Re: New Solaris CBE release available

[Tony Rodriguez]

Hi All,

Since I originally reported the issue with the previous CBE on the SPARC 
S7‑2—and provided a temporary workaround—I wanted to follow up. During a 
recent OS installation, I confirmed that the latest Solaris CBE resolves 
the SATA hard‑drive problem on the S7‑2. Under the older CBE, a 
ZFS‑related error prevented the system from booting when a SATA drive 
was present, leaving SAS as the only viable option.


Thanks again for resolving this issue.

Regards

Tony Rodriguez

On 3/7/26 4:45 PM, Alan Coopersmith wrote:

On 3/7/26 01:12, John Paul Adrian Glaubitz wrote:

Hi Alan,

first of all, thanks a lot for providing another CBE release within 
such a short
interval after the last one! I have already updated Debian's SPARC 
servers without

any issues.

Being able to run a recent version of Solaris puts peace to the mind 
as we don't
have to worry about potential unfixed CVEs. Plus, there are users who 
have reported
issues with the previous CBE release on their SPARC S7, so I hope 
that issue will now

be fixed as well.

So, huge thanks to you and your colleagues for making this possible!


You're welcome.

I think that date was also pushed into the future already, wasn't it? 
I think the

original plan foresaw the end of the service life in 2034, didn't it?


Yes, it was pushed out from 2034 to 2037 by request of customers who 
plan their
IT infrastructure many years into the future.  Whether it gets 
extended again
is still TBD.  (And to be pedantically correct, the original plan for 
Solaris 11
had a support life ending in 2024, which was then extended to 2034 
when it was
decided to turn our Solaris 12 work-in-progress into an 11.4 release 
and stay on

Solaris 11 for the long term.)

Btw, there are some open-source projects involving SPARC that Oracle 
previously worked
on but that never got upstreamed. For example, there is an incomplete 
port of libunwind
to SPARC [1] that I extracted from the Oracle Linux 6 sources. I 
would love to get these
upstreamed and have been trying to reach out to Oracle engineers to 
make this happen. If

you have the possibility to talk to some people, that would be great.


Unfortunately, the Linux engineers no longer work on SPARC ports, and 
the Solaris engineers mainly work on things we need for Solaris, so 
things like

that may fall through the cracks.

-alan-

Re: New Solaris CBE release available

[Alan Coopersmith]

On 3/7/26 01:12, John Paul Adrian Glaubitz wrote:

Hi Alan,

first of all, thanks a lot for providing another CBE release within such a short
interval after the last one! I have already updated Debian's SPARC servers 
without
any issues.

Being able to run a recent version of Solaris puts peace to the mind as we don't
have to worry about potential unfixed CVEs. Plus, there are users who have 
reported
issues with the previous CBE release on their SPARC S7, so I hope that issue 
will now
be fixed as well.

So, huge thanks to you and your colleagues for making this possible!


You're welcome.


I think that date was also pushed into the future already, wasn't it? I think 
the
original plan foresaw the end of the service life in 2034, didn't it?


Yes, it was pushed out from 2034 to 2037 by request of customers who plan their
IT infrastructure many years into the future.  Whether it gets extended again
is still TBD.  (And to be pedantically correct, the original plan for Solaris 11
had a support life ending in 2024, which was then extended to 2034 when it was
decided to turn our Solaris 12 work-in-progress into an 11.4 release and stay on
Solaris 11 for the long term.)


Btw, there are some open-source projects involving SPARC that Oracle previously 
worked
on but that never got upstreamed. For example, there is an incomplete port of 
libunwind
to SPARC [1] that I extracted from the Oracle Linux 6 sources. I would love to 
get these
upstreamed and have been trying to reach out to Oracle engineers to make this 
happen. If
you have the possibility to talk to some people, that would be great.


Unfortunately, the Linux engineers no longer work on SPARC ports, and the 
Solaris engineers mainly work on things we need for Solaris, so things like

that may fall through the cracks.

-alan-

Re: New Solaris CBE release available

[John Paul Adrian Glaubitz]

Hi Alan,

first of all, thanks a lot for providing another CBE release within such a short
interval after the last one! I have already updated Debian's SPARC servers 
without
any issues.

Being able to run a recent version of Solaris puts peace to the mind as we don't
have to worry about potential unfixed CVEs. Plus, there are users who have 
reported
issues with the previous CBE release on their SPARC S7, so I hope that issue 
will now
be fixed as well.

So, huge thanks to you and your colleagues for making this possible!

On Fri, 2026-03-06 at 17:27 -0800, Alan Coopersmith wrote:
> On 3/6/26 13:46, Peter Tribble wrote:
> > 
> > Given the huge amount of effort they've put in to fix Y2038 issues, I would 
> > go 
> > for coincidence.
> > (And I imagine the support lifetime isn't set by engineering.)
> 
> We do have some input to it, and have paid attention to the 2038 issue when
> providing our input, but the current dates are indeed set by customer demand,
> and there are limits to how far forward we can reliably forecast that on the
> business side.

I think that date was also pushed into the future already, wasn't it? I think 
the
original plan foresaw the end of the service life in 2034, didn't it?

> We are assuming that even if all our support contracts are finished by the end
> of 2037 that some customers will want to keep running the software past that,
> based on our experience with customers popping up to ask about SunOS 4 or
> Solaris 2.6 many years after their support lifetimes ended, and have been
> working for more than a decade now to make that possible, though we're not
> 100% complete yet.

That makes me smile though. I have heard similar stories from other software 
vendors,
including my employer. It's surprising what kind of old systems people are 
still using
in production.

> For instance, this CBE release makes the snoop command and the underlying 
> bufmod
> streams module it uses become Y2038 safe (but broken instead in 2106, as 
> explained
> under "Enhancements for Developers" in
> https://blogs.oracle.com/solaris/whats-new-in-oracle-solaris-11-4-sru-90 .)
> 
> I gave some further explanation/examples on Mastodon last year:
> https://hachyderm.io/@alanc/114519310064497360
> https://hachyderm.io/@alanc/114519356149693816

Thanks for the links. Definitely interested in such reads.

Btw, there are some open-source projects involving SPARC that Oracle previously 
worked
on but that never got upstreamed. For example, there is an incomplete port of 
libunwind
to SPARC [1] that I extracted from the Oracle Linux 6 sources. I would love to 
get these
upstreamed and have been trying to reach out to Oracle engineers to make this 
happen. If
you have the possibility to talk to some people, that would be great.

Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer
`. `'   Physicist
  `-GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913

Re: New Solaris CBE release available

[Alan Coopersmith]

On 3/6/26 13:46, Peter Tribble wrote:



On Fri, Mar 6, 2026 at 9:37 PM Gregor Riepl > wrote:


 > https://blogs.oracle.com/solaris/announcing-a-new-version-of-our-oracle-
solaris-environment-for-developers 

 >> ... the committed support for Oracle Solaris until at least 2037, ...

Do I smell a "we can't guarantee our stuff will work beyond the year-2038
problem" here, or is that just a coincidence? ;-)


Given the huge amount of effort they've put in to fix Y2038 issues, I would go 
for coincidence.

(And I imagine the support lifetime isn't set by engineering.)


We do have some input to it, and have paid attention to the 2038 issue when
providing our input, but the current dates are indeed set by customer demand,
and there are limits to how far forward we can reliably forecast that on the
business side.

We are assuming that even if all our support contracts are finished by the end
of 2037 that some customers will want to keep running the software past that,
based on our experience with customers popping up to ask about SunOS 4 or
Solaris 2.6 many years after their support lifetimes ended, and have been
working for more than a decade now to make that possible, though we're not
100% complete yet.   For instance, this CBE release makes the snoop command
and the underlying bufmod streams module it uses become Y2038 safe (but broken
instead in 2106, as explained under "Enhancements for Developers" in
https://blogs.oracle.com/solaris/whats-new-in-oracle-solaris-11-4-sru-90 .)

I gave some further explanation/examples on Mastodon last year:
https://hachyderm.io/@alanc/114519310064497360
https://hachyderm.io/@alanc/114519356149693816

--
-Alan Coopersmith- alan.coopersm...@oracle.com
 Oracle Solaris Engineering - https://blogs.oracle.com/solaris

Re: New Solaris CBE release available

[Peter Tribble]

On Fri, Mar 6, 2026 at 9:37 PM Gregor Riepl  wrote:

> >
> https://blogs.oracle.com/solaris/announcing-a-new-version-of-our-oracle-solaris-environment-for-developers
>
> >> ... the committed support for Oracle Solaris until at least 2037, ...
>
> Do I smell a "we can't guarantee our stuff will work beyond the year-2038
> problem" here, or is that just a coincidence? ;-)
>

Given the huge amount of effort they've put in to fix Y2038 issues, I would
go for coincidence.
(And I imagine the support lifetime isn't set by engineering.)

Certainly Solaris are way ahead of where we are in the illumos world wrt
Y2038.

-- 
-Peter Tribble
https://www.petertribble.co.uk/ - https://ptribble.blogspot.com/

Re: New Solaris CBE release available

[Gregor Riepl]

https://blogs.oracle.com/solaris/announcing-a-new-version-of-our-oracle-solaris-environment-for-developers



... the committed support for Oracle Solaris until at least 2037, ...


Do I smell a "we can't guarantee our stuff will work beyond the year-2038 
problem" here, or is that just a coincidence? ;-)