DSXchange

Our system admin has asked me to find out which processes they should monitor for Datastage Server health

Apart from doing a ps -ef | grep IBM, is there anything more reliable that I should ask them to check ?

There are a number of processes. Go to the Launchpad web page and open the Operations Console. It has a dashboard that is great for monitoring the server health. It also lists several important server processes in green or red.

I have opened a call with IBM to check this.

I find the best way to run a crude health check is to run a RowGen->Peek job.

Kick it off in the background. (should run super fast since you would only need about a hand full of rows to prove the point that you can run a job).
sleep 1 or 2 mins.
Kick off the job again.

Test return code of second submit.

The reason you want a 1 -> 2 punch is because if job number 1 hangs for over a min, job #2 will fail because it is not in a runnable state (still running)... you have a health issue.

If job #1 and 2 both fail, you have a health issue.


Your health of the environment basically boils down to "Can I run a job in a timely fashion?" if you submit a 10 second job and it takes over a min or so to come back... you may have your answer. You would submit the job at the highest priority (if grid) so as to not wait in the queue for to long.

Also investigate the capabilities of ISA Lite which, among other things, can perform a health check of your Information Server installation.