AW: Re: EAV bug or feature?

[Peter Hunkeler]

>So if you want to use EAV's, make them SMS. ...



... and be prepared to have to deal with strange errors with software which is 
not EAV-savvy, i.e. which show strange behaviour with cylinder managed block 
addresses. E.g. code written with SAS-C may not like them.


--
Peter Hunkeler



--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Glenn Wilcock]

Besides the Redbooks, a good place to start is the 'z/OS DFSMS Using New 
Functions' manual.

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Klaus Stanislawiak]

On 5/10/2016 10:30 AM, Tom Conley wrote:
> I ran many different searches trying to find this information, also 
> using the term "cylinder-managed" space.  While it directed me to the 
> EAV section of the DFSMS Using Datasets manual, not one example of how 
> to allocate a dataset in the cylinder-managed space.

Tom,

You may also want to take a look at some Redbooks.
These two provide a comprehensive introduction to EAVs, even with examples:
DFSMS V1.10 and EAV Technical Guide, SG24-7617-00 (Chapter 18)
z/OS Version 1 Release 11 Implementation, SG24-7729-00 (Chapter 4)

And those two provide information about more recent enhancements:
z/OS Version 1 Release 13 Implementation, SG24-7946-00 (Chapter 24)
z/OS V1.13 DFSMS Technical Update, SG24-7961-00 (Chapter 10)

Regards,
Klaus Stanislawiak

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Tom Conley]

On 5/11/2016 1:35 AM, Barbara Nitz wrote:

Tom,


I'll be covering this in the Bit Bucket at SHARE in Atlanta.  Thanks for
giving me some more information about this.  If you go to zTechU in
Munich, I'm presenting there, so please look me up.


I had planned on it when I saw your name on the roster...



LOL.  See you there!

Tom

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Barbara Nitz]

Tom,

>I'll be covering this in the Bit Bucket at SHARE in Atlanta.  Thanks for
>giving me some more information about this.  If you go to zTechU in
>Munich, I'm presenting there, so please look me up.

I had planned on it when I saw your name on the roster...

Barbara

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Ron Hawkins]

All,

As with SDB, I have a catch all DATACLAS that assigns EATTR=OPT to everything 
that hasn't been assigned another DATACLAS.

Ron

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf 
Of Mike Schwab
Sent: Tuesday, May 10, 2016 1:22 PM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: [IBM-MAIN] EAV bug or feature?

I often use this clist for Non-SMS recalls.

2 DSLIST xx.CLIST(HRECN) - 01.00Columns 1 00072
+1+2+3+4+5+6+7--
* Top of Data ** PROC 2 
D V
  CONTROL   PROMPT NOSYMLIST NOLIST NOCONLIST NOCAPS   MSG NOFLUSH
  HSEND RECALL  VOL() UNIT(3390) FORCENON END /* PROC */
 Bottom of Data 

On a TSO ISPF 3.4 Dataset list,
HRECN / volser;  data.set.name  ...
works well.

On Tue, May 10, 2016 at 12:52 PM, nitz-ibm <nitz-...@gmx.net> wrote:
>> One of the things I found appalling is that there is not a single 
>> example of how to allocate an EAV dataset in the cylinder-managed 
>> space anywhere in any IBM doc.  You can find EATTR=OPT if you know to 
>> look for it, but do a search on how to allocate an EAV dataset.  Nothing.
> When I tried that on an ADCD system about two years ago with my own 10cyl 
> EAV, I found the same thing. I got lucky in that I had some presentation 
> handy that directed me to the EATTR attribute. I remember that I spend almost 
> an hour trying to figure it out.
>
>> So if you want to use EAV's, make them SMS.  That way you can set
>> EATTR(OPT) in all your DATACLASses and fuhgeddabotit.  Non-SMS EAV's 
>> don't make a whole lotta sense, unless updating thousands of JCL 
>> members is  your thing.
> I *was* using an SMS-managed EAV. I went through several iterations in the 
> dataclass until I got it right. I promptly forgot how I did it, though. All I 
> needed was the track managed space filled with a huge (but empty) data set, 
> and then a small data set in cylinder managed space, just to have an example 
> for testing. Of course the product(s) were unable to deal with DSCB8/9's.
>
> I remember that I had not gotten the management class right, so the 
> huge thing got migrated to ML1 (since it was basically empty, it 
> apparently got severely compressed). Let me tell you that it was a 
> real problem trying to recall that data set, because it didn't want to 
> go back to the one and only EAV and failed allocation on my small 
> mod9's. I think I ended up hdeleting it and reallocating. :-(
>
> Barbara
>
> --
> For IBM-MAIN subscribe / signoff / archive access instructions, send 
> email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN



--
Mike A Schwab, Springfield IL USA
Where do Forest Rangers go to get away from it all?

--
For IBM-MAIN subscribe / signoff / archive access instructions, send email to 
lists...@listserv.ua.edu with the message: INFO IBM-MAIN

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Mike Schwab]

I often use this clist for Non-SMS recalls.

2 DSLIST xx.CLIST(HRECN) - 01.00Columns 1 00072
+1+2+3+4+5+6+7--
* Top of Data **
PROC 2 D V
  CONTROL   PROMPT NOSYMLIST NOLIST NOCONLIST NOCAPS   MSG NOFLUSH
  HSEND RECALL  VOL() UNIT(3390) FORCENON
END /* PROC */
 Bottom of Data 

On a TSO ISPF 3.4 Dataset list,
HRECN / volser;  data.set.name  ...
works well.

On Tue, May 10, 2016 at 12:52 PM, nitz-ibm  wrote:
>> One of the things I found appalling is that there is not a single
>> example of how to allocate an EAV dataset in the cylinder-managed space
>> anywhere in any IBM doc.  You can find EATTR=OPT if you know to look for
>> it, but do a search on how to allocate an EAV dataset.  Nothing.
> When I tried that on an ADCD system about two years ago with my own 10cyl 
> EAV, I found the same thing. I got lucky in that I had some presentation 
> handy that directed me to the EATTR attribute. I remember that I spend almost 
> an hour trying to figure it out.
>
>> So if you want to use EAV's, make them SMS.  That way you can set
>> EATTR(OPT) in all your DATACLASses and fuhgeddabotit.  Non-SMS EAV's
>> don't make a whole lotta sense, unless updating thousands of JCL members
>> is  your thing.
> I *was* using an SMS-managed EAV. I went through several iterations in the 
> dataclass until I got it right. I promptly forgot how I did it, though. All I 
> needed was the track managed space filled with a huge (but empty) data set, 
> and then a small data set in cylinder managed space, just to have an example 
> for testing. Of course the product(s) were unable to deal with DSCB8/9's.
>
> I remember that I had not gotten the management class right, so the huge 
> thing got migrated to ML1 (since it was basically empty, it apparently got 
> severely compressed). Let me tell you that it was a real problem trying to 
> recall that data set, because it didn't want to go back to the one and only 
> EAV and failed allocation on my small mod9's. I think I ended up hdeleting it 
> and reallocating. :-(
>
> Barbara
>
> --
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN



-- 
Mike A Schwab, Springfield IL USA
Where do Forest Rangers go to get away from it all?

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Tom Conley]

On 5/10/2016 1:53 PM, nitz-ibm wrote:

One of the things I found appalling is that there is not a single
example of how to allocate an EAV dataset in the cylinder-managed space
anywhere in any IBM doc.  You can find EATTR=OPT if you know to look for
it, but do a search on how to allocate an EAV dataset.  Nothing.

When I tried that on an ADCD system about two years ago with my own 10cyl 
EAV, I found the same thing. I got lucky in that I had some presentation handy 
that directed me to the EATTR attribute. I remember that I spend almost an hour 
trying to figure it out.


So if you want to use EAV's, make them SMS.  That way you can set
EATTR(OPT) in all your DATACLASses and fuhgeddabotit.  Non-SMS EAV's
don't make a whole lotta sense, unless updating thousands of JCL members
is  your thing.

I *was* using an SMS-managed EAV. I went through several iterations in the 
dataclass until I got it right. I promptly forgot how I did it, though. All I 
needed was the track managed space filled with a huge (but empty) data set, and 
then a small data set in cylinder managed space, just to have an example for 
testing. Of course the product(s) were unable to deal with DSCB8/9's.

I remember that I had not gotten the management class right, so the huge thing 
got migrated to ML1 (since it was basically empty, it apparently got severely 
compressed). Let me tell you that it was a real problem trying to recall that 
data set, because it didn't want to go back to the one and only EAV and failed 
allocation on my small mod9's. I think I ended up hdeleting it and 
reallocating. :-(

Barbara



Barbara,

I'll be covering this in the Bit Bucket at SHARE in Atlanta.  Thanks for 
giving me some more information about this.  If you go to zTechU in 
Munich, I'm presenting there, so please look me up.


Regards,
Tom Conley

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[nitz-ibm]

> One of the things I found appalling is that there is not a single 
> example of how to allocate an EAV dataset in the cylinder-managed space 
> anywhere in any IBM doc.  You can find EATTR=OPT if you know to look for 
> it, but do a search on how to allocate an EAV dataset.  Nothing. 
When I tried that on an ADCD system about two years ago with my own 10cyl 
EAV, I found the same thing. I got lucky in that I had some presentation handy 
that directed me to the EATTR attribute. I remember that I spend almost an hour 
trying to figure it out.

> So if you want to use EAV's, make them SMS.  That way you can set 
> EATTR(OPT) in all your DATACLASses and fuhgeddabotit.  Non-SMS EAV's 
> don't make a whole lotta sense, unless updating thousands of JCL members 
> is  your thing.
I *was* using an SMS-managed EAV. I went through several iterations in the 
dataclass until I got it right. I promptly forgot how I did it, though. All I 
needed was the track managed space filled with a huge (but empty) data set, and 
then a small data set in cylinder managed space, just to have an example for 
testing. Of course the product(s) were unable to deal with DSCB8/9's.

I remember that I had not gotten the management class right, so the huge thing 
got migrated to ML1 (since it was basically empty, it apparently got severely 
compressed). Let me tell you that it was a real problem trying to recall that 
data set, because it didn't want to go back to the one and only EAV and failed 
allocation on my small mod9's. I think I ended up hdeleting it and 
reallocating. :-(

Barbara

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Ed Jaffe]

On 5/10/2016 7:30 AM, Tom Conley wrote:

I suspect you'll have better luck if you search for EAS. Unfortunately,
that overloaded acronym is also a VTAM term, so you might want to use
something like "EAS data set" or "data set EAS" instead.



I ran many different searches trying to find this information, also 
using the term "cylinder-managed" space.  While it directed me to the 
EAV section of the DFSMS Using Datasets manual, not one example of how 
to allocate a dataset in the cylinder-managed space.


OK. I just tried searching for 'EAS' (no other words) using the 
_dreaded_ knowledge center 
http://www.ibm.com/support/knowledgecenter/SSLTBW_2.2.0/com.ibm.zos.v2r2/en/homepage.html


As predicted, there are several matches having to do with the VTAM APPL 
usage. Ignoring those, the 6th hit is the doc on EATTR= JCL parameter 
http://www.ibm.com/support/knowledgecenter/SSLTBW_2.2.0/com.ibm.zos.v2r2.ieab600/iea3b6_Subparameter_definition19.htm?lang=en 
which contains:



EATTR = OPT
Extended attributes are optional. The data set can have extended 
attributes and reside in EAS. This is the default value for VSAM data sets.

EATTR = NO
No extended attributes. The data set cannot have extended 
attributes (format 8 and 9 DSCBs) or reside in EAS. This is the default 
value for non-VSAM data sets.



The 7th hit is the EATTR doc for SVC 99 callers.

HTH

--
Edward E Jaffe
Phoenix Software International, Inc
831 Parkview Drive North
El Segundo, CA 90245
http://www.phoenixsoftware.com/

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Tom Conley]

On 5/10/2016 10:26 AM, Ed Jaffe wrote:

On 5/10/2016 6:59 AM, Tom Conley wrote:

You can find EATTR=OPT if you know to look for it, but do a search on
how to allocate an EAV dataset.  Nothing.


There is no such thing as an EAV data set. EAV is a type of volume with
both track-managed and cylinder-managed areas. The cylinder-managed area
is called the Extended Address Space or EAS.

I suspect you'll have better luck if you search for EAS. Unfortunately,
that overloaded acronym is also a VTAM term, so you might want to use
something like "EAS data set" or "data set EAS" instead.



Edward,

I ran many different searches trying to find this information, also 
using the term "cylinder-managed" space.  While it directed me to the 
EAV section of the DFSMS Using Datasets manual, not one example of how 
to allocate a dataset in the cylinder-managed space.


Regards,
Tom Conley

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Ed Jaffe]

On 5/10/2016 6:59 AM, Tom Conley wrote:
You can find EATTR=OPT if you know to look for it, but do a search on 
how to allocate an EAV dataset.  Nothing.


There is no such thing as an EAV data set. EAV is a type of volume with 
both track-managed and cylinder-managed areas. The cylinder-managed area 
is called the Extended Address Space or EAS.


I suspect you'll have better luck if you search for EAS. Unfortunately, 
that overloaded acronym is also a VTAM term, so you might want to use 
something like "EAS data set" or "data set EAS" instead.


--
Edward E Jaffe
Phoenix Software International, Inc
831 Parkview Drive North
El Segundo, CA 90245
http://www.phoenixsoftware.com/

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Tom Conley]

On 5/10/2016 3:04 AM, Tom Conley wrote:

On 5/10/2016 1:23 AM, Pinnacle wrote:

FYI,

I allocated a non-SMS EAV on my system with 111300 cylinders.  I try to
run three jobs with SPACE=(CYL,(1,5000),RLSE),DSNTYPE=LARGE.  The
first job runs, the other two fail with IEF257I SPACE REQUESTED NOT
AVAILABLE, on the EAV when it has 9 CYLS avail.  The allocation
message indicates that it attempted allocation on the EAV.  If I single
thread the jobs, no problem.  Is this a bug or a feature?  And no, I
can't submit a PMR.

Regards,
Tom Conley



I fingered it out.  In the failing cases, I was using an IBM Futility
which apparently wasn't updated with EAV code.  It was failing because
it was allocating in the track-managed space.  I'm going to verify that
with a Generic Tracker run, then I'll be pursuing an RFE with the culprits.

Regards,
Tom Conley

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN


I must thank the redoubtable James Mulder for the solution to this 
problem.  In addition to DSNTYPE=LARGE and an allocation over the BPV 
(break point value) in IGDSMSxx, you need to have EATTR=OPT in the JCL, 
DYNALLOC, TSO ALLOC, etc.  Why my other allocations worked is anyone's 
guess, but once I added EADDR=OPT to the failing JCL, that drove the 
failing jobs to allocate in the cylinder managed space.


One of the things I found appalling is that there is not a single 
example of how to allocate an EAV dataset in the cylinder-managed space 
anywhere in any IBM doc.  You can find EATTR=OPT if you know to look for 
it, but do a search on how to allocate an EAV dataset.  Nothing. 
Another futile email to mhvr...@us.ibm.com coming up.


So if you want to use EAV's, make them SMS.  That way you can set 
EATTR(OPT) in all your DATACLASses and fuhgeddabotit.  Non-SMS EAV's 
don't make a whole lotta sense, unless updating thousands of JCL members 
is  your thing.


Regards,
Tom Conley

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Tom Conley]

On 5/10/2016 1:23 AM, Pinnacle wrote:

FYI,

I allocated a non-SMS EAV on my system with 111300 cylinders.  I try to
run three jobs with SPACE=(CYL,(1,5000),RLSE),DSNTYPE=LARGE.  The
first job runs, the other two fail with IEF257I SPACE REQUESTED NOT
AVAILABLE, on the EAV when it has 9 CYLS avail.  The allocation
message indicates that it attempted allocation on the EAV.  If I single
thread the jobs, no problem.  Is this a bug or a feature?  And no, I
can't submit a PMR.

Regards,
Tom Conley



I fingered it out.  In the failing cases, I was using an IBM Futility 
which apparently wasn't updated with EAV code.  It was failing because 
it was allocating in the track-managed space.  I'm going to verify that 
with a Generic Tracker run, then I'll be pursuing an RFE with the culprits.


Regards,
Tom Conley

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: EAV bug or feature?

[Lizette Koehler]

Could you share the error message(s) you get?

Lizette


> -Original Message-
> From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
> Behalf Of Pinnacle
> Sent: Monday, May 09, 2016 10:24 PM
> To: IBM-MAIN@LISTSERV.UA.EDU
> Subject: EAV bug or feature?
> 
> FYI,
> 
> I allocated a non-SMS EAV on my system with 111300 cylinders.  I try to run
> three jobs with SPACE=(CYL,(1,5000),RLSE),DSNTYPE=LARGE.  The first job
> runs, the other two fail with IEF257I SPACE REQUESTED NOT AVAILABLE, on the
> EAV when it has 9 CYLS avail.  The allocation message indicates that it
> attempted allocation on the EAV.  If I single thread the jobs, no problem.  Is
> this a bug or a feature?  And no, I can't submit a PMR.
> 
> Regards,
> Tom Conley
> 

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

EAV bug or feature?

[Pinnacle]

FYI,

I allocated a non-SMS EAV on my system with 111300 cylinders.  I try to 
run three jobs with SPACE=(CYL,(1,5000),RLSE),DSNTYPE=LARGE.  The 
first job runs, the other two fail with IEF257I SPACE REQUESTED NOT 
AVAILABLE, on the EAV when it has 9 CYLS avail.  The allocation 
message indicates that it attempted allocation on the EAV.  If I single 
thread the jobs, no problem.  Is this a bug or a feature?  And no, I 
can't submit a PMR.


Regards,
Tom Conley

--
 



--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN