02-24-2015 11:58 AM
I have a Process flow that works absolutely fine if i run it from a starting Task node.
Im trying to convert this process flow into a Stored procedure(SP) for scheduling purposes.
The process flow basically reads a flat file (on z/os) and splits that file in to 7 SAS datasets on mainframe. The process flow allocates the mainframe sas datasets and then creatses the SAS dataset. When I go through the wizard of creating a stored process then end result is that the SP will stall out on creating 3rd dataset. and the spawner just hangs. i wont get a Log because its just frozen. but if i would submit the Job regularly it would work fine.
Im not sure what client side excution options i should enable but i've done every setting for the same result.
I dont know if there is some setting in Management Console I need to enable or Parameter.
I do have the possiblility of trying to convert this Eguide project to a Data intergration studio(DIS) project. I read that DIS has enhanced Z/os capabilities.
02-25-2015 01:40 PM
Another option would be to export your EG process flow to a SAS program and then schedule that. For such a simple process, turning it into a stored process just for scheduling seems to be overkill.
02-25-2015 08:43 PM
If you have an EG flow you like, you can schedule that.
The scheduling process depends on your SAS server etc. With our BI server, I usually end up writing a SAS program, then using DI studio to create a DI job which has basically one user-generate coded node that %includes the .sas file, then deploying that job as a flow, then using SMC to schedule the flow via LSF. (It feels like far more steps than necessary, back in the day on 'nix it was just one .sas file and crontab, but perhaps this is the price of storing so much in metadata).
This link has some scheduling tips: