A long while ago, someone much smarter than I said that specifying
freespace as anything other than (0 0) was a waste of otherwise good DASD
space along with the time required to format it, back it up, restore it,
reorg it, etc. Leaving freespace evenly distributed through the dataset
implies that inserts will be evenly distributed through the dataset. I did
some looking and discovered that keys tend to force inserts to cluster.
Specifying FREESPACE(0 0) meant that initial inserts (after dataset load
and/or reorg) resulted in some split activity, but the act of splitting
caused VSAM to put freespace right in the neighborhood of where freespace
was needed.
----------------------------------------------------------------------------------------
This is a PRIVATE message. If you are not the intended recipient, please
delete without copying and kindly advise us by e-mail of the mistake in
delivery. NOTE: Regardless of content, this e-mail shall not operate to
bind CSC to any order or other contract unless pursuant to explicit written
agreement or government initiative expressly permitting the use of e-mail
for such purpose.
----------------------------------------------------------------------------------------
Bruce Black
<bblack To: [email protected]
@FDRINNOVATION.C cc:
OM> Subject: Re: VSAM freespace 0
Sent by: IBM
Mainframe
Discussion List
<IBM-MAIN
01/10/2006 09:48
AM
Please respond
to IBM Mainframe
Discussion List
>
>
>VSAM sees the need to perform a CI split. Since the CA has no extra CIs,
it
>must perform a CA split first. It formats a new cylinder at the end of the
>used space with physical records. It may extend the data set first, if
>necessary. Then half of the CIs from the original CA are moved to the new
>CA. The index is updated appropriately. The necessary CI is split into two
>approximately equal-sized CIs. The index is again updated appropriately.
>The new record is inserted into the appropriate CI.
>
Sounds correct to me. The good news is that this process creates free
space. The split CA now has lots of free CIs, so another insert into
another CI in the CA does a CI split but no CA split (unless it fills
up, of course). The split CIs now have free space, so inserts in the
same CIs can be accomodated without splits (until the CI fills up again).
Very random inserts will cause a lot of splitting initially, but the
split activity will split decrease over time as records are inserted.
The split is time-consuming, especially CA splits, but once the splits
have occured, they do not cause any particular performance penalty on
modern cached control units.
--
Bruce A. Black
Senior Software Developer for FDR
Innovation Data Processing 973-890-7300
personal: [EMAIL PROTECTED]
sales info: [EMAIL PROTECTED]
tech support: [EMAIL PROTECTED]
web: www.innovationdp.fdr.com
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
