Skip to content

Latest commit

 

History

History
174 lines (132 loc) · 7.24 KB

development.md

File metadata and controls

174 lines (132 loc) · 7.24 KB

Development guide

Requirements

  • JDK 8 (JDK 11 is recommended)
  • Maven 3.5.x

Backend development

Prerequisites

Build and start Log Search in docker container

# to see available commands: run logsearch-docker without arguments
cd docker
./logsearch-docker build-and-run # build mvn project locally, build docker image, start containers

If you run the script at first time, it will generate you a new Profile file or an .env file inside docker directory (run twice if both missing and you want to generate Profile and .env as well), in .env file you should set MAVEN_REPOSITORY_LOCATION (point to local maven repository location, it uses ~/.m2 by default). These will be used as volumes for the docker container. Profile file holds the environment variables that are used inside the containers, the .env file is used outside of the containers

Then you can use the logsearch-docker script to start the containers (start command). Also you can use docker-compose manually to start/manage the containers.

docker-compose up -d

After the logsearch container is started you can enter to it with following commands:

docker exec -it docker_logsearch_1 bash

In case if you started the containers separately and if you would like to access Solr locally with through your external ZooKeeper container, then point solr to localhost in your /etc/hosts file.

Enable debug for Log Search Server or Log Feeder

By default remote debug is enabled in docker containers for Log Search server and Log Feeder (in order to use IDEs for debugging). Log Search server port for debug is 5005 and 5006 for Log Feeder.

To suspend components in debug mode, you need to edit your Profile file (that was generated in docker folder) and set LOGSEARCH_DEBUG_SUSPEND or LOGFEEDER_DEBUG_SUSPEND to true.

Run applications from IDE / maven

Start Log Search server locally from maven / IDE

Other services (like zookeeper, solr, logfeeder) can be started with docker-compose

cd ambari/ambari-logsearch/docker
docker-compose up -d zookeeper solr logfeeder

Then you can start Log Search server from maven

cd ambari/ambari-logsearch/ambari-logsearch-server
./run.sh
# or
mvn clean package -DskipTests spring-boot:run

You can also start Log Search server from an IDE as well. One thing is important: the config set location that the server tries to upload to ZooKeeper. By default config sets are located at ${LOGSEARCH_SERVER_RELATIVE_LOCATION:}src/main/configsets in logsearch.properties. Based or from where you run LogSearch.java, you need to set LOGSEARCH_SERVER_RELATIVE_LOCATION env variable properly. (or just simply use the ambari-logsearch-server as the working directory)

Start Log Feeder locally from maven / IDE

First you need to start every required service (except logfeeder), go to ambari-logsearch/docker folder and run:

docker-compose up -d zookeeper solr logsearch

Secondly, if you are planning to run Log Feeder from an IDE, for running the LogFeeder main method, you will need to set the working directory to ambari/ambari-logsearch/ambari-logsearch-logfeeder or set LOGFEEDER_RELATIVE_LOCATION env variable. With Maven, you won't need these steps, just run this command from the ambari-logsearch-logfeeder folder:

mvn clean package -DskipTests spring-boot:run

For Log Feeder, it is also important to use the ambari-logsearch-logfeeder as a working directory if you are trying to run the application from an IDE.

Package build process

  1. Check out the code from GIT repository

  2. On the logsearch root folder (ambari/ambari-logsearch), please execute the following make command to build RPM/DPKG:

make rpm
# or for jdk11
export LOGSEARCH_JDK_11=true
make rpm

or

make deb
# or for jdk11
export LOGSEARCH_JDK_11=true
make deb
  1. Generated RPM/DPKG files will be found in ambari-logsearch-assembly/target folder

Running Integration Tests

By default integration tests are not a part of the build process, you need to set -Dbackend-tests or -Dselenium-tests (or you can use -Dall-tests to run both). For running the tests you will need docker here as well (right now docker-for-mac and unix are supported by default, for boot2docker you need to pass -Ddocker.host parameter to the build).

# from ambari-logsearch folder
mvn clean integration-test -Dbackend-tests failsafe:verify
# or run selenium tests with docker for mac, but before that you need to start xquartz
open -a XQuartz
# then in an another window you can start ui tests
mvn clean integration-test -Dselenium-tests failsafe:verify
# you can specify the folder that contains the story files with -Dbackend.stories.location and -Dui.stories.location (absolute file path) in the commands

Also you can run from the IDE, but make sure all of the ambari logsearch modules are built.

Update version (for release or specific builds)

make update-version new-version="2.8.0.0-11"

Update REST API docs

make update-rest-api-docs

Update markdown docs

make prop-docs

UI development

After running yarn install or npm install you can run npm start or yarn start to start the dev server. So you can navigate to http://localhost:4200/. The app will automatically reload if you change any of the source files.

Webpack Development Config

In order to use the UI you'll need a running service instance too. You can setup a local development env or if you have running instance elsewhere you can set the proxy to that URL. In order to keep the main webpack config file intact we use the webpack.config.dev.js file so you can set a service URL proxy here.

The content of the webpack.config.dev.js (when you have a local service with docker):

const merge = require('webpack-merge');
const baseConfig = require('./webpack.config.js');

module.exports = merge(baseConfig, {
  devServer: {
    historyApiFallback: true,
    proxy: {
      '/api': 'http://localhost:61888/', // proxying the api requests
      '/login': 'http://localhost:61888/', // proxying the login action
      '/logout': 'http://localhost:61888/' // proxying the the logout action
    }
  }
});

And you can start the client this way: yarn start --config webpack.config.dev.js