IBM Mainframe Forum

by **jaggz** » Thu Dec 27, 2012 4:11 pm

Hi,

I have allocated a file with space(cyl(1,0)). While writing to the file it continues to grow until it reaches 16 extents, but the secondary Space zero is not honoured. I am just curious to know if Secondary space with the value Zero does grows ?

Select all

Data Set Name . . . . : JAGS.ASM

General Data Current Allocation
Management class . . : **None** Allocated cylinders : 16
Storage class . . . : **None** Allocated extents . : 16
Volume serial . . . : LT3RT1
Device type . . . . : 3390
Data class . . . . . : **None**
Organization . . . : PS Current Utilization
Record format . . . : VB Used cylinders . . : 16
Record length . . . : 32756 Used extents . . . : 16
Block size . . . . : 32760
1st extent cylinders: 1
Secondary cylinders : 0 Dates
Data set name type : Creation date . . . : 2012/06/27
SMS Compressible. . : NO Referenced date . . : 2012/06/27
Expiration date . . : ***None***

Jaggz

by **BillyBoyo** » Thu Dec 27, 2012 4:34 pm

How are you using the file? What happens when it gets to 16 extents? Does your research of that lead to any clues?

Why do you have such a large LRECL and therefor BLKSIZE? It is wasteful of space on DASD, but if you need it, you need it :-)

by **enrico-sorichetti** » Thu Dec 27, 2012 4:39 pm

hard to believe!

by **enrico-sorichetti** » Thu Dec 27, 2012 5:18 pm

if the secondary space is 0,
after having filled the primary allocation an abend SD37 will occur

(on a badly fragmented volume, the primary allocation could have been satisfied by multiple extents )

better to start over, being sure of what is entered on the keyboard :mrgreen:

by **BillyBoyo** » Thu Dec 27, 2012 5:40 pm

Primary allocation is one cylinder, can't be split due to fragmentation...

I'd suspect that the program is doing multiple open/close, with the primary extent being available each time, or perhaps DISP=OLD with the SPACE, but I haven't tried it (or done it).

Otherwise, on the information(!) so far provided, I can't think of why this would happen.

by **enrico-sorichetti** » Thu Dec 27, 2012 6:05 pm

Primary allocation is one cylinder, can't be split due to fragmentation...

that' s what the TS told us, as I said before - hard to believe

Nobody by looking at the VTOC after the allocation is completed or after dataset usage will ever know what was requested ! :mrgreen:

TESTED by
filling up an unused volume leaving just two NON CONTIGUOUS cylinders
allocating a dataset requesting a PRIMARY allocation of TWO cylinders

and here is what 3.2 tells

Select all

Data Set Information
Command ===>

Data Set Name . . . . : ENRICO.FRAGMNTD.DATASET

General Data Current Allocation
Management class . . : **None** Allocated cylinders : 2
Storage class . . . : **None** Allocated extents . : 2
Volume serial . . . : ZAPRD4
Device type . . . . : 3390
Data class . . . . . : **None** Current Utilization
Organization . . . : PS Used cylinders . . : 0
Record format . . . : FB Used extents . . . : 0
Record length . . . : 80
Block size . . . . : 27920
1st extent cylinders: 1
Secondary cylinders : 2
Data set name type : SMS Compressible : NO

Creation date . . . : 2012/12/27 Referenced date . . : 2012/12/27
Expiration date . . : ***None***

by **steve-myers** » Thu Dec 27, 2012 10:44 pm

Just because secondary space is 0 does not mean it will be honored. I ran this JCL -

Select all

//A EXEC PGM=IEFBR14
//A DD DISP=(MOD,DELETE),UNIT=SYSALLDA,SPACE=(TRK,0),
// DSN=XXXXXX.TESTSEC.NOSEC
//B EXEC PGM=IEFBR14
//A DD DISP=(,CATLG),UNIT=SYSDA,SPACE=(CYL,(1,0,500)),
// DCB=(RECFM=FB,LRECL=80),DSN=*.A.A
//C EXEC PGM=IEBUPDTE
//SYSPRINT DD SYSOUT=*
//SYSUT1 DD DISP=SHR,DSN=SYS1.MACLIB
//SYSUT2 DD DISP=OLD,DSN=*.A.A,SPACE=(CYL,(1,1))
//SYSIN DD *
./ REPRO NAME=ABEND
./ REPRO NAME=ACB
./ REPRO NAME=ACBVS
./ REPRO NAME=ACYAPHDH
./ REPRO NAME=ADREID0
./ REPRO NAME=ADRMCLVL
./ REPRO NAME=ADSR
./ REPRO NAME=ADYDSTAT
./ REPRO NAME=ADYENF
./ REPRO NAME=AHLFFAP
./ REPRO NAME=AHLWKAL
./ REPRO NAME=AHLZGTO
./ REPRO NAME=AHLZGTS
./ REPRO NAME=ALESERV
./ REPRO NAME=AMDSADMP
./ REPRO NAME=ANFUEXTP
./ REPRO NAME=ANTPQMAP
./ REPRO NAME=ANTPSL10
./ REPRO NAME=ANTPSL11
./ REPRO NAME=ANTPSL20
./ REPRO NAME=ANTPSL30
./ REPRO NAME=ANTQFRVL
./ REPRO NAME=ANTRQST
./ REPRO NAME=ANTRQSTL
./ REPRO NAME=ANTRQTA
./ REPRO NAME=ANTRQTB
./ REPRO NAME=ANTRQTC
./ REPRO NAME=ANTRQTD
...

I got this result -

Select all

Data Set Name . . . . : XXXXXX.TESTSEC.NOSEC

General Data Current Allocation
Management class . . : **None** Allocated cylinders : 15
Storage class . . . : USER Allocated extents . : 15
Volume serial . . . : FUSR27 Maximum dir. blocks : 500
Device type . . . . : 3390
Data class . . . . . : **None**
Organization . . . : PO Current Utilization
Record format . . . : FB Used cylinders . . : 15
Record length . . . : 80 Used extents . . . : 15
Block size . . . . : 27920 Used dir. blocks . : 5
1st extent cylinders: 1 Number of members . : 76
Secondary cylinders : 0
Data set name type : PDS

Creation date . . . : 2012/12/27 Referenced date . . : 2012/12/27
Expiration date . . : ***None***

The secondary space was not honored because the secondary space in SPACE=(CYL,(1,1)) in the IEBUPDTE step overrode the secondary space in the initial allocation. There are products in the market place - something called STOPX37 may be the best known - that will do this without JCL changes.

by **dick scherrer** » Thu Dec 27, 2012 11:15 pm

Hello,

Suggest Jaggz speak with the storage managent people and learn which rules apply to this particular dataset.

by **enrico-sorichetti** » Fri Dec 28, 2012 1:39 am

But IMO even if it works people using that trick should have their fingers cut

guess about using it on SYS1.LINKLIB :mrgreen:

by **steve-myers** » Fri Dec 28, 2012 5:04 am

enrico-sorichetti wrote::oops:

But IMO even if it works people using that trick should have their fingers cut

guess about using it on SYS1.LINKLIB

Cut? Amputated might be more appropriate!

Seriously, my sole goal was to demonstrate that a secondary allocation specification is not cast in stone and it is possible to defeat it. Notice that even though it was defeated for the DD statement, the original specification is still there.

It appears to me that the original data set appears to be sequential, so the only practical way to extend it is to use DISP=MOD.

IBM Mainframe Forum

Secondary Space at ZERO

Secondary Space at ZERO

Re: Secondary Space at ZERO

Re: Secondary Space at ZERO

Re: Secondary Space at ZERO

Re: Secondary Space at ZERO

Re: Secondary Space at ZERO

Re: Secondary Space at ZERO

Re: Secondary Space at ZERO

Re: Secondary Space at ZERO

Re: Secondary Space at ZERO