Update/upgrade cluster through one node?
 

I'm pretty much a newbie at Linux. I've got four rather old PCs, each running Debian 3.1, all connected to a system running XP with freeproxy serving to allow access to the 'net. I will shortly reconfigure this cluster so that one Debian system will run Apache 2 as a web server, and - hopefully - provide IP Masquerading for the internal cluster.

My question is this: right now, I regularly use apt-get update and apt-get upgrade to keep the software current. It occurred to me today (slow learner, sorry) that each system is going out on the net and contacting the repositories, even though the first system upgraded pretty much has all the software the subsequent systems need. When I reconfigure the cluster, is there a way I can configure it so that the "Apache" system - the one providing the link between the external web and the internal network - serves as the repository for subsequent systems?

I tried searching, but I don't know enough to know what exactly I should be search for, so even advice along that line would be appreciated.

TIA.

Information about this you might find in the Debian Reference Manual, chapter 6 "Debian Package Management"

I think you should try to create something that is named a local repository. I tried it once and I failed, but that might be because I did something wrong. I would be interested in your experiences if you succeeded, since I have more or less the same situation as you with multiple machines running Debian on the LAN

jlinkels

Thanks for the pointer. I will let you know if I get it working, but I probably won't have time until Christmas to work on this.

I did find "How to build a local apt-get mirror", which describes what I think is a possible solution. And the apt-get info suggests that I should be able to installing an http or ftp daemon on the "main" node - and it will be running Apache2. However, I'm more interested in the possibility of using the file options in the sources.list files on the "other" nodes.

I'll let you know.

Something else you could give a shot is sharing out /var/cache/apt via NFS to the other machines in the same location. This way only one machine is actually holding onto the cached files. I have no idea whether or not this will work, however, so if it doesn't work, you may want to have the old directory lying around somewhere so that you can repair the damage this may cause.

This is more in line with what I'd like to do, if I'm able to figure it out. The docs say that the sources.list can include nfs: repositories as well as http: and ftp: repositories. I just haven't seen any examples of that actually being done.

Look at apt-proxy

You just have to change the /etc/apt/sources.list on the clients to point to the apt-proxy machine.

This sounds very good. Thanks a lot.

Oooh...nice! This looks like an automated version of what I was talking about, sort of.

I haven't tried it, but it looks very good... exactly the thing I was looking for.

jlinkels