Using a command-line tool to submit queries to your Neptune DB cluster - Amazon Neptune
Using HTTPSUsing awscurl
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Using a command-line tool to submit queries to your Neptune DB cluster

Having a command-line tool for submitting queries to your Neptune DB cluster is very handy, as illustrated in many of the examples in this documentation. The curl tool is an excellent option for communicating with Neptune endpoints if IAM authentication is not enabled.

However, to keep your data secure it is best to enable IAM authentication.

When IAM authentication is enabled, every request must be signed using Signature Version 4 (Sig4). The third-party awscurl command-line tool uses the same syntax as curl, and can sign queries using Sig4 signing. The Using awscurl section below explains how to use awscurl securely with temporary credentials.

Setting up a command-line tool to use HTTPS

Neptune requires that all connections use HTTPS. Any command line tool like curl or awscurl needs access to appropriate certificates in order to use HTTPS. As long as curl or awscurl can locate the appropriate certificates, they handle HTTPS connections just like HTTP connections, without needing extra parameters. Examples in this documentation are based on that scenario.

To learn how to obtain such certificates and how to format them properly into a certificate authority (CA) certificate store that curl can use, see SSL Certificate Verification in the curl documentation.

You can then specify the location of this CA certificate store using the CURL_CA_BUNDLE environment variable. On Windows, curl automatically looks for it in a file named curl-ca-bundle.crt. It looks first in the same directory as curl.exe and then elsewhere on the path. For more information, see SSL Certificate Verification.

Using awscurl with temporary credentials to securely connect to a DB cluster with IAM authentication enabled

The awscurl tool uses the same syntax as curl, but needs additional information as well:

  • --access_key   –   A valid access key. If not supplied using this parameter, it must be supplied in the AWS_ACCESS_KEY_ID enviroment variable, or in a configuration file.

  • --secret_key   –   A valid secret key corresponding to the access key. If not supplied using this parameter, it must be supplied in the AWS_SECRET_ACCESS_KEY enviroment variable, or in a configuration file.

  • --security_token   –   A valid session token. If not supplied using this parameter, it must be supplied in the AWS_SECURITY_TOKEN enviroment variable, or in a configuration file.

In the past, it was common practice to use persistent credentials with awscurl, such as IAM user credentials or even root credentials, but this is not recommended. Instead, generate temporary credentials using one of the Amazon Security Token Service (STS) APIs, or one of their Amazon CLI wrappers.

It is best to place the AccessKeyId, SecretAccessKey, and SessionToken values that are returned by the STS call into appropriate environment variables in your shell session rather than into a configuration file. Then, when the shell terminates, the credentials are automatically discarded, which is not the case with a configuration file. Similarly, don't request a longer duration for the temporary credentials than you are likely to need.

The following example shows the steps you might take in a Linux shell to obtain temporary credentials that are good for half an hour using sts assume-role, and then place them in environment variables where awscurl can find them:

aws sts assume-role \
    --duration-seconds 1800 \
    --role-arn "arn:aws:iam::(account-id):role/(rolename)" \
    --role-session-name AWSCLI-Session > $output
AccessKeyId=$(cat $output | jq '.Credentials''.AccessKeyId')
SecretAccessKey=$(cat $output | jq '.Credentials''.SecretAccessKey')
SessionToken=$(cat $output | jq '.Credentials''.SessionToken')

export AWS_ACCESS_KEY_ID=$AccessKeyId
export AWS_SECRET_ACCESS_KEY=$SecretAccessKey
export AWS_SESSION_TOKEN=$SessionToken

You could then use awscurl to make a signed request to your DB cluster something like this:

awscurl (your cluster endpoint):8182/status \
    --region us-east-1 \
    --service neptune-db