Hi Jason,

Thanks for pointing some things that I think most non-Sun folks have been 
thinking about..

I do question the "linux familiarity" goal. It's as if Sun expects the linux 
folks to stop drinking their kool-aid and come over to our camp. The reality is 
that most shops have SA's that have managed Solaris/AIX/HP-UX/Tru64/etc and are 
now having to deal with linux. The reality is that linux is not familiar, if 
anything it's confusing. I can't even begin to list the number of complaints I 
hear from SA's who have managed UNIX platforms and are now forced to manage 
linux. A great example is how difficult it can be to configure SAN storage in 
Linux compared to Solaris (Leadville and MPXIO have made soo many peoples lives 
easier). Enough mistakes have been made that have opened the flood gates for 
companies to explore alternatives that are *cheaper*, only to find they are 
more costly when you taken into account TCO. 

With Solaris being free and OpenSolaris being open sourced, we are left with 
one definitive stance to prevent the continual bleeding of our users and 
customers.. we must seek to be better than linux or the past UNIX brands. I 
don't think copying linux is the answer. If linux customers are going to cross 
(back?) over to Solaris or goto OpenSolaris, there has to be something better 
than just the kernel or the logo. Surrounding our OS with lots of freeware and 
tools from linux isn't the answer.

Here's some food for thought. Redhat is very popular in North America, while 
Suse is more popular in Europe. Ever thought of what will happen with 
globalization as companies using linux continue to consolidate? It's like 
trying to sell pepsi and coke in the same vending machine, doesn't work out:) 
The reason why soo many companies and governments use Solaris today is because 
of its stability and the standards it delivers on. The rush to create something 
new, OpenSolaris, has its merits, but it must be tempered. Do people really 
want an OpenSolaris core in Linux clothes?

Ultimately, it's the standards and the differentiation that a product exhibits 
that enables it to last. Believe me, many linux distros have come and gone over 
the years. Why are we trying to follow? I thought we were innovators?

I applaud the Indiana project for creating new features and methods of doing 
things. There are definitely many things that were getting stale and crufty in 
Solaris. It was time for a good house cleaning. But I'm not getting a sense of 
standards or direction. I see a lot of requests to fast track everything in the 
open source kitchen sink into OpenSolaris. Would it not be easier to just have 
a pkg-src/ports converter to IPS tool, make sure things work, and be done with 
that? I'm all for having pkgs A-Z available from the IPS repo, but do I want 
everything in /usr/bin? I haven't seen guidelines on where things will install 
or what defines the core of OpenSolaris. Do customers want a large bloated 
/usr/bin, or do they want non-core items to be in /opt/local ?? Do customers 
want a linux, bsd, or Solaris personality in their OS? Having a well organized 
and standardized OS would be ideal and highly desirable. When are these 
conversations going to happen?

Hopefully these issues will be addressed. I know many people in the community 
are eager to participate, but right now feel a lack of communication and 
direction. It's bad enough that 2008.05 doesn't run on SPARC! It's nice to see 
it work on x86, but many folks have SPARC gear (parity is important). I hope 
it'll work well on my Sun Blade 2000 this year.

 *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
Octave J. Orgeron
Solaris Systems Engineer
http://www.opensolaris.org/os/community/sysadmin/
http://unixconsole.blogspot.com
unixconsole at yahoo.com
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*


----- Original Message ----
From: Jason King <[email protected]>
To: John Plocher <John.Plocher at sun.com>
Cc: Octave Orgeron <unixconsole at yahoo.com>; sysadmin-discuss at 
opensolaris.org; Alan Coopersmith <Alan.Coopersmith at sun.com>; Dean Roehrich 
<dr161460 at sac.sfbay.sun.com>; lsarc-ext at sun.com
Sent: Thursday, June 5, 2008 11:16:27 PM
Subject: Re: [sysadmin-discuss] logrotate

On Thu, Jun 5, 2008 at 10:57 PM, John Plocher <John.Plocher at sun.com> wrote:
> Octave Orgeron wrote:
>> 1. We stick with logadm, forget logrotate, and perhaps enhance logadm?
>> 2. We drop logadm and switch to logrotate.
>> 3. Standardize on logadm and make it handle logrotate.conf syntax in 
>> addition to the logadm.conf syntax.
>> 4. Standardize on logrotate and make it handle logadm.conf syntax in 
>> addition to the logrotate.conf syntax.
>
> and maybe
>
> 5.  Ship logadm as usual.  Put logrotate into the IPS non-core
>     repo for people to pull if desired (with all the caveats and
>     warning comments mentioned in this thread...), but ensure
>     that it does not get installed by default.
>
> The project team needs to consider why we need to invest
> resources into putting this project into core opensolaris.
> I haven't seen any rationale for /why/ we need this - and
> being on a list somewhere isn't sufficient reason stop
> using common sense...

I believe the justification was 'linux familiarity'.  Unfortunately, I
haven't found any sort of discussion in terms of policy or guidelines
around this (if there are, pointers would be welcome).  All I've seen
is 'anything that makes things more like Linux is ok', even though
that's a bit of a vague target -- different Linux distros ship
different stuff.  It ends up at least having the appearance of being
used a trump card over all objections.

If the overriding goal is 'linux familiarity', we can all save a lot
of time and effort and rename nexenta opensolaris and call it done --
by it's nature it's going to have the most 'linux familiarity' of
anything else using opensolaris.  If not, we should decide how much is
enough/too much/too little.

>
> As Richard says, logrotate does not seem to be one of those
> "must have" things whose lack makes OpenSolaris unusable
> compared to Ubuntu.  Yes, it /is/ a "nice to have" convenience,
> and users will be glad that it is easily available and just
> works on OpenSolaris.  But, IMO, it is just desert, and we
> need to focus on the main course.
>
>   -John
> _______________________________________________
> sysadmin-discuss mailing list
> sysadmin-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/sysadmin-discuss
>



      

Reply via email to