Simple Web Service that saves messages to Datomic.
Note: the web app is currently a work-in-progress. Right now, it runs locally but fails to initialize under Tomcat when deployed via a war file. Hope to have it fixed up in a few days.
This is a demo web application that saves data sent to the /write endpoint to Datomic, and reads from Datomic for queries to the /read endpoint.
$ curl 'http://localhost:3001/write?id=1&msg=HelloWorld'
1006
$ curl 'http://localhost:3001/read?id=1'
HelloWorld
This project provides a reference implementation of deploying Datomic in AWS as a follow-up to my presentation at Clojure/West 2014: DevOps Done Right: Room Key's Datomic Deployment in AWS
The provided CloudFormation template (config/cfn-template.json) builds a stack,
peers and transactors, without using the tools included in the Datomic
distribution. The web application included is very basic, but the template
includes examples of many advanced CloudFormation features, including
Conditions, Auto Scaling Update Policy, and CloudFormation MetaData.
The template should provide a refernce for those who may want to run a custom Datomic deployment in Amazon Web Services.
- AWS Account and Access Keys
- Datomic Pro Starter license and credentials
- leiningen
- AWS CLI tools
- Java7 JDK Linux RPM downloaded from Oracle (jdk-7u67-linux-x64.rpm)
First setup environment variables in your shell profile. Tools in the scripts
directory will look for MY_DATOMIC_USERNAME and MY_DATOMIC_PASSWORD, as will
leiningen when it tries to download the Datomic peer library from Cognitect's
maven repo (see project.clj). These are the credentials assigned to your
Datomic account after registering for a license key.
AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY aren't explicitly used by this project,
but other AWS tools and libraries look for them, so they're useful to have set.
For example, add the following to your ~/.bashrc
export [email protected]
export MY_DATOMIC_PASSWORD=XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
export AWS_ACCESS_KEY_ID=AKI...
export AWS_SECRET_ACCESS_KEY=<SECRET>The tools in the scripts directory rely on the AWS CLI tools being installed and configured.
Installation requires Python, and can typically be done via pip:
pip install awscli
Configuration is just as simple. For example:
$ aws configure
AWS Access Key ID [None]: AKI...
AWS Secret Access Key [None]: ...
Default region name [None]: us-east-1
Default output format [None]: table
This tells the tools to use us-east-1 by default, and the preferred output type is table format.
You can use the describe-regions command to verify it has been successfully configured:
$ aws ec2 describe-regions
----------------------------------------------------------
| DescribeRegions |
+--------------------------------------------------------+
|| Regions ||
|+-----------------------------------+------------------+|
|| Endpoint | RegionName ||
|+-----------------------------------+------------------+|
|| ec2.eu-west-1.amazonaws.com | eu-west-1 ||
|| ec2.sa-east-1.amazonaws.com | sa-east-1 ||
|| ec2.us-east-1.amazonaws.com | us-east-1 ||
|| ec2.ap-northeast-1.amazonaws.com | ap-northeast-1 ||
|| ec2.us-west-2.amazonaws.com | us-west-2 ||
|| ec2.us-west-1.amazonaws.com | us-west-1 ||
|| ec2.ap-southeast-1.amazonaws.com | ap-southeast-1 ||
|| ec2.ap-southeast-2.amazonaws.com | ap-southeast-2 ||
|+-----------------------------------+------------------+|
Cognitect recommends that production Datomic transactors run using the JRE from Oracle. In fact, the official transactor AMI they build includes the Oracle JDK.
Go to Oracle.com and download the Linux 64-bit Java7 JDK RPM (jdk-7u67-linux-x64.rpm).
Keep this handy, the script that creates resources in AWS will upload it to S3.
After your register for Datomic Pro Starter, save the license key they send you to a file. Don't include license-key= when you save it - just save the key.
Before you can run the service locally, you need to download Datomic and start up a transactor.
Assuming you setup your environment correctly, you should be able to run scripts/download-datomic-pro.sh.
For example:
mkdir ~/download
cd ~/download
../clojure-west-2014-demo/scripts/download-datomic-pro.shUnzip the downloaded Datomic Pro zip file in the filesystem location of your choice.
Included is a helper script that will populate a properties file for a dev transactor and start it up.
It takes the path to the Datomic Pro distribution, as well as the path to your license key file, as arguments.
For example:
./scripts/transactor.sh ~/datomic-pro-0.9.4956 ~/download/datomic-license-keyThe output should look something like this:
Creating dev-transactor.propertie ...
Executing: /home/var/datomic-pro-0.9.4956/bin/transactor dev-transactor.properties
Launching with Java options -server -Xms1g -Xmx1g -XX:NewRatio=4 -XX:SurvivorRatio=8 -XX:+UseConcMarkSweepGC -XX:+UseParNewGC -XX:+CMSParallelRemarkEnabled -XX:CMSInitiatingOccupancyFraction=60 -XX:+UseCMSInitiatingOccupancyOnly -XX:+CMSScavengeBeforeRemark
Starting datomic:dev://localhost:4334/<DB-NAME>, storing data in: data ...
System started datomic:dev://localhost:4334/<DB-NAME>, storing data in: dataOnce the transactor is online, you can run lein ring server
lein ring server-headlessFor write & query examples, see Overview.
Deploying to Amazon requires you to create some AWS resources - DynamoDB tables, S3 buckets, and an EC2 Security Group. This is handled by scripts/mk-aws-resources.sh.
NOTE: Creating DynamoDB tables cost you money. If you don't delete the tables created by scripts/mk-aws-resources.sh, they will cost you roughly $15/mo.
The mk-aws-resources.sh saves an environment file in the scripts directory named .demo.env. The other scripts read in details about your environment from it.
To remove the resources mk-aws-resources.sh created, make sure there are no
CloudFormation stacks running, then run scripts/rm-aws-resources.sh.
Run scripts/mk-aws-resources.sh, supplying it with the path to the JDK RPM, as well as the path to the license key file you created.
./scripts/mk-aws-resources.sh ~/download/jdk-7u67-linux-x64.rpm ~/download/datomic-license-keyIt will prompt for S3 bucket names (it generates defaults), as well as an EC2 key-pair to associate with instances launched in the CloudFormation stack. (You may choose not to have a SSH key, but you will not be able to access those instances.)
If successful, your AWS account will now have two DynamoDB tables for this app: one for 'staging' and one for 'production'. You will also have two EC2 Security Groups for yet-to-be-created Datomic peers, and two S3 buckets.
mk-aws-resources.sh uploads your license-key and JDK RPM to the first S3 bucket, and this S3 bucket will be used for war file uploads.
The second S3 bucket is where transactors will save their logs.
Again, all of these resources can be deleted via scripts/rm-aws-resources.sh.
After the required AWS resources are created, deploying a CloudFormation stack that runs this demo is as simple as building a war file with lein and running scripts/create-stack.sh:
lein ring uberwarThe output of lein ring uberwar will look something like
$ lein ring uberwar
Created /home/var/clojure-west-2014-demo/target/clj-west-0.0.1-SNAPSHOT-standalone.warPass in the path to the war file the lein ring uberwar generated, as well as a stack name:
./scripts/create-stack.sh \
/home/var/clojure-west-2014-demo/target/clj-west-0.0.1-SNAPSHOT-standalone.war \
datomic-demo-stagingNOTE: The EC2 instance-types defined in the CloudFormation template are not in the free tier, so creating this stack will cost you money.
The war file you specified will be uploaded to S3 by create-stack.sh, and then passed in as a parameter to the stack it creates.
Watch your stack come online by logging into the AWS Console and selecting CloudFormation under Services. Once the stack's status is CREATE_COMPLETE, lookup the TransLaunchGroup output value in the Outputs tab.
Then scale-up the transactor Auto Scaling Group to one:
aws autoscaling set-desired-capacity --desired-capacity 1 \
--no-honor-cooldown \
--auto-scaling-group-name datomic-demo-staging-TransactorLaunchGroup-XXXXXXXXXXXXXIf you want high-availability, just specify --desired-capacity 2.
Also listed in the stack outputs is the hostname of the peer Elastic Load Balancer.
We can use this hostname to write data to Datomic, and to query it:
TODO: Examples with curl hitting /query & /read against ELBDNS
If you'd like to log in to the instances, look up their hostname in the EC2 console.
If we make an update to the source code or we make update to the CloudFormation template, and we need to deploy a new version, this can be accomplished two ways.
After we build a new war file, we can either run scripts/update-stack.sh, or we can create a new stack using create-stack.sh as described above.
If you choose to update the existing stack, update-stack.sh takes the same arguments as create-stack.sh.
For example:
./scripts/update-stack.sh \
/home/var/clojure-west-2014-demo/target/clj-west-0.0.2-SNAPSHOT-standalone.war \
datomic-demo-stagingThe CloudFormation template has Update Policies defined for the transactor and peer Auto Scaling groups, so CloudFormation will take care of cycling instances. You can watch the actions CloudFormation is taking by viewing the stack's Events tab in the CloudFormation console.
If you choose to create a new stack, you'll need to scale-up the transactor auto scaling group in the new stack while scaling down the transactor group in the old stack.
Both methods were described in my presentation, and I leave it as an exercise to the user to work through them.
As mentioned above, creating DynamoDB tables using mk-aws-resources.sh will incur charges on your AWS account, as will running a CloudFormation stack using the included template.
To delete resources from your AWS account, first delete any CloudFormation stack you created:
./scripts/delete-stack.sh datomic-demo-stagingAfter the stack is completely deleted, run rm-aws-resources.sh:
./scripts/rm-aws-resources.shThese tools should remove AWS resources they created. To be safe, you should poke around your AWS console to make sure.