Towards the end of 2008, we brought online a new datacenter to serve the over 5.5 million blogs now hosted on the WordPress.com platform.  Adding the data center in Chicago, IL gives us a total of 3 data centers across the US which serve live content at any given time.  We have decommissioned one of our facilities in the Dallas, TX area.  Our friends at Layered Technologies were kind enough to shoot this footage for us (think The Blair Witch Project) and the always awesome Michael Pick took care of the editing.  Here’s a peak at what a typical WordPress data center installation looks like…
For those interested in technical details here is a hardware overview of the installation:
150 HP DL165s dual quad-core AMD 2354 processors 2GB-4GB RAM
50 HP DL365s dual dual-core AMD 2218 processors 4GB-16GB RAM
5 HP DL185s dual quad-core AMD 2354 processors 4GB RAM
And here is a graph of what the current CPU usage looks like across about 700 CPU cores. Â As you can see there is plenty of idle CPU for those big spikes or in case one of the other 2 data centers fail and we have to route more traffic to this one.
Barry on WordPress 
That’s very cool Barry.
I’m in a similar position to you, but I can’t imagine managing and planning that amount of hardware.
Good work.
Great post Barry. Thanks for sharing the technical details; I am sure many people find this information useful.
Thanks for this perspective.
How many k-watts/rack? Servers/rack? AC power only?
nice…very nice
Boy, a guy can dream!
Wow!
Can i find more details, somewhere or are they secret?
For example, are you using a Load Balanced cluster?
I’m very interested in the WP infrastructure
GrG
How do you do to keep wordpress.com free ?
I see you’re using munin to gather your data. how do you sum up the values? in the munin config file or is there some automagic way? is a pain to add each new server to the sum paragraphs in the munin file – that’s how we do it atm.
We have a simple-ish script which reads the sever info from Servermattic and generates the munin configs automatically. This includes adding/removing hosts and producing the aggregate graph definitions.
thanks Barry will have a look
Hi Barry,
I really like the fact that you seem to heavily rely on open source software.
Whenever I’m hoping to establish a new service, I try to base them on open source software. However I’ve found it hard to find any good realtime file replication services for Linux.
Would you mind telling how you guys handle it? I assume you do file replication between the datacenters.
regards,
Erik
Thats crazy!
So much to keep track of..
Glad Im not responsible
Reblogged this on fajarramd.
we love wordpress it has made easier our lives