It’s amazingly easy to run an Exasol Cluster on Amazon Web Services (AWS).
Subscribe Exasol in AWS marketplace
After having registered and having logged in to your AWS account, go to the AWS marketplace and search for Exasol:
Click on the Exasol Single Node and Cluster BYOL link and then on Continue to Subscribe:
After having reviewed the T&C, click on Accept Terms. It shows this message afterwards:
Create Key Pair
Now login to the AWS Management Console, select a region close to your location and open the EC2 Dashboard. Click on Key Pairs:
Click on Create Key Pair now and enter a name for the new Key Pair, then click on Create:
Now you are ready to use the Exasol Cloud Deployment Wizard. Stay logged in with AWS Management Console as you will be routed back there by the Deployment Wizard soon.
Using the Cloud Deployment Wizard
Put this URL into your browser: https://cloudtools.exasol.com/ and click on AWS then:
Select a region close to your location and click on Continue:
Click on Advanced Configuration and specify
License Model Bring-your-own-license, System Type Enterprise Cluster, Instance Family Memory Optimized, Instance Type r5, Instance Model r5 large, Number of DB Node 1 then click Continue.
BYOL works without license file with a limit of 20 GB memory for the database. Means no costs are charged by Exasol (But by Amazon) for this environment.
Select create new VPC and click on Launch Stack on this page now:
This takes you to the Quick create stack page of CloudFormation in AWS Management Console:
Enter these details on the page:
Key Pair (select the key pair created previously)
SYS User Password
ADMIN User Password
Public IPs (true)
Tick the acknowledge box and click on Create stack
Now go to the EC2 details page and copy the Public IP of the management node:
Put that with an https:// prefix into a browser and click on Advanced:
Then you should see a progress bar like this:
That screen changes after about 30 Minutes to the EXAoperation login screen.
Login as user admin with the password, you specified previously on the CloudFormation Quick create stack page. There should be a database running:
As you can see now, you have a database, a remote archive volume using an Amazon S3 bucket ready for backup & restore and a log service to monitor your system.
This database is limited to 20 GB memory only unless a license file is uploaded to the license server aka management node. For educational purposes, I don’t need more.
Use Elastic IPs
The public IPs of your data nodes will change upon every restart, which is probably not convenient.
Therefore, click on Elastic IPs in the EC2 dashboard, then click on Allocate new address:
Select Amazon pool then click on Allocate:
Click on the IP on the following screen:
Select the action Associate address on the next screen:
Select the data node from the Select instance pull down menu and click on Associate:
Close the next screen and go to the EC2 instance page. You should see the elastic IP assigned to the data node there:
Connect with a SQL Client to your Exasol database on AWS
This is how that looks with DbVisualizer:
And that’s it: Now you have an Exasol 1+0 cluster running on AWS. That’s not the same as a single node system, because this 1+0 cluster can be enlarged with more data nodes. I will show how to do that in future posts.
A word about costs: Instead of using our corporate AWS account, I registered myself to see how much that will take. It was less than 80 Euro with a 2+1 cluster environment I used for about one month, shutting down the EC2 instances whenever I didn’t need them for testing and for creating courseware. It should be well below 10 Euro per day with the very moderate resource consumption configured for the environment subject to my postings.
Stay tuned for some more to come about Exasol on AWS 🙂