DSXchange

i want to know when the datastage run 1 record/second , and what's the reason ?


database: oracle
datastage:6.0.1

That is not a question that can be answered, it is like asking "why does my car only go 10 Km/H?". There are too many variables.

If you run a job that copies from one sequential file to another in DataStage at the same time does it go at 1 row/second? or 50,000 rows/second? Does reading from your Oracle to a sequential file go at 1 row/second?

What is your job design?
How many stages does you job has?
What is the load on the server when your job runs?
What is the load of your job (how big is the input and outfiles are)?
How many joins, lookups, Pivoting does your job perform?
Was it a read or write to the oracle?
What is the commit rate of the database?
Whether the database reside in the server or different?
If it was connected to network, is you network bandwidth ok by the time of you job run?
..........

Sometimes, the DataStage director shows that its processing the rows at the speed of 1row/sec though it really isn't doing that. It refreshes again at once showing at the end of the job how many rows/sec did it really process. Is this what is happening in your case or Do you really want to know "when" the datastage processes 1row/sec and "why" as you have put in your question? You need to give more details as suggested by Arnd and Kumar.

I guess its for 'us' to guess his job design and deliver a niche solution.

Possibly this is running on the MVS (mine's very slow) edition?

if the database is down, can the datastage inspect the status of the database?
if the database log is full, what's the datastage running status?
when the database log is clear , is the datastage running status normal? i ask the question that the job occurs to run 1 record/second in the unnormal condition.

The 'abnormal condition' is again not particular to a database or datastage or to operating system, or to netwrok...
If you assume it is database, yes you will get the log out from the job. If the database is full or transaction log is full, load to the database will be aborted and hence the job, and hence you will find the relevent logs in the director.
If you feel you job runs very slow, first try to isolate you issue.
Try to run the same command either extraction or loading in to database from the command prompt (direct load). If the load is unexpectedly slow, you can focus on the same. Else locate the source and target and the IO process requried for that. Look into job design as mention earliar. Or revert back with the updates.

elvishada wrote:i want to know when the datastage run 1 record/second , and what's the reason ?

database: oracle
datastage:6.0.1

It means that, for whatever reason, you are processing approximately 60 rows per minute.

There are myriad possible causes, including very large rows. See earlier comments for some ideas.