New public SPARC and Solaris roadmap

[Calum]

Posted yesterday.

<http://www.oracle.com/us/products/servers-storage/servers/sparc/oracle-sparc/sparc-roadmap-slide-2076743.pdf>

[YTC#1]

On 18/01/2017 12:58, Calum wrote:
> Posted yesterday.
>
> <http://www.oracle.com/us/products/servers-storage/servers/sparc/oracle-sparc/sparc-roadmap-slide-2076743.pdf>
>

Ah well, it is official then.

Now the marketing bods have to work out what foolishness it was to drop
the 2.

If they had kept it, then S2.10U11 would have been the last Solaris 2
release.

S11 was so new (hell we could not even upgrade) it should have been 3.

Therefore we would now beon S3.3 and awaiting S3.4. World panic
resolved. :-)


--
Bruce Porter
"The internet is a huge and diverse community but mainly friendly"
http://ytc1.blogspot.co.uk/
There *is* an alternative! http://www.openoffice.org/

[Gary R. Schmidt]

On 18/01/2017 23:58, Calum wrote:
> Posted yesterday.
>
> <http://www.oracle.com/us/products/servers-storage/servers/sparc/oracle-sparc/sparc-roadmap-slide-2076743.pdf>
>

Interesting, some sort of OS group inside Fujitsu Japan have been
fiddling with Solaris 12 for a while now, "such-and-such on S12 is
different, please make (your thing) handle it" messages come our way
every now and then.

Cheers,
Gary B-)

--
When men talk to their friends, they insult each other.
They don't really mean it.
When women talk to their friends, they compliment each other.
They don't mean it either.

[John D Groenveld]

In article <9rm9ld-...@paranoia.mcleod-schmidt.id.au>,

Gary R. Schmidt <grsc...@acm.org> wrote:
>Interesting, some sort of OS group inside Fujitsu Japan have been
>fiddling with Solaris 12 for a while now, "such-and-such on S12 is
>different, please make (your thing) handle it" messages come our way
>every now and then.

Fujitsu now trapped in the future with an OS version that will
never be certified against the current versions of Oracle's database
product stack. :)

John
groe...@acm.org

[YTC#1]

On 20/01/2017 15:27, Gary R. Schmidt wrote:
> On 18/01/2017 23:58, Calum wrote:
>> Posted yesterday.
>>
>> <http://www.oracle.com/us/products/servers-storage/servers/sparc/oracle-sparc/sparc-roadmap-slide-2076743.pdf>
>>
>>
> Interesting, some sort of OS group inside Fujitsu Japan have been
> fiddling with Solaris 12 for a while now, "such-and-such on S12 is
> different, please make (your thing) handle it" messages come our way
> every now and then.

Aye, but S12 (Sorry Solaris Next) has been in dev for a number of years,
so their requests have been correct.

I expect these "differences" will feed back into S11.x at some future
point and the messages from Fuji will change to "such and such on S11.x
is different ...."

[Andrew Gabriel]

In article <9rm9ld-...@paranoia.mcleod-schmidt.id.au>,

"Gary R. Schmidt" <grsc...@acm.org> writes:
> On 18/01/2017 23:58, Calum wrote:
>> Posted yesterday.
>>
>> <http://www.oracle.com/us/products/servers-storage/servers/sparc/oracle-sparc/sparc-roadmap-slide-2076743.pdf>
>>
> Interesting, some sort of OS group inside Fujitsu Japan have been
> fiddling with Solaris 12 for a while now, "such-and-such on S12 is
> different, please make (your thing) handle it" messages come our way
> every now and then.

Although Fujitsu did release a new SPARC chip last year (although
I'm not sure it's in any products which run Solaris), they have
indicated their future direction is going to be ARM.
They don't run Solaris on their SPARC supercomputers anyway, only
on their M10 range. Their supercomputers run Linux.

--
Andrew Gabriel
[email address is not usable -- followup in the newsgroup]

[YTC#1]

On 21/01/2017 14:25, Andrew Gabriel wrote:
> In article <9rm9ld-...@paranoia.mcleod-schmidt.id.au>,
> "Gary R. Schmidt" <grsc...@acm.org> writes:
>> On 18/01/2017 23:58, Calum wrote:
>>> Posted yesterday.
>>>
>>> <http://www.oracle.com/us/products/servers-storage/servers/sparc/oracle-sparc/sparc-roadmap-slide-2076743.pdf>
>>>
>> Interesting, some sort of OS group inside Fujitsu Japan have been
>> fiddling with Solaris 12 for a while now, "such-and-such on S12 is
>> different, please make (your thing) handle it" messages come our way
>> every now and then.
>
> Although Fujitsu did release a new SPARC chip last year (although

> I'm not sure it's in any products which run Solaris), they have

The Sparc 64 X+ ?

I'm installing some M10-1+s at the moment. Quite strange playing with
LDoms on non Sun4v architecture :-)

> indicated their future direction is going to be ARM.
> They don't run Solaris on their SPARC supercomputers anyway, only
> on their M10 range. Their supercomputers run Linux.
>

Solaris on ARM ??? :-)

[Andrew Gabriel]

In article <o61vdm$6fr$1...@dont-email.me>,

YTC#1 <b...@ytc1-spambin.co.uk> writes:
> On 21/01/2017 14:25, Andrew Gabriel wrote:
>> In article <9rm9ld-...@paranoia.mcleod-schmidt.id.au>,
>> "Gary R. Schmidt" <grsc...@acm.org> writes:
>>> On 18/01/2017 23:58, Calum wrote:
>>>> Posted yesterday.
>>>>
>>>> <http://www.oracle.com/us/products/servers-storage/servers/sparc/oracle-sparc/sparc-roadmap-slide-2076743.pdf>
>>>>
>>> Interesting, some sort of OS group inside Fujitsu Japan have been
>>> fiddling with Solaris 12 for a while now, "such-and-such on S12 is
>>> different, please make (your thing) handle it" messages come our way
>>> every now and then.
>>
>> Although Fujitsu did release a new SPARC chip last year (although
>
>> I'm not sure it's in any products which run Solaris), they have
>
>
> The Sparc 64 X+ ?

Looking back at last year's Hot Chips, that might have been it.

> I'm installing some M10-1+s at the moment. Quite strange playing with
> LDoms on non Sun4v architecture :-)

A client has some. They were the cheapest SPARC systems available
before Sonoma, so useful when you simply needed something with
SPARC instruction set to run a legacy application.

My concern for SPARC for some years is that volumes have been way
below what I consider sustainable. Oracle and Fujitsu had done some
excellent work on the processor development, but on the other hand,
Oracle did everything they could to stop people buying them.

>> indicated their future direction is going to be ARM.
>> They don't run Solaris on their SPARC supercomputers anyway, only
>> on their M10 range. Their supercomputers run Linux.
>
> Solaris on ARM ??? :-)

If ARM forges itself a base which gives significant benefits over
x86 in areas which are of interest to Illumos users, then they might
take up an ARM port again. However, I don't see this at the moment -
Joyent and Delphix are quite tied to running on x86, Nexenta and the
other ZFS storage platforms don't have anything to be gained by moving
to ARM. ARM is big in the minature low power and HPC spaces, and
Solaris/Illumos are not in those spaces. (That's not to say an
enthusiast won't port Illumos so they can run it on their tablet, but
I don't see a commercial market for a port.)

ZFS is clearly the main driver for using Solaris/Illumos over, say,
linux at the moment, and combined with the virtualisation foundations
in Solaris/Illumos, it made it worth Joyent implementing lots of Linux
compatibility in Illumos (e.g. Joyent's SmartOS virtualisation platform).

However, ZFS ports on other OS's are coming along at an astonishing
rate - particularly Linux where there are now many hundreds of
thousands of systems running ZFS (300,000 in commercial use by one
company alone), and a large number of ZFS Linux developers spread
across many companies, but all working together.

There are still some functionality gaps which limit adoption for
some use cases, but those are all being worked. I think the
functionality gaps will be gone well before the end of this year.
That would give you fully functioning ZFS on Linux/ARM, if you want
to use it on ARM, and reduces any likelyhood of porting Illumos to ARM.

[YTC#1]

On 22/01/2017 12:49, Andrew Gabriel wrote:
> In article <o61vdm$6fr$1...@dont-email.me>,
> YTC#1 <b...@ytc1-spambin.co.uk> writes:
>> On 21/01/2017 14:25, Andrew Gabriel wrote:
>>> In article <9rm9ld-...@paranoia.mcleod-schmidt.id.au>,
>>> "Gary R. Schmidt" <grsc...@acm.org> writes:
>>>> On 18/01/2017 23:58, Calum wrote:
>>>>> Posted yesterday.
>>>>>
>>>>> <http://www.oracle.com/us/products/servers-storage/servers/sparc/oracle-sparc/sparc-roadmap-slide-2076743.pdf>
>>>>>
>>>> Interesting, some sort of OS group inside Fujitsu Japan have been
>>>> fiddling with Solaris 12 for a while now, "such-and-such on S12 is
>>>> different, please make (your thing) handle it" messages come our way
>>>> every now and then.
>>>
>>> Although Fujitsu did release a new SPARC chip last year (although
>>
>>> I'm not sure it's in any products which run Solaris), they have
>>
>>
>> The Sparc 64 X+ ?
>
> Looking back at last year's Hot Chips, that might have been it.
>
>> I'm installing some M10-1+s at the moment. Quite strange playing with
>> LDoms on non Sun4v architecture :-)
>
> A client has some. They were the cheapest SPARC systems available

I think they are cheap, current client (a council) thinks they are
expensive :-)

> before Sonoma, so useful when you simply needed something with
> SPARC instruction set to run a legacy application.
>
> My concern for SPARC for some years is that volumes have been way
> below what I consider sustainable. Oracle and Fujitsu had done some
> excellent work on the processor development, but on the other hand,
> Oracle did everything they could to stop people buying them.

I know what you mean, I am seeing less and less install and config work
for new projects. Recent years the focus has been P2V of legacy to new,
while customer is deciding on future project direction (ie possibly move
away from Solaris).

>
>>> indicated their future direction is going to be ARM.
>>> They don't run Solaris on their SPARC supercomputers anyway, only
>>> on their M10 range. Their supercomputers run Linux.
>>
>> Solaris on ARM ??? :-)
>
> If ARM forges itself a base which gives significant benefits over
> x86 in areas which are of interest to Illumos users, then they might
> take up an ARM port again. However, I don't see this at the moment -
> Joyent and Delphix are quite tied to running on x86, Nexenta and the
> other ZFS storage platforms don't have anything to be gained by moving
> to ARM. ARM is big in the minature low power and HPC spaces, and
> Solaris/Illumos are not in those spaces. (That's not to say an
> enthusiast won't port Illumos so they can run it on their tablet, but
> I don't see a commercial market for a port.)

A challenge for Casper :-)

>
> ZFS is clearly the main driver for using Solaris/Illumos over, say,
> linux at the moment, and combined with the virtualisation foundations
> in Solaris/Illumos, it made it worth Joyent implementing lots of Linux
> compatibility in Illumos (e.g. Joyent's SmartOS virtualisation platform).
>

I notice Joyent was bought by Samsung recently.
I am curious to where SmartOS will end up.

> However, ZFS ports on other OS's are coming along at an astonishing
> rate - particularly Linux where there are now many hundreds of
> thousands of systems running ZFS (300,000 in commercial use by one
> company alone), and a large number of ZFS Linux developers spread
> across many companies, but all working together.

Yet Oracle (unless I am mistaken) don't support it on their version of
Linux ?

>
> There are still some functionality gaps which limit adoption for
> some use cases, but those are all being worked. I think the
> functionality gaps will be gone well before the end of this year.
> That would give you fully functioning ZFS on Linux/ARM, if you want
> to use it on ARM, and reduces any likelyhood of porting Illumos to ARM.
>



--

[Andrew Gabriel]

In article <o62vab$8vr$1...@dont-email.me>,

YTC#1 <b...@ytc1-spambin.co.uk> writes:
> On 22/01/2017 12:49, Andrew Gabriel wrote:
>> ZFS is clearly the main driver for using Solaris/Illumos over, say,
>> linux at the moment, and combined with the virtualisation foundations
>> in Solaris/Illumos, it made it worth Joyent implementing lots of Linux
>> compatibility in Illumos (e.g. Joyent's SmartOS virtualisation platform).
>>
>
> I notice Joyent was bought by Samsung recently.
> I am curious to where SmartOS will end up.

Running Samsung's cloud.
However, they've been very clear that they won't be interfering
with the way Joyent works with SmartOS, so it will remain available
and developed just as it is now.

>> However, ZFS ports on other OS's are coming along at an astonishing
>> rate - particularly Linux where there are now many hundreds of
>> thousands of systems running ZFS (300,000 in commercial use by one
>> company alone), and a large number of ZFS Linux developers spread
>> across many companies, but all working together.
>
> Yet Oracle (unless I am mistaken) don't support it on their version of
> Linux ?

Oracle Linux was driving the BTRFS project at the time, although those
staff left some time ago, and the project now lives inside Facebook.
BTRFS saw ZFS (and Solaris) as their competitor, so there was no way
they would have considered using ZFS in Oracle Linux.

If Oracle Linux changed its mind, it would be interesting to see which
ZFS they included in Oracle Linux. I suspect there would be no customer
interest in using Oracle ZFS because it's now proprietary/lock-in.
That would mean Oracle adopting OpenZFS, which would be vastly easier
because it already works on Red Hat. (Oracle Linux is rebuilt/repackaged
Red Hat.)

[Calum]

On 18/01/2017 12:58, Calum wrote:
> Posted yesterday.
>
> <http://www.oracle.com/us/products/servers-storage/servers/sparc/oracle-sparc/sparc-roadmap-slide-2076743.pdf>

And a video update from John Fowler:
https://www.youtube.com/watch?v=SDR5hyQ_1VE

[YTC#1]

Software on silicon

How long before "Solaris on Sparc" means the OS is on the chip ? "