Re: SuSe and Oracle11- why SLES9 is notsupportedwhile RHEL4 (and OL4) are supported?

Here comes another problem. It's why I argued with Novell on all 'version 
1', explaining that they publish system of 'Open Beta' quality as 'Release' 
and so shorten life time dramatically.

SLES9 - release was not usable (virtual memory problems, iSCSI problems, 
stability problems) - only SLES9 SP1 became usable (and only SLES9 SP3 
became rock solid, after upgrade 283). SLES10 release was not usable - only 
SLES10 SP1 became partially usable.
If we look onto RHEL5, we will see that they had a very long (almost a year) 
Open Beta period when everyone could download system, try it, write a 
software, but system was not counted as a production so these 5 years was 
not counting yet. As a result, their '-release' comes in time of 'SLES* SP1' 
but have already good software and hardware vendors support and have a rock 
solid quality of SLES* SP1. Just because ythey don't pretend that their 
'Beta' is 'Release'.

If we look into the SLES life span - they release useless -release but start 
counting 5 years from this time. And, worst, they abandon hardware support 
for the old version since this very moment (purchase DELL. You can order 
RHEL4, RHEL5, SLES10 but not SLES9. Why? Take ET64T server - SLES8 never was 
modified to support this platform /RHEL3 was modified/.)

I (and many, MANY, users) DONT NEED SLES10 on a servers. I don't need all 
these new changes - they all are weird and not stable. New ISCSI is 
terrible. New YOU system is terrible. New installation system is very 
flexible but extremely slow - I was 100% satisfied with SLES9 installation 
system on servers /only on servers/. New libraries lost old thread support. 
System is not Oracle9 compatible. Oracle have a strange errors with standby 
logs when running on SLES10. I don't want to be a beta tester on production 
systems. All I need is a system of SLES9 SP3 quality (not of SLES10 SP1 
quality), with full support on new hardware (including new DELL servers, new 
SUN servers, new HP servers - exceptions possible but they must be 
exceptions), compatible with all 3 major Oracle versions, and having at 
least some basic software updated (heartbeat2 - why it is not installed as 
an option? It's pretty simple).

If we are talking about SLES10, I can count 5 years from this June - when 
SP1 was released and system became usable. Not from SLES10 - release - it 
was pretty useless system. I can understand if Novell stop migrating SLES9 
onto new platforms when SLES10 is (counting from June) 2 years old. But 
now... hmm, I don't want to be an experimental rabbit (to be honest, I like 
to be - in the lab, so I wil test it all - but not in prod).

And I am software engineer (not just sysadmin) so I can bear _Beta in 
production_...  Many other sysadmins (esp who are not software engineers) do 
not have enginering skills and have only one choice after all. And you can 
guess, which one.

PS. OpenSuSe is a great system (I put it ahead of Ubuntu after few 
experiments). SLED is a great system. Problem is that SLES is for _SERVERS_ 
and can't bear 'SLED' or 'OpenSUSe' quality and variability. SO it should 
not be so volative as SLED or OpenSUSe.

PPS. Strange, but FreeBSD life span is much better, too. They support 2 
systems at once all the time (FreeBSD 5 and FreeBSD 6 for now), and switch 
from the old system on approximately 10-th - 15-th release of the new 
system.

----- Original Message ----- 
From: "Tom Sightler" <[email protected]>
To: "Alexei_Roudnev" <[email protected]>
Cc: "Arun Singh" <[email protected]>; <[email protected]>
Sent: Saturday, August 11, 2007 7:50 AM
Subject: Re: [suse-oracle] SuSe and Oracle11- why SLES9 is notsupportedwhile 
RHEL4 (and OL4) are supported?


> On Fri, 2007-08-10 at 15:33 -0700, Alexei_Roudnev wrote:
>> Sysadmins are not dumb - seen all this, they select RHEL4 or UL4 - these 
>> 2
>> systems are well compatible with all Oracle versions (I can even run
>> Oracle8i on RHEL4, with some hacks) and are rock-solid.
>
> Your right, sysadmins are not dumb, unfortunately, software companies
> are dumb.  They seem to be pushing shorter and shorter product
> lifecycles and have forgone "stable" products for simply "current"
> products.  Last time I looked Novell's official lifecycle for SLES9 was
> five years.  Oracle's support for their database products is not much
> better, although at least you can pay extra.
>
> Now, 5 years may sound like a long time, but the problem is that modern
> software products are complex and generally full of bugs when they are
> released.  It usually takes 1-2 years before a product truly becomes
> stable.  Many companies won't even look at a product until the 1st or
> 2nd service release, which is usually between 1-2 years after a product
> release.  Then it usually takes another few months before it's approved
> for deployment and another year to be deployed within the company.  That
> means that a product doesn't see major market penetration until 2-3
> years into it's lifecycle.  If the product only has a 5 year lifecycle
> that means only 2-3 years of truly usable time in most companies.
> That's far too short for critical software like databases and OS's.
>
> My guess is that this is what lead to RHEL4 being certified while SLES9
> is not.  Redhat has a 7 year lifecycle on it's products, but Novell only
> officially guarantees 5 years (as far as I know anyway).  That means
> that SLES9 is scheduled to be desupported on August 2009 only two years
> from now, and it already feels abandoned.
>
> RHEL4 on the other hand, will be supported until Feb 2012, nearly 4 1/2
> years from now, even though it's initial release only trailed the SLES9
> release by 6 months.  Those two years might not seem like much, but they
> are huge for most companies and have a huge TCO impact.  It was one of
> the major reasons that we switched from SUSE to Redhat several years
> ago.
>
> Later,
> Tom
>
> 


-- 
To unsubscribe, email: [email protected]
For additional commands, email: [email protected]
Please see http://www.suse.com/oracle/ before posting