Page 1 of 1

To trace the time taken for execution for a particular job

Posted: Sat Nov 08, 2008 4:56 am
by G.K.K
hi,

i want to list the name of job which takes more time for execution through unix shell scripting.

first, i want to know how to read the RT_LOG FILE which will be stored as hash . whether this file can be readable or not through unix scripting ? and how to find the particular log file for a job name.

if there is any websites which explains about the DATASTAGE UNIVERSE.
PLEASE LET ME KNOW....


THANKS & REGARDS,
KAMAL :D

Posted: Sat Nov 08, 2008 5:55 am
by ray.wurlod
What you seek to do is cumbersome, though possible. The dsjob command includes options for reading the job log; you must provide the project name and job name whose log you wish the command to access.

To get the job number associated with a job name from UNIX (which you don't need to use the dsjob command), be in the project directory and execute dssh with the relevant query, for example

Code: Select all

dssh "SELECT JOBNO FROM DS_JOBS WHERE NAME = 'JobName';"
There are various commands for reading hashed files from the operating system; usually SQL executed in a dssh shell is easiest to manage.

Apart from what you can glean on DSXchange, I don't believe there is any website that describes the contents of the Repository database associated with a DataStage project.