Discussion:
Accessing 65536 devices
(too old to reply)
Jesse 1 Robinson
2018-01-04 22:35:33 UTC
Permalink
Raw Message
We would like to be able to access >65535 device addresses (UCBs) from a single LPAR via a single IODF. The need is for bringing a new DASD subsystem online while retaining the old subsystem until all volumes can be copied across. We currently have spare UCBs available, but not enough to have all old and new devices online at the same time. Is this possible?

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office <===== NEW
***@sce.com<mailto:***@sce.com>


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Jerry Whitteridge
2018-01-04 23:03:20 UTC
Permalink
Raw Message
By using the CSSID/SS* in the IODF you can get way more than the
basic number

Example from a z13

CSS Devices in SS0 Devices in SS1 Devices in SS2 Devices
in SS3
/ ID Maximum + Actual Maximum + Actual Maximum + Actual Maximum
+ Actual
_ 0 65280 12384 65535 0 65535 0 65535 0
_ 1 65280 0 65535 0 65535 0 65535 0
_ 2 65280 0 65535 0 65535 0 65535 0
_ 3 65280 0 65535 0 65535 0 65535 0
_ 4 65280 0 65535 0 65535 0 65535 0
_ 5 65280 0 65535 0 65535 0 65535 0


--
Jerry Whitteridge

IBM Global Services
Delivery Manager
e-Mail: ***@ibm.com
Cell: 602 527 4871 <---- Note New Phone Number

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Jesse 1 Robinson
2018-01-04 23:19:29 UTC
Permalink
Raw Message
So this would involve 5-digit UCBs? We could probably manage that as long as all devices are concurrently accessible. Aren't commands like VARY limited to four digits?

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jerry Whitteridge
Sent: Thursday, January 04, 2018 3:05 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Fw: Accessing 65536 devices

By using the CSSID/SS* in the IODF you can get way more than the basic number

Example from a z13

CSS Devices in SS0 Devices in SS1 Devices in SS2 Devices
in SS3
/ ID Maximum + Actual Maximum + Actual Maximum + Actual Maximum + Actual
_ 0 65280 12384 65535 0 65535 0 65535 0
_ 1 65280 0 65535 0 65535 0 65535 0
_ 2 65280 0 65535 0 65535 0 65535 0
_ 3 65280 0 65535 0 65535 0 65535 0
_ 4 65280 0 65535 0 65535 0 65535 0
_ 5 65280 0 65535 0 65535 0 65535 0


--
Jerry Whitteridge

IBM Global Services
Delivery Manager
e-Mail: ***@ibm.com
Cell: 602 527 4871 <---- Note New Phone Number

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Nims,Alva John , Al
2018-01-04 23:32:00 UTC
Permalink
Raw Message
Ah, UCB's are HEX, so x'FFFF' = 65535, correct?

Al Nims
Systems Admin/Programmer 3
UFIT
University of Florida
(352) 273-1298

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jesse 1 Robinson
Sent: Thursday, January 04, 2018 6:20 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: Re: Accessing 65536 devices

So this would involve 5-digit UCBs? We could probably manage that as long as all devices are concurrently accessible. Aren't commands like VARY limited to four digits?

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jerry Whitteridge
Sent: Thursday, January 04, 2018 3:05 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Fw: Accessing 65536 devices

By using the CSSID/SS* in the IODF you can get way more than the basic number

Example from a z13

CSS Devices in SS0 Devices in SS1 Devices in SS2 Devices
in SS3
/ ID Maximum + Actual Maximum + Actual Maximum + Actual Maximum + Actual
_ 0 65280 12384 65535 0 65535 0 65535 0
_ 1 65280 0 65535 0 65535 0 65535 0
_ 2 65280 0 65535 0 65535 0 65535 0
_ 3 65280 0 65535 0 65535 0 65535 0
_ 4 65280 0 65535 0 65535 0 65535 0
_ 5 65280 0 65535 0 65535 0 65535 0


--
Jerry Whitteridge

IBM Global Services
Delivery Manager
e-Mail: ***@ibm.com
Cell: 602 527 4871 <---- Note New Phone Number

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

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Jesse 1 Robinson
2018-01-04 23:38:18 UTC
Permalink
Raw Message
Yes. 65,535 is the maximum number of devices that can be represented with four hex digits. If you add one more device, you need an extra hex digit. I understand that channel sets can allow for more than 64K UCBs in an IODF, but I thought that no single LPAR could address all of them at the same time. In order to copy old volumes to new, we have to have all of them online concurrently.

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Nims,Alva John (Al)
Sent: Thursday, January 04, 2018 3:33 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Re: Accessing 65536 devices

Ah, UCB's are HEX, so x'FFFF' = 65535, correct?

Al Nims
Systems Admin/Programmer 3
UFIT
University of Florida
(352) 273-1298

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jesse 1 Robinson
Sent: Thursday, January 04, 2018 6:20 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: Re: Accessing 65536 devices

So this would involve 5-digit UCBs? We could probably manage that as long as all devices are concurrently accessible. Aren't commands like VARY limited to four digits?

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jerry Whitteridge
Sent: Thursday, January 04, 2018 3:05 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Fw: Accessing 65536 devices

By using the CSSID/SS* in the IODF you can get way more than the basic number

Example from a z13

CSS Devices in SS0 Devices in SS1 Devices in SS2 Devices
in SS3
/ ID Maximum + Actual Maximum + Actual Maximum + Actual Maximum + Actual
_ 0 65280 12384 65535 0 65535 0 65535 0
_ 1 65280 0 65535 0 65535 0 65535 0
_ 2 65280 0 65535 0 65535 0 65535 0
_ 3 65280 0 65535 0 65535 0 65535 0
_ 4 65280 0 65535 0 65535 0 65535 0
_ 5 65280 0 65535 0 65535 0 65535 0


--
Jerry Whitteridge

IBM Global Services
Delivery Manager
e-Mail: ***@ibm.com
Cell: 602 527 4871 <---- Note New Phone Number


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Jesse 1 Robinson
2018-01-04 23:56:16 UTC
Permalink
Raw Message
I think I out-clevered myself. x'FFFF' is the highest numbered device, but since x'0000' is also a valid device number, the total number is 65,536 UCBs. In any case, we need to go beyond that limit in a single LPAR. It's only temporary, but the alternative is drawn out and laborious.

1. Identify some subset of the new subsystem that can be accommodated in the current IODF.
2. Activate a new IODF with that subset defined.
3. Copy old volumes to new within that range.
4. Activate a new IODF that deletes the old volumes already copied.
5. Rinse and repeat until the last of the old volumes have been copied.
6. Activate a new IODF that deletes the old subsystem entirely.

We've been through this house of mirrors many times in the past. It takes months. I'm hoping for a linear approach that allows us to copy all volumes in a single exercise.


.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jesse 1 Robinson
Sent: Thursday, January 04, 2018 3:40 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Re: Accessing 65536 devices

Yes. 65,535 is the maximum number of devices that can be represented with four hex digits. If you add one more device, you need an extra hex digit. I understand that channel sets can allow for more than 64K UCBs in an IODF, but I thought that no single LPAR could address all of them at the same time. In order to copy old volumes to new, we have to have all of them online concurrently.

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Nims,Alva John (Al)
Sent: Thursday, January 04, 2018 3:33 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Re: Accessing 65536 devices

Ah, UCB's are HEX, so x'FFFF' = 65535, correct?

Al Nims
Systems Admin/Programmer 3
UFIT
University of Florida
(352) 273-1298

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jesse 1 Robinson
Sent: Thursday, January 04, 2018 6:20 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: Re: Accessing 65536 devices

So this would involve 5-digit UCBs? We could probably manage that as long as all devices are concurrently accessible. Aren't commands like VARY limited to four digits?

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jerry Whitteridge
Sent: Thursday, January 04, 2018 3:05 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Fw: Accessing 65536 devices

By using the CSSID/SS* in the IODF you can get way more than the basic number

Example from a z13

CSS Devices in SS0 Devices in SS1 Devices in SS2 Devices
in SS3
/ ID Maximum + Actual Maximum + Actual Maximum + Actual Maximum + Actual
_ 0 65280 12384 65535 0 65535 0 65535 0
_ 1 65280 0 65535 0 65535 0 65535 0
_ 2 65280 0 65535 0 65535 0 65535 0
_ 3 65280 0 65535 0 65535 0 65535 0
_ 4 65280 0 65535 0 65535 0 65535 0
_ 5 65280 0 65535 0 65535 0 65535 0


--
Jerry Whitteridge

IBM Global Services
Delivery Manager
e-Mail: ***@ibm.com
Cell: 602 527 4871 <---- Note New Phone Number


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Ward, Mike S
2018-01-05 15:21:58 UTC
Permalink
Raw Message
Innovation has a product called FDRPAS that can be rented that does the copies while the system is up and running. We have used it several times and it works great.

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jesse 1 Robinson
Sent: Thursday, January 04, 2018 5:58 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: Re: Accessing 65536 devices

I think I out-clevered myself. x'FFFF' is the highest numbered device, but since x'0000' is also a valid device number, the total number is 65,536 UCBs. In any case, we need to go beyond that limit in a single LPAR. It's only temporary, but the alternative is drawn out and laborious.

1. Identify some subset of the new subsystem that can be accommodated in the current IODF.
2. Activate a new IODF with that subset defined.
3. Copy old volumes to new within that range.
4. Activate a new IODF that deletes the old volumes already copied.
5. Rinse and repeat until the last of the old volumes have been copied.
6. Activate a new IODF that deletes the old subsystem entirely.

We've been through this house of mirrors many times in the past. It takes months. I'm hoping for a linear approach that allows us to copy all volumes in a single exercise.


.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jesse 1 Robinson
Sent: Thursday, January 04, 2018 3:40 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Re: Accessing 65536 devices

Yes. 65,535 is the maximum number of devices that can be represented with four hex digits. If you add one more device, you need an extra hex digit. I understand that channel sets can allow for more than 64K UCBs in an IODF, but I thought that no single LPAR could address all of them at the same time. In order to copy old volumes to new, we have to have all of them online concurrently.

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Nims,Alva John (Al)
Sent: Thursday, January 04, 2018 3:33 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Re: Accessing 65536 devices

Ah, UCB's are HEX, so x'FFFF' = 65535, correct?

Al Nims
Systems Admin/Programmer 3
UFIT
University of Florida
(352) 273-1298

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jesse 1 Robinson
Sent: Thursday, January 04, 2018 6:20 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: Re: Accessing 65536 devices

So this would involve 5-digit UCBs? We could probably manage that as long as all devices are concurrently accessible. Aren't commands like VARY limited to four digits?

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jerry Whitteridge
Sent: Thursday, January 04, 2018 3:05 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Fw: Accessing 65536 devices

By using the CSSID/SS* in the IODF you can get way more than the basic number

Example from a z13

CSS Devices in SS0 Devices in SS1 Devices in SS2 Devices
in SS3
/ ID Maximum + Actual Maximum + Actual Maximum + Actual Maximum + Actual
_ 0 65280 12384 65535 0 65535 0 65535 0
_ 1 65280 0 65535 0 65535 0 65535 0
_ 2 65280 0 65535 0 65535 0 65535 0
_ 3 65280 0 65535 0 65535 0 65535 0
_ 4 65280 0 65535 0 65535 0 65535 0
_ 5 65280 0 65535 0 65535 0 65535 0


--
Jerry Whitteridge

IBM Global Services
Delivery Manager
e-Mail: ***@ibm.com
Cell: 602 527 4871 <---- Note New Phone Number


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

==========================
This email, and any files transmitted with it, is confidential and intended solely for the use of the individual or entity to which it is addressed. If you have received this email in error, please notify the system manager. This message contains confidential information and is intended only for the individual named. If you are not the named addressee, you should not disseminate, distribute or copy this e-mail. Please notify the sender immediately by e-mail if you have received this message by mistake and delete this e-mail from your system. If you are not the intended recipient, you are notified that disclosing, copying, distributing or taking any action in reliance on the contents of this information is strictly prohibited.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
John Eells
2018-01-05 20:05:33 UTC
Permalink
Raw Message
Post by Ward, Mike S
Innovation has a product called FDRPAS that can be rented that does the copies while the system is up and running. We have used it several times and it works great.
<snip>

I have no experience with FDRPAS or its IBM counterparts (like TDMF, if
I recall correctly). However, unless they are able to talk to devices
in subchannel sets other than SCS0, none would solve Skip's problem.

I'm afraid I don't know whether any of them talk to devices outside
SCS0, and nobody I know who might be able to answer that question seems
to be around this afternoon.

I am not really a Storage guy, so I don't know whether it fits the bill
from an ease-of-use (or cost) point of view, but as PPRC secondaries and
FlashCopy targets *can* be defined in SCSs other than zero, one could
for example potentially establish (for example) the needed PPRC pairs,
let PPRC do the copies, and then take the primaries offline. Whether
this would work, and whether it could even be done with HyperSwap so
there was no required outage, is another question I'd need those same
people for...the ones that aren't around today, I mean.
--
John Eells
IBM Poughkeepsie
***@us.ibm.com

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Edward Finnell
2018-01-05 21:18:00 UTC
Permalink
Raw Message
Ain't they gots no eMail?


In a message dated 1/5/2018 2:07:03 PM Central Standard Time, ***@US.IBM.COM writes:

 
there was no required outage, is another question I'd need those same

people for...the ones that aren't around today, I mean.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
John Eells
2018-01-09 12:04:04 UTC
Permalink
Raw Message
Post by Edward Finnell
Ain't they gots no eMail?
Of course they do. But, with the Storage and Replication teams spread
as they are across several IBM sites and no fewer than 8 timezones, it's
usually faster (both in real time and my time) to deal with people who
are available in the moment to find the right one or ones to answer a
particular question.

And, all appearances to the contrary, I do have a day job. ;-)
--
John Eells
IBM Poughkeepsie
***@us.ibm.com

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Edward Finnell
2018-01-09 12:29:54 UTC
Permalink
Raw Message
Guess I could envision something like a virtual BRICK for hot topics leading to potential sales or upgrades.


In a message dated 1/9/2018 6:05:35 AM Central Standard Time, ***@US.IBM.COM writes:

 
to deal with people who

are available in the moment to find the right one or ones to answer a
particular question.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Edward Gould
2018-01-09 12:43:39 UTC
Permalink
Raw Message
Post by Edward Finnell
Ain't they gots no eMail?
Of course they do. But, with the Storage and Replication teams spread as they are across several IBM sites and no fewer than 8 timezones, it's usually faster (both in real time and my time) to deal with people who are available in the moment to find the right one or ones to answer a particular question.
And, all appearances to the contrary, I do have a day job. ;-)
John,

And you do a great job! Seriously, I think that you contribute greatly to the list and you do respond to issues that are relevant to you in an extremely short amount of time, with real answers.
Thank you for all of your hard work.

Ed


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Peter Hunkeler
2018-01-09 18:31:58 UTC
Permalink
Raw Message
Post by Edward Gould
And you do a great job! Seriously, I think that you contribute greatly to the list and you do respond to issues that are relevant to you in an extremely short amount of time, with real answers.
Thank you for all of your hard work.


+1 vote (assuming I only have got one :-)


--
Peter Hunkeler







----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Edward Gould
2018-01-10 03:50:00 UTC
Permalink
Raw Message
Of course they do. But, with the Storage and Replication teams spread as they are across several IBM sites and no fewer than 8 timezones, it's usually faster (both in real time and my time) to deal with people who are available in the moment to find the right one or ones to answer a particular question.
And, all appearances to the contrary, I do have a day job. ;-)
--
John Eells
IBM Poughkeepsie
John,

I don’t expect you to answer this, but here it goes. IBM seems to have gotten out of the answering questions mode and seemingly wants to charge for answering questions. Isn’t this extremely shortsighted on IBM?
IBM SE’s used to field these questions and gave answers back to the customer same day or less.
IBM dropped SE’s so now the customer is left trying to decide what to do. This seems to be not only stupid but bad for sales. We now have companies that compete with IBM and the only person that management talks to is sales reps. We know sales reps aren’t all that reputable (I won’t repeat the half truths and quarter lies I have heard from them). To top it off they have the customer calling a 1-800 number who knows ZERO about the customer.
If I walked into a car showroom and if I had a question to ask about a car and were told that for every question I had I had rot pay say $100, I would walk out of the showroom and avoid that car type again. I see no difference than an SE being a sales rep that a car dealer.
Ed


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Tom Brennan
2018-01-10 04:31:40 UTC
Permalink
Raw Message
... and now back to watching Star Trek. It's the one with "The world is
hollow, and I have touched the sky".

P.S. Thanks to all the IBMer's who help people as best they can every day.
Post by Edward Gould
John,
I don’t expect you to answer this, but here it goes. IBM seems to have gotten out of the answering questions mode and seemingly wants to charge for answering questions. Isn’t this extremely shortsighted on IBM?
IBM SE’s used to field these questions and gave answers back to the customer same day or less.
IBM dropped SE’s so now the customer is left trying to decide what to do. This seems to be not only stupid but bad for sales. We now have companies that compete with IBM and the only person that management talks to is sales reps. We know sales reps aren’t all that reputable (I won’t repeat the half truths and quarter lies I have heard from them). To top it off they have the customer calling a 1-800 number who knows ZERO about the customer.
If I walked into a car showroom and if I had a question to ask about a car and were told that for every question I had I had rot pay say $100, I would walk out of the showroom and avoid that car type again. I see no difference than an SE being a sales rep that a car dealer.
Ed
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Seymour J Metz
2018-01-10 18:58:45 UTC
Permalink
Raw Message
It used to be that SE's were free and, like SHARE, offered IBM valuable market research. The SE generally saw his job as making it easier for the customer to use IBM products and forestalling temptation to look at other vendors.


--
Shmuel (Seymour J.) Metz
http://mason.gmu.edu/~smetz3

________________________________________
From: IBM Mainframe Discussion List <IBM-***@listserv.ua.edu> on behalf of Edward Gould <***@COMCAST.NET>
Sent: Tuesday, January 9, 2018 10:51 PM
To: IBM-***@listserv.ua.edu
Subject: Re: Accessing 65536 devices
Of course they do. But, with the Storage and Replication teams spread as they are across several IBM sites and no fewer than 8 timezones, it's usually faster (both in real time and my time) to deal with people who are available in the moment to find the right one or ones to answer a particular question.
And, all appearances to the contrary, I do have a day job. ;-)
--
John Eells
IBM Poughkeepsie
John,

I don’t expect you to answer this, but here it goes. IBM seems to have gotten out of the answering questions mode and seemingly wants to charge for answering questions. Isn’t this extremely shortsighted on IBM?
IBM SE’s used to field these questions and gave answers back to the customer same day or less.
IBM dropped SE’s so now the customer is left trying to decide what to do. This seems to be not only stupid but bad for sales. We now have companies that compete with IBM and the only person that management talks to is sales reps. We know sales reps aren’t all that reputable (I won’t repeat the half truths and quarter lies I have heard from them). To top it off they have the customer calling a 1-800 number who knows ZERO about the customer.
If I walked into a car showroom and if I had a question to ask about a car and were told that for every question I had I had rot pay say $100, I would walk out of the showroom and avoid that car type again. I see no difference than an SE being a sales rep that a car dealer.
Ed


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

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Ed Jaffe
2018-01-06 03:37:57 UTC
Permalink
Raw Message
Post by John Eells
Post by Ward, Mike S
Innovation has a product called FDRPAS that can be rented that does
the copies while the system is up and running. We have used it
several times and it works great.
<snip>
I have no experience with FDRPAS or its IBM counterparts (like TDMF,
if I recall correctly).  However, unless they are able to talk to
devices in subchannel sets other than SCS0, none would solve Skip's
problem.
Yes, they would! What Skip should do is upgrade from whatever ancient
DASD configuration has been carried forward at SCE for decades to
modern, 21st-century DASD geometries. FDRPAS and similar products will
help him do that seamlessly. Even being past the half-way point on UCB
count won't be an issue if he consolidates his mod-3 DASD pools to
mod-27s, which are nine times larger. Even an extremely-conservative
conversion to (still quite ancient) mod-9s can cut the device address
footprint by 2/3!
--
Phoenix Software International
Edward E. Jaffe
831 Parkview Drive North
El Segundo, CA 90245
http://www.phoenixsoftware.com/

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Feller, Paul
2018-01-06 19:42:15 UTC
Permalink
Raw Message
Using FDRPAS or TDMF to go from smaller volumes to larger volumes will take you only so far. You can move one MOD3 to one MOD9 (or MOD whatever). After that you have to move datasets. I know there is another FDR product that will help move datasets, but even that has limitations. Naturally you could use SMS to help by disabling new allocations to the MOD3 and over time things may move as datasets are reallocated. At some point you will have to have subsystems down (or lpars down) to move datasets.

As we have swapped out DASD subsystems we have gone through the pain of eliminating as many MOD3 volumes as we could. There came a point when we had to get business unit approval to take things like CICS or DB2 or IMS or MQ or even whole lpars down to move datasets.

Thanks..

Paul Feller
AGT Mainframe Technical Support


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Ed Jaffe
Sent: Friday, January 05, 2018 21:39
To: IBM-***@LISTSERV.UA.EDU
Subject: Re: Accessing 65536 devices
Post by John Eells
Post by Ward, Mike S
Innovation has a product called FDRPAS that can be rented that does
the copies while the system is up and running. We have used it
several times and it works great.
<snip>
I have no experience with FDRPAS or its IBM counterparts (like TDMF,
if I recall correctly).  However, unless they are able to talk to
devices in subchannel sets other than SCS0, none would solve Skip's
problem.
Yes, they would! What Skip should do is upgrade from whatever ancient
DASD configuration has been carried forward at SCE for decades to
modern, 21st-century DASD geometries. FDRPAS and similar products will
help him do that seamlessly. Even being past the half-way point on UCB
count won't be an issue if he consolidates his mod-3 DASD pools to
mod-27s, which are nine times larger. Even an extremely-conservative
conversion to (still quite ancient) mod-9s can cut the device address
footprint by 2/3!
--
Phoenix Software International
Edward E. Jaffe
831 Parkview Drive North
El Segundo, CA 90245
https://urldefense.proofpoint.com/v2/url?u=http-3A__www.phoenixsoftware.com_&d=DwIDaQ&c=9g4MJkl2VjLjS6R4ei18BA&r=eUhu3PeeWy6RTndlJVKembFjFsvwCa8eeU_gm45NyOc&m=7YFFfIHl6r7-leHPbtq0ChsKfyZLk4xwNbMMmmj3wVo&s=DVbod8a0Hfy-mVgiikk-AOW5mJut8kVcuXk8CMZ5T_8&e=

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

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Mike Schwab
2018-01-06 20:26:13 UTC
Permalink
Raw Message
TDMF does include a logical data mover. When a database is closed it
will swap over. During shutdown for an IPL it should get the
remaining datasets.
If your new dasd allows dynamic growing a volume, you can grow a
non-EAV to the maximum EAV volume size, but no bigger.

On Sat, Jan 6, 2018 at 1:43 PM, Feller, Paul
Post by Feller, Paul
Using FDRPAS or TDMF to go from smaller volumes to larger volumes will take you only so far. You can move one MOD3 to one MOD9 (or MOD whatever). After that you have to move datasets. I know there is another FDR product that will help move datasets, but even that has limitations. Naturally you could use SMS to help by disabling new allocations to the MOD3 and over time things may move as datasets are reallocated. At some point you will have to have subsystems down (or lpars down) to move datasets.
As we have swapped out DASD subsystems we have gone through the pain of eliminating as many MOD3 volumes as we could. There came a point when we had to get business unit approval to take things like CICS or DB2 or IMS or MQ or even whole lpars down to move datasets.
Thanks..
Paul Feller
AGT Mainframe Technical Support
-----Original Message-----
Sent: Friday, January 05, 2018 21:39
Subject: Re: Accessing 65536 devices
Post by John Eells
Post by Ward, Mike S
Innovation has a product called FDRPAS that can be rented that does
the copies while the system is up and running. We have used it
several times and it works great.
<snip>
I have no experience with FDRPAS or its IBM counterparts (like TDMF,
if I recall correctly). However, unless they are able to talk to
devices in subchannel sets other than SCS0, none would solve Skip's
problem.
Yes, they would! What Skip should do is upgrade from whatever ancient
DASD configuration has been carried forward at SCE for decades to
modern, 21st-century DASD geometries. FDRPAS and similar products will
help him do that seamlessly. Even being past the half-way point on UCB
count won't be an issue if he consolidates his mod-3 DASD pools to
mod-27s, which are nine times larger. Even an extremely-conservative
conversion to (still quite ancient) mod-9s can cut the device address
footprint by 2/3!
--
Phoenix Software International
Edward E. Jaffe
831 Parkview Drive North
El Segundo, CA 90245
https://urldefense.proofpoint.com/v2/url?u=http-3A__www.phoenixsoftware.com_&d=DwIDaQ&c=9g4MJkl2VjLjS6R4ei18BA&r=eUhu3PeeWy6RTndlJVKembFjFsvwCa8eeU_gm45NyOc&m=7YFFfIHl6r7-leHPbtq0ChsKfyZLk4xwNbMMmmj3wVo&s=DVbod8a0Hfy-mVgiikk-AOW5mJut8kVcuXk8CMZ5T_8&e=
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
--
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 ***@listserv.ua.edu with the message: INFO IBM-MAIN
Ron hawkins
2018-01-07 10:53:03 UTC
Permalink
Raw Message
Mike,

The 2nd part of your reply will require 3390-A to support DVE.

Hopefully, that is all that used for new DASD subsystems, for a few years now.


Ron

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Mike Schwab
Sent: Saturday, January 6, 2018 12:27 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: Re: [IBM-MAIN] Accessing 65536 devices

TDMF does include a logical data mover. When a database is closed it will swap over. During shutdown for an IPL it should get the remaining datasets.
If your new dasd allows dynamic growing a volume, you can grow a non-EAV to the maximum EAV volume size, but no bigger.
Post by Feller, Paul
Using FDRPAS or TDMF to go from smaller volumes to larger volumes will take you only so far. You can move one MOD3 to one MOD9 (or MOD whatever). After that you have to move datasets. I know there is another FDR product that will help move datasets, but even that has limitations. Naturally you could use SMS to help by disabling new allocations to the MOD3 and over time things may move as datasets are reallocated. At some point you will have to have subsystems down (or lpars down) to move datasets.
As we have swapped out DASD subsystems we have gone through the pain of eliminating as many MOD3 volumes as we could. There came a point when we had to get business unit approval to take things like CICS or DB2 or IMS or MQ or even whole lpars down to move datasets.
Thanks..
Paul Feller
AGT Mainframe Technical Support
-----Original Message-----
Sent: Friday, January 05, 2018 21:39
Subject: Re: Accessing 65536 devices
Post by John Eells
Post by Ward, Mike S
Innovation has a product called FDRPAS that can be rented that does
the copies while the system is up and running. We have used it
several times and it works great.
<snip>
I have no experience with FDRPAS or its IBM counterparts (like TDMF,
if I recall correctly). However, unless they are able to talk to
devices in subchannel sets other than SCS0, none would solve Skip's
problem.
Yes, they would! What Skip should do is upgrade from whatever ancient
DASD configuration has been carried forward at SCE for decades to
modern, 21st-century DASD geometries. FDRPAS and similar products will
help him do that seamlessly. Even being past the half-way point on UCB
count won't be an issue if he consolidates his mod-3 DASD pools to
mod-27s, which are nine times larger. Even an extremely-conservative
conversion to (still quite ancient) mod-9s can cut the device address
footprint by 2/3!
--
Phoenix Software International
Edward E. Jaffe
831 Parkview Drive North
El Segundo, CA 90245
https://urldefense.proofpoint.com/v2/url?u=http-3A__www.phoenixsoftwar
ecom_&d=DwIDaQ&c=9g4MJkl2VjLjS6R4ei18BA&r=eUhu3PeeWy6RTndlJVKembFjFsvw
Ca8eeU_gm45NyOc&m=7YFFfIHl6r7-leHPbtq0ChsKfyZLk4xwNbMMmmj3wVo&s=DVbod8
a0Hfy-mVgiikk-AOW5mJut8kVcuXk8CMZ5T_8&e=
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions, send
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions, send
--
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 ***@listserv.ua.edu with the message: INFO IBM-MAIN

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Ed Jaffe
2018-01-07 16:07:57 UTC
Permalink
Raw Message
Post by Mike Schwab
If your new dasd allows dynamic growing a volume, you can grow a
non-EAV to the maximum EAV volume size, but no bigger.
That was my favorite feature of our old DS8100! We lost that feature
when we upgraded to DS8870, but so far it hasn't been a problem because
we planned fairly well.

Having said that, we're running a bit short on mod-216 volumes. I was
thinking about removing eight of our mod-27s to make room for one more
of these bad boys! Of course, that kind of re-allocation can be done
with the DS8870, just not quite as easily as with the DS8100 because you
can't grow a device in place.

The downside to in-place growth is varying levels of support from, and
different procedures for, Linux for Z (we use RHEL), z/VM, z/VSE, and
older releases of z/OS. You need to be careful or things can go FUBAR in
a hurry! (Volume backups are a wonderful thing...)
--
Phoenix Software International
Edward E. Jaffe
831 Parkview Drive North
El Segundo, CA 90245
http://www.phoenixsoftware.com/

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
John Eells
2018-01-09 12:01:20 UTC
Permalink
Raw Message
I am told that TDMF provides access to devices via SCS0 only.
Post by John Eells
Post by Ward, Mike S
Innovation has a product called FDRPAS that can be rented that does
the copies while the system is up and running. We have used it several
times and it works great.
<snip>
I have no experience with FDRPAS or its IBM counterparts (like TDMF, if
I recall correctly). However, unless they are able to talk to devices
in subchannel sets other than SCS0, none would solve Skip's problem.
I'm afraid I don't know whether any of them talk to devices outside
SCS0, and nobody I know who might be able to answer that question seems
to be around this afternoon.
I am not really a Storage guy, so I don't know whether it fits the bill
from an ease-of-use (or cost) point of view, but as PPRC secondaries and
FlashCopy targets *can* be defined in SCSs other than zero, one could
for example potentially establish (for example) the needed PPRC pairs,
let PPRC do the copies, and then take the primaries offline. Whether
this would work, and whether it could even be done with HyperSwap so
there was no required outage, is another question I'd need those same
people for...the ones that aren't around today, I mean.
--
John Eells
IBM Poughkeepsie
***@us.ibm.com

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Ravi Gaur
2018-01-05 08:48:38 UTC
Permalink
Raw Message
Why can't you define the CSS id more than 1 and then have partition access to that as well ....so example as below

CSS Devices in SS0 Devices in SS1 Devices in SS2 Devices in SS3
/ ID Maximum + Actual Maximum + Actual Maximum + Actual Maximum + Actual
_ 0 65280 10684 65535 0 65535 0 65535 0
_ 1 65280 0 65535 0 65535 0 65535 0
_ 2 65280 364 65535 0 65535 0 65535 0
_ 3 65280 3218 65535 0 65535 0 65535 0
_ 4 65280 0 65535 0 65535 0 65535 0
_ 5 65280 0 65535 0 65535 0 65535 0

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
John Eells
2018-01-05 11:51:23 UTC
Permalink
Raw Message
Post by Jesse 1 Robinson
We would like to be able to access >65535 device addresses (UCBs) from a single LPAR via a single IODF. The need is for bringing a new DASD subsystem online while retaining the old subsystem until all volumes can be copied across. We currently have spare UCBs available, but not enough to have all old and new devices online at the same time. Is this possible?
You can have a maximum of 65,280* devices (not quite 64K) in Subchannel
Set 0. There are some things that can be accessed via other SCSs, but
IIRC the exceptions don't allow you to have more than 65,280 primary
devices online concurrently. That includes all devices, of course,
including networking, tape, consoles, and such.

If replication and PAVs are eating into SCS0, you can move the
secondaries to another SCS to "make room" in SCS0 for additional primary
devices. When last I looked, these could be defined in other SCSs:

- PPRC secondary devices
- PAV aliases
- FlashCopy target devices

Perhaps that would free up enough slots in SCS0 for you. (It seems to
work for almost everyone.)

There is a very narrow exception to the above for IPLing from nonzero
subchannel sets for a small number of volumes. (I don't have that list
handy, but IIRC it includes at least the IPL and IODF volumes but not
many more than that.)

* We reserve 256 subchannels in SCS0 for future use.
--
John Eells
IBM Poughkeepsie
***@us.ibm.com

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Feller, Paul
2018-01-05 15:24:01 UTC
Permalink
Raw Message
We started putting PAV aliases into SS1 to save on UCBs. It does make for an interesting IODF. The only issue was with z/VM. We have several z/VM lpars along with our z/OS lpars. z/VM currently can't handle PAV aliases in SS1, they have to be in SS0.

Thanks..

Paul Feller
AGT Mainframe Technical Support

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of John Eells
Sent: Friday, January 05, 2018 05:53
To: IBM-***@LISTSERV.UA.EDU
Subject: Re: Accessing 65536 devices
Post by Jesse 1 Robinson
We would like to be able to access >65535 device addresses (UCBs) from a single LPAR via a single IODF. The need is for bringing a new DASD subsystem online while retaining the old subsystem until all volumes can be copied across. We currently have spare UCBs available, but not enough to have all old and new devices online at the same time. Is this possible?
You can have a maximum of 65,280* devices (not quite 64K) in Subchannel
Set 0. There are some things that can be accessed via other SCSs, but
IIRC the exceptions don't allow you to have more than 65,280 primary
devices online concurrently. That includes all devices, of course,
including networking, tape, consoles, and such.

If replication and PAVs are eating into SCS0, you can move the
secondaries to another SCS to "make room" in SCS0 for additional primary
devices. When last I looked, these could be defined in other SCSs:

- PPRC secondary devices
- PAV aliases
- FlashCopy target devices

Perhaps that would free up enough slots in SCS0 for you. (It seems to
work for almost everyone.)

There is a very narrow exception to the above for IPLing from nonzero
subchannel sets for a small number of volumes. (I don't have that list
handy, but IIRC it includes at least the IPL and IODF volumes but not
many more than that.)

* We reserve 256 subchannels in SCS0 for future use.
--
John Eells
IBM Poughkeepsie
***@us.ibm.com

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

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Ed Jaffe
2018-01-05 22:10:57 UTC
Permalink
Raw Message
Post by John Eells
Post by Jesse 1 Robinson
We would like to be able to access >65535 device addresses (UCBs)
from a single LPAR via a single IODF. The need is for bringing a new
DASD subsystem online while retaining the old subsystem until all
volumes can be copied across. We currently have spare UCBs available,
but not enough to have all old and new devices online at the same
time. Is this possible?
You can have a maximum of 65,280* devices (not quite 64K) in
Subchannel Set 0.  There are some things that can be accessed via
other SCSs, but IIRC the exceptions don't allow you to have more than
65,280 primary devices online concurrently.  That includes all
devices, of course, including networking, tape, consoles, and such.
Skip, you might consider standardizing on larger capacity devices where
it makes sense to do so. We were *very* constrained on the total number
of UCBs we could have, based on architectural point-to-point (no switch)
FICON limits, and so we eliminated ALL of our mod-3 3390s a couple of
years ago. We now have only mod-9, mod-27 and mod-216 devices. (FWIW,
the most popular -- and most numerous -- ended up being the mod-27s.)

It's like a breath of fresh air!
--
Phoenix Software International
Edward E. Jaffe
831 Parkview Drive North
El Segundo, CA 90245
http://www.phoenixsoftware.com/

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Seymour J Metz
2018-01-05 21:16:30 UTC
Permalink
Raw Message
UCBs are not devices. There is no way for a single z/OS system to access more than 64Ki minus 256 devices, although you can add additional exposures to existing devices.


--
Shmuel (Seymour J.) Metz
http://mason.gmu.edu/~smetz3

________________________________________
From: IBM Mainframe Discussion List <IBM-***@listserv.ua.edu> on behalf of Jesse 1 Robinson <***@SCE.COM>
Sent: Thursday, January 4, 2018 5:35 PM
To: IBM-***@listserv.ua.edu
Subject: Accessing 65536 devices

We would like to be able to access >65535 device addresses (UCBs) from a single LPAR via a single IODF. The need is for bringing a new DASD subsystem online while retaining the old subsystem until all volumes can be copied across. We currently have spare UCBs available, but not enough to have all old and new devices online at the same time. Is this possible?

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office <===== NEW
***@sce.com<mailto:***@sce.com>


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

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Jesse 1 Robinson
2018-01-05 22:26:42 UTC
Permalink
Raw Message
We have device count reduction on our wish list. Unfortunately we know of no way to accomplish that without taking (multiple) application outages to consolidate data on fewer larger volumes. High impact and high risk with the benefit going mainly to infrastructure care-and-feeders. Our stated goal of rolling-IPLs whenever possible would be severely compromised.

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Ed Jaffe
Sent: Friday, January 05, 2018 2:12 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Re: Accessing 65536 devices
Post by John Eells
Post by Jesse 1 Robinson
We would like to be able to access >65535 device addresses (UCBs)
from a single LPAR via a single IODF. The need is for bringing a new
DASD subsystem online while retaining the old subsystem until all
volumes can be copied across. We currently have spare UCBs available,
but not enough to have all old and new devices online at the same
time. Is this possible?
You can have a maximum of 65,280* devices (not quite 64K) in
Subchannel Set 0.  There are some things that can be accessed via
other SCSs, but IIRC the exceptions don't allow you to have more than
65,280 primary devices online concurrently.  That includes all
devices, of course, including networking, tape, consoles, and such.
Skip, you might consider standardizing on larger capacity devices where it makes sense to do so. We were *very* constrained on the total number of UCBs we could have, based on architectural point-to-point (no switch) FICON limits, and so we eliminated ALL of our mod-3 3390s a couple of years ago. We now have only mod-9, mod-27 and mod-216 devices. (FWIW, the most popular -- and most numerous -- ended up being the mod-27s.)

It's like a breath of fresh air!

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


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Ed Jaffe
2018-01-05 22:44:05 UTC
Permalink
Raw Message
Post by Jesse 1 Robinson
We have device count reduction on our wish list. Unfortunately we know of no way to accomplish that without taking (multiple) application outages to consolidate data on fewer larger volumes. High impact and high risk with the benefit going mainly to infrastructure care-and-feeders. Our stated goal of rolling-IPLs whenever possible would be severely compromised.
All of the popular z/OS "disk migration" ISV tools allow you to do this
without taking down applications. Track by track migration of extents
that are not currently "open" is child's play. The sophistication is in
the handling of the "in use" tracks on the old volumes. I/Os against
them are intercepted and transparently redirected to the new volumes
until such time as those extents are eventually closed. After that, it's
BAU.
--
Phoenix Software International
Edward E. Jaffe
831 Parkview Drive North
El Segundo, CA 90245
http://www.phoenixsoftware.com/

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Ron hawkins
2018-01-06 02:48:17 UTC
Permalink
Raw Message
FDRMOVE anyone.



-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Jesse 1 Robinson
Sent: Friday, January 5, 2018 2:28 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: Re: [IBM-MAIN] Accessing 65536 devices

We have device count reduction on our wish list. Unfortunately we know of no way to accomplish that without taking (multiple) application outages to consolidate data on fewer larger volumes. High impact and high risk with the benefit going mainly to infrastructure care-and-feeders. Our stated goal of rolling-IPLs whenever possible would be severely compromised.

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Ed Jaffe
Sent: Friday, January 05, 2018 2:12 PM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Re: Accessing 65536 devices
Post by John Eells
Post by Jesse 1 Robinson
We would like to be able to access >65535 device addresses (UCBs)
from a single LPAR via a single IODF. The need is for bringing a new
DASD subsystem online while retaining the old subsystem until all
volumes can be copied across. We currently have spare UCBs available,
but not enough to have all old and new devices online at the same
time. Is this possible?
You can have a maximum of 65,280* devices (not quite 64K) in
Subchannel Set 0. There are some things that can be accessed via
other SCSs, but IIRC the exceptions don't allow you to have more than
65,280 primary devices online concurrently. That includes all
devices, of course, including networking, tape, consoles, and such.
Skip, you might consider standardizing on larger capacity devices where it makes sense to do so. We were *very* constrained on the total number of UCBs we could have, based on architectural point-to-point (no switch) FICON limits, and so we eliminated ALL of our mod-3 3390s a couple of years ago. We now have only mod-9, mod-27 and mod-216 devices. (FWIW, the most popular -- and most numerous -- ended up being the mod-27s.)

It's like a breath of fresh air!

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


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

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Jesse 1 Robinson
2018-01-08 22:31:08 UTC
Permalink
Raw Message
Just to clarify a few points here. The old DASD we're moving off of is not crusty ancient tired iron: it's DS8800 (2424). The new DASD is DS8886 (2834). The motivation for moving is primarily business (fiscal), not technical.

We have millions of customers who are being converted to paperless billing, so more and more of them access the system more and more often at all hours of every day. There is virtually no time when all data bases happen to be closed. Closing them is not impossible, but it constitutes a 'total outage' to Customer Service.

Besides striving to provide continuous availability for our customers' sake, we are also mandated by the California Public Utilities Commission to service customers' issues in a timely way that gets translated to dollars and cents. An extra incentive to minimize total outages.

We maintain three sets of DASD volumes: primary (production), secondary (XRC mirror), and tertiary (flash copy from secondary to run DR). These add up to a lot of UCBs in addition to lots of virtual tapes and lots of FICON CTCs.

We're looking at ways to move PAVs to SCS1. Maybe even also the tertiary DR volumes. The goal is free up enough SCS0 addresses to put old and new boxes online. Meanwhile the new DASD is on the floor, and we need to utilize it.

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-543-6132 Office ⇐=== NEW
***@sce.com


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@LISTSERV.UA.EDU] On Behalf Of Ed Jaffe
Sent: Sunday, January 07, 2018 8:09 AM
To: IBM-***@LISTSERV.UA.EDU
Subject: (External):Re: Accessing 65536 devices
Post by Mike Schwab
If your new dasd allows dynamic growing a volume, you can grow a
non-EAV to the maximum EAV volume size, but no bigger.
That was my favorite feature of our old DS8100! We lost that feature when we upgraded to DS8870, but so far it hasn't been a problem because we planned fairly well.

Having said that, we're running a bit short on mod-216 volumes. I was thinking about removing eight of our mod-27s to make room for one more of these bad boys! Of course, that kind of re-allocation can be done with the DS8870, just not quite as easily as with the DS8100 because you can't grow a device in place.

The downside to in-place growth is varying levels of support from, and different procedures for, Linux for Z (we use RHEL), z/VM, z/VSE, and older releases of z/OS. You need to be careful or things can go FUBAR in a hurry! (Volume backups are a wonderful thing...)

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


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@listserv.ua.edu with the message: INFO IBM-MAIN
Tom Marchant
2018-01-09 20:14:13 UTC
Permalink
Raw Message
Post by Jesse 1 Robinson
Just to clarify a few points here. The old DASD we're moving off
of is not crusty ancient tired iron: it's DS8800 (2424). The new
DASD is DS8886 (2834). The motivation for moving is primarily
business (fiscal), not technical.
Are you using Exended Address Volumes? EAV volumes can hold a
terabyte of data, and most data sets can go in the cylinder-managed
space on an EAV.
Post by Jesse 1 Robinson
We have millions of customers who are being convertedoto
paperless billing, so more and more of them access the system
more and more often at all hours of every day. There is virtually
no time when all data bases happen to be closed. Closing them
is not impossible, but it constitutes a 'total outage' to Customer
Service.
I think that FDRPAS and FDRMOVE can handle the migration without
downtime, while consolidating to fewer, larger volumes. And I think
that there are other tools that can also be used. I don't know how
well these tools work in a parallel Sysplex environment.
--
Tom Marchant

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