hsds-nsrdb-test

NodeJS service to download NSRDB data via HSDS data

Quickstart

git clone git@github.com:PjEdwards/hsds-nsrdb-test.git
cd hsds-nsrdb-test
yarn

HSDS requirement

This service requires a running instance of HSDS. BUCKET_NAME can be set as below, or you can create a bucket as described in the HSDS installation instructions.

For best performance it's desirable that the service run in the AWS uus-west-2 reegion.

BUCKET_NAME=nrel-pds-hsds
AWS_REGION=us-west-2
AWS_S3_GATEWAY=http://s3.us-west-2.amazonaws.com

Configure

edit the constants file setting the appropriate HSDS_URI

Start

run yarn start (or yarn dev if you want to support hot reloading)

Intialization routine

on startup the service will attempt to fetch the metadata from HSDS. If successful you will see a series of messages in the console shortly after startup which, if successful, will culminate in a message that states:

NSRDB meta fetch complete after 4.755 seconds

If this initialization fails then the most likely culprit is the connection to the HSDS service, or the HSDS service does not see the NSRDB data bucket. In any case, this service can't work if initialization of the metadata fails.

To Loadtest

Start the service as above, then

yarn ltest

To customize the test, edit the parameters in the loadtest script

* Keep in mind that the number of attributes requested in the URL will equal the number of parallel requests sent to HSDS for each single request sent by loadtest

For more info, see https://github.com/alexfernandez/loadtest

NSRDB Target

• 6 requests per second
• sustained for 24 hours
• each request for 4 attributes
• response time for each < 30 seconds
• request years and attributes randomized across full range