I say it is a move in the right direction. I'm sure it's going to step on
some bodies toes at some point, but if the community as a whole moves in
this direction and can settle on CentOS as the base OS of choice it will
only make us all stronger since we would all be on the same OS whether it's
32 or 64 bit.

Not to mention the time saved with coding for so many different OS's that
you coders support right now.

I propose we move in that direction and make CentOS the OS of choice as soon
as possible.

Just my 2 cents.

Joel

-----Original Message-----
From: Eric Shubert [mailto:[email protected]] 
Sent: Monday, February 13, 2012 11:44 AM
To: [email protected]
Subject: [qmailtoaster] Future Distros - RHEL/CentOS ONLY

I've done a good deal of thinking about this, and think that it'd be 
best to run it by the community at large (not just the developers) for 
everyone's consideration. This is not really new, and is not much 
different than what Jake had committed to some time ago. I just want to 
be sure that everyone is on board with this, and explain a few things.

Due to various changes in the IT landscape over the past several years, 
I think it's best that future QMT development be limited to the 
RHEL/CentOS platform. There are several factors involved.

First is that we'll be changing the method of distribution from source 
rpms to binary rpms, using yum to install packages (qtp-newmodel will be 
modified accordingly). We can do this because the qmail (et al) 
licensing was changed to public domain a couple years ago, so there is 
no restriction to distribute source-only any more. We also have mirrors 
in place that eliminate the need to have a single distribution point 
with high bandwidth capability. Using binary rpms for distribution not 
only simplifies installs and upgrades, but it also substantially reduces 
the disk space required, in addition to making QMT more secure due to 
the absence of a compiler and build tools. All in all, this is a win-win 
change.

Secondly, the industry in general is moving toward virtual hosts, and 
QMT is making this move as well (many of us already run QMT as one or 
more VM guests). One of the advantages of virtualization is that 
multiple machines can coexist on the same host hardware, concurrently 
running entirely different operating systems and versions of languages 
and software. There's little need any more for QMT to coexist on the 
same machine with other applications or services. In fact, things are 
moving in a direction such that QMT itself will become divided into 
logical roles that will be able to implemented on separate hosts, 
allowing for more flexible and scalable QMT configurations. Stay tuned 
for that development, which is a ways off yet.

So let's take a look briefly at the prominent distros that QMT will be 
discontinuing.

Mandriva is on the ropes, struggling to survive. If you presently have a 
QMT running on Mandy, I would seriously consider a migration in the near 
future.

SUSE does not use yum, it has yast instead. When I looked at yast some 
time ago it had no CLI, which was a big drawback to me. While I expect 
that yum could be installed and used, it goes against the "When in Rome" 
philosophy. The source rpms will of course continue to be available, so 
if someone cares to adapt them for SUSE, they may do so.

While Fedora contains a great deal of what's in store for future 
RHEL/CentOS releases, it's not well suited as a QMT platform, simply 
because it changes too often (a new release twice a year), and most 
often none of the changes provide any benefit to QMT. If there happens 
to be something that would benefit QMT, it would most likely be 
available for RHEL/CentOS in the EPEL repo. So there is really no sense 
in packaging QMT for Fedora.

I think this covers the distros worth mentioning. If I missed one, 
please let me know.

In summary, going forward QMT will be available only on RHEL/CentOS 
platforms, for both x86 and x86_64 architectures. This will simplify 
spec files, documentation and installation/utility scripts 
substantially. For all other distros, the existing build options in the 
spec files will no longer be included. They will however be archived in 
a source code repository before being removed, so that they'll be 
available should anyone want to reference them at some point in the future.

If you have a problem with or question about any of this, or you'd 
simply like to comment about something, please don't hesitate to reply.

Thanks to everyone for their continued support and participation.

-- 
-Eric 'shubes'


----------------------------------------------------------------------------
-----
Qmailtoaster is sponsored by Vickers Consulting Group
(www.vickersconsulting.com)
    Vickers Consulting Group offers Qmailtoaster support and installations.
      If you need professional help with your setup, contact them today!
----------------------------------------------------------------------------
-----
     Please visit qmailtoaster.com for the latest news, updates, and
packages.
     
      To unsubscribe, e-mail: [email protected]
     For additional commands, e-mail:
[email protected]



-----
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2012.0.1913 / Virus Database: 2112/4807 - Release Date: 02/13/12


---------------------------------------------------------------------------------
Qmailtoaster is sponsored by Vickers Consulting Group 
(www.vickersconsulting.com)
    Vickers Consulting Group offers Qmailtoaster support and installations.
      If you need professional help with your setup, contact them today!
---------------------------------------------------------------------------------
     Please visit qmailtoaster.com for the latest news, updates, and packages.
     
      To unsubscribe, e-mail: [email protected]
     For additional commands, e-mail: [email protected]


Reply via email to