3

DataStage version is 8.1 - I have no direct access but need to give instructions to extract some job runtime stats for me. I believe that repository is in DB2 database or maybe in flat files if that's still supported in 8.1. I can't install any tools there.

There is a master sequence with few dozen jobs. Some of them are parallel. The sequence itself runs one at a time and it runs constantly during the day.

For 24 hours period, I need to extract performance stats for this master sequence. Something like that:

job_name, start_time, finish_time, completion_status

I really just need timing info. I can work with any format but what I'm really after in the end is a csv-style file with timings as above.

I get some hints on dsjob -report but I don't have enough background in DataStage to provide good instructions. Hope to get access myself but again, I'm a noob at DataStage.

Thanks in advance.

Alex
  • 31
  • 1
  • 3

2 Answers2

2

You can export the metadata of the job's last run to a XML file. 

#!/usr/bin/ksh

project="yourProjectName"
job="yourJobName"

dsjob -report $project $job XML >"$project_$job.xml"
Dan
  • 31
  • 5
1

You could use Director to export the sequence log and parse when jobs are kicked off and complete. Alternatively, you could export the log via command line (see dsjob -logsum command) and then do the parsing in awk (if on UNIX or AIX system).

If you have the ability to modify the sequence, you could add a Activity Stage and write to a file before and after a job in the sequence.

Mensur
  • 1,196
  • 3
  • 18
  • 30