system sub-directory. You should see a variety of files appearing, with names of different formats. You may see temporary files prefaced by uppercase T. Typically there will also be files with names of the form yyyymmddhhmmss_asXXnnnnnnnn.h and yyyymmddhhmmss_oknnnnnnnn.h where XX is the monitor source, as follows.
(Files in brackets may or may not be present depending on the variant of UNIX).
ps from ps
[sa from sar]
[vm from vmstat]
[io from iostat]
[mp from mpstat]
[nw from netstat]
[sp from swap]
ac from acctcom
nnnnnnnn is the Target Number.
(See below for important information about the Target Number.)
h is the interval number, starting at 1001, in hexadecimal notation. The first interval number is an odd number and the following interval numbers should all ascend in twos, e.g., 1001, 1002, 1004, 1006, 1008, 100A, 100C, 100E, 1010 etc. At the start of a new Measurement Period (i.e., a new period of continuous data capture), another odd interval number is generated, again followed by even numbers. The odd interval numbers act as indicators that a new Measurement Period has started.
Target Number
Each target system is identified by a unique 8-digit number called the Target Number. The default number that is embedded in the captured data files is obtained from the APMconfg file, from the statement:
TargetNo=nnnnnnnn
You will need to know this number when you define your UNIX system to Syncsort Capacity Management’s System Manager.
Note that when Control Center makes contact with the UNIX Acquire in the normal way, it will send whatever Target Number you used when you defined the target system to System Manager. This number will replace the original one used in the manual data capture. If you use a different number from the one you used for capturing the test files, Control Center will not be able to retrieve those files containing the old number.