Is it kosher to manual change the version number in the header, for a particular service/package, after back porting it to patch the previous version, without necessarily uninstalling(or overwriting) the old version and installing the new version?

(I know its possible and probably best to just block the version altogether.)

TIA!

Let's return the question: what valid and compelling reasons do you (think you) have that makes you resist performing the upgrade?

An insane amout of customizations, that make reinstalling the package extremely time consuming and definately impractical every time there is a security update. When the code to fix the security issue, is usually just a few lines (as I don't need the other changes within the update), and doesn't affect the other customizations.

There's two sides of this: cost-benefit analysis and efficiency. As far as cost goes, if we're talking about software packages distributed by RHEL or Centos, then you know they take care of the whole release process: required programming manpower standing by, up to date package knowledge available, tracking vulnerabilities, figuring out and incorporating the fixes, QA testing and releasing the package. They do all of that. And timely. You could possibly make releasing your own version of a package, because that's basically what you're doing, easier, more efficient, if you could group and split out customizations (sub packages even?) and use versioning for storing differences from the standard. That's all I can add because, as usual, you excel at providing no detail at all.