Driver Database Updates

The updates and driver blogs have now been running on the new server for a few days. During the night the server refused to co-operate and I had to reboot it. Looking at the logs it appeared to have ran out of memory. One of the upsides of being on a cloud server is that there’s no need to physically add more RAM to the server so after a few clicks of the mouse I added another 512 MB RAM to the system.

Will monitor it closely to see if that’s enough. I think it ought to be, unless there’s a memory leak somewhere.

On another note I guess you may have seen that we suffered some data loss last Thursday. While trying to optimize a slow database table it in some weird way imploded and delete two thirds of the data. Not what I have intended, eventhough it did speed up things.

The latest backup (which was four days old) was in place a few hours later and I spent all of Friday reconstructing lost connections in the database. We lost a few championship points and positions as well as team info that was added in the last couple of days but fortunately there was no major harm done.

Next up is to move the news.driverdb.com blog to the new server – hopefully without having to upgrade RAM again.

All the driver blogs are now located on the new server and are therefor way faster than on the old server. There should also be a bit less load for the main site that is still located on the old server (that move is a bit more complicated).

In the coming weeks I’ll open up the driver blogs for new registrations again – something which I removed over a year ago due to the extreme amounts of spam bloggers signing up (I had over 2000 spam blogs signed up in less than a week). Got a new anti-splog plugin that hopefully will take care of things.

Now work will continue on making the fastest Driver Database ever seen (shouldn’t be too hard, right?) on the new server. There will be some layout changes as well…

If you can read this message it means that the DNS changes have come into effect and updates.driverdb.com is actually located on the new cloud server. The plan next is to move the driver blogs over to the new server and in step three move the news blog as well. There are two reasons I’m moving the blogs first:

1. It’s much easier and quicker to migrate them.
2. Having the blogs on the new server will take some load off the main site (which is still struggling on the old hardware) and therefor hopefully make the main site cope a bit better until I’m ready for the big move.

Since WordPress blogs – and especially WordPress MU – can be quite resource heavy I decided to actually use two cloud servers instead of the original plans to use just one. The server that this blog now resides on will be dedicated to the three blogs (updates, news and driver blogs) while a second server will deal with the main site.

It’s been a good while since I announced that I was looking to move Driver Database to a cloud server. Lots of things got in my way since then though and the site has been stuck on it’s old server, going slower and slower as the visitor numbers keep getting bigger.

Last month I decided it was time to start the work on the move. I haven’t really been into Linux since the late nineties so getting from a clean install to something in the regions of the current setup was trial and error hell but now everything on the server side is set up and migration of the code has begun.

There a lots of pages on Driver Database that load slowly, not just because of the current server that is not quite coping but also because of lots of database queries that when the database is as huge as it is and when there is much load on the server takes too long to process. As most of these queries are being repeated lots of time the answer is to cache them so they don’t have to go through the database server unless changes have been made. Caching is already in place on some pages (the driver overview page, for example) but there are lots of places where it’s needed.

Therefor I’ve decided not to just move the current code to the new server but to go through every part of the code to try and optimize it as much as possible. I’ll also make a few much needed changes to a couple of things. More on that later…

My son Albin, who’s now one year and four months, has had a feature request. Well, he hasn’t really been aware of the request, but the solution saves him from getting fed up.

Albin likes the Driver Datbase photo gallery as he’s really into looking at race cars (and making sound effects to them). The problem is he thinks it takes too long with loading the gallery, selectiong a photo, then going back to the gallery to select another photo. Patience is not something he’s been blessed with.

So, from now on, there’s a little icon of a die down to the right of every racing photo. One click on that icon takes you to another photo, selected at random.

Perhaps not the most important feature ever added to the website, but it sure is in the top five.

Driver Database has grown considerably this past year, which is great. What’s not so great is that it also basically means the Driver Database server needs considerable updates too. I’ve been contemplating on the way forward and I almost came to a descicion on the path ahead. The idea was to keep the current server (which now does everything) and rent a second and faster one that would only serve as the MySQL server (which is the biggest bottleneck when it comes to speed on the website). This would have the advantage that I only need to migrate the database and not the entire website. I would also stay at the current provider, JTLnet, which I only have good things to say about.

However I’m currently reconsidering this and am thinking about moving into a cloud based hosting setup from City Cloud. It would mean migrating the complete website to a whole new continent, but the benefits will probably be greater. The major benefit of cloud servers is that I would be able to upgrade the servers on the fly, adding memory, CPU cores, disk space etc when it’s needed. If the traffic, for example, is huge one day I can add an extra core, only to remove it the day after when the traffic is back to normal.

I’m going to set up a new server and evaluate it for a while, and then make a decision.

David Fleming at Grand Prix Shootout has made an incredible exclusive offer to members of Driver Database. If you’re a young racer with potential, then this is the chance for you. If you sign up before midnight on April 25 you’ll get a full entry for just £1,950 (that’s saving £2,000!!) AND you might be the lucky entrant who gets chosen at random to get the place for FREE!

Read more about it here!

The long awaited premier profiles are now ready for a few days of testing. Everything is now implemented, I just need to test it out a bit before I’ll let drivers have the option of upgrading their regular boring profile into a premier profile.

With a premier profile the driver (or manager/racing mom or dad) can add their biography, presentations of their sponsors, YouTube clips,  latest tweets, links to Facebook fan-page, Q&A’s, external blog feeds, favourite links, published press releases and perhaps most importantly get top priority when it comes to updates of data and photos.

My goal with the top priority is to be able to add press releases the same business day that they are received (if they are received before 17:00 CET) and all photos and data (if they are sent in correctly) within two business days. This is just a goal however, not a guarantee, as most of the functions on Driver Database relies on me alone and I can’t guarantee that I won’t be home sick or out of the office some days of the year.

The price for a premier profile will be €40 for six months or €65 for one year (may be subject to change).

Five drivers will have the chance though to try it out right now for free though, before the proper launch (which will be a few days or so from now).

The first five drivers to comment on this blog entry with something in the likes of “Yeah! Let me try this thing out!” will get their Driver Database profile upgraded to a Premier Profile for 6 months for FREE (and there are off course no strings attached once those six months run out)!

(The drivers on Driver Database who have a nice and shiny gold medal on their profile page has already had their profiles upgraded (for one year) as a thanks for their support .)

Today is a big day. It’s Driver Database’s birthday. The website has now turned six years old!

It all started in early 2004 when I was doing a little research during the F1 silly season. I found it hard to get all the information I wanted from one (or in some cases even from several sites). Formula 1 results wasn’t hard to find, but earlier career statistics was a bit more difficult to find and the driver’s own web pages obviously forgets about those years where there were no good results.

I tried to get some tips about good web pages over at the Swedish Formula 1 forum F1BB.com (now F1BB.se), but there really wasn’t anything that came close to what I was looking for. I then asked “if there’s no site like this, why not create one?” and got very pessimistic answers about the amount of work involved, can’t be done etc… Those responses obviously made me more motivated to start the website. I registered the domain name driverdb.com on the 20th of February 2004, as a little tribute to hockeydb.com which sort of was the main inspiration at that point.

In retrospect I should have listened to the pessimists – it really was way too much work.

The first decent working version of Driver Database wasn’t really up and running until 2005 or something, but work began right away.

Here’s the first image of what Driver Database looked like in early 2004. It was just an ugly mock-up basically, complete with phony driver names.

We’ve come a long way.

Driver Database was never really meant to be my number one project. At this time I had just started developing a massively multiplayer racing management game. The backbone of Driver Database was then going to be used for driver statistics in the game and non-player characters would be randomly created from real world data. The management game was quickly put on ice though, a very basic version still exists on my hard drive somewhere.