Linux update server?

[Red Squirrel]

Like windows WSUS, is there a way to setup a Linux update server?

One where if I go yum update, it would actually pull the results from that server?

How would I go about setting that up? I know the basics, just not sure how I'd go about

1: setting up proper folder structure/index files etc

2: how to update the update server

3: which domain yum update goes to so I can do host file redirect to my own server


Also would it be possible to make it so this server can update for multiple distros? Like FC*, CentOS, etc.

 

[jaqie]

My expertise lies more with fBSD, but the process is probably very similar for most linucies. Please be aware I am speaking of FreeBSD here.

Basically, you have to set up CVSup to an automatic pulldown mode (push mode is optional, where the central servers push updates to your box instead of it pulling them down), set up NFS, and then set all the other machines to CVSup from that one in their supfile. Optionally, you could also set that machine to automatically compile all the ports for the other machines as well.

 

[Brazen]

You just copy down the rpms to to a folder then run "createrepo ." in that folder (with an account that has write access in the folder. Then share the folder via a webserver (ie. Apache) and there's your update server. Then you just have to edit the repo files under /etc/yum.repos.d/ (or something like that, to tired to check) to get rid of the remote repos and add in your local repo.

Just make a cron or anacron script to download the rpms and run createrepo to keep your repo packages updated.

 

[drag]

If your using deb package system (Debian or Ubuntu) you can install a apt-proxy. There are two versions that I've used, one is called apt-proxy and the other is approx. They act as a sort of web proxy. Target your clients to use those as the repository and they will cache packages as you update your clients saving WAN bandwidth and accelerating later updates. Very effective and has very little in the way of administrative overhead.

With RPM systems I don't know. I have much less experience with those. A quick google found this though:
http://blog.softdux.com/everyt...-on-sme-server-70.html

It describes how to use a cach'ng http proxy to capture and save files as you download them. And how to use that to cache files for yum.