Run DynamoDB locally with Docker:
docker run -p 8000:8000 dwmkerr/dynamodb
open http://localhost:8000/shell
This container has full support for all of the commandline parameters in the DynamoDB Documentation.
Note - there is now an Official AWS Docker Image for DynamoDB:
To run an ephemeral instance of DynamoDB:
docker run -p 8000:8000 dwmkerr/dynamodb
If you want to have persistent data, just mount a volume from your host and set it as a data directory for DynamoDB:
docker run -v /data:/data -p 8000:8000 dwmkerr/dynamodb -dbPath /data
If you want to access tables and data created by the AWS CLI through a language SDK (Node, Java, etc), you will want to use the -sharedDb
option as described in this AWS forum post:
docker run -p 8000:8000 dwmkerr/dynamodb -sharedDb
Without this option, each connection will get is own database and the data will not be accessible between different clients.
The code is structured like this:
Dockerfile # the important thing, the actual dockerfile
makefile # commands to build, test deploy etc
tests/ # bash scripts to test how the container works
The Dockerfile is based on OpenJDK and essentially just runs a jar file in a JRSE 7 environment.
The makefile contains commands to build, test and deploy. Parameters can be passed as environment variables or through the commandline.
Command | Notes |
---|---|
make build |
Builds the image dwmkerr/dynamodb:latest . If a BUILD_NUM parameter is provided, also builds dwmkerr/dynamodb:BUILD_NUM . |
make test |
Runs the test scripts. |
make deploy |
Deploys the images to the docker hub. If you are not logged in, you're gonna have a bad time. |
The tests are simple bash scripts which check for basic capabilties which relate to the image. This means they're not there to make sure DynamoDB Local works, they're there to make sure the docker features work with the image. For example, mounting a volume to provide a persistent data directory.
CI is provided currently by Circle. Ensure you provide AWS credentials as we are using the AWS CLI (they are not used, but the CLI still checks that they are present).
A basic sample showing how to build an image with custom test data is at ./samples/test-data
.
- Go to the sample:
cd ./samples/test-data
- Create some sample data:
make create-test-data
. This creates sample data files at./data
. - Build a new docker image called
sample-test-data
, withmake build
. - The newly created image has the test data built in. Verify with
make test
.
There's a great blog post on this here:
Missing required key 'ProvisionedThroughput' in params Unexpected key 'BillingMode' found in params
The issue is the underlying DynamoDB Local jar provided by AWS requires read and write capacity units to be specified (https://docs.aws.amazon.com/amazondynamodb/latest/developerguide/DynamoDBLocal.UsageNotes.html).
Provisioned throughput settings are ignored in downloadable DynamoDB, even though the CreateTable operation requires them. For CreateTable, you can specify any numbers you want for provisioned read and write throughput, even though these numbers are not used. You can call UpdateTable as many times as you want per day. However, any changes to provisioned throughput values are ignored.
Here is an example CreateTable call which will work:
// Note we have to specify ProvisionedThroughput, but it will not be used.
var params = {
TableName: 'table',
ProvisionedThroughput: {
ReadCapacityUnits: 1,
WriteCapacityUnits: 1
},
KeySchema: [
{
AttributeName: 'id',
KeyType: 'HASH',
}
],
AttributeDefinitions: [
{
AttributeName: 'id',
AttributeType: 'S',
}
],
};
dynamodb.createTable(params, function(err, data) {
if (err) ppJson(err); // an error occurred
else ppJson(data); // successful response
});
Please help out! Here are some areas I'd like to improve upon:
- Cleaner or colourised output for the tests. Is there any simple tool to do assets in a shell script?
If you contribute, include a README.md
update in your PR and I'll list your contributions here.