View previous topic :: View next topic |
Author |
Message |
Smith Beginner
Joined: 07 Apr 2005 Posts: 13 Topics: 5
|
Posted: Mon Oct 16, 2006 7:12 pm Post subject: VSAM parallel process |
|
|
Is there any way you can process VSAM file updates in parallel? We have a stream of VSAM jobs with updates in our shop, trying to see whether it is possible to split them into two streams to reduce the processing time. I have heard about tools like vsam transparency that could be used, but trying to see whether there are any other ways through which we can accomplish this. All suggestions are welcome.
Thanks
Smith. |
|
Back to top |
|
 |
warp5 Intermediate

Joined: 02 Dec 2002 Posts: 429 Topics: 18 Location: Germany
|
Posted: Tue Oct 17, 2006 1:19 am Post subject: |
|
|
Many issues are involved here, and an analysis would be complicated to say the least. You also do not provide enough information, like what kind of VSAM files, what kind of processing (seq. or random), what do your jobs look like, do you have RLS, etc? An important issue is deadlocks. Post detailed information if you want to get an answer that can help you. |
|
Back to top |
|
 |
Smith Beginner
Joined: 07 Apr 2005 Posts: 13 Topics: 5
|
Posted: Tue Oct 17, 2006 2:04 pm Post subject: |
|
|
We have both seq and random processing with no RLS. Both KSDS and ESDS are there in the job stream.
Thanks. |
|
Back to top |
|
 |
warp5 Intermediate

Joined: 02 Dec 2002 Posts: 429 Topics: 18 Location: Germany
|
Posted: Wed Oct 18, 2006 12:00 am Post subject: |
|
|
With the information you supplied (which does not even scratch the surface) I would say no. 2 batch jobs could access the same datasets and the same CIs and cause a deadlock. You might be updating from 2 jobs and corrupt your datasets. You might actually slow down your processing because dataset access from 2 jobs may be delaying each other. Maybe you should review improving your performance of the jobs by increasing buffers, etc. Actually, this is something that should be researched by the experts in your shop, if you have any. |
|
Back to top |
|
 |
|
|