very slow executing of job with db2 stage

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

Post Reply
ppalka
Participant
Posts: 118
Joined: Thu Feb 10, 2005 7:25 am
Contact:

very slow executing of job with db2 stage

Post by ppalka »

hi,

I have a problem with jobs which are using a DB2 stage. It is the same in server and parallel job. These jobs are executing very slowly. For example simple job:
seq file -> DB2
source data: two columns and one thousand records
It finished after 7min !!!
What can I do to speed up that jobs?

Regards,
Piotrek
ArndW
Participant
Posts: 16318
Joined: Tue Nov 16, 2004 9:08 am
Location: Germany
Contact:

Post by ArndW »

Piotrek,

you have the option of doing bulk inserts in DB/2, are you doing this or using some form of update/insert? Is this table partitioned - if so then the PX load ought to be faster than a server job. Do other tables load faster? Does this table have constraints/triggers or is it some form of a view? Do you have many columns and how many average bytes per row? How did you set the transaction and array sizes in DB/2

As you can see there many factors involved that affect performance and I have only touched upon a couple of them. It would also be helpful if you posted what you have already tried and if it had any effect on the speed.

Since all sites have different speeds, perhaps you could get a baseline top speed by creating a table with only one column and do a simple load to that from a sequential file or row generator using the bulk loader. If that only performs at the speed of your original job then you know you have either a connectivity speed issue (if DB/2 is on another system) or very serious database configuration problems that the successor of your current DBA needs to resolve
Post Reply