User guide

balance. However, this configuration does not overwrite mapping that the
Job gathers from Proviso. Therefore, following is the partition rule:
Mapping from Priviso DB → collector_mapping.csv → Job key hash.
Changing the collection period for SAM statistics
This section provides general guidelines on how to change the collection period for
SAM statistics:
v “General guidelines on changing the collection period for SAM statistics”
General guidelines on changing the collection period for SAM
statistics
The default collection period for all SAM statistics is 15 minutes. If you change the
collection period, ensure that you follow these guidelines:
1. You must set the APP.FILE_PERIOD parameter on each UBA component and the
WINDOW_PERIOD parameter on the IBM InfoSphere Information Server to
equivalent values.
For example, if you set the APP.FILE_PERIOD parameter to 300 (that is, 300
seconds, or 5 minutes), then you must set the WINDOW_PERIOD parameter to 300.
Note: The APP.FILE_PERIOD parameter is specified in seconds, and the
WINDOW_PERIOD parameter is specified as a series of values that are expressed in
minutes.
2. The collection period for any type of SAM statistics must be set to the same
value as the collection period for the corresponding Tivoli Netcool Performance
Manager requests.
DataStage settings
You can configure the SAM classes for 5 minutes, 15 minutes, or in mix periods.
This configuration is done at the SAM Server. However, you must specify this
period information in the Job parameter to enable Job, which produces the CSV file
at the correct period interval.
In the mix periods, you can configure some Alcatel-Lucent SAM 5620 classes at 15
minutes, and the remaining classes at 5 minutes. You can use the mix period to
increase the collection frequency for some classes that you want, while you can
keep the other classes at the default collection period, which is 15 minutes. As a
result, the
v Processing load on the DataStage and UBA is reduced
v Resource usage efficiency of the underlying machine is improved
DataStage receives and processes the metric XML files both at 5 minutes and 15
minutes at the respective frequency. However, the output CSV files are generated
after every 15 minutes. These CSV files contain one datapoint for the classes that
are at 15-minutes period, and 3 datapoints for the classes that are at 5-minutes
period. The UBA can receive the CSV files, and process the single and three
datapoints accordingly.
Restriction: This configuration is useful when you want a higher collection
frequency to obtain detailed reporting. However, you will experience an additional
latency of 5 minutes because the data is collected for 15 minutes, and then it is
passed to the UBA. If you want to achieve minimum latency at higher collection
66 Alcatel-Lucent 5620 SAM LogToFile 2.3.0.0 Technology Pack Installation and User Guide