[csw-users] Why support Solaris 8 onwards?

[Trygve Laugstøl]

Dennis Clarke wrote:
>> On 20/09/07, 16:37:00, Torrey McMahon <tmcmahon2 at yahoo.com> wrote regarding
>> Re: [csw-users] Why support Solaris 8 onwards?:
>>
>>>>>> http://www.sun.com/service/eosl/solaris/solaris_vintage_eol_5.2005.xml
>>>>>>
>>>>> 2009-03-31 looks like end of support to me.
>>>>>
>>>>> Phase 2 to us might mean keeping the archive on the mirrors but
>>>>> not updating, "access to existing".
>>>>>
>>>>>
>>>> Solaris 8 End Of Service Life is 03/31/12
>>>>
>>> I would assert that Sun has more resources in the support space then
>>> Blastwave does and that aligning with the Sun support documents isn't
>>> the best thing to do.  Do we really want to maintain Solaris 8 packages
>>> for an other 4+ years?
> 
> That has no bearing here at all.  If I choose to compile gzip on Solaris 8,
> and I do just that, then I am happy to report that it builds and passes all
> tests and I can even optimize it a bit. It can be released for any Solaris
> revisoion from 8 upwards.
> 
> If some other package is murder to compile on Solaris 8, for whatever
> reasons, then it makes sense to just move over to Solaris 10 Update 4 and
> make a release for that Solaris only.  Maybe. Depends on the situation.
> 
> This has nothing to do at all with Sun's resources or the resources here at
> Blastwave.
> 
>> 2009-03-31 - today => 1 1/2 years
>>
>> Sun supports the OS; we provide packages built on the OS.  We're doing
>> different jobs.  It's just a date that represents the last time anyone
>> will reasonably use S8.
>>
> 
>   The word "reasonably" is a tough one to bring up here. There are a lot of
> really really large corporations out there running Solaris 8 and they will
> not move until they are pushed real hard. They have no reason to move. It
> works just fine and is very stable and very secure.
> 
>   I would go so far, right out on a limb here based on my many years of
> business experience that Sun will be supporting Solaris 8 for as long as
> those customers dictate. Its a simple matter of money and some
> corporations will write a check for $5M a year for support just to keep
> there E4500 servers running for one more year.
> 
>   If anyone out there thinks I am completely wrong, and you have been in the
> IT business over a decade ( at least ) then please speak up. To me .. its
> just simple business. While the number of Solaris 8 users is slowly going
> away there will *always* be Solaris 8 users.
> 
>> I strongly believe we should set a date, one that gives people due
>> notice, announce it and stick to it.
> 
>   I agree.
> 
>   No more discussion or debate.
> 
>   We should pick that date that Sun has set for end of phase 1 support and
> that looks to be 3/31/09
> 
> See http://www.sun.com/service/eosl/solaris/solaris_vintage_eol_5.2005.xml
> 
>>> What about Solaris 9?
>> Good question, most of the augments put forward against S8 apply to S9,
>> (some even apply to S10, eg OpenSolaris libs).
>>
>> Same rational?  Although personally I've skipped S9 completely so
>> after S8 is dropped I can start moaning. :-)
> 
> So, let's tackle one mountain at a time here. What are your thoughts on end
> of phase 1 support ?  31st March 2009 ?

How about forking the existing Blastwave into a solaris 8 branch and a 
solaris 10 branch until 2009 and then leave the 8 branch for dead?

The purpose for the Solaris 8 would be to mainly fix bugs and upgrade 
existing packages. I'm pretty sure that most of the developers would 
move on to the Solaris 10 branch, but it seems that Solaris 8 still is 
hanging around, and then there should be enough maintainers to bugfix 
and upgrade a package now and then.

--
Trygve