Consolidating glFusion Sites

I've been using some form of content manager for years now - it's really the only way to easily allow the owner of a site to make daily or periodic changes to their site without incurring huge webmaster fees - and changing contents on a site is the ONLY long-term proven way to keep your site's position on the search engines!

Most of the sites I do personally today (I oversee lots of systems and sites but these are the ones I personally work on and put together) are done with either Geeklog, or the "fork" of this open source software calle glFusion

Over the years I've created several such sites for my own writings and musings - and today I decided it was about time I simply consolidated them under the one domain, Digital-Rag.com - which I recently acquired after many years of it's being owned by a paper-making company. You see I'd used the "Digital Rag" as the name of my first webzine while we had our ISP, Wimsey.com and had had to settle for it as a sub-domain or just the masthead of the various incarnations it had gone through over the years. Well, today they're all being consolidated and I want to share with you how I'm doing it.

I have 2 sites:

• blog.pacdat.net
• digital-rag.com

Each of these has mostly unique content - some duplicate but not much.

Both have been updated to the latest glfusion 1.1.5 version

They are hosted on the same virtual Linux machine, completely controlled by me including mysql and shell access. If you're trying to do this without either of these you may have to jump through some extra hoops.

One major item of note is that I'll be using a "RedirectPermanent" line in the web server configuration file for the domains I'm consolidating to the one Digital-Rag.COM domain - so searches and such that try for a page in the old domain will end up going to the page in the new one - so the page should be there or the viewer will get a 404 Page not found error and have to look again. In fact, this is very straight forward with glFusion since the link to the story (or static page) is via the ID field, so if the ID field for a story is identical in the new database, the story should just come up. Duplicate IDs are not allowed - so if I try to copy a story from one of the old domains to the new one I'll get an error and the story won't be copied.

I've decided to move all the stories to the blog.pacdat.net site and then re-name it to Digital-Rag.com since it has the most stories (over 250) in it compared to the other two sites (a couple of dozen each). I'm going to do it manually since this will give me some time to do minor corrections and check for duplicates, etc. If I were to do it via SQL I'd have to ensure that the topic ID for each story was either modifiec to fit the different list in the destination site, or to be unique (and add the relevant topic to the new site too). This is what I'll be doing for a customer site that we're doing a similar job on - they have several hundred complex stories to move to the new (old) site and integrate with the 5000 or so there. I'll post details on how we do this too.

So I have two browser windows open - one on each of the old and new sites (old is copy from, new is copy to)

I've opened up each of the 20 stories I'm copying in a separate tab in the "edit" mode. As I do each one I'll close that tab so I know it is done. I have two monitors side-by-side (actually, I have 5 but that's a different story) so I can see both edit sessions at once. You can do this on a single screen but it is more awkward and a bit slower.

Open the "Create New" screen - copy the Title and ID (blank out the pre-generated ID in the new story screen - this is the key to getting things working correctly when searches come to the old site and get the new one instead!)

Click on "Show All" tab on both new and old - you'll want to modify the date on the new one to match the old and this tab will likely have scrolled off the screen by the time you need it.

Pick the topic to match the old one. This can be a concern if someone comes to the site to a "search" screen as the topic ID will be different unless you make new topics with exactly the same topic ID. Generally this is not an issue since most searches will come directly to the page via its own ID.

Copy the Intro (CTRL-A CTRL-C) and place your mouse in the INTRO on the new screen and paste (CTRL-V)

Do the same for the Body Text. Note that you normally will NOT have to do this as HTML (i.e. switch the edit mode to HTML from WYSIWYG mode) - the HTML should just "come along" with the copy. If you do have to copy in HTML mode make sure the destination is also in HTML mode or you'll simply see the HTML when you want to really see the resulting bold and formatted text. Fixing this later is actually as simple as copying the text version of HTML, switching to HTML mode, deleting what is there and pasting the "real" HTML in instead.

Don't forget to turn off the "Draft" flag if it is set on by default (mine is)

Check the permissions and other settings and click "Save" - done!

Do this for all the other articles in the old site.

Now for the interesting parts - Search Engine Optimization (SEO) and how to consolidate two sites so the search engines don't get upset.

The first thing to understand is that you can't just have two sites with the same content - nor can you just alias the old site's name to the new one - the search engines will consider them identical and start discounting both sites' ranking.

The way to have a site properly "aliased" is to use the "RedirectPermanent" construct in the web server (in this case Apache)

In my case the line for the changeover went into the config include file for the "old" domain of blog.pacdat.net as follows:

RedirectPermanent / http://digital-rag.com/

This does two things - it immediately redirects all requests to the new site exactly as if the URL had been typed with the name of the new site - and it tells the requestor this is a permanent change (301 return code) so the web crawlers know to update their tables and drop the reference to the old site and replace it with the new one.

The key here is that since the URL that would retrieve the transferred pages refers to them with their story ID - the page should display correctly with no other work needed - and it does. Google Analytics is showing all the traffic going to the Digital-Rag.com site now - and the instances of 404 Page Not Found errors has not gone up. Now I'll take the old site out of analytics and the webmaster tools sites and other places. I'll also change the DNS entries so the old domain name is a CNAME for the new one - so if there are any changes in the future it will simply follow along. And of course there are notes in the various config files telling me what I did and why - leave breadcrumbs because memory is not always 100% :)

See Consolidating glFusion Sites Part II for how to do this with a site that has a lot of content - in this case thousands of articles.

Tag: glfusion linux redirectpermanent cname geeklog content management