Hi Keith,
I have solid word from Red Hat that each minor rev to their major
releases are 100% binary compatible, and yes, they lock the
version numbers for the entire release. If you look at the RH
version numbers, you'll see something like this:
5.3.3-27.el6_5
Everything after the dash is Red Hat's patch. So even after they
backport a fix, the version (5.3.3) remains the same, but the
patch number increases. So in this case, this is the 27th Red Hat
patch to PHP 5.3.3.
I had fun with that when this high-falutin' Washington DC Beltway
Bandit risk assessment team came rolling in to do an assessment.
They grabbed the SSL banner (0.9.8 something) off some web servers
and called an OMG emergency meeting with the system administrators
and management about why we're running outdated versions of Apache
and SSL. After they presented their "findings" they all looked at
me, and I said flatly "We don't use Apache here. We use IHS."
(IBM HTTP Server - based on Apache, but with IBM secret sauce.)
You could have heard a pin drop as they huddle and whisper and
look silly. Yeah, that was fun. They hate me. They should have
done their research and asked a couple questions first. Oh well.
Then I had to research the SSL thing and show the Red Hat Errata
demonstrating the "old" version of SSL was patched against known
vulnerabilities.
As far as Centos and RHEL, I don't know why you assume CentOS
would be a year or two later than RHEL. This article indicates
CentOS will be tightly coupled and more fluid than RHEL:
http://www.zdnet.com/red-hat-reveals-centos-plans-7000027812/
"
However, there's a firewall between
RHEL and CentOS developers. The net effect is that CentOS will
continue to lag a bit behind RHEL in releases. Even so, CentOS
releases will be coming out on RHEL heels rather than weeks or
months behind."
I'm amused that you are trying to plan 6-10 years out in the IT
field.
</sarcasm>
Regards,
George Toft
On 5/6/2014 10:23 AM, keith smith wrote: