Platform Analytics 7 Dataflow

Cluster Capacity datamart

Main ETL XML file

cluster_capacity.xml

Output CSV files

cluster_usage.csv

Record sample

TIMELOCAL,DATE,HOUR,WEEKDAY,CATEGORY,CLUSTERNAME,LSFHOST_TYPE,LSFHOST_MODEL,PROJECTNAME,QUEUENAME,USERNAME,CPUQTY,SLOTS

2010-02-03 14:00:00","2010-02-03",14,4,"RUN","lsf705_MC_left","X86_64","Intel_EM64T","my_default","normal","jxzhu",13.0,0.75

"2010-02-03 15:00:00","2010-02-03",15,4,"RUN","lsf705_MC_left","LINUX86","PC6000","my_default","normal","jxzhu",13.0,2.5

Column description

The Cluster Capacity datamart get its information from the wi_jobs_vs_slots table and wi_cluster_slots table.

The following table briefly describes each column and its source:
Column name
Source and description
DATE
This is the date portion of the localized time_stamp column in the wi_numberofjobs table, and the format is "YYYY-MM-DD".
HOUR
This is the hour portion of the localized time_stamp column in the wi_numberofjobs table.
WEEKDAY
This is the weekday number from the localized time_stamp column in the wi_numberofjobs table, where "1" is Sunday, "2" is Monday, etc.
TIMELOCAL
The source is the localized time_stamp.
CATEGORY  
CLUSTERNAME
The source is dimension_name in the wi_dimensioncode table, which is based on cluster_code in the wi_jobs_vs_slots table.
LSFHOST_TYPE
This is the Host type of job execution host .
LSFHOST_MODEL
This is the Host model of job execution host .
PROJECTNAME
The source is dimension_name in the wi_dimensioncode table, which is based on project_code in the wi_jobs_vs_slots table.
QUEUENAME
The source is dimension_name in the wi_dimensioncode table, which is based on queue_code in the wi_jobs_vs_slots table.
USERNAME The source is dimension_name in the wi_dimensioncode table, which is based on user_code in the wi_jobs_vs_slots table.

CPUQTY

This is the # of execution processor of a slot which is using CPU.
SLOTS
This is the number of slots.