Re: Reusing deleted records -- MIDRANGE-L

At 04:39 PM 9/23/98 -0400, you wrote:
>Al Barsa, Jr. wrote:
>> 
>> At 12:26 AM 9/23/98 -0700, you wrote:
>> >
>> >I don't believe that this could be correct from what I understand the
>> >relationship is between job ODP's and the underlying RDBMS.
>> 
>> Sorry.  I spent a great deal of time testing and researching this function
>> when it was added in V2R1M1.  At the time, we beta tested the release, and
>> were (re) writing an application for a client where the outage caused by
>> RGZPFM was intolerable.
>> >
>
>Is it possible that IBM has since corrected this in subsequent releases
>?

Anything is possible, however I do stay fairly wired in the area of the
database.  I think that if there was a significant change in this function,
I would have been briefed.

>If not how do you access this deleted records map from your program ?

You don't.  OS/400 handles the re-use for you automatically.
>
>> >In a layered OS like OS/400, any program can, through the feed back DS,
>> know how
>> >many other programs are using a particular file.  So I have to conclude
>> that the
>> >OS is aware of multi job interaction with a particular file.
>> 
>> True, but not true for the deleted record map which is a function of any
>> individual ODP's file open.
>> >
>> >I missed the original post so I'm going out on a limb here .... Does the
>> program
>> >perform record deletes and block adds?  The reason I ask is that if the
disk
>> >controller receives a record delete it will do so immediately (or appears
>> to) but
>> >blocked writes would be delayed and possibly be treated as a group that
the
>> >controller may attempt to search for a contiguous block write area.
>> 
>> The function has to do with using the REUSEDLT parameter on the creation of
>> the physical file.  Writes to the LF are ultimately reflected in the PF.
>> >
>> >So although the program may delete x records then write x records, if the
>> delete
>> >was through a logical, and the writes through a physical there is no
>> relationship
>> >to the deleted space (record numbers) and the attempted block write.
>> 
>> Different ODP, different deleted records map.
>> >
>> 
>> Al
>> 
>
>Back to the original problem with the system not using deleted records -
>is it possible that when another job opens the file it copies the
>existing deleted records map and does not create a new one?  Therefore
>the deleted records map would never be recreated until All applications
>that had opened the file ended.  This could have the effect of the
>deleted records never being reused.
Anything is possible.  Bruce Vining, do you have any comment??


Al

Al Barsa, Jr. - Account for Midrange-L
Barsa Consulting, LLC.  
400 > 390

Phone:  914-251-9400
Fax:    914-251-9406





Private mail should be sent to barsa@ibm.net
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---