digression
IBMs Power may continue. Whether that is to support both AIX and RedHat is another issue. Muttering from users suggests a strong push in IBM to encourage moves to Redhat Linux.
Oracle has created an additional version of the Solaris operating system it acquired in 2009, when it bought Sun Microsystems. The new cut of the OS is called a Common Build Environment (CBE). As explained by Oracle senior software engineer Darren Moffat this week, a CBE is akin to a beta because it includes prerelease builds …
Those are all consumer hardware manufacturers. The Power architecture seems to sell mainly in the more serious space. Here you go, a cluster with 31424 Power cores spread over 1964 processors.
Those are all consumer hardware manufacturers. The Power architecture seems to sell mainly in the more serious space. Here you go, a cluster with 31424 Power cores spread over 1964 processors.
You may be missing the point here. In the mists of history, the Power architecture used to have some interesting volume markets, such as Macintosh, OS/2, and Nintendo consoles. 31,424 cores is super-cool and awesome, but 1,964 is fewer than the number of Arm processors sold every three seconds not so long ago ("842 Chips Per Second: 6.7 Billion Arm-Based Chips Produced in Q4 2020", Tom's Hardware). I imagine the numbers have grown since then.
You've got to have a hell of a lot of margin to make that kind of market math work in your favor. You can be serious AF, but at some point you need to make money.
Is there anything worthwhile in SPARC or Oracle CBE that cant be completely replaced by Linux built for SPARC?
IBM have Redhat even if all the AIX olden goldies get shoved out the door. Many of the truly useful stuff in AIX are much dead and not relevant any which way you look at it.
Why even bother at this point? Oracle and IBM are just "Holding" the pursestrings and not really driven like before.
Yes. Solaris is stable. It also has very functional working SMF unlike LInux's systemd cancer.
LDOM/containers and virtualisation works very well and you can live migrate VMs with no impact to what is running.
IMHO running Solaris (in corporate environment) on Sun^WOracle hardware is a no brainer while Solaris is being supported.
Disclaimer: Yes I am slightly biased with house full of old Sun kit, but in my defense nothing new enough to have been done under Oracle's label.
Also forgot live update. On a Solaris 11 box w/a ZFS root, updating a system became a breeze. One benifit of Oracle buying Sun was that Sun's Ops Mgr became included with the support for the OS. The older versions were quirky, but I set it up in an environment I inherited, and keeping it updated was but a click.
-> Is there anything worthwhile in SPARC or Oracle CBE that cant be completely replaced by Linux built for SPARC?
Sun Clustering for one thing. I have seen and heard Oracle DBAs insisting on Sun Cluster rather than Oracle clustering. They know it inside out and trust it.
Before you write something like 'port it to Linux', why bother? Linux is not actually all that good. It's not even cheap if you want a supported version. The fact that it (used to be able to) runs on low end hardware is not even slightly relevant in the corporate field.
One feature where Linux has regressed horribly is the stupidly different packaging systems there are available. yum, apt, pkg, dnf, etc, etc, etc, etc, etc, etc.
Oh it's just a matter of learning apt instead of yum. Yes, like I want to waste my time learning YAPS yet another packing system. The Linux 'community', that is the people who keep thinking up these entirely incompatible and competing packaging systems are stupid. They're clever enough to waste time developing a new packaging system, but they're not clever enough to see how much time this wastes for everyone.
Many people would say that the rapid development of alternative systems is the advantage Linux had over Solaris.
With Solaris, you get what Sun/Oracle gives you and that’s the way it will always be.
Of course: yes, Linux is hugely messy compared to a top-down controlled Solaris or Mac…
The use of the word "learn" is a little bit of an overstatement here. This is nothing more than reading a one screen help page since all packaging systems are doing the same thing: install/update/remove. If you find this a time consuming and challenging activity then maybe you should reconsider your career as a sysadmin.
Worked with Solaris for almost 25 years, damn fine O/S stable as hell and if you're using Oracle on top it's hard to beat. However, the end is near and Solaris is going the way of Tru64 and HP/UX, only a handful of places will keep it simply because "it ain't broke, don't fix it". No problems with that.
The shop I work for has been using Solaris since 1997 and from around 80 SPARC servers in the heyday around 2000, we're now down to running just 3 hosts as global hosts with just 6 containers ( think VMs on ESX ). We've moved to Windows, Linux or cloud native services in the last 2-3 years and we're planning to kill the last bit of SPARC over the next 18 months, also looking to replace Oracle with Postgres and SQL Server. MS licensing isn't great but you're made to feel slighly less like a criminal when you mess up an MS license over Oracle!
You might try SmartOS.
The whole thing runs on a USB flash drive, and lets you offer all your storage as ZFS pools with no footprint of the hypervisor OS.
KVM has been grafted into the Illumos kernel for running Linux VMs, but you can also run Zones.
It's likely possible to run an Oracle database on it, but you probably have other preferences in that arena.
There are two ways to interpret your question depending on whether you're asking whether Solaris has anything Linux doesn't, or whether Solaris on SPARC machines has anything Linux on SPARC machines doesn't. The answer in both cases is certainly yes, but for very different reasons.
Sun allowed their SPARC processors to fall far behind the competition about 15-20 years ago in much the same way Intel have today, and they never recovered. SPARC was irrelevant even before Oracle bought it, but it was a long goodbye for Oracle and a longer one for Fujitsu. If you're stuck with a hefty investment in it, Solaris is the only real game in town; Linux doesn't support any modern machines; by modern here I mean machines that anyone who is economically sensitive would even consider using: nostalgic hobbyists aren't included. Even the most recent processors on the supported list are 10 years old, and if you follow the mailing list traffic you'll see that anything much newer than 2005 is likely to be broken and rarely used -- again, because hobbyists can't afford that equipment and the people who can need it to keep working until it's finally retired.
As for the more general Solaris vs GNU/Linux question, that as always is a matter of taste. Solaris was built with a clear architectural vision: the pieces of system software fit together, work together, and feel coherent and solid. It is an operating system, not just a kernel. The atrocities one finds in the GNU/Linux world are mostly absent. On the other hand, Oracle ships every copy of Solaris (again, assuming you are not a hobbyist but using these things for something real) with a fat bill and the promise of license audits and support costs for many years to come. Unless you prefer Linux in the abstract, you're better off with illumos which comes from the same technology base as Solaris but is open source and free to use for everyone. OpenIndiana was an obsolete distribution of illumos; modern alternatives such as SmartOS and OmniOS exist.
OpenIndiana - not really. Due to lack of available hardware (and maybe people able/willing to do the work) OpenIndiana Hipster no longer supports SPARC.
https://www.openindiana.org/es/documentation/faq/#does-openindiana-provide-a-sparc-release
There seem to be a few (more) obscure spins that still support SPARC.
I loved Solaris. Still do. It is the best commercial Unix available, in my opinion. And it is better than Linux and *BSD.
But I won't touch it for my own use anymore. I remember the early 90s when it was free for personal use, then it wasn't. Then it just was not available. Then it was again. And you want me to trust Oracle not to yank access at some VPs whim? If access was bad under Sun Microsystems, it was worse under Oracle and I, for one, will never go back.