Software update 'master' grace period

Hi all,

I would like to have 1 master server that pules, stores and installs software updates automatically. After some grace period then makes the updates available to other NS servers within the reach of this master server.

Objective is to test updates before they are deployed to production servers.

Has this been discussed before, does it makes sense?



Hi, it is correct to check first and test update for production server. I can advise you use a Virtual Machine before for your tests

1 Like

Hi, Thanks for you reply. More specifically I am looking for the mast to be:

Central update repository for other NS servers within reach
Manually/Auto tests updates on the master
Flip the switch so the updates are available to the other NS servers.


It makes sense, but I’m not sure it’s been discussed before here. Implementing it, though, is likely to be tricky, as there’s a variety of repos involved, and their URLs are hardcoded into individual files in /etc/yum/repos.d. To make it work, it seems you’d need to either:

  • Spoof DNS on your network, such that any of the hostnames for the repos resolve to your local “master” instead; or
  • Hand-edit the repo files so that they point to your “master” server rather than to the original repos.

Setting up a local mirror seems to be pretty well-documented–see:

Handling the “grace period” in some automated way is probably something you’d need to sort out yourself, though.

1 Like

Interesting and simple approach, will look into it. Thanks!

Or really, a much simpler solution: buy a subscription for the production servers. That gives you access to the “stable” repos, which pretty much work as you describe.


Taking a closer look at some of the links I posted, it looks like there’s another option:

  • Create a separate repo file pointing to your mirror, and replace the existing repo files with that one

Likely better than either of the two I came up with off the top of my head.