Hello!

Today somebody linked an interesting article in a German Newsgroup.

https://dissociatedpress.net/2023/06/24/red-hat-and-the-clone-wars/
https://www.redhat.com/en/blog/furthering-evolution-centos-stream

RedHat will stop providing public access to their source code.

This will make building binary-compatible distributions like the former
CentOS or Rocky Linux much harder.

Marco Moock wrote:
> Hello!
>
> Today somebody linked an interesting article in a German Newsgroup.
>
> https://dissociatedpress.net/2023/06/24/red-hat-and-the-clone-wars/
> https://www.redhat.com/en/blog/furthering-evolution-centos-stream
>
> RedHat will stop providing public access to their source code.
>
> This will make building binary-compatible distributions like the former
> CentOS or Rocky Linux much harder.
>

Also Alma, Oracle and Springdale...

Marco Moock <mo01@posteo.de> writes:

> Hello!
>
> Today somebody linked an interesting article in a German Newsgroup.
>
> https://dissociatedpress.net/2023/06/24/red-hat-and-the-clone-wars/
> https://www.redhat.com/en/blog/furthering-evolution-centos-stream
>
> RedHat will stop providing public access to their source code.
>
> This will make building binary-compatible distributions like the former
> CentOS or Rocky Linux much harder.

IBM evidently still has to provide access to their code to RHEL
users. They're trying to intimidate those users into thinking that the
users will get in trouble if they re-share that code. However, the GPL
makes such sharing allowable as part of the license, so I don't imagine
that IBM will get very far with preventing this.

I suppose IBM could start suing RHEL users that share the code, but it
doesn't look very good for a vendor to start suing their customers over
things like this.

Rocky and Alma Linux have said that they have other means to continue
their distros, but I don't know if they'll be able to keep the high
level of compatibility with RHEL that has proven so attractive in the
past (as it did with CentOS), and I suspect it could take more time and
resources, so taking this tack won't be without costs for Rocky and
Alma.

Bud Frede <frede@mouse-potato.com> writes:
> IBM evidently still has to provide access to their code to RHEL
> users. They're trying to intimidate those users into thinking that the
> users will get in trouble if they re-share that code. However, the GPL
> makes such sharing allowable as part of the license, so I don't imagine
> that IBM will get very far with preventing this.

Depends how much their customers know about the GPL. Many organizations
use Linux because it meet their technical needs rather than because of
its licensing model, which they may not care about at all.

> I suppose IBM could start suing RHEL users that share the code, but it
> doesn't look very good for a vendor to start suing their customers over
> things like this.

AIUI the idea is if you share the code onwards then RH/IBM terminate
your support contract (assuming they find out). An organization that’s
willing and able to manage with that contract may not care, but maybe
that’s not the kind of organization that choose RHEL in the first place.

--
https://www.greenend.org.uk/rjk/

Am 01.07.2023 um 09:57:46 Uhr schrieb Bud Frede:

> Marco Moock <mo01@posteo.de> writes:

> > Today somebody linked an interesting article in a German Newsgroup.
> >
> > https://dissociatedpress.net/2023/06/24/red-hat-and-the-clone-wars/
> > https://www.redhat.com/en/blog/furthering-evolution-centos-stream
> >
> > RedHat will stop providing public access to their source code.
> >
> > This will make building binary-compatible distributions like the
> > former CentOS or Rocky Linux much harder.
>
> IBM evidently still has to provide access to their code to RHEL
> users. They're trying to intimidate those users into thinking that the
> users will get in trouble if they re-share that code. However, the GPL
> makes such sharing allowable as part of the license, so I don't
> imagine that IBM will get very far with preventing this.
>
> I suppose IBM could start suing RHEL users that share the code, but it
> doesn't look very good for a vendor to start suing their customers
> over things like this.

At least for GPL and other licenses, they are not allowed to sue people
for that.

I also don't think they will do this.
I assume they will just make it much, much harder to get their code and
to build it.

Marco Moock <mo01@posteo.de> writes:

> Am 01.07.2023 um 09:57:46 Uhr schrieb Bud Frede:
>
>> Marco Moock <mo01@posteo.de> writes:
>
>> > Today somebody linked an interesting article in a German Newsgroup.
>> >
>> > https://dissociatedpress.net/2023/06/24/red-hat-and-the-clone-wars/
>> > https://www.redhat.com/en/blog/furthering-evolution-centos-stream
>> >
>> > RedHat will stop providing public access to their source code.
>> >
>> > This will make building binary-compatible distributions like the
>> > former CentOS or Rocky Linux much harder.
>>
>> IBM evidently still has to provide access to their code to RHEL
>> users. They're trying to intimidate those users into thinking that the
>> users will get in trouble if they re-share that code. However, the GPL
>> makes such sharing allowable as part of the license, so I don't
>> imagine that IBM will get very far with preventing this.
>>
>> I suppose IBM could start suing RHEL users that share the code, but it
>> doesn't look very good for a vendor to start suing their customers
>> over things like this.
>
> At least for GPL and other licenses, they are not allowed to sue people
> for that.
>
> I also don't think they will do this.
> I assume they will just make it much, much harder to get their code and
> to build it.

It's really aimed at Oracle. I don't think IBM really cares about Rocky
and Alma Linux. They just want to try to prevent Oracle from taking the
RHEL source, filing off the serial numbers, and making money off it as
Oracle Linux.

I guess if your company spends a lot money with Oracle for their other
software might go with Oracle for Linux support too. Oracle's support is
pretty crappy though, so I don't quite understand anyone wanting it.

I used to be a Solaris admin, but after Oracle bought them I saw the
writing on the wall and pursued other career paths (Linux and
AIX). Sun's support was decent, but it turned into Oracle-level support
quite rapidly. If you ran into some issue with Solaris or Sun hardware,
you'd call Oracle and get nowhere. The Oracle DBAs were used to that
level of non-support, but none of the Unix admins were.

It made IBM pretty happy though. Lots of POWER systems were brought in
to replace the old Sun stuff. :-)

Am 01.07.2023 um 17:48:57 Uhr schrieb Bud Frede:

> I used to be a Solaris admin, but after Oracle bought them I saw the
> writing on the wall and pursued other career paths (Linux and
> AIX). Sun's support was decent, but it turned into Oracle-level
> support quite rapidly. If you ran into some issue with Solaris or Sun
> hardware, you'd call Oracle and get nowhere. The Oracle DBAs were
> used to that level of non-support, but none of the Unix admins were.
>
> It made IBM pretty happy though. Lots of POWER systems were brought in
> to replace the old Sun stuff. :-)

I wonder that people still buy them. I think AIX is also on the way
down. They don't produce workstations anymore and AIX only runs on
special IBM stuff - not on normal machines.

Bud Frede <frede@mouse-potato.com> writes:
> I guess if your company spends a lot money with Oracle for their other
> software might go with Oracle for Linux support too. Oracle's support is
> pretty crappy though, so I don't quite understand anyone wanting it.

AIUI the sensible reason for buying Oracle Linux is as a platform
optimized for Oracle’s DB.

--
https://www.greenend.org.uk/rjk/

Am 02.07.2023 um 10:02:45 Uhr schrieb Richard Kettlewell:

> Bud Frede <frede@mouse-potato.com> writes:
> > I guess if your company spends a lot money with Oracle for their
> > other software might go with Oracle for Linux support too. Oracle's
> > support is pretty crappy though, so I don't quite understand anyone
> > wanting it.
>
> AIUI the sensible reason for buying Oracle Linux is as a platform
> optimized for Oracle’s DB.

They have Solaris.

Marco Moock <mo01@posteo.de> writes:
> Am 02.07.2023 um 10:02:45 Uhr schrieb Richard Kettlewell:
>> Bud Frede <frede@mouse-potato.com> writes:
>> > I guess if your company spends a lot money with Oracle for their
>> > other software might go with Oracle for Linux support too. Oracle's
>> > support is pretty crappy though, so I don't quite understand anyone
>> > wanting it.
>>
>> AIUI the sensible reason for buying Oracle Linux is as a platform
>> optimized for Oracle’s DB.
>
> They have Solaris.

Obviously on its last legs, almost nobody is going to use it if they
have a free choice. It might last a little longer than AIX and HPUX but
it’s obviously doomed.

--
https://www.greenend.org.uk/rjk/

Richard Kettlewell <invalid@invalid.invalid> writes:

> Bud Frede <frede@mouse-potato.com> writes:
>> I guess if your company spends a lot money with Oracle for their other
>> software might go with Oracle for Linux support too. Oracle's support is
>> pretty crappy though, so I don't quite understand anyone wanting it.
>
> AIUI the sensible reason for buying Oracle Linux is as a platform
> optimized for Oracle’s DB.

They officially support RHEL too.

Also, I don't know if this is still the case, but it used to be less
expensive to get your Oracle licenses from IBM and run it on POWER/AIX
than to buy it directly from Oracle with one of their platforms.

But for Linux, unless you're running only Oracle DBs, I'm not sure it
makes much sense to standardize on Oracle Linux. RHEL is more of the
industry standard, and there's a lot more info available for it too.

Yes, you could argue that Oracle Linux is RHEL, but is Oracle going to
be as helpful as Red Hat itself when it comes to supporting
non-Oracle-DB-related OS issues?

Marco Moock <mo01@posteo.de> writes:

> Am 02.07.2023 um 10:02:45 Uhr schrieb Richard Kettlewell:
>
>> Bud Frede <frede@mouse-potato.com> writes:
>> > I guess if your company spends a lot money with Oracle for their
>> > other software might go with Oracle for Linux support too. Oracle's
>> > support is pretty crappy though, so I don't quite understand anyone
>> > wanting it.
>>
>> AIUI the sensible reason for buying Oracle Linux is as a platform
>> optimized for Oracle’s DB.
>
> They have Solaris.

I thought I had read that they were simply maintaining the current
version of Solaris now for existing customers, and were not planning to
release a new version.

I don't know what the status of SPARC hardware is, but it seemed to me
that for a long time now Fujitsu was where things were happening with
SPARC and not at Oracle.

Marco Moock <mo01@posteo.de> writes:

> Am 01.07.2023 um 17:48:57 Uhr schrieb Bud Frede:
>
>> I used to be a Solaris admin, but after Oracle bought them I saw the
>> writing on the wall and pursued other career paths (Linux and
>> AIX). Sun's support was decent, but it turned into Oracle-level
>> support quite rapidly. If you ran into some issue with Solaris or Sun
>> hardware, you'd call Oracle and get nowhere. The Oracle DBAs were
>> used to that level of non-support, but none of the Unix admins were.
>>
>> It made IBM pretty happy though. Lots of POWER systems were brought in
>> to replace the old Sun stuff. :-)
>
> I wonder that people still buy them. I think AIX is also on the way
> down. They don't produce workstations anymore and AIX only runs on
> special IBM stuff - not on normal machines.

That "special IBM stuff" was pretty special the last time I worked with
it. It had a lot of mainframe features for RAS (Reliability,
availability and serviceability).

I really haven't had much to do with real enterprise hardware for a
while, since my last few jobs have been working with Linux in the
cloud. I kind of miss being able to get my hands on the big datacenter
servers, etc.

VMWare has allowed for a lot of mainframe features on commodity
hardware. I'm not sure if it's completely caught up or not. It's
probably at least close enough for most purposes.

When using IBM POWER hardware, the IBM Field Engineers often called up
to schedule a service call before anyone onsite even knew there was an
issue or potential issue. Almost everything was hot-swap. Bad stick of
RAM? Hot-swap. Bad CPU? Hot-swap. If worst came to worst, you could move
the LPAR to a different box while you were replacing a backplane or
something. (Similar to a vmotion, but IBM had it before VMWare did
AFAIK.)

I suppose that HP and Dell might offer some of this now, but their style
of onsite service used to be "send in the clowns." IBM FEs were more
professional and skilled than those from HP and Dell. (EMC had really
good support in the field too. I don't know if they still do.)

Richard Kettlewell <invalid@invalid.invalid> writes:

> Marco Moock <mo01@posteo.de> writes:
>> Am 02.07.2023 um 10:02:45 Uhr schrieb Richard Kettlewell:
>>> Bud Frede <frede@mouse-potato.com> writes:
>>> > I guess if your company spends a lot money with Oracle for their
>>> > other software might go with Oracle for Linux support too. Oracle's
>>> > support is pretty crappy though, so I don't quite understand anyone
>>> > wanting it.
>>>
>>> AIUI the sensible reason for buying Oracle Linux is as a platform
>>> optimized for Oracle’s DB.
>>
>> They have Solaris.
>
> Obviously on its last legs, almost nobody is going to use it if they
> have a free choice. It might last a little longer than AIX and HPUX but
> it’s obviously doomed.

I'll kind of miss it. It was nice to work with. AIX was not so nice to
work with. :-)

I'd say that about 2/3 of what I touch these days is Ubuntu Server. Most
of the rest is RHEL, with a bit of SuSE on systems in Germany.

On Sun, 02 Jul 2023 08:52:08 -0400
Bud Frede <frede@mouse-potato.com> wrote:
> That "special IBM stuff" was pretty special the last time I worked with
> it. It had a lot of mainframe features for RAS (Reliability,
> availability and serviceability).

[...]

> Almost everything was hot-swap. Bad stick of
> RAM? Hot-swap. Bad CPU? Hot-swap.

How is it possible to do such things , I mean even in theory ? Bad
RAM could mean that you can't read it so if you don't know what's in
the addresses that you can't read then how can you transfer it to
different RAM ?

Similar question for CPUs. Does the hardware offer the capability to
copy the whole state of the CPU you're going to replace to a different
CPU ? And if the CPU is bad , how can you even know the state ?

--
vlaho.ninja/prog

Spiros Bousbouras wrote:

> Bud Frede wrote:
>
>> Bad stick of RAM? Hot-swap. Bad CPU? Hot-swap.
>
> How is it possible to do such things , I mean even in theory ? Bad
> RAM could mean that you can't read it so if you don't know what's in
> the addresses that you can't read then how can you transfer it to
> different RAM ?

mirrored-RAM, not unique to IBM, HPE do it too and presumably others ...

On 2023-07-03 11:03, Spiros Bousbouras wrote:
> On Sun, 02 Jul 2023 08:52:08 -0400
> Bud Frede <frede@mouse-potato.com> wrote:
>> That "special IBM stuff" was pretty special the last time I worked with
>> it. It had a lot of mainframe features for RAS (Reliability,
>> availability and serviceability).
>
> [...]
>
>> Almost everything was hot-swap. Bad stick of
>> RAM? Hot-swap. Bad CPU? Hot-swap.
>
> How is it possible to do such things , I mean even in theory ? Bad
> RAM could mean that you can't read it so if you don't know what's in
> the addresses that you can't read then how can you transfer it to
> different RAM ?
>
> Similar question for CPUs. Does the hardware offer the capability to
> copy the whole state of the CPU you're going to replace to a different
> CPU ? And if the CPU is bad , how can you even know the state ?
>

Mirrored hardware.

I have not worked with that on computers, but I worked with a telephone
exchange (5ESS) which had a mirrored computer (AM). It could do things
like run cpu 1 using memory of cpu 2. And that machine is ancient design.

--
Cheers, Carlos.

On Mon, 3 Jul 2023 14:29:33 +0200, Carlos E.R. wrote:
> On 2023-07-03 11:03, Spiros Bousbouras wrote:
>> On Sun, 02 Jul 2023 08:52:08 -0400
>> Bud Frede <frede@mouse-potato.com> wrote:
>>> That "special IBM stuff" was pretty special the last time I worked with
>>> it. It had a lot of mainframe features for RAS (Reliability,
>>> availability and serviceability).
>>
>> [...]
>>
>>> Almost everything was hot-swap. Bad stick of
>>> RAM? Hot-swap. Bad CPU? Hot-swap.
>>
>> How is it possible to do such things , I mean even in theory ? Bad
>> RAM could mean that you can't read it so if you don't know what's in
>> the addresses that you can't read then how can you transfer it to
>> different RAM ?
>>
>> Similar question for CPUs. Does the hardware offer the capability to
>> copy the whole state of the CPU you're going to replace to a different
>> CPU ? And if the CPU is bad , how can you even know the state ?
>>
>
> Mirrored hardware.

https://en.wikipedia.org/wiki/Tandem_Computers

Non-stop computer systems back in the Dark Ages.

Allodoxaphobia <trepidation@example.net> wrote:
>
> https://en.wikipedia.org/wiki/Tandem_Computers
>
> Non-stop computer systems back in the Dark Ages.

I had a short contract with Citibank, way back in the late 1970s,
programming on a Tandem, under its TOS operating system.
Yeah, the hardware was very reliable, indeed. But that TOS operating
system was crashing all the time. And down is down, no matter
how you slice/dice/rice it. You'd think Tandem would've figured
that out, and gotten the bugs out of TOS so that uptime was
on a par with their hardware's reliability. Never happened,
as far as I know.
--
John Forkosh ( mailto: j@f.com where j=john and f=forkosh )

Andy Burns <usenet@andyburns.uk> writes:

> Spiros Bousbouras wrote:
>
>> Bud Frede wrote:
>>
>>> Bad stick of RAM? Hot-swap. Bad CPU? Hot-swap.
>> How is it possible to do such things , I mean even in theory ? Bad
>> RAM could mean that you can't read it so if you don't know what's in
>> the addresses that you can't read then how can you transfer it to
>> different RAM ?
>
> mirrored-RAM, not unique to IBM, HPE do it too and presumably others ...

A place I worked had one HP9000 server, but I never did anything with
it. (I'm not sure they did either. The DR system for this had never even
been powered-on.) I did own an HP 9000 workstation at one point - I
think it came with HP UX 9.something? I sold it on pretty quickly
because HP was not hobbyist-friendly. I couldn't get ssh to compile on
the HP UX I had and HP told me the only way I could get a newer version
would be to pay them several thousand dollars. I raised a finger in
their general direction, sold it on, and got a DEC Alphastation instead
that I ran Linux on.

I've been told that Linux LVM was based on or modeled after LVM from HP
UX? LVM is kind of nice, but also kind of ugly. I guess I got spoiled by
ZFS. :-)

Bud Frede wrote:

> A place I worked had one HP9000 server, but I never did anything with
> it.

At work we had a HP9000-800/F10, we ported our system to it, but never
managed to sell to any HP customers, I ended up taking the server home,
it was an extremely densely packed box!

On 6/26/23 10:02 AM, Marco Moock wrote:
> Hello!
>
> Today somebody linked an interesting article in a German Newsgroup.
>
> https://dissociatedpress.net/2023/06/24/red-hat-and-the-clone-wars/
> https://www.redhat.com/en/blog/furthering-evolution-centos-stream
>
> RedHat will stop providing public access to their source code.
>
> This will make building binary-compatible distributions like the former
> CentOS or Rocky Linux much harder.

Look, RH went 'public' and IBM bought them out.
They aren't selling "linux" per-se, but all their
"enhancements"/support/expertise, so it's all legit.

This devolution has been going on for a couple of
years now. CENTOS is now the beta-code version of
RHEL and you are their free test users. Apparently
now Centos won't even be that.

So - DUMP yer RPM distros. Deb/Arch/Slack are still
the future of OS freedom (for now).

I think the most popular system for IBM minis/mainframes
(yes, there IS still a profitible market for mainframes)
is basically Linux - even ahead of IBM's own OS. This is
gonna be re-branded RHEL for all intents and purposes.

So FORGET anything based on RHEL. It's not for "us"
anymore.

On 02/07/2023 09:00, Marco Moock wrote:
> Am 01.07.2023 um 17:48:57 Uhr schrieb Bud Frede:
>
>> I used to be a Solaris admin, but after Oracle bought them I saw the
>> writing on the wall and pursued other career paths (Linux and
>> AIX). Sun's support was decent, but it turned into Oracle-level
>> support quite rapidly. If you ran into some issue with Solaris or Sun
>> hardware, you'd call Oracle and get nowhere. The Oracle DBAs were
>> used to that level of non-support, but none of the Unix admins were.
>>
>> It made IBM pretty happy though. Lots of POWER systems were brought in
>> to replace the old Sun stuff. :-)
>
> I wonder that people still buy them. I think AIX is also on the way
> down. They don't produce workstations anymore and AIX only runs on
> special IBM stuff - not on normal machines.
>

AIX is IBMs x86 Unix.

Whether it needs custom hardware or not is perhaps not germane to its
users who are, as you say overwhelmingly running IBM applications -
usually legacy database code, RPG and COBOL - on it, and the annual
support contract probably exceeds the cost of the PC hardware it runs on
after the first few months :-)

--
Renewable energy: Expensive solutions that don't work to a problem that
doesn't exist instituted by self legalising protection rackets that
don't protect, masquerading as public servants who don't serve the public.

The Natural Philosopher <tnp@invalid.invalid> writes:
> On 02/07/2023 09:00, Marco Moock wrote:

>> I wonder that people still buy them. I think AIX is also on the way
>> down. They don't produce workstations anymore and AIX only runs on
>> special IBM stuff - not on normal machines.
>
> AIX is IBMs x86 Unix.

It’s been supported on a range of platforms in the past but it’s been
Power architecture only for a long time now.

--
https://www.greenend.org.uk/rjk/

On 05/07/2023 18:28, Richard Kettlewell wrote:
> The Natural Philosopher <tnp@invalid.invalid> writes:
>> On 02/07/2023 09:00, Marco Moock wrote:
>
>>> I wonder that people still buy them. I think AIX is also on the way
>>> down. They don't produce workstations anymore and AIX only runs on
>>> special IBM stuff - not on normal machines.
>>
>> AIX is IBMs x86 Unix.
>
> It’s been supported on a range of platforms in the past but it’s been
> Power architecture only for a long time now.
>
Well Richard, you know when I last saw it ;-)

--
“The fundamental cause of the trouble in the modern world today is that
the stupid are cocksure while the intelligent are full of doubt."

- Bertrand Russell