Features

• Zero-downtime update deployment
• Fast page loading due client-side rendering
• Fulltext search by posts
• Updating posts through web STOMP on main page
• Draft posts that visible only for author and administrator
• User locking
• User deletion (with migrating posts to special deleted user)
• Pages prerendering for crawlers with rendertron
• Dynamically setting header, subheader and background image without server restart
• Auto cleaning "orphanned" images from PostgreSQL, and "orphaned" posts from Elasticsearch
• Cluster out from the box - simple scale it with docker service scale BLOGSTACK_blog=4
• Login through Facebook, Vkontakte OAuth2 providers
• Binding several OAuth2 account to same blog account
• Simply installation with docker swarm
• Applications like Vkontakte/Facebook apps. Example storage application on Go
• Self-sufficient frontend asset. No CDN used
• Simply backup of everything to one .sql file

Requirements

Run

• Docker 18.09.0+

Development

• JDK 13
• docker-compose 1.24.1 +
• Google Chrome (as default browser for webdriver-test). Just dnf install chromium in latest Fedora.
• disable SELinux

FAQ

Q: Can I run it without docker ?

A: Yes, you can achieve it by manually install PostgreSQL, RabbitMQ, Redis, Elasticsearch and configure it's connections in config or through commandline. See Spring Boot documentation https://docs.spring.io/spring-boot/docs/2.1.1.RELEASE/reference/html/boot-features-external-config.html.

Q: How to build frontend if I am backend developer ?

A:

./mvnw -P frontend generate-resources

Q: How to build full jar (with static) ?

A:

./mvnw -P frontend clean package

It will download java dependencies and nodejs with frontend dependencies.

Q: Why does blog wait for PostgreSQL, Elasticsearch, Redis, RabbltMQ port availability on boot?

A: Primarily for deploy tests run inside Github CI. When there isn' t these waits, I had spuriously tests fails due to unpredictable time of Elasticsearch boot.

Embedded API documentation

Embedded documentation are available at http://127.0.0.1:8080/docs/index.html

Request version info

This will available after full package, e. g. after resource filtering of git.template.json and renaming result in target/classes/static dir to git.json

curl -i http://127.0.0.1:8080/git.json

Running on Windows without docker

First you should install Redis, PostgreSQL, Rabbit MQ, Elasticsearch and manually setup them (create database, schema, user for PostgreSQL, install web stomp plugin and create user for RabbitMQ).

Redis Windows x86 which works on my PC (Windows 7 x86) http://bitsandpieces.it/redis-x86-32bit-builds-for-windows

2.8.2104 http://fratuz610.s3.amazonaws.com/upload/public/redis-builds/x86/redis-windows-x86-2.8.2104.zip - requires enabled swapfile.

run

redis-server.exe --maxheap 8Mb

Next you should use localhost IP addresses and disable asciidoctor:

mvnw -P local -Dasciidoctor.skip=true clean test

Demo Run / Installation

cd docker
./swarm-init.sh

I strongly recommend copy and rename docker-compose.template.yml to docker-compose.stack.yml. Next I'll use renamed file.

Copy files on your server:

scp -r /path/to/blog/docker/* user@blog.test:/path/to/blog/
chmod 600 traefik/acme.json

Manual changes

Let' s assume cd docker.

a) ./swarm-init.sh

b) In docker-compose.template.yml or docker-compose.stack.yml:.

Change tag in service blog image: nkonev/blog:current-test -> image: nkonev/blog:latest

Also you can remove demo profile

c) Change next properties:

      - SPRING_MAIL_HOST=smtp.yandex.ru
      - CUSTOM_EMAIL_FROM=username@yandex.ru
      - SPRING_MAIL_USERNAME=username
      - SPRING_MAIL_PASSWORD=password
      - CUSTOM_BASE-URL=http://blog.test
 

And remove explicit ports definition where it's don't need - postgres, redis, rabbit, because of docker publishes ports by add it to iptables chain. If you very want, you can skip setting these properties, but you'll have non-working email, wrong links in emails and so on.

d) Generating monitoring grafana & prometheus password

sudo yum install -y httpd-tools
# generate login and hash with replaced $ with $$ sign for able to copy-paste to docker-compose.stack.yml
htpasswd -nb admin admin | sed -e 's/\$/\$\$/g'

e) Set journald logging with appropriate tag for all services

    logging:
      driver: "journald"
      options:
        tag: blog

f) Uncomment & change SSL setting in ./traefik/traefik.toml

g) Configure notifications in ./alertmanager/alert.yml

i) For able to http(s) request your domain registrar name with curl from container ensure that

cat /proc/sys/net/ipv4/ip_forward

returns non-zero

next

Option a)

firewall-cmd --permanent --zone=public --add-port=80/tcp
firewall-cmd --permanent --zone=public --add-port=443/tcp
firewall-cmd --reload

Check

firewall-cmd --list-all-zones
iptables -t nat --line-numbers --numeric --list

Option b) insert iptables rule

iptables -I INPUT -i docker_gwbridge -p tcp -m multiport --dports 80,443 -j ACCEPT

If all ok, you should do it persistent by

chmod +x /etc/rc.local
vim /etc/rc.local

iptables -I INPUT -i docker_gwbridge -p tcp -m multiport --dports 80,443 -j ACCEPT
echo "Successful inserted docker_gwbridge rule"

Starting with docker swarm

Next you can

docker stack deploy --compose-file docker-compose.stack.yml BLOGSTACK
docker service scale BLOGSTACK_blog=4
docker service ls

See postgres volume

docker volume inspect BLOGSTACK_postgresql_blog_dev_data_dir

See logs of jars via journalctl (see applied tags in docker-compose.stack.yml):

journalctl -f CONTAINER_TAG=blog
journalctl -f CONTAINER_TAG=blog -o verbose
journalctl -f CONTAINER_TAG=blog CONTAINER_TAG=postgresql CONTAINER_TAG=redis CONTAINER_TAG=rabbitmq

or via docker

docker service logs -f BLOGSTACK_blog

Remove

docker stack rm BLOGSTACK

Remove exited containers

docker rm $(docker ps -aq -f name=BLOGSTACK_blog -f status=exited)

Test on local machine

curl

curl -H "Host: blog.test" http://127.0.0.1:8088
curl -H "Host: grafana.blog.test" -u "admin:admin" http://127.0.0.1:8088
curl -H "Host: prometheus.blog.test" -u "admin:admin" http://127.0.0.1:8088
curl -H "Host: alertmanager.blog.test" -u "admin:admin" http://127.0.0.1:8088

Browser

We add domains to /etc/hosts for browser sends correct Host header

sudo tee --append /etc/hosts <<'EOF'
127.0.0.1 blog.test
127.0.0.1 grafana.blog.test
127.0.0.1 prometheus.blog.test
127.0.0.1 alertmanager.blog.test
EOF

Maintenance

docker ps -aq | xargs docker rm
docker volume ls -q | xargs docker volume rm
docker images -q -a | xargs  docker rmi

Open PostgreSQL

docker exec -it $(docker ps --filter label=com.docker.swarm.service.name=BLOGSTACK_postgresql -q) psql -U blog
docker exec -it $(docker ps --filter label=com.docker.swarm.service.name=TESTBLOGSTACK_postgresql -q) psql -U blog

Open blog

docker exec -it $(docker ps --filter label=com.docker.swarm.service.name=BLOGSTACK_blog -q | head -n 1) bash

SEO

First configure custom.rendertron.serviceUrl - setup correct url of Rendertron installation. See also dockerized build.

How to add SEO metrics scripts

Just prepend file: location which contains index.html, and copy modified index.html to there folder.

spring.resources.static-locations: file:/var/www/, file:backend/src/main/resources/static/, classpath:/static/

So firstly Spring Mvc will looking in /var/www, next in $PWD/backend/src/main/resources/static/...

If your search(Yandex Metrics for example) checks for existence script - request will passed through rendertron, which wipes <script> tags.

In order to solve it, use custom.seo.script=file:/var/www/seo.html - Rendertron filter will inject content of this file before closing </head>.

Grafana

Fix disk usage in https://grafana.com/dashboards/1860

Set query 100 - ((node_filesystem_avail_bytes{mountpoint="/rootfs"} * 100) / node_filesystem_size_bytes{mountpoint="/rootfs"})

Set Instant

TODO

• re-implement buttons css
• sitemap for SEO
• edit metainfo for SEO by user
• change post owner by admin
• change comment owner by admin
• LDAP
• Google OAuth2 login
• search by comments

Generate configs

./mvnw -pl configs-generator generate-sources