On 4/1/2022 1:50 PM, Simon Clubley wrote:
> On 2022-04-01, Arne Vajhøj <arne@vajhoej.dk> wrote:
>> Of course they could also wait just a few months and go for
>> VMS x86-64 in a VM under ESXi. But given that it is on Alpha
>> not Itanium then that may not be the plan.
>
> Isn't that timetable just a little optimistic Arne ? :-)

Maybe.

> I doubt that anyone will be running live production payloads on
> the production-level version of x86-64 VMS until a good year or
> so after it is released.

They could start test with 9.1 now.

I know the "let us wait to put it in production until 12 months
after release" mantra. But there is this little problem that
if everybody follows that then it will not have been used in
production after 12 month.

Arne

On 4/1/2022 2:07 PM, Arne Vajhøj wrote:
> On 4/1/2022 1:50 PM, Simon Clubley wrote:
>> On 2022-04-01, Arne Vajhøj <arne@vajhoej.dk> wrote:
>>> Of course they could also wait just a few months and go for
>>> VMS x86-64 in a VM under ESXi. But given that it is on Alpha
>>> not Itanium then that may not be the plan.
>>
>> Isn't that timetable just a little optimistic Arne ? :-)
>
> Maybe.
>
>> I doubt that anyone will be running live production payloads on
>> the production-level version of x86-64 VMS until a good year or
>> so after it is released.
>
> They could start test with 9.1 now.
>
> I know the "let us wait to put it in production until 12 months
> after release" mantra. But there is this little problem that
> if everybody follows that then it will not have been used in
> production after 12 month.

Yeah, and if one is waiting for "everyone else" to test it, perhaps "everyone
else" is waiting for you to test it.

Perhaps better if it is used in parallel for testing, and if it works, then work it.

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

Den 2022-04-01 kl. 19:20, skrev Dave Froble:
> On 4/1/2022 11:40 AM, Arne Vajhøj wrote:
>> On 4/1/2022 10:48 AM, Bill Gunshannon wrote:
>>> On 4/1/22 09:43, Jan-Erik Söderholm wrote:
>>>> We have other issues. Alpha only supports up to 2 Gb FC cards.
>>>> The SAN group wants to go to 8/16/32 FC switches. They have
>>>> now kept 2 pcs of 2/4/8 Gb switches just for our servers, but
>>>> these are end-of-support in April 2024, and the SAN group do not
>>>> want to run FC switches that are unsupported...
>>>
>>> And that's how it's done.  Push VMS out the door without ever
>>> even mentioning its name.
>>
>> They don't mention VMS because they are not targeting VMS specific -
>> they are targeting all systems that cannot run with modern/supported
>> hardware.
>>
>> Alpha was fine, but it is more than 20 years since HP killed
>> (announced EOL) Alpha. This is very old systems. No surprise
>> that they do not support current (up to 20 years newer) technology.
>
> At first I thought Jan-Erik was joking about an "upgrade" to DS25 systems.

Not at all. It would be a simple box-swap. Just boot up the (dual 1 GHz)
DS25's from the same SAN disks as the current (single 666 Mhz) DS20 boxes.

> Not all that familiar, but, I thought the DS25 was just a DS20 with faster
> CPUs.

Faster CPUs. Faster memory interface. 16 GB max instead of 4 GB max.
My feeling (had one DS25 as my "play system" for some years) is that
they are somehere to 2-3 times faster then DS20/666.

And yes, adding the second CPU to the current boxes has also been
discussed. We can have them for free with a low installation fee.
But we'll still be locked at 4 GB max.

>
> As for the networking, I've got to believe that there is something
> available, that is new, and supported, that could talk to the Alpha FC
> cards.  But, perhaps not.

OK, right. I'm also think so. But then it has to be from the same brand
and prefered supplier that my customer works with.

>
> I do understand wanting support.  Open the hood on a new auto, and one will
> understand why one might not want to keep it past the warranty period.
>

On 2022-04-01, Arne Vajhøj <arne@vajhoej.dk> wrote:
> On 4/1/2022 1:50 PM, Simon Clubley wrote:
>> On 2022-04-01, Arne Vajhøj <arne@vajhoej.dk> wrote:
>>> Of course they could also wait just a few months and go for
>>> VMS x86-64 in a VM under ESXi. But given that it is on Alpha
>>> not Itanium then that may not be the plan.
>>
>> Isn't that timetable just a little optimistic Arne ? :-)
>
> Maybe.
>
>> I doubt that anyone will be running live production payloads on
>> the production-level version of x86-64 VMS until a good year or
>> so after it is released.
>
> They could start test with 9.1 now.
>

You should be doing the tests against the actual version you are
thinking of putting into production and 9.1 doesn't even have native
compilers yet. :-)

> I know the "let us wait to put it in production until 12 months
> after release" mantra. But there is this little problem that
> if everybody follows that then it will not have been used in
> production after 12 month.
>

My choice of wording was precise when I talked about live production
payloads. :-)

There's nothing to stop them testing the production version for
a year or so, until users trust it enough to start using it in
production.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

On Thursday, March 31, 2022 at 9:57:04 AM UTC-7, Jan-Erik Söderholm wrote:
> Hi.
>
> Is it my feed, or has it been quite quiet on the list a few days?
>
> Regards,
> Jan-Erik.

Perhaps this will stimulate a discussion

https://highassurance.rs

On 4/2/2022 10:32 PM, Simon Clubley wrote:
> On 2022-04-01, Arne Vajhøj <arne@vajhoej.dk> wrote:
>> On 4/1/2022 1:50 PM, Simon Clubley wrote:
>>> On 2022-04-01, Arne Vajhøj <arne@vajhoej.dk> wrote:
>>>> Of course they could also wait just a few months and go for
>>>> VMS x86-64 in a VM under ESXi. But given that it is on Alpha
>>>> not Itanium then that may not be the plan.
>>>
>>> Isn't that timetable just a little optimistic Arne ? :-)
>>
>> Maybe.
>>
>>> I doubt that anyone will be running live production payloads on
>>> the production-level version of x86-64 VMS until a good year or
>>> so after it is released.
>>
>> They could start test with 9.1 now.
>
> You should be doing the tests against the actual version you are
> thinking of putting into production and 9.1 doesn't even have native
> compilers yet. :-)

If it works on 9.1 then they should not need a year to test 9.1 to 9.2.
So get started.

>> I know the "let us wait to put it in production until 12 months
>> after release" mantra. But there is this little problem that
>> if everybody follows that then it will not have been used in
>> production after 12 month.
>
> My choice of wording was precise when I talked about live production
> payloads. :-)
>
> There's nothing to stop them testing the production version for
> a year or so, until users trust it enough to start using it in
> production.

Testing a year seems excessive and possible pointless as VSI
will undoubtedly release lots of patches during that year.

Arne

On 4/4/22 8:07 AM, Arne Vajhøj wrote:
> On 4/2/2022 10:32 PM, Simon Clubley wrote:
>> On 2022-04-01, Arne Vajhøj <arne@vajhoej.dk> wrote:

>>> They could start test with 9.1 now.
>>
>> You should be doing the tests against the actual version you are
>> thinking of putting into production and 9.1 doesn't even have native
>> compilers yet. :-)
>
> If it works on 9.1 then they should not need a year to test 9.1 to 9.2.
> So get started.

Unless you can't, perhaps because the compiler or product you need to
test with doesn't exist yet.

On 4/4/2022 12:05 PM, Craig A. Berry wrote:
> On 4/4/22 8:07 AM, Arne Vajhøj wrote:
>> On 4/2/2022 10:32 PM, Simon Clubley wrote:
>>> On 2022-04-01, Arne Vajhøj <arne@vajhoej.dk> wrote:
>>>> They could start test with 9.1 now.
>>>
>>> You should be doing the tests against the actual version you are
>>> thinking of putting into production and 9.1 doesn't even have native
>>> compilers yet. :-)
>>
>> If it works on 9.1 then they should not need a year to test 9.1 to 9.2.
>> So get started.
>
> Unless you can't, perhaps because the compiler or product you need to
> test with doesn't exist yet.

Of course.

What are missing of compilers and platform software?

Is Rdb available for test yet?

Arne

On 2022-04-04, Arne Vajhøj <arne@vajhoej.dk> wrote:
> On 4/2/2022 10:32 PM, Simon Clubley wrote:
>>
>> There's nothing to stop them testing the production version for
>> a year or so, until users trust it enough to start using it in
>> production.
>
> Testing a year seems excessive and possible pointless as VSI
> will undoubtedly release lots of patches during that year.
>

When you are talking about systems that allow your business to
function, it's not something you decide to do within a couple of months.

How long do you consider it acceptable to wait before you consider the
production version of VMS ready to use in a live production environment ?

Simon.

PS: Those patches will be fixing all the problems everyone else
found during that first year. :-)

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

On 4/4/2022 1:38 PM, Simon Clubley wrote:
> On 2022-04-04, Arne Vajhøj <arne@vajhoej.dk> wrote:
>> On 4/2/2022 10:32 PM, Simon Clubley wrote:
>>>
>>> There's nothing to stop them testing the production version for
>>> a year or so, until users trust it enough to start using it in
>>> production.
>>
>> Testing a year seems excessive and possible pointless as VSI
>> will undoubtedly release lots of patches during that year.
>
> When you are talking about systems that allow your business to
> function, it's not something you decide to do within a couple of months.

If you can't test it in a couple of months then you are toast.

You have a system. There are coming in updates to your application,
there are coming in updates to third party software, there are coming
in VMS updates from VSI.

If you test a year then what was tested at the beginning of the year
is not what is going live.

You don't want to when after 8 months there come a fix to a critical
security bug to say "we will not take that fix".

> How long do you consider it acceptable to wait before you consider the
> production version of VMS ready to use in a live production environment ?

Do a thorough test on 9.1 and then a final test on 9.2. There is no
magic in waiting N months. And unless ones workload is very standard
(which it is not on VMS) then other customers test may not be relevant.

Arne

On 4/4/22 11:49 AM, Arne Vajhøj wrote:
> On 4/4/2022 12:05 PM, Craig A. Berry wrote:
>> On 4/4/22 8:07 AM, Arne Vajhøj wrote:
>>> On 4/2/2022 10:32 PM, Simon Clubley wrote:
>>>> On 2022-04-01, Arne Vajhøj <arne@vajhoej.dk> wrote:
>>>>> They could start test with 9.1 now.
>>>>
>>>> You should be doing the tests against the actual version you are
>>>> thinking of putting into production and 9.1 doesn't even have native
>>>> compilers yet. :-)
>>>
>>> If it works on 9.1 then they should not need a year to test 9.1 to 9.2.
>>> So get started.
>>
>> Unless you can't, perhaps because the compiler or product you need to
>> test with doesn't exist yet.
>
> Of course.
>
> What are missing of compilers and platform software?

At least BASIC.

> Is Rdb available for test yet?

Dunno.

On 4/4/2022 3:02 PM, Craig A. Berry wrote:
> On 4/4/22 11:49 AM, Arne Vajhøj wrote:
>> On 4/4/2022 12:05 PM, Craig A. Berry wrote:
>>> On 4/4/22 8:07 AM, Arne Vajhøj wrote:
>>>> On 4/2/2022 10:32 PM, Simon Clubley wrote:
>>>>> On 2022-04-01, Arne Vajhøj <arne@vajhoej.dk> wrote:
>>>>>> They could start test with 9.1 now.
>>>>>
>>>>> You should be doing the tests against the actual version you are
>>>>> thinking of putting into production and 9.1 doesn't even have native
>>>>> compilers yet. :-)
>>>>
>>>> If it works on 9.1 then they should not need a year to test 9.1 to 9.2.
>>>> So get started.
>>>
>>> Unless you can't, perhaps because the compiler or product you need to
>>> test with doesn't exist yet.
>>
>> Of course.
>>
>> What are missing of compilers and platform software?
>
> At least BASIC.

OK - that means some people will need to wait a bit longer.

Arne

On 4/4/2022 1:38 PM, Simon Clubley wrote:
> On 2022-04-04, Arne Vajhøj <arne@vajhoej.dk> wrote:
>> On 4/2/2022 10:32 PM, Simon Clubley wrote:
>>>
>>> There's nothing to stop them testing the production version for
>>> a year or so, until users trust it enough to start using it in
>>> production.
>>
>> Testing a year seems excessive and possible pointless as VSI
>> will undoubtedly release lots of patches during that year.
>>
>
> When you are talking about systems that allow your business to
> function, it's not something you decide to do within a couple of months.
>
> How long do you consider it acceptable to wait before you consider the
> production version of VMS ready to use in a live production environment ?
>
> Simon.
>
> PS: Those patches will be fixing all the problems everyone else
> found during that first year. :-)
>

So suppose one is installing a new software application for a customer. How
long might you suggest the customer wait before using the new application?

What is normally done, at least in my experience, is the system is tested with
live data, and if everything works, then you turn it on for live operations.

Ok, the software application may have many users, and been rather thoroughly
tested prior to the new customer getting it. But that's similar to VMS V9.2.
VSI is testing V9.1, and customers are also testing V9.1.

That mentioned, I may have mentioned in the past that I don't consider x86 VMS
ready for use until native compilers are available. So, we ain't there yet.

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

Den 2022-03-31 kl. 18:57, skrev Jan-Erik Söderholm:

> Hi.
>
> Is it my feed, or has it been quite quiet on the list a few days?
>
> Regards,
> Jan-Erik.

Right... Maybe it was quite before, but this post seems to
have created some traffic and forked threads... :-)

I'm at least happy that it wasn't a news provider outage.

On the philosophical side, hadn't the "Rust as a HS language" thread
existed if I hadn't asked if it was quiet here? Who knows… :-)

On 2022-04-09, Jan-Erik Söderholm <jan-erik.soderholm@telia.com> wrote:
> Den 2022-03-31 kl. 18:57, skrev Jan-Erik Söderholm:
>
>> Hi.
>>
>> Is it my feed, or has it been quite quiet on the list a few days?
>>
>> Regards,
>> Jan-Erik.
>
> Right... Maybe it was quite before, but this post seems to
> have created some traffic and forked threads... :-)
>
> I'm at least happy that it wasn't a news provider outage.
>

Quite a lot of names seem to have disappeared so the newsgroup doesn't
have the continuous churn of activity that it used to have. There's
no way to tell if those people are now lurking only, or if it's due
to VMS dying off in their companies and organisations.

There's also the problem that nothing new is happening in the VMS world
at the moment, so there isn't something new to talk about.

> On the philosophical side, hadn't the "Rust as a HS language" thread
> existed if I hadn't asked if it was quiet here? Who knows? :-)

:-)

BTW, I've just had an idea for some fun.

Coming up shortly...

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.