Louwrentius

I was looking at open-source load balancing software and it seems that there isn't a nice overview except from this website, although many of the listed projects seem dead.

I've made a selection of products that seem to be relevant. The biggest problem with open-source software is that projects are abandoned or unmaintained. So I created this table and added a column 'last product update' which gives you a feel for how active the project is.

I currently don't have hands-on experience with these products. Some of those products are briefly discussed at this blog - worth a visit.

There are many more products but most seem to be abandoned years ago. If you feel there are more products that are noteworthy but not in this list, feel free to contact me or comment about it.

It seems that the top-3 web servers like ngnix, Apache and Lighttpd all have support for load balancing. It depends on your needs, time and knowledge if you want to invest in other products or stick with the web server software you know.

At this location some people are talking about the pro's and con's of commercial off-the-shelve products vs. home-grown open-source solutions.

This blog is a static website, which makes it fast, simple and secure. It was generated by Blogofile but I switched to Pelican.

Blogofile has seen almost no updates over the years and I consider the project dead. Realising that blogofile is dead, I decided to look around for different open source static blog generators.

There are many other static blog generators than pelican. But Pelican is well-documented, is based on Python, is very actively maintained (good track record) and supports all features that I wanted. Some of those features are Atom/RSS, Disqus and Google analytics support.

My blog posts are written using Markdown. This makes it very easy to migrate away from Blogofile to Pelican, as Pelican also supports Markdown. Blogofile uses a different header format not recognised by Pelican, so you have to search and replace some key words in all your files before Pelican can actually generate your new website.

I wrote this horrible bash shell for-loop to process all my blog posts:

    :::bash
    for x in *.md
    do
        TITLE=`grep -i "title:" "$x"`
        TITLEFIXED=`echo $TITLE | sed s/\"//g`
        DATE=`grep -i "date:" "$x"`
        DATEFIXED=`echo $DATE | sed "s/\//-/g" | cut -d ":" -f 1,2,3`
        CATEGORY=`grep -i "categories:" "$x"`
        CATEGORYFIXED=`echo $CATEGORY | sed s/\"//g | sed s/categories/category/g | cut -d "," -f 1 | /usr/local/bin/sed -e "s/\b\(.\)/\u\1/g"`
        echo "$TITLEFIXED" > tmp.txt
        echo "$CATEGORYFIXED" >> tmp.txt
        echo "$DATEFIXED" >> tmp.txt
        grep -v "title:" "$x" | grep -v -e '---' | grep -v -i "date:" | grep -v -i "categories:" >> tmp.txt
        mv tmp.txt "$x"
    done

Notice how Build-in syntax highlighting of Pelican applies nice colors to this horrible code. Regarding this horrible code: I had to use GNU sed as the Mac OS X sed did not support the regular expression I used.

To enable comments for my blog posts I always used Disqus with Blogofile. Pelican generates web pages in a different way compared to blogofile, so all old pages need to be redirected to the new location. I used the redirect functionality of Lighttpd to redirect all existing pages to the new location.

The cool thing is that Disqus has a tool called "Redirect Crawler". If you have configured 301 "permanent redirects" for all pages and run this tool, Disqus will automatically update all existing links to the new locations, so your comments are migrated to the new web pages.

Furthermore, I've implemented a Pelican plugin called titlecase which capitalizes the first letter of words in the title of your article. It's just that I think it looks better.

I think I'm really happy with Pelican.