|
View previous topic :: View next topic
|
| Author |
Message |
razesh84
New User
Joined: 05 Apr 2010
Posts: 41
Location: Kolkata,India
|
|
|
|
With reference to this post.
I have few doubts on VSAM buffering.can you please answer below questions
1>Is there any way of checking what buffering technique(lsr/nsr/rls) is used for a VSAM file in a jobstep?
2>Does buffering technique depends on some parameter(DATACLASS?) of that VSAM during its definition?
3>Suppose, I have a KSDS .In one progam it used randomly and in other sequentially.Will the buffering technique be different for both cases?
4>What are the factors that influence buffering technique.
Can i change it(nsr to lsr or others) from jcl or i'm only allowed to increase buffers through bufni/bufnd? |
|
| Back to top |
|
 |
dbzTHEdinosauer
Global Moderator
Joined: 20 Oct 2006
Posts: 6966
Location: porcelain throne
|
|
|
|
The Redbook VSAM Demystified
is something you could/should read this weekend.
I think you will find the answers to your questions there,
as well as learn the BUF.. parms do not affect VSAM files,
check out the AMP parm. |
|
| Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
Please do not tailgate an old topic. Please start a new one, referring to one or more posts if appropriate.
Also, avoid soliciting answers. If you have a question, ask it of everyone, please.
Your topic has been split, and edited. |
|
| Back to top |
|
|
Robert Sample
Global Moderator
Joined: 06 Jun 2008
Posts: 8697
Location: Dubuque, Iowa, USA
|
|
|
|
1. I'm not aware of any external way to tell, although there are some general guidelines. If the job has SUBSYS=BLSR specified, then I would expect it to be using batch LSR; otherwise, since LSR is specific to CICS and some select vendor programs, a batch job won't be using LSR.
2. Default buffering is set during VSAM definition -- the BUFFERSPACE parameter indicates the amount of buffer space allocated by default, and is set to 2 times data CI size plus 1 times index CI size unless overridden. This is grossly inadequate in most cases. DATACLASS may be used to define an extended format VSAM file, but has little role to play in buffer allocation.
3. If the option is there, LSR works better with random access while NSR is best for sequential access. Random access requires more index buffers than data buffers since you're using the index values to retrieve the record, while sequential access requires plenty of data buffers and not many index buffers -- one to be precise.
4. Program access is the guiding factor -- random versus sequential versus mixed (AKA skip sequential). Allocate plenty of index buffers in the first case, plenty of data buffers in the second case, and plenty of both in the third case. If your site has BLSR installed, you can use it but otherwise your basically limited to AMP=(BUFND=, BUFNI=) to control buffers.
All this and much more is explained in VSAM Demystified -- you need to start reading. |
|
| Back to top |
|
|
Pete Wilson
Active Member
Joined: 31 Dec 2009
Posts: 582
Location: London
|
|
|
|
You should enquire if your site has any automatic buffering tools installed.
It would still be worth reading the VSAM Demystified and possibly the Using Datasets manuals anyway, for a better understanding.
DATACLAS can affect buffering via the RECORD ACCESS BIAS field and SYSTEM MANAGED BUFFER field settings although I don't fully understand how, not having used these myself.
Your choice of CISIZE can be significant as it may or may not have a corresponding CICS buffer pool |
|
| Back to top |
|
|
|
|
