On Fri, Jun 1, 2012 at 3:41 PM, James Wilkinson <fedora(a)aprilcottage.co.uk&gt;wrote: ... snip ... ... snip ... Just because there is no internet access, doesn't mean there isn't network access, and sometimes you need need to update to newer stuff when you change/update your own applications. There's lot of reasons.

If you set keepcache=1 in /etc/yum.conf, yum will download the RPMs to /var/cache/yum/$basearch/$releasever/$reponame/packages, install them, but not delete them: you can then transport these files to the target site. Might I ask, though: if the target computer has no Internet connectivity, why do you want to update it? Are you after bugfixes or security, and if the latter, is it possible to restrict access instead? Hope this helps, James. -- E-mail: james@ | “Drums must never stop. Very bad if drums stop.” aprilcottage.co.uk | “Why? What will happen if the drums ever stop?” | “Bass solo.”

Thanks Sam - I will give that method a spin.

leaving aside the reading necessary to come up do speed with mrepo, createrepo and such (wich I would recommend anyway if someone deals with the situation of maintaining a setup alike...), the setup takes maybe 10 minutes more. Daily business of installing updates is cp/rsync * followed by "createrepo" vs. "mrepo -ug" - no difference. If in _any_ case the number of systems raises to two, you will be hopping mad :) regards Jens Neu

Jens, RIchard Thanks for the info - much appreciated. Is there any way to download all the required updates (on a machine that has Internet access, obviously) and save these downloads for the creation of the local repository you speak of? Let me clarify. The target system (with no Internet access) is in a customer site. The development system, which essentially 'mirrors' the target installation, is in our own lab and *does* have Internet access. So for us to update our (development) system is straightforward, and this system automatically 'knows' which updates to download as part of FC16's package update software. Since this system mirrors the target one, what would be ideal is for it to not just automatically install the updates, but merely download them as a first step - such that they can then be copied to a local repository that I will create. Then I can use this repository to update the development system as well as the target system. In other words, if the development system can be used to automatically download but *not* install the updates, it saves me the trouble of having to 'manually' determine which updates are required for the systems.

On Thu, May 31, 2012 at 10:24 AM, Peter Butler <PButler(a)pt.com&gt; wrote: It will be up to you to figure out what specifically needs updating, but yes. I do this, not for lack of internet access, but because I do a lot of packaging on my own, both test packages and a few that are not available in Fedora or RPM Fusion. I setup directory structure like this: /var/local/packages/local/16/{x86_64,i686,noarch} I also create a: /var/local/packages/install/16 and dump all the packages from the DVD on that so if I want to install something that's unchanged it pulls it from there instead of downloading it. After you put all the packages in the directories just run createrepo /path/to/root/of/repo i.e.: createrepo /var/local/packages/local/16 Then when you add packages run: createrepo --update /var/local/packages/local/16 which only looks for changes so it's faster than creating the repo from scratch. Then just create a local.repo file pointing to that directory. Since you don't have internet access it won't matter, you'll have to set all the other ones "enabled=0", but if that wasn't the case you can set "cost=500" Default is 1000 I think so that way the local repo is preferred over the online repo in the case of duplicate packages. Richard