Does anybody have a "good" way to enable users to schedule or initiate a DataStage job? Ideally, this would work with no intervention on the part of developers, operators, or any other IT staff people.
The only ways I can think of would be by using a DB table which a user could update and then having a job which would periodically wake up and query this table to see if any user requests exist, or to periodically run a job which would look for a trigger file of some kind which the user would create.
Are there better ways? (Environment is Unix, DataStage 7.1)
Larry
On request jobs
Moderators: chulett, rschirm, roy
Thanks for your response Ketfos. We are using the scheduler in DataStage, but that does not give the non-IT person any facility, it only allows an IT person to schedule a job. Unless, that is we installed the Director on the user's pc, which does not seem like a good option.
A script which would look for a trigger file is an option. I am just looking for more "elegant" solutions. You know, like somehow the user could say "ala ca dabra" three times and then job x would run!
How are others handling this?
Larry
A script which would look for a trigger file is an option. I am just looking for more "elegant" solutions. You know, like somehow the user could say "ala ca dabra" three times and then job x would run!
How are others handling this?
Larry
Simple - we don't have any "user scheduled" jobs. 
Actually, I've never had a need to do this, so I would be curious what kind of jobs these people would be scheduling.
Now, we do have "loosely scheduled" jobs that are waiting for files from various people... but they run over a specific window and poll for the arrival of a file. For example, I've got one that runs on every other Wednesday and waits for a payroll file. Are you doing something like that?
Actually, I've never had a need to do this, so I would be curious what kind of jobs these people would be scheduling.Now, we do have "loosely scheduled" jobs that are waiting for files from various people... but they run over a specific window and poll for the arrival of a file. For example, I've got one that runs on every other Wednesday and waits for a payroll file. Are you doing something like that?
-craig
"You can never have too many knives" -- Logan Nine Fingers
"You can never have too many knives" -- Logan Nine Fingers
What kinds of jobs indeed!
These are not data warehouse type jobs. They are more like normal data processing jobs. Things like General Ledger are we in balance yet jobs where the user runs a special job to see if various accounts balance and if not, makes some changes and then runs the thing again before closing the books for that accounting period.
I know, this type of thing should be done within the application, but my client has opted to use DataStage for quite a bit of this. Actually, many of these jobs are running under an out dated windows batch system and DataStage is being used to replace it.
I do have a DS job which looks for a file, sleeps if it's not there, etc. until the file either shows up or a certain amount of time goes by. The only problem is that we don't have a good way for a user to create a file in Unix. Plus, I can see this getting out of hand and becoming an administration nightmare.
Any more thoughts?
Larry
These are not data warehouse type jobs. They are more like normal data processing jobs. Things like General Ledger are we in balance yet jobs where the user runs a special job to see if various accounts balance and if not, makes some changes and then runs the thing again before closing the books for that accounting period.
I know, this type of thing should be done within the application, but my client has opted to use DataStage for quite a bit of this. Actually, many of these jobs are running under an out dated windows batch system and DataStage is being used to replace it.
I do have a DS job which looks for a file, sleeps if it's not there, etc. until the file either shows up or a certain amount of time goes by. The only problem is that we don't have a good way for a user to create a file in Unix. Plus, I can see this getting out of hand and becoming an administration nightmare.
Any more thoughts?
Larry
So, in other words you are emulating a BI tool with DataStage? :D Yerk.
Do your users have a methodology whereby they could insert records into a database table? I'd probably create a set of jobs that periodically polled that table for unprocessed records that indicate specific requests and then run the associated job(s). Some sort of scheduler could run it every X minutes during certain hours, or you could launch it once at the beginning of the day and have it poll internally until it becomes time to shut down for the night.
Something like that, anyway.
Do your users have a methodology whereby they could insert records into a database table? I'd probably create a set of jobs that periodically polled that table for unprocessed records that indicate specific requests and then run the associated job(s). Some sort of scheduler could run it every X minutes during certain hours, or you could launch it once at the beginning of the day and have it poll internally until it becomes time to shut down for the night.
Something like that, anyway.
-craig
"You can never have too many knives" -- Logan Nine Fingers
"You can never have too many knives" -- Logan Nine Fingers
-
ray.wurlod
- Participant
- Posts: 54607
- Joined: Wed Oct 23, 2002 10:52 pm
- Location: Sydney, Australia
- Contact:
