Tag Archives: repositories

Awstats 7.2 + extras RPMs

I’ve been a long term user of Awstats for reporting on visitor traffic to my websites. Whilst it’s a little dated, it’s simplicity and reliance only on the web server logs makes it ideal for any application, including general websites such as blog, but also more specialised sites such as my package repositories which can’t make use of more sophisticated client-side Javascript tracking methods as files are being downloaded by non-browser clients.

Simple web 1.0 goodness. No fancy AJAX graphs here son!

Simple web 1.0 goodness. No fancy AJAX graphs here son!

That repository server in particular (repos.jethrocarr.com) is now pushing 20-40GB of traffic per month to around 2500-3000 servers. Unfortunately Awstats doesn’t differentiate between general purpose file grabbers and the Yum downloader for RPM-based distributions, and it makes it difficult to see if downloads are from machines vs mirror scripts scanning and re-downloading files.

I also run dual-stack IPv4 and IPv6 – Awstats includes some useful GeoIP modules to lookup where user traffic comes from, but it doesn’t support mixed IPv4 and IPv6 by default and as my IPv6 traffic usage increases, this could be a problem as the “Unknown” country counter increases.

To fix this, I’ve written a patch for adding Yum user agent support and also merged in a patch by Sven Strickroth which adds a geoip6 module that does both IPv4 and IPv6 country lookups using the popular MaxMind GeoLite databases.

I’ve built packages for CentOS/RHEL/etc 5 & 6, which are available at my repositories at repos.jethrocarr.com. The awstats package I’ve built includes these two patches and also pulls in a current copy of MaxMind’s GeoIP database and required dependencies, so you’re all good to go immediately.

If you’re after the patches themselves, you can download them directly:

Updated Repositories

I’ve gone and updated my GNU/Linux repositories with a new home page – some of you may have been using this under my previous Amberdms branding, but it’s more appropriate that it be done under my own name these days and have it’s own special subdomain.

I want to unify the branding of a bit more of the stuff I have out there on the internet and also make sure I’m exposing it in a way that makes it easy for people to find and use, so I’m going through a process of improving site templates, linking between places and improving documentation/wording with the perspective of viewing as an outside user.

CSS3 shinyness! And it even mostly works in IE.

Been playing with new HTML5/CSS3 functionality for this site, have to say, it’s pretty awesome.

You can check out the new page at repos.jethrocarr.com, I’ve tried to make it as easy as possible to add my repositories to your servers -I’ll be refining this a little more in coming weeks, such as adding a decent package search function to the site to make it easier to grab some of the goodies hidden away in distribution directories.

I’m currently providing packages for RHEL & clones, Debian and Ubuntu. Whilst my RHEL repos are quite sizable now, the Debian & Ubuntu repositories are much sparser, so I’m going to make an effort to bring them to a level where they at least have all my public software (see projects.jethrocarr.com) available as well tested packages for current Debian Stable and Ubuntu LTS releases.

There’s some older stuff archived on the server if you go hunting as well, such as Fedora and ancient RHEL version packages, but I’m keeping them in the background for archival purposes only.

And yes, all packages are signed with my Amberdms/Jethro Carr GPG signing key. You should never be using any repositories without GPG signed packages, since they’re ideal attack vectors to use to install malicious content with a man-in-the-middle attack otherwise.