Hi together,
we are running some large LSMW Jobs.
This jobs are working parallel on different batch work processes.
When I am looking to the CPU I can see 15% load on the cpu pool.
When I am looking at the disks arm usage, I can see between 1 to 20% usage, depends on the kind of the job.
When I am loking at the Page faults in the Memory of all pools, I can see that there are not so much page faults. This should also not be the problem.
We don't have the dbmon active.
We have XSM running in backround, but the disks in the sysbas are not working much.
When I am looking to sm50 in the R/3 4.70 I see that some jobs are waiting for update, but not on the fact the update jobs are full. There are enough update jobs free.
When I am looking to sm21 there is all ok.
When I am looking to st22 there are only 20 lockentries.
When I am making an SQL trace I see very fast SQL Updates or inserts.
I have asked the developer, if the LSMW is reading from IFS, and he is saying no. This job are reading from DB and writing to DB.
Why is the CPU load not going up, if I/Os are not so much?
Or why are the I/Os not at a maximum?
So, I have no idea, where the speed is gone. Normaly a good running iSeries job pushes up the CPU Load or is going to a maximum at DISK I/Os per disk arm.
Best regards,
Carsten Schulz