Future of Solaris
4Space
What are peoples thoughts on the future of Solaris? Sun have begun to
include their own flavour of Linux on some of their servers in place of
Solaris. Coupled with Sun's interest in Gnome, does this perhaps signal an
intent to direct R&D towards Linux and away from Solaris??
I'm looking to get some *nix development under my belt, I'm already used to
using and developing on Linux, but have begun looking at Sun workstations
and Solaris. Is this wasted effort? Is Solaris here to stay?
What do people see as the relatively merits of Linux over Solaris and vice
versa?
Cheers,
4Space
Thomas Dehn
"4Space" <4Sp...@nospam.com> wrote:
> I'm sure this question has been asked before, but....
>
> What are peoples thoughts on the future of Solaris?
> Sun have begun to include their own flavour of
> Linux on some of their servers in place of
> Solaris. Coupled with Sun's interest
> in Gnome, does this perhaps signal an
> intent to direct R&D towards Linux and away from Solaris??
I think its just making a few extra bucks with Linux.
A few popular open source applications and tools
such as gnome will be integrated into Solaris,
and thats it. Sun has a long history of contributing
to the open source community, such as NFS, and
math libraries used in gcc. This is nothing new.
And Solaris is the core of Sun's business, they won't
abandon it.
Thomas
Thomas Glanzmann
> I'm sure this question has been asked before, but....
Stop that stupid thread now and look at other equivalant threads at this
group.
> What are peoples thoughts on the future of Solaris? Sun have begun to
> include their own flavour of Linux on some of their servers in place of
> Solaris. Coupled with Sun's interest in Gnome, does this perhaps signal an
> intent to direct R&D towards Linux and away from Solaris??
No, it will not.
> I'm looking to get some *nix development under my belt, I'm already used to
> using and developing on Linux, but have begun looking at Sun workstations
> and Solaris. Is this wasted effort? Is Solaris here to stay?
Yes, Solaris will stay.
> What do people see as the relatively merits of Linux over Solaris and vice
> versa?
Linux servers are stable and more flexible. Solaris servers are more
reliable. :-) Everyone happy?
bolt thrower
: reliable. :-) Everyone happy?
What is the difference between stable and reliable?
Thomas Glanzmann
>: Linux servers are stable and more flexible. Solaris servers are more
>: reliable. :-) Everyone happy?
> What is the difference between stable and reliable?
I can't find a definition of reliable. But stable is [1].
[1] http://www.investorwords.com/cgi-bin/getword.cgi?4675
But to answer your question ... just ignore stable. :-)
Thomas
Rich Teer
> What are peoples thoughts on the future of Solaris? Sun have begun to
> include their own flavour of Linux on some of their servers in place of
Not true. Sun has offered Linux IN ADDITION TO, not instead
of, Solaris x86. Hey, even if you offer people caviar (Solaris),
some of them will still just want beans on toast (Linux).
> Solaris. Coupled with Sun's interest in Gnome, does this perhaps signal an
> intent to direct R&D towards Linux and away from Solaris??
Absolutely not.
> I'm looking to get some *nix development under my belt, I'm already used to
> using and developing on Linux, but have begun looking at Sun workstations
If you're developing for Linux, you're not (necessarily)
developing for UNIX. See the thread on comp.unix.programmer.
Specifically, Linux is not standards compliant, and from what
I gather, has no real intentions or interest in being so.
Hmm. Sounds just like everyone's favourite convicted monopolist...
In other words, if you want to be a UNIX developer, you're
starting from the wrong place.
> and Solaris. Is this wasted effort? Is Solaris here to stay?
Solaris is definately here to stay.
> What do people see as the relatively merits of Linux over Solaris and vice
> versa?
Linux has a lot of mindshare due to it being the subject of
much hype. On a server that has a light workload, it is
more stable than Windoze. On x86, it also supports more
hardware.
Solaris is a standards compliant, data centre ready UNIX OS.
It is stable, reliable, and secure.
Solaris is to Linux like Linux is to Windoze.
--
Rich Teer
President,
Rite Online Inc.
Voice: +1 (250) 979-1638
URL: http://www.rite-online.net
Paul Eggert
> Linux is not standards compliant,
Sure it is. It's just a different standard. :-)
http://www.linuxbase.org/
> Sounds just like everyone's favourite convicted monopolist...
Not at all. GNU/Linux is freely available and if you want it to
behave like POSIX you're quite free to make it do so.
> Solaris is a standards compliant,
Not exactly. Solaris strives for POSIX compliance harder than
GNU/Linux does, but it often falls short. For example, Solaris 9 NFS
breaks GNU make, because it causes 'stat' to fail with errno == EINTR
even if you've specified via SA_RESTART that the call be restarted.
See, for example, a bug report from Ted Stern of Cray at
<http://mail.gnu.org/archive/html/bug-make/2003-01/msg00034.html>.
Stern suggested that someone file a bug report with Sun. I did so, in
Sun service order 63351622. Sun replied that this is not a standards
compliance issue, partly because Solaris 9 does not claim to conform
to POSIX 1003.1-2001 (it conforms only to an older POSIX version,
which doesn't have SA_RESTART), and partly because (I'll paraphrase
:-) NFS doesn't conform to POSIX anyway so who cares?
So what good does Solaris's "standards compliance" buy us here? Not
much. The code works on GNU/Linux, OpenBSD, etc., where "stat" works
as expected. But it doesn't work on Solaris.
Dragan Cvetkovic
> Rich Teer <rich...@rite-group.com> writes:
>
> > Linux is not standards compliant,
>
> Sure it is. It's just a different standard. :-)
> http://www.linuxbase.org/
>
I had hard time there looking for shmctl(2) definition and at the end it was
pointing to "CAE Specification, January 1997, System Interfaces and Headers
(XSH),Issue 5 (ISBN: 1-85912-181-0, C606)" which I assume is the same as in
POSIX. But it doesn't behave like that. :-(
Bye, Dragan
--
Dragan Cvetkovic,
To be or not to be is true. G. Boole No it isn't. L. E. J. Brouwer
CJT
> I'm sure this question has been asked before, but....
Yes, so do a Google search.
Joerg Schilling
Paul Eggert <egg...@twinsun.com> wrote:
>> Solaris is a standards compliant,
>
>Not exactly. Solaris strives for POSIX compliance harder than
>GNU/Linux does, but it often falls short. For example, Solaris 9 NFS
>breaks GNU make, because it causes 'stat' to fail with errno == EINTR
>even if you've specified via SA_RESTART that the call be restarted.
>See, for example, a bug report from Ted Stern of Cray at
><http://mail.gnu.org/archive/html/bug-make/2003-01/msg00034.html>.
And BASH has a much worse bug with ^C and make...
Bash illegally makes Jobcontrol for "sh -c 'command'" things.
Infortunately /bin/sh is bash on Lunux :-(
This results in the following problem:
If you have a set of structured recursive make calls and one or more
of the higher level make scripts executes a for loop that calls sub-makes
while you hit ^C, you will only abort the top level make because bash
illegaly did change the process group of the sub processes.
So Linux hit back for make environments and we are 1:1....
--
EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
j...@cs.tu-berlin.de (uni) If you don't have iso-8859-1
schi...@fokus.fhg.de (work) chars I am J"org Schilling
URL: http://www.fokus.fhg.de/usr/schilling ftp://ftp.berlios.de/pub/schily
Joerg Schilling
Dragan Cvetkovic <dcvet...@gmx.net> wrote:
>Paul Eggert <egg...@twinsun.com> writes:
>
>> Rich Teer <rich...@rite-group.com> writes:
>>
>> > Linux is not standards compliant,
>>
>> Sure it is. It's just a different standard. :-)
>> http://www.linuxbase.org/
>
>I had hard time there looking for shmctl(2) definition and at the end it was
>pointing to "CAE Specification, January 1997, System Interfaces and Headers
>(XSH),Issue 5 (ISBN: 1-85912-181-0, C606)" which I assume is the same as in
>POSIX. But it doesn't behave like that. :-(
Linux did illegaly include a definition for union semun in sys/sem.h for a long
time.
Casper H.S. Dik
>In article <7wisuuc...@twinsun.com>,
>Paul Eggert <egg...@twinsun.com> wrote:
>>> Solaris is a standards compliant,
>>
>>Not exactly. Solaris strives for POSIX compliance harder than
>>GNU/Linux does, but it often falls short. For example, Solaris 9 NFS
>>breaks GNU make, because it causes 'stat' to fail with errno == EINTR
>>even if you've specified via SA_RESTART that the call be restarted.
>>See, for example, a bug report from Ted Stern of Cray at
>><http://mail.gnu.org/archive/html/bug-make/2003-01/msg00034.html>.
>And BASH has a much worse bug with ^C and make...
The POSIX standard actually allows EINTR returns from system calls
even if SA_RESTART is given. (SA_RESTART is understood to only
affect a limited set of system calls, the "blocking" system calls;
file operations are not blocking. It is perfectly legal to continue
to return EINTR for other system calls as long as you document that
fact. (which we haven't, but I trust you get the point)
However, NFS is a bit to quick in returning EINTR so I've
logged bug 4826958. NFS tries to block all interrupts save some
but appears not to do so consistently.
Casper
Ian
>Yes, so do a Google search.
And unnecessarily reply copying the whole original message :-(
4Space
Strong stuff.
Thanks for the reply. It makes a refreshing change, usually when these types
of question are asked, the signal to noise ratio is obliterated by
sycophantic rants. Probably indicative of Solaris' intended audience. In the
whole thread, there was only CJT who posted without contributing anything.
Time to start scanning ebay methinks. Some good prices on their.
Cheers,
4Space
Paul Eggert
> The POSIX standard actually allows EINTR returns from system calls
> even if SA_RESTART is given. (SA_RESTART is understood to only
> affect a limited set of system calls, the "blocking" system calls;
OK, but the standard uses the term "interuptible function" rather than
"blocking system call", and it defines an "interruptible function" to
be one that can fail with errno==EINTR.
> file operations are not blocking. It is perfectly legal to continue
> to return EINTR for other system calls as long as you document that
> fact. (which we haven't, but I trust you get the point)
I think I get your point, but I think your interpretation of POSIX is
somewhat dubious. POSIX 1003.1-2001 says that SA_RESTART
affects the behavior of interruptible functions; that is, those
specified to fail with errno set to [EINTR]. If set, and a function
specified as interruptible is interrupted by this signal, the
function shall restart and shall not fail with [EINTR] unless
otherwise specified. If the flag is not set, interruptible functions
interrupted by this signal shall fail with errno set to [EINTR].
-- <http://www.opengroup.org/onlinepubs/007904975/functions/sigaction.html>
Now, what you seem to be saying is that Sun can say for any system
call (e.g., 'stat'), that it can fail with errno==EINTR. So far I
agree -- POSIX explicitly says this. But you also seem to be saying
either (1) that Sun's 'stat' is not specified as interruptible --
which doesn't make sense, as the definition of an interruptible
function is one that can fail with errno==EINTR -- or (2) that Sun can
define 'stat' to be one of the "otherwise specified" functions that
fails with errno==EINTR even if SA_RESTART is set -- which doesn't
seem right to me, as POSIX never says that an implementation can do
such a thing.
Anyway, as a practical matter, if an implementation has the ability to
ignore the SA_RESTART flag on any system call that it pleases, then
SA_RESTART is useless. Surely this is not the intent.
> However, NFS is a bit to quick in returning EINTR so I've
> logged bug 4826958. NFS tries to block all interrupts save some
> but appears not to do so consistently.
Thanks for helping to fix this.
Rev. Don Kool
Paul Eggert wrote:
> Rich Teer <rich...@rite-group.com> writes:
>
>
>>Linux is not standards compliant,
>
>
> Sure it is. It's just a different standard. :-)
> http://www.linuxbase.org/
>
>
>>Sounds just like everyone's favourite convicted monopolist...
>
>
> Not at all. GNU/Linux is freely available and if you want it to
> behave like POSIX you're quite free to make it do so.
>
>
>>Solaris is a standards compliant,
>
>
> Not exactly. Solaris strives for POSIX compliance harder than
> GNU/Linux does, but it often falls short. For example, Solaris 9 NFS
> breaks GNU make,
So what do you suppose running GNU's Not UNIX's (GNU) "gmake" program
have to do with POSIX compliancy? SOLARIS is POSIX compliant and LINTEL
never will be.
Hope this helps,
Don
--
*************************** Q: How many frenchmen does it take
* Rev. Don McDonald, SCNA * to defend Paris?
* Baltimore, MD * A: No one knows, it's never been tried.
***************************
http://mywebpages.comcast.net/oldno7/TheFrench.mpg
Reluctant Yamaha YZF-R1 Owner
Alan Coopersmith
|> Not exactly. Solaris strives for POSIX compliance harder than
|> GNU/Linux does, but it often falls short. For example, Solaris 9 NFS
|> breaks GNU make,
|
| So what do you suppose running GNU's Not UNIX's (GNU) "gmake" program
|have to do with POSIX compliancy? SOLARIS is POSIX compliant and LINTEL
|never will be.
The whole point of the POSIX standards is so that you can write an
application to the standards and it will run correctly on all POSIX
compliant platforms. It shouldn't matter whether the app is GNU make,
Oracle, or some company's custom in-house application. (I don't know
the intimate details of the POSIX specs in the matter in question, so
I don't know if Solaris fails to comply or not. There is unfortunately
quite a fair bit of system behavior outside the scope of the standards,
but upon which programs depend.)
Whether or not Linux is compliant is completely irrelevant - GNU make
ran on SunOS before Linus ever dreamed up his project, and runs on
many non Linux systems today.
--
________________________________________________________________________
Alan Coopersmith al...@alum.calberkeley.org
http://www.CSUA.Berkeley.EDU/~alanc/ aka: Alan.Coo...@Sun.COM
Working for, but definitely not speaking for, Sun Microsystems, Inc.
Rev. Don Kool
Alan Coopersmith wrote:
> "Rev. Don Kool" <old...@comcast.net> writes in comp.unix.solaris:
> |> Not exactly. Solaris strives for POSIX compliance harder than
> |> GNU/Linux does, but it often falls short. For example, Solaris 9 NFS
> |> breaks GNU make,
> |So what do you suppose running GNU's Not UNIX's (GNU) "gmake" program
> |have to do with POSIX compliancy? SOLARIS is POSIX compliant and LINTEL
> |never will be.
>
> The whole point of the POSIX standards is so that you can write an
> application to the standards and it will run correctly on all POSIX
> compliant platforms. It shouldn't matter whether the app is GNU make,
> Oracle, or some company's custom in-house application. (I don't know
> the intimate details of the POSIX specs in the matter in question, so
> I don't know if Solaris fails to comply or not. There is unfortunately
> quite a fair bit of system behavior outside the scope of the standards,
> but upon which programs depend.)
>
> Whether or not Linux is compliant is completely irrelevant -
LINUX isn't POSIX compliant and never will be.
> GNU make
> ran on SunOS before Linus ever dreamed up his project, and runs on
> many non Linux systems today.
And as was pointed out earlier, "gmake" has nothing to do with POSIX
compliancy.