ozym / haz

haz

Projects related to haz messaging and the haz DB.

Working with Forks

To work with a fork of this repo you will need to clone the fork in such a way to preserve the import paths. Fork and clone to preserve the organization name e.g.,

 git clone [email protected]:YOUR_FORK/haz.git ${GOPATH}/src/github.com/GeoNet/haz

This could possibly include reseting GOPATH if you need to keep origin and a fork

Sub Projects

Messaging Applications

Producers

• haz-sc3-producer - produces haz messges from SeisComPML. Need file system access in the container. See Container Testing below.

Consumers

Subprojects *-consumer consume msg.Haz or msg.Impact messages from SQS and process them.

Web Applications

• geonet-rest - the server for api.geonet.org.nz
• sc3ml-to-quakeml - web services to return QuakeML from SeisComPML.

Support Applications

• haz-aws-messaging - creates AWS resources for the haz messaging. impact-intensity-consumer has a CFN template for it's resources.
• haz-db-loader - used to load SeisComPML into the db. See below.

Protobufs

Compile protobufs

Database

Uses postgis. Pull and run the image (which already has the hazard db initialised and ready to use):

docker run --name hazdb -p 5432:5432 -d 862640294325.dkr.ecr.ap-southeast-2.amazonaws.com/haz-db:9.5

A Postgres 9.4 with Postgis 2.2 image can be built and pushed using:

docker build --rm=true -t 862640294325.dkr.ecr.ap-southeast-2.amazonaws.com/haz-db:9.5 -f database/Dockerfile database
docker push 862640294325.dkr.ecr.ap-southeast-2.amazonaws.com/haz-db:9.5

There is also a script to (re)initialise the DB ./database/scripts/initdb-93.sh

Loading Quake Data

Quake data can be back loaded from SeisComPML. Download SeisComPML from the S3 bucket and then load it to the DB using haz-db-loader:

aws s3 sync s3://seiscompml07 /work/seismcompml07 --exclude "*"  --include "2015p*"
cd haz-db-loader
go run haz-db-loader.go

Loading Quake Data - Origin Web Servers

Init the database with the qrt schema from the geonet project.

Quake data can be back loaded from SeisComPML. Download SeisComPML from the S3 bucket and then load it to the DB using haz-db-loader:

aws s3 sync s3://seiscompml07 /work/seismcompml07 --exclude "*"  --include "2015p*"
cd haz-db-origin-loader
go run haz-db-origin-loader.go

Tests

With the DB up run all tests

./all.sh

Docker Builds

./build.sh proj-name [proj-name]...

./build-push proj-name [proj-name]...

Also refer to .travis.yml.

Container Testing

docker-compose can be used to run a number of the containers together to test messaging.

Edit sc3-producer.env, db-consumer.env, and geonet-rest.env adding the outputs from running haz-aws-messaging.

mkdir /work/spool
sudo chown nobody /work/spool

Run the containers with docker-compose up.

Visit http://localhost:8080/soh/esb and check that heartbeat messages are arriving.

Send a quake and check it arrived (you should get GeoJSON for the quake):

rsync /work/seismcompml07/2015p494191.xml /work/spool/
curl http://localhost:8080/quake/2015p494191

Deployment

AWS Elastic Beanstalk

geonet-rest, quakesearch, sc3ml-to-quakeml, wfs.

There are files for EB - both to deploy the application and also set up logging from the container (application) to CloudWatch Logs. Create a zip file and then upload the zip to EB.

cd deploy
zip wfs.zip Dockerrun.aws.json .ebextensions/*