Implementation Diffferences Shell Script vs Datastage

[vskr72]

We have a requirement to do the following in our project:
- Poll for file. If it does not come by X secs, then send a DList , then continue to poll. If it does not come by Y secs send an email to different Dlist and continue to poll. If it does not come by Z secs, then continue to poll and exit.
- Once file received, then trigger a series of jobs.
- File archive/ backup
- finally exit

Is it best to do this in Datastage (is it possble at all) or thru a script. Thank you

[ray.wurlod]

What is "best"? You can do it either way (a job (sequence) or a script). What's best for you? What will be easier to maintain (for example parameterising the wait periods)?

[SURA]

I would prefer to do it in DS.


DS User

[PaulVL]

Are you running in a GRID environment? If so, I would not submit a grid job and have it remain active for X amount of time just waiting around for the file.

Having a polling process (could be on a remote server) do the sleep and poll.
executing the dsjob -run from that script would be my vote.

You are going to be expending far more system resources by executing the polling in the datastage environment / job rather than the shell script.