[csw-maintainers] Our infrastructure on Solaris 9

Peter FELECAN pfelecan at opencsw.org
Tue Aug 6 11:03:37 CEST 2013


Laurent Blume <laurent at opencsw.org> writes:

> On 06/08/13 00:43, Peter Bonivart wrote:
>> Me too. I suspect that subset will be quite large, I had to stop
>> building pkgutil for Solaris 8 because basically nothing worked in GAR
>> for Solaris 8 and the same will happen to Solaris 9. It's a wasteful
>> effort, Solaris 10 was released in 2005...
>>
>
> I generally concur with all of the above.
>
> I would like to add a suggestion: outline precisely what somebody
> would need to do if they really wanted to get S9 builds running.
> There are still so many people whining when their S8 boxes stops
> running, I can't help thinking there must be an awful lot of
> mission-critical and yet fully unsupported S9 out there (unless it was
> really little used, because I've known sysadmins avoiding odd-numbered
> Solaris versions).
> And that, in my opinion, could be either in participating directly, of
> course, or in the form of paying a fee or contacting an organization,
> an individual who would do the job of keeping those S9 packages
> running for a fee, and support OpenCSW in another way.
>
> Is that reasonable? Doable?

Laurent,

This reasonable and doable, However, from my standpoint, it is outside
the scope of OpenCSW.

BTW, can somebody explain what hinders the adoption of Solaris 10 in a
private or an enterprise environment after 8 years of availability.

The only case I know from direct experience is in a manufacturing unit
where a multi million euro machine tool is controlled by a Solaris 2.5.1
running system and the supplier ask an equivalent amount to update the
controller by replacing it with an unit running an old version of
Microsoft Windows (what a deal). There is even worse, in a similar
environment there is a SunOS 4.3 controller unit. Now, should we support
all this? The only software stack that they are using is specific to
that machine tool. They don't need the last shiny Python interpreter or
a 64 bit Guile library.

What bothers me is that all this energy is wasted instead of spending it
on other venues, such as Solaris 11 native support, documentation, web
site, &c.

-- 
Peter


More information about the maintainers mailing list