Catching up with DataStax & Cassandra 2.0 – Initial DSE Installation blog posting was created by Colin Clark. To view more postings by Colin, check out his blog Concurrent Me.
BACK TO THE FUTURE
As one of the early adopters of Cassandra, most of my work was in the 1.x series of releases with dev, qa, ua, performance and production clusters in 1.2. It’s been a while since taking the summer to produce a Bikes-n-Blues concert to assist rural development so I decided to get up to speed with all of the new, wholesome goodness promised by Datastax’s release of Cassandra, OpsCenter, and DevCenter.
I maintain a fairly robust environment in the home lab – from time to time, I’m called on to process relatively large amounts of data in connection with pre and post audit of HFT trading operations. That means someone sends me a bunch-o-data and I need to analyze it to see if there are potential abuses occurring. I have an edge up in this area as I get to use the same Darkstar system previously used by NYSE Technologies (pre ICE merger) as I retain the IP to that system. If it’s in there, I’ll find it.
For this exercise, I thought I’d use a small cluster of 3 nodes running Ubuntu 12.10 with 4 cores and 8 gig. These are small machines I use for development clusters. This will be the Cassandra cluster.
For my OpsCenter, DevCenter, and development, I’ll be using a MacBook Pro. I feel equally at home with Ubuntu, MacOSX, and Windows 7 but I’m not a big fan of Windows 8. I hadn’t used a Mac for this type of operation before and thought I’d give it a try to maximize the number of moving parts to increase my frustration level.
OFF TO THE RACES
First, select some great music to listen to. I chose David Bowie’s latest album, “The Next Day.” You can give that a listen here.
After verifying the latest and greatest components of 12.10 were installed and running correctly on the nodes, I also verified the Java version. I did this by typing this at the command line:
This was the output on each node:
java version “1.7.0_51″
Java(TM) SE Runtime Environment (build 1.7.0_51-b13)
Java HotSpot(TM) 64-Bit Server VM (build 24.51-b03, mixed mode)
Also, make sure you have curl installed:
sudo apt-get install curl
Next, register to download the DataStax Enterprise edition by going here – you’ll need the username and password for the next step.
Let’s add the DataStax repository to each node – make sure to use your username and password obtained above.
echo “deb http://username:firstname.lastname@example.org/enterprise stable main” | \
sudo tee -a /etc/apt/sources.list.d/datastax.sources.list
And let’s trust the DataStax repository by:
curl -L https://debian.datastax.com/debian/repo_key | sudo apt-key add -
and install the DataStax Enterprise packages – I’m only going to install DataStax Enterprise on the nodes. I’m going to put OpsCenter and DevCenter somewhere else. So, do this:
sudo apt-get update
sudo apt-get install dse-full
One mistake I made was to install repository with username:password, and then reinstalling without removing the bad entry. If you make the same mistake I did, just remove the wrong entry and try again. The file is edited by:
sudo vi /etc/apt/sources.list.d/datastax.sources.list
Once I had the right information there, updating the sources and installed DSE ran just fine.
You can run this on one node, or as many as you might have lying around. After getting everything installed on one node, I installed everything on the other two nodes in my cluster. In the next post, I’ll configure the nodes and get the cluster up and running.
THANKS FOR READING!