On my dev server I am running Percona (5.5.27), and have just spun up another server to act as a staging area for public access to a new project. I am trying to control package versions by using yum plugin 'versioncontrol', but for some packages this approach doesn't seem to be working and I think the reason is; the upstream source for the specified version is no longer available (current version is 5.5.32). For packages like httpd this really isn't an issue and versionlock works fine, as the CentOS upstream doesn't update too often and the version I installed on the dev server (over a year ago) is still available.

But what do I do to ensure the same Percona version (5.5.27) is available to all my existing and future servers? Do I build an rpm, or is there another approach?

If I am to create an RPM, is it possible to do it for a specific package, or series of packages? Can I create an rpm with the following packages, resulting in a single rpm?


A lot of questions, but as a lesser skilled sys admin I have put off building rpms for as long as I could.


If you are already using yum to install your packages then the RPMs already exist and you could just grab the RPMs you want and store them locally for later use... or you could get a bit fancier and actually set up a local repository with your desired packages (or even the entire distribution) and use that for your servers (though if you go this route you need to be careful about getting security updates and the like if these are things that need a reasonable security status).

  • Looks like I can use the yum-plugin-downloadonly to grab the raw rpms. I'll search around for a way to get the client servers to pull these rpms instead of upstream. Thanks for the advice. – Mike Purcell Jul 16 '13 at 5:45
  • You can use that and also the yumdownloader tool as well as simply pulling up the repository URL in your browser and downloading the files manually (the URL can be found in the yum repo configuration files). – Etan Reisner Jul 16 '13 at 5:53
  • What if I wanted to take it a step further, is it good practice to download ALL rpms of new build; ` yum reinstall $(cat ~/software.log) --downloadonly --downloaddir=.`, then moving these files into rpm accessible folder, then scp the software.log to the newest server, use the custom repo and build from it? – Mike Purcell Jul 17 '13 at 4:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.