Re: How set CVTLSO?

[Giliad Wilf]

On Wed, 2 Nov 2016 16:03:10 -0500, George Kozakos  wrote:
.
>The "epoch" on z/OS systems is 1900.
>
.
Right. It corresponds to a zeroed TOD.

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

Re: How set CVTLSO?

[Tony Harminc]

On 2 November 2016 at 17:03, George Kozakos  wrote:
>>Not any computer systems I work with. They use either 1900 or 1970
>>as their epoch. What uses 1972?
>
> The "epoch" on z/OS systems is 1900.

The TOD epoch is 1900. The UNIX epoch -- z/OS UNIX or any other -- is 1970.

Tony H.

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

Re: How set CVTLSO?

[George Kozakos]

>Not any computer systems I work with. They use either 1900 or 1970
>as their epoch. What uses 1972?

The "epoch" on z/OS systems is 1900.

George Kozakos
z/OS Software Service, Level 2 Supervisor


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

Re: How set CVTLSO?

[Paul Gilmartin]

On Tue, 1 Nov 2016 15:25:28 -0400, Tony Harminc wrote:
>
>And in the familiar case where the numerator is in a length unit (say,
>m), and the denominator in a time unit (say, s), we have names for
>each level: distance, speed, acceleration, jerk.
>
"jerk"?  Is anyone much concerned with "jerk" nowadays except for the
Dean Drive jerks?

>In this case we have time/time/time, which just fails to jump out at
>me. Maybe my imagination, visual or otherwise, is lacking.
>
Earlier, I posted the reciprocal of that number with the dimensions
collapsed to about 6x10^9 years, which is a plausible Fermi estimate
for the time it will take the moon to drag the earth to a halt.

>> The epoch being used in computer systems is 1972, with 26 leap seconds 
>> counted since then.
>
>Not any computer systems I work with. They use either 1900 or 1970 as
>their epoch. What uses 1972?
>
1972 was the inception of UTC.  At that time, GMT was 10 seconds behind
TAI.  MVS chose to set its TOD clock to GMT at that instant and run it
at the TAI rate thereafter, which is why TAI is (now) 36 seconds ahead of
UTC but TOD is only 26 seconds ahead of UTC.  All the curves intersect in
1972.

The units of Hubble's Constant are km/sec/megaparsec.  Inverse time
in some units.

-- gil

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

Re: How set CVTLSO?

[Giliad Wilf]

On Tue, 1 Nov 2016 15:25:28 -0400, Tony Harminc  wrote:

.
>
>Sure - I understand what's going on. It's just that, typically, one
>can "see" the nature of the units involved in such a statement. Often
>enough, when some politician or news reporter makes a statement like
>"Ontario exported 2.5 GW of electricity last year", or "an electric
>kettle uses about 1.5 kWh", the meaninglessness jumps right out
>because the units make no sense in the context.
>
>And in the familiar case where the numerator is in a length unit (say,
>m), and the denominator in a time unit (say, s), we have names for
>each level: distance, speed, acceleration, jerk.
>
>In this case we have time/time/time, which just fails to jump out at
>me. Maybe my imagination, visual or otherwise, is lacking.
>
.
The unit is 1/sec, or Hz. The value is ~ 5.134753814886006906326E-18.
. 
>
>Not any computer systems I work with. They use either 1900 or 1970 as
>their epoch. What uses 1972?
>
.
If you use leap seconds, then you should see the figure 26 on some 
HMC STP panel on your z machine, if your epoch is 1972.
If you see 36, then your epoch is 1957.
I know of no other epoch.
Look at page 83 (not the 83rd page but where page footing reads 83) on
the below publication:
http://www.redbooks.ibm.com/redbooks/pdfs/sg247281.pdf
In figure 3-5 there, you can see that leap second count or offset is 25.
The manual was printed June 2013 when the count relative to 1972
was 25.

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

Re: How set CVTLSO?

[Charles Mills]

> the units make no sense in the context

That job used 5 MSUs.

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf 
Of Tony Harminc
Sent: Tuesday, November 01, 2016 12:25 PM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

On 1 November 2016 at 13:44, Giliad Wilf 
<00d50942efa9-dmarc-requ...@listserv.ua.edu> wrote:
> On Tue, 1 Nov 2016 13:03:27 -0400, Tony Harminc <t...@harminc.net> wrote:

>>I'm a little confused about what kind of units "1.4 milliseconds a day 
>>per century" would be in.
>>
>>Tony H.
>>
> This means that  every 100 years, the day gets about 1.4 thousandths 
> of a second longer, compared to the length of the day measured the 
> moment atomic clocks became available commercially, at 1957, and since 
> then, 36 leap seconds were counted.

Sure - I understand what's going on. It's just that, typically, one can "see" 
the nature of the units involved in such a statement. Often enough, when some 
politician or news reporter makes a statement like "Ontario exported 2.5 GW of 
electricity last year", or "an electric kettle uses about 1.5 kWh", the 
meaninglessness jumps right out because the units make no sense in the context.

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

Re: How set CVTLSO?

[Tony Harminc]

On 1 November 2016 at 13:44, Giliad Wilf
<00d50942efa9-dmarc-requ...@listserv.ua.edu> wrote:
> On Tue, 1 Nov 2016 13:03:27 -0400, Tony Harminc  wrote:

>>I'm a little confused about what kind of units "1.4 milliseconds a day
>>per century" would be in.
>>
>>Tony H.
>>
> This means that  every 100 years, the day gets about 1.4 thousandths of a 
> second
> longer, compared to the length of the day measured the moment atomic clocks
> became available commercially, at 1957, and since then, 36 leap seconds were
> counted.

Sure - I understand what's going on. It's just that, typically, one
can "see" the nature of the units involved in such a statement. Often
enough, when some politician or news reporter makes a statement like
"Ontario exported 2.5 GW of electricity last year", or "an electric
kettle uses about 1.5 kWh", the meaninglessness jumps right out
because the units make no sense in the context.

And in the familiar case where the numerator is in a length unit (say,
m), and the denominator in a time unit (say, s), we have names for
each level: distance, speed, acceleration, jerk.

In this case we have time/time/time, which just fails to jump out at
me. Maybe my imagination, visual or otherwise, is lacking.

> The epoch being used in computer systems is 1972, with 26 leap seconds 
> counted since then.

Not any computer systems I work with. They use either 1900 or 1970 as
their epoch. What uses 1972?

Tony H.

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

Re: How set CVTLSO?

[Giliad Wilf]

On Tue, 1 Nov 2016 13:03:27 -0400, Tony Harminc  wrote:

.
>
>I'm a little confused about what kind of units "1.4 milliseconds a day
>per century" would be in.
>
>Tony H.
>
This means that  every 100 years, the day gets about 1.4 thousandths of a 
second 
longer, compared to the length of the day measured the moment atomic clocks
became available commercially, at 1957, and since then, 36 leap seconds were
counted.
The epoch being used in computer systems is 1972, with 26 leap seconds counted
since then.


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

Re: How set CVTLSO?

[Paul Gilmartin]

On Tue, 1 Nov 2016 13:03:27 -0400, Tony Harminc wrote:
>
>I'm a little confused about what kind of units "1.4 milliseconds a day
>per century" would be in.
> 
Does this help?:

530 $ units
500 units, 54 prefixes
You have: 1.4 msec / day / century
You want: Hz
* 5.1347535e-18
/ 1.9475131e+17

You have: day * century / 1.4 / msec
You want: year
* 6.1714286e+09
/ 1.6203704e-10

-- gil

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

Re: How set CVTLSO?

[Tony Harminc]

On 31 October 2016 at 09:04, Giliad Wilf
<00d50942efa9-dmarc-requ...@listserv.ua.edu> wrote:
> Atomic clock ticks at an absolutely constant rate of some 9192631770 
> oscillations
> per second (Cesium-133 atom oscillating between two energy ground levels),
> while Earth sidereal day gets slower(*) by 1.4 milliseconds a day per century.
> This lagging behind accumulates over time and mounts to almost a complete
> second over approx. 500 days.

I'm a little confused about what kind of units "1.4 milliseconds a day
per century" would be in.

Tony H.

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

Re: How set CVTLSO?

[Giliad Wilf]

On Mon, 31 Oct 2016 10:37:14 -0500, Paul Gilmartin  wrote:

.
>Similarly, when Daylight Saving Time ends (next week) we "add" ond hour
>to our civil time clocks in the sense that that day will be 25 hours long
>rather than the ordinary 24, by replicating the hour [01:00,02:00).  Yet
>z/OS chooses to subtract one hour from CVTLDTO.
>
. 
I did not mention CVTLDTO because it only affects local time, and most z/OS 
major
subsystems or components, such as JES2, XCF, GRS, DB2, LOGR, CICS (current 
versions), IMS (current versions) log events using UTC time, not local time.
Other major subsystems or components, such as RMF, use local time and may show
duplicate time-stamped recordings or one hour gap due to applying DST.

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

Re: How set CVTLSO?

[Paul Gilmartin]

On Mon, 31 Oct 2016 08:04:31 -0500, Giliad Wilf wrote:
>.
>Why do we have to add CVTLDTO (which could be negative if East of GMT), but
>subtract CVTLSO?
>Because CVTLSO represents by how much TAI is greater than Earth time, and
>ETS probably feeds you TAI values.
>  
Whereas CVTLDTO represents how much UTC is less than civil time, the opposite
choice of convention.

>Atomic clock ticks ...
> 
All of which I agree with.

>When this is about to happen (once in approximately 500 days), we have to 
>"add" one second to earth clocks, to "postpone" midnight.
> ...  
>  
"Add" in the sense that the day when that occurs is 86,401 seconds long
rather than the ordinary 86,400.

Similarly, when Daylight Saving Time ends (next week) we "add" ond hour
to our civil time clocks in the sense that that day will be 25 hours long
rather than the ordinary 24, by replicating the hour [01:00,02:00).  Yet
z/OS chooses to subtract one hour from CVTLDTO.

>Below you can find animation of two clocks. The upper is TAI (the atomic 
>clock).
>The lower is earth clock. The difference at this moment is 36 (relative to 
>1957,
>but only 26 relative to 1972, which is an alternative epoch on some systems).
>  
z/OS (and I believe GPS) use the 1972 convention.

>https://www.timeanddate.com/time/leap-seconds-background.html

I suspect the best explanation is historical.  OS/360 chose its convention
(which I prefer) for CVTLDTO before the advent of UTC.  BIPM/IERS
independently chose the opposite convention.  z/OS chose not to
complement the BIPM/IERS published value but only to offset it when
setting CVTLSO.  The 10-second offset provides continuity at the
transition from GMT to TAI.

(Topic drift; related to sign conventions and time):
I once debated the misdesign of a 10,000 pound rose granite equatorial
sundial which was constructed with the hours numbered clockwise on
both north and south sides.  My opponent argued that it was impossible
for the sun to appear to move counterclockwise in the northern hemisphere
and clockwise in the southern.  The sundial has since been repaired.
http://sundials.org/index.php/component/sundials/onedial/172

-- gil

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

Re: How set CVTLSO?

[Giliad Wilf]

On Tue, 11 Oct 2016 12:59:47 -0500, Paul Gilmartin  wrote:
.
>
>It's not clear why the convention is to add CVTLDTO to ETOD but subtract
>CVTLSO...
>
.
Why do we have to add CVTLDTO (which could be negative if East of GMT), but
subtract CVTLSO?
Because CVTLSO represents by how much TAI is greater than Earth time, and
ETS probably feeds you TAI values.
 
Atomic clock ticks at an absolutely constant rate of some 9192631770 
oscillations
per second (Cesium-133 atom oscillating between two energy ground levels), 
while Earth sidereal day gets slower(*) by 1.4 milliseconds a day per century. 
This lagging behind accumulates over time and mounts to almost a complete 
second over approx. 500 days.
 
As earth rotates and midnight is about to be reached, observatories might notice
that the correct position of a fixed point on earth, in relation to very 
distant 
stars, is still far away due to the slowing of earth, and midnight must be 
"postponed" a bit to let earth catch up with distant stars.
 
When this is about to happen (once in approximately 500 days), we have to 
"add" one second to earth clocks, to "postpone" midnight.
Postponement of midnight is attained by stepping earth clock from 23:59:59 
through 23:59:60, to 00:00:00, rather than the normal sequence 23:59:59 to
 00:00:00. Internally, CVTLSO is increased by one and the accumulated sum
 is subtracted from the TAI fed from ETS.  
 
As I've said, TAI is greater than Earth time by the accumulated leap seconds 
(a negative leap second is possible in principle, but has not been observed so
far).
 
Below you can find animation of two clocks. The upper is TAI (the atomic clock).
The lower is earth clock. The difference at this moment is 36 (relative to 1957,
but only 26 relative to 1972, which is an alternative epoch on some systems).
 
https://www.timeanddate.com/time/leap-seconds-background.html
 
 
(*) Why does earth slow down?
This slowing down is caused because earth is not a rigid body. Rather, it has 
liquid core, has tectonic plates floating and drifting astray, or being shifted 
during earthquakes, is covered by seas and oceans that undergo cyclic high 
and low tides, which convey some of earth's angular momentum to the moon, 
which in turn gets further away from earth at a rate of 1.5 inch a year.

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

Re: How set CVTLSO?

[George Kozakos]

On 12/10/2016 07:53 PM, Paul Gilmartin wrote:
> I understand that the TOD clock continues to be updated during that spin.
> Are processes otherwise quiesced or may they continue to execute until
> they do (e.g.) STCK?
>
> What happens to the clock comparator and to other STIMER-queued events?
> Someone may want to wait for two seconds physical time; someone else may
> want to wait until two seconds after midnight.

All CPUs are interrupted to process the external interrupt for the leap
second change.

The leap second change is applied to TOD (LT) and GMT real/wait TQEs.

George Kozakos
z/OS Software Service, Level 2 Supervisor



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

Re: How set CVTLSO?

[Paul Gilmartin]

On Wed, 12 Oct 2016 18:52:09 -0400, George Kozakos wrote:
>
>z/OS does not affect the hardware TOD clock. If the leap second change is
>positive, Timer processing spins on all CPUs for the amount of the leap
>second change and CVTLSO is updated. STCK time does not change but UTC/GMT
>jumps backward due to the change in CVTLSO.
>
I understand that the TOD clock continues to be updated during that spin.
Are processes otherwise quiesced or may they continue to execute until they
do (e.g.) STCK?

What happens to the clock comparator and to other STIMER-queued events?
Someone may want to wait for two seconds physical time; someone else may
want to wait until two seconds after midnight.

-- gil

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

Re: How set CVTLSO?

[Charles Mills]

Got it! Thanks much.

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of George Kozakos
Sent: Wednesday, October 12, 2016 3:52 PM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

On 12/10/2016 05:45:48 PM, , Charles Mills wrote:
> @George, thanks. I'm looking at the from a software development point 
> of view, not a sysprog point of view.
>
> So if a shop is using an ETS and does not schedule the change then the
TOD
> clock will be steered into reflecting the leap seconds, right? The TOD 
> clock will slow down until it falls back to the correct time, right?

Yes, the TOD is steered to reflect the change in the ETS. This takes about
7 hours, the steering rate is 7 hours per second of adjustment.

> For a "Time Control Parameter Change event" where does the leap second
get
> reflected? Still in the hardware TOD clock? It effectively stops for 
> one second? Or in CVTLSO?

z/OS does not affect the hardware TOD clock. If the leap second change is
positive, Timer processing spins on all CPUs for the amount of the leap
second change and CVTLSO is updated. STCK time does not change but UTC/GMT
jumps backward due to the change in CVTLSO.

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

Re: How set CVTLSO?

[George Kozakos]

On 12/10/2016 05:45:48 PM, , Charles Mills wrote:
> @George, thanks. I'm looking at the from a software development point of
> view, not a sysprog point of view.
>
> So if a shop is using an ETS and does not schedule the change then the
TOD
> clock will be steered into reflecting the leap seconds, right? The TOD
> clock will slow down until it falls back to the correct time, right?

Yes, the TOD is steered to reflect the change in the ETS. This takes about
7 hours, the steering rate is 7 hours per second of adjustment.

> For a "Time Control Parameter Change event" where does the leap second
get
> reflected? Still in the hardware TOD clock? It effectively stops for one
> second? Or in CVTLSO?

z/OS does not affect the hardware TOD clock. If the leap second change is
positive, Timer processing spins on all CPUs for the amount of the leap
second change and CVTLSO is updated. STCK time does not change but UTC/GMT
jumps backward due to the change in CVTLSO.

Regards,
George Kozakos


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

Re: How set CVTLSO?

[Paul Gilmartin]

On 2016-10-12 15:45, Charles Mills wrote:
> @George, thanks. I'm looking at the from a software development point of
> view, not a sysprog point of view.
> 
> So if a shop is using an ETS and does not schedule the change then the TOD
> clock will be steered into reflecting the leap seconds, right? The TOD clock
> will slow down until it falls back to the correct time, right?
>  
(which takes about a third of a day.)  Provided that CVTLSO was initially set
to 0.  If it's set to any other value, TIME GMT will give a correct result.
but STCK will give a surprising value.

> For a "Time Control Parameter Change event" where does the leap second get
> reflected? Still in the hardware TOD clock? It effectively stops for one
> second? Or in CVTLSO?
>
The latter.  z/OS adds one second to CVTLSO to "make the leap second adjustment"
while it has the CPUs spinning for one second so no user process ever perceives
an anachronistic sequence of STCK values.  The TOD clock inexorably runs at
4,096,000,000 ticks per second (as best STP can steer it).

(This would appear to be a one-second timed spin loop?.)

> -Original Message-
> From: George Kozakos
> Sent: Wednesday, October 12, 2016 1:29 PM
...
> If you do schedule the leap second, then STP generates a "Time Control
> Parameter Change event" external interrupt that gets processed by z/OS to
> make the leap second adjustment. If it is a positive change, z/OS spins on
> all CPUs for the amount of the positive leap second change to ensure there
> are no duplicate UTC time stamps and updates CVTLSO. No steering is required
> as STP UTC time remained accurate with the ETS.

-- gil

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

Re: How set CVTLSO?

[Charles Mills]

@George, thanks. I'm looking at the from a software development point of
view, not a sysprog point of view.

So if a shop is using an ETS and does not schedule the change then the TOD
clock will be steered into reflecting the leap seconds, right? The TOD clock
will slow down until it falls back to the correct time, right?

For a "Time Control Parameter Change event" where does the leap second get
reflected? Still in the hardware TOD clock? It effectively stops for one
second? Or in CVTLSO?

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of George Kozakos
Sent: Wednesday, October 12, 2016 1:29 PM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

> Ah! I am starting to understand. Leap Second steering is accomplished
with
> the PTFF instruction and is independent of CVTLSO. PTFF appears to 
> slow
down
> the physical clock.
>
> So ... steering and CVTLSO are essentially alternatives, right? CVTLSO 
> should not include any leap seconds that were previously (or were 
> about
to
> be) "steered" -- is that right? And if a shop is using STP it is 
> probably not modifying CVTLSO: CVTLSO is probably either zero, or at least
stable.
Am
> I getting this right?
>
> Charles

No, you have it wrong. Steering has nothing to do with leap seconds. STP
checks the external time source (ETS) at regular intervals and makes
adjustments via steering to keep accurate with the ETS.

It just so happens that if you don't schedule the leap second at the
appropriate time via the STP panel, when the leap second occurs, the STP UTC
time will be 1 second ahead of the ETS and so steering will occur to correct
it.

If you do schedule the leap second, then STP generates a "Time Control
Parameter Change event" external interrupt that gets processed by z/OS to
make the leap second adjustment. If it is a positive change, z/OS spins on
all CPUs for the amount of the positive leap second change to ensure there
are no duplicate UTC time stamps and updates CVTLSO. No steering is required
as STP UTC time remained accurate with the ETS.

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

Re: How set CVTLSO?

[George Kozakos]

> Ah! I am starting to understand. Leap Second steering is accomplished
with
> the PTFF instruction and is independent of CVTLSO. PTFF appears to slow
down
> the physical clock.
>
> So ... steering and CVTLSO are essentially alternatives, right? CVTLSO
> should not include any leap seconds that were previously (or were about
to
> be) "steered" -- is that right? And if a shop is using STP it is probably
> not modifying CVTLSO: CVTLSO is probably either zero, or at least stable.
Am
> I getting this right?
>
> Charles

No, you have it wrong. Steering has nothing to do with leap seconds. STP
checks
the external time source (ETS) at regular intervals and makes adjustments
via
steering to keep accurate with the ETS.

It just so happens that if you don't schedule the leap second at the
appropriate time via the STP panel, when the leap second occurs, the STP
UTC
time will be 1 second ahead of the ETS and so steering will occur to
correct it.

If you do schedule the leap second, then STP generates a "Time Control
Parameter
Change event" external interrupt that gets processed by z/OS to make the
leap
second adjustment. If it is a positive change, z/OS spins on all CPUs for
the
amount of the positive leap second change to ensure there are no duplicate
UTC
time stamps and updates CVTLSO. No steering is required as STP UTC time
remained
accurate with the ETS.

George Kozakos
z/OS Software Service, Level 2 Supervisor



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

Re: How set CVTLSO?

[George Kozakos]

> Mostly a curiosity and "long-term" question: where do you set the leap
> second offset? Can it be set without an IPL?
>
> I searched System Commands for "leap" and did not find anything.
>
> Charles

Via STP or Sysplex Timer

Regards,
George Kozakos
z/OS Software Service, Level 2 Supervisor


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

Re: How set CVTLSO?

[Paul Gilmartin]

On Tue, 11 Oct 2016 10:39:20 -0700, Charles Mills wrote:

>Ah! I am starting to understand. Leap Second steering is accomplished with
>the PTFF instruction and is independent of CVTLSO. PTFF appears to slow down
>the physical clock.
>
>So ... steering and CVTLSO are essentially alternatives, right? CVTLSO
>should not include any leap seconds that were previously (or were about to
>be) "steered" -- is that right? And if a shop is using STP it is probably
>not modifying CVTLSO: CVTLSO is probably either zero, or at least stable. Am
>I getting this right?
> 
CVTLSO should be the difference between ETOD and UTC regardless of how
prior leap second adjustments were accomplished.

If a shop chooses to keep CVTLSO zero, STP will steer ETOD to match UTC
within a few hours after a leap second.

If a shop chooses to use CVTLSO non-zero, STP (perhaps by non-disclosed
techniques) will be signalled by the HMC at a leap second to make user
tasks non-dispatchable for a second during which the CPU will add one
second to CVTLSO.

In principle, a shop could elect to run ETOD at TAI, always ten seconds
ahead of IBM's recommendation and maintain CVTLSO at a corresponding
ten second higher value.  But why?

It's not clear why the convention is to add CVTLDTO to ETOD but subtract
CVTLSO.  Perhaps a designer had a phobia of negative numbers.

I believe either Peter Relson or John Eells has explained all this with more
detail and probabaly more accurately than I in some previous June or
December.

-- gil

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

Re: How set CVTLSO?

[Charles Mills]

Ah! I am starting to understand. Leap Second steering is accomplished with
the PTFF instruction and is independent of CVTLSO. PTFF appears to slow down
the physical clock.

So ... steering and CVTLSO are essentially alternatives, right? CVTLSO
should not include any leap seconds that were previously (or were about to
be) "steered" -- is that right? And if a shop is using STP it is probably
not modifying CVTLSO: CVTLSO is probably either zero, or at least stable. Am
I getting this right?

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of Charles Mills
Sent: Tuesday, October 11, 2016 9:00 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

*Somebody* sets it ... who and how?

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of Mark Jacobs - Listserv
Sent: Tuesday, October 11, 2016 8:48 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

I don't think that CVTLSO is used by STP.

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

Re: How set CVTLSO?

[Paul Gilmartin]

On Tue, 11 Oct 2016 12:13:16 -0400, Mark Jacobs  wrote:

>CVTLSO is all zero on our system, and we're using STP.
> 
According to that white paper, it depends on how the HMC is configured.
A number of shops choose to set CVTLSO=0 and ignore leap seconds
(the "stering" option) because a (decreasing?) number of ISVs, including
(once?) IBM are inconsistent with each other and even within their own
product sets on conventions to deal with CVTLSO.  If CVTLSO=0:

o It doesn't matter.

o Response-sensitive peripherals are not disturbed

http://members.iinet.net.au/~nathanael/ntpd/leap-second.html

Search for:
amazon google leap second

... even spreading the leap second over several hours cause perceptible
disruption.

>> -Original Message-
>> From: Mark Jacobs - Listserv
>> Sent: Tuesday, October 11, 2016 7:21 AM
>>
>> STP will recognize that a leap second has occurred, and slowly steer zOS
>> time to the actual time. Here's a good writeup on the process.
>>
>> https://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP102081

-- gil

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

Re: How set CVTLSO?

[Paul Gilmartin]

On Tue, 11 Oct 2016 08:43:46 -0700, Charles Mills  wrote:

>Okay, I read the paper. Good paper. Thanks. (I don't feel too bad -- the
>paper only came out a couple of months ago.)
> 
It's hardly automatic; according to the reference:
Lastly, through the Hardware Management Console, manually schedule STP to
insert the new leap seconds when they are announced and due to occur.

"schedule" implies a comfortable lead time.

>The paper is short on low level details. Does anyone know: in a "steering"
>situation, does z/OS basically "change CVTLSO a little, spin a little, go
>back to business as usual for a while -- repeat as necessary for seven
>hours"?
> 
Any clock will drift with respect to a standard.  STP compares the ETOD clock
with an external standard and loads a steering register that adjusts the ETOD
rate in a manner that the CPU never perceives a discontinuity.  At a leap second
STP observes that ETOD is a second fast and adjusts the ETOD rate to the
minimum.  It takes about a third of a day to fall into sync.

In the non-steering case, all (user?) processes are non-dispatchable for
one second.  This could pose a problem for certain (obsolete?) peripheral
devices.  Ed Gould commented on this several months ago.

Both Amazon and Google have conventions to spread the leap second
over several hours centered on midnight.  Unsurprisingly, they chose
different durations for the spread.

>-Original Message-
>From: Mark Jacobs - Listserv
>Sent: Tuesday, October 11, 2016 7:21 AM
>
>STP will recognize that a leap second has occurred, and slowly steer zOS
>time to the actual time. Here's a good writeup on the process.
>
>https://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP102081

-- gil

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

Re: How set CVTLSO?

[Mark Jacobs - Listserv]

CVTLSO is all zero on our system, and we're using STP.

  EXT2: 00FDF670
 +0004  NUCLS F1FLGBT 00IOCID F0F3
 +0008  DEBVR 00FD93C2  CVAF. 00D5F000  MMVT. 00FD9688
 +0014  NCVP.   QIDA. 00OLTEP 
 +0024  AVVT. 8000  CCVT.   SKTA. 
 +0030  ICB..   FBYT1 00
 +0038  LDTO. CA5B  1700ATCVT 80C8B000
 +0048  BCLMT 0064  LSO..   

Mark Jacobs


Charles Mills <mailto:charl...@mcn.org>
October 11, 2016 at 12:00 PM
*Somebody* sets it ... who and how?

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of Mark Jacobs - Listserv
Sent: Tuesday, October 11, 2016 8:48 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

I don't think that CVTLSO is used by STP.

Mark Jacobs


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


Please be alert for any emails that may ask you for login information 
or directs you to login via a link. If you believe this message is a 
phish or aren't sure whether this message is trustworthy, please send 
the original message as an attachment to 'phish...@timeinc.com'.


Mark Jacobs - Listserv <mailto:mark.jac...@custserv.com>
October 11, 2016 at 11:47 AM
I don't think that CVTLSO is used by STP.

Mark Jacobs


Charles Mills <mailto:charl...@mcn.org>
October 11, 2016 at 11:43 AM
Okay, I read the paper. Good paper. Thanks. (I don't feel too bad -- the
paper only came out a couple of months ago.)

The paper is short on low level details. Does anyone know: in a "steering"
situation, does z/OS basically "change CVTLSO a little, spin a little, go
back to business as usual for a while -- repeat as necessary for seven
hours"?

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of Mark Jacobs - Listserv
Sent: Tuesday, October 11, 2016 7:21 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

STP will recognize that a leap second has occurred, and slowly steer zOS
time to the actual time. Here's a good writeup on the process.

https://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP102081

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


Please be alert for any emails that may ask you for login information 
or directs you to login via a link. If you believe this message is a 
phish or aren't sure whether this message is trustworthy, please send 
the original message as an attachment to 'phish...@timeinc.com'.


Charles Mills <mailto:charl...@mcn.org>
October 11, 2016 at 10:00 AM
So could in fact change without an IPL? That is really the core of my
curiosity.

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of Mark Jacobs - Listserv
Sent: Tuesday, October 11, 2016 6:45 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

I believe that if you're using STP to control system time, it all happens
automagically.

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


Please be alert for any emails that may ask you for login information 
or directs you to login via a link. If you believe this message is a 
phish or aren't sure whether this message is trustworthy, please send 
the original message as an attachment to 'phish...@timeinc.com'.


Mark Jacobs - Listserv <mailto:mark.jac...@custserv.com>
October 11, 2016 at 9:45 AM
I believe that if you're using STP to control system time, it all 
happens automagically.





--

Mark Jacobs
Time Customer Service
Global Technology Services

The standard you walk past is the standard you accept.
Lt. Gen. David Morrison


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

Re: How set CVTLSO?

[Charles Mills]

*Somebody* sets it ... who and how?

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of Mark Jacobs - Listserv
Sent: Tuesday, October 11, 2016 8:48 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

I don't think that CVTLSO is used by STP.

Mark Jacobs

> Charles Mills <mailto:charl...@mcn.org> October 11, 2016 at 11:43 AM 
> Okay, I read the paper. Good paper. Thanks. (I don't feel too bad -- 
> the paper only came out a couple of months ago.)
>
> The paper is short on low level details. Does anyone know: in a "steering"
> situation, does z/OS basically "change CVTLSO a little, spin a little, 
> go back to business as usual for a while -- repeat as necessary for 
> seven hours"?

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

Re: How set CVTLSO?

[Charles Mills]

Okay, I read the paper. Good paper. Thanks. (I don't feel too bad -- the
paper only came out a couple of months ago.)

The paper is short on low level details. Does anyone know: in a "steering"
situation, does z/OS basically "change CVTLSO a little, spin a little, go
back to business as usual for a while -- repeat as necessary for seven
hours"?

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of Mark Jacobs - Listserv
Sent: Tuesday, October 11, 2016 7:21 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

STP will recognize that a leap second has occurred, and slowly steer zOS
time to the actual time. Here's a good writeup on the process.

https://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP102081

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

Re: How set CVTLSO?

[Mark Jacobs - Listserv]

I don't think that CVTLSO is used by STP.

Mark Jacobs


Charles Mills <mailto:charl...@mcn.org>
October 11, 2016 at 11:43 AM
Okay, I read the paper. Good paper. Thanks. (I don't feel too bad -- the
paper only came out a couple of months ago.)

The paper is short on low level details. Does anyone know: in a "steering"
situation, does z/OS basically "change CVTLSO a little, spin a little, go
back to business as usual for a while -- repeat as necessary for seven
hours"?

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of Mark Jacobs - Listserv
Sent: Tuesday, October 11, 2016 7:21 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

STP will recognize that a leap second has occurred, and slowly steer zOS
time to the actual time. Here's a good writeup on the process.

https://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP102081

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


Please be alert for any emails that may ask you for login information 
or directs you to login via a link. If you believe this message is a 
phish or aren't sure whether this message is trustworthy, please send 
the original message as an attachment to 'phish...@timeinc.com'.


Charles Mills <mailto:charl...@mcn.org>
October 11, 2016 at 10:00 AM
So could in fact change without an IPL? That is really the core of my
curiosity.

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of Mark Jacobs - Listserv
Sent: Tuesday, October 11, 2016 6:45 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

I believe that if you're using STP to control system time, it all happens
automagically.

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


Please be alert for any emails that may ask you for login information 
or directs you to login via a link. If you believe this message is a 
phish or aren't sure whether this message is trustworthy, please send 
the original message as an attachment to 'phish...@timeinc.com'.


Mark Jacobs - Listserv <mailto:mark.jac...@custserv.com>
October 11, 2016 at 9:45 AM
I believe that if you're using STP to control system time, it all 
happens automagically.



Charles Mills <mailto:charl...@mcn.org>
October 11, 2016 at 9:42 AM
Mostly a curiosity and "long-term" question: where do you set the leap
second offset? Can it be set without an IPL?

I searched System Commands for "leap" and did not find anything.

Charles

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


Please be alert for any emails that may ask you for login information 
or directs you to login via a link. If you believe this message is a 
phish or aren't sure whether this message is trustworthy, please send 
the original message as an attachment to 'phish...@timeinc.com'.




--

Mark Jacobs
Time Customer Service
Global Technology Services

The standard you walk past is the standard you accept.
Lt. Gen. David Morrison


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

Re: How set CVTLSO?

[Mark Jacobs - Listserv]

STP will recognize that a leap second has occurred, and slowly steer zOS 
time to the actual time. Here's a good writeup on the process.


https://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP102081



Charles Mills <mailto:charl...@mcn.org>
October 11, 2016 at 10:00 AM
So could in fact change without an IPL? That is really the core of my
curiosity.

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of Mark Jacobs - Listserv
Sent: Tuesday, October 11, 2016 6:45 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

I believe that if you're using STP to control system time, it all happens
automagically.

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


Please be alert for any emails that may ask you for login information 
or directs you to login via a link. If you believe this message is a 
phish or aren't sure whether this message is trustworthy, please send 
the original message as an attachment to 'phish...@timeinc.com'.


Mark Jacobs - Listserv <mailto:mark.jac...@custserv.com>
October 11, 2016 at 9:45 AM
I believe that if you're using STP to control system time, it all 
happens automagically.



Charles Mills <mailto:charl...@mcn.org>
October 11, 2016 at 9:42 AM
Mostly a curiosity and "long-term" question: where do you set the leap
second offset? Can it be set without an IPL?

I searched System Commands for "leap" and did not find anything.

Charles

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


Please be alert for any emails that may ask you for login information 
or directs you to login via a link. If you believe this message is a 
phish or aren't sure whether this message is trustworthy, please send 
the original message as an attachment to 'phish...@timeinc.com'.




--

Mark Jacobs
Time Customer Service
Global Technology Services

The standard you walk past is the standard you accept.
Lt. Gen. David Morrison


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

Re: How set CVTLSO?

[Charles Mills]

So could in fact change without an IPL? That is really the core of my
curiosity. 

Charles

-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of Mark Jacobs - Listserv
Sent: Tuesday, October 11, 2016 6:45 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: How set CVTLSO?

I believe that if you're using STP to control system time, it all happens
automagically.

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

Re: How set CVTLSO?

[Mark Jacobs - Listserv]

I believe that if you're using STP to control system time, it all 
happens automagically.



Charles Mills 
October 11, 2016 at 9:42 AM
Mostly a curiosity and "long-term" question: where do you set the leap
second offset? Can it be set without an IPL?

I searched System Commands for "leap" and did not find anything.

Charles

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


Please be alert for any emails that may ask you for login information 
or directs you to login via a link. If you believe this message is a 
phish or aren't sure whether this message is trustworthy, please send 
the original message as an attachment to 'phish...@timeinc.com'.




--

Mark Jacobs
Time Customer Service
Global Technology Services

The standard you walk past is the standard you accept.
Lt. Gen. David Morrison


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

How set CVTLSO?

[Charles Mills]

Mostly a curiosity and "long-term" question: where do you set the leap
second offset? Can it be set without an IPL?

I searched System Commands for "leap" and did not find anything.

Charles 

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