You can check last images here and use flash tool to flash your RaspberryPi SD:
flash --hostname your-hostname https://github.com/hypriot/image-builder-rpi/releases/download/v1.4.0/hypriotos-rpi-v1.4.0.img.zip
SSH into each RPI:
ssh pirate@you-rpi-ip
As of version 1.4, default credentials are pirate/hypriot. You can use arp-scan to guess the IP. You can also use:
function getip() { (traceroute $1 2>&1 | head -n 1 | cut -d\( -f 2 | cut -d\) -f 1) }
Change default password:
passwd
You can also set up paswordless access with:
ssh-copy-id -i ~/.ssh/your-key_rsa.pub pirate@your-rpi -o "IdentitiesOnly yes"
And also add an entry to you ~/.ssh/config file:
Host your-rpi-1 your-rpi-2 ...
Hostname %h.local
User pirate
IdentityFile ~/.ssh/your-key_rsa
IdentitiesOnly yes
StrictHostKeyChecking no
If you want, you can also add this config snippet to all your nodes and add your private key to each ~/.ssh
folder to be able to connect from one RPI to another.
(?) Add regular user to docker group
sudo usermod -aG docker pirate
(Optional) In case you see annoying warning messages about locales from perl:
sudo dpkg-reconfigure locales
(Optional) Install some useful packages
sudo aptitude update && sudo aptitude install rsync zsh
sudo aptitude install cryptsetup
sudo fdisk /dev/sdX
sudo cryptsetup --verify-passphrase luksFormat /dev/sdX1 -c aes -s 256 -h sha256
sudo cryptsetup luksOpen /dev/sdX1 volumes
sudo mkfs -t ext4 -m 1 -O dir_index,sparse_super /dev/mapper/volumes
#mount -t auto /dev/mapper/volumes /media/volumes
sudo dd if=/dev/urandom of=/root/volumes_luks_pwd bs=1024 count=4
sudo chmod 0400 /root/volumes_luks_pwd
sudo cryptsetup luksAddKey /dev/sdX1 /root/volumes_luks_pwd
Add to /etc/crypttab
:
volumes /dev/disk/by-uuid/uuid-of-your-drive /root/volumes_luks_pwd luks
and add to /etc/fstab
:
/dev/mapper/volumes /media/volumes ext4 defaults 0 2
Install server on main host:
sudo aptitude install nfs-kernel-server
sudo mkdir -p /export/volumes
sudo mount --bind /media/volumes /export/volumes
And add the following line to /etc/fstab
to avoid repeating it on startup:
/media/volumes /export/volumes none bind 0 0
And to /etc/exports
:
/export 192.168.1.0/24(rw,fsid=0,insecure,no_subtree_check,async)
/export/volumes 192.168.1.0/24(rw,nohide,insecure,no_subtree_check,async,no_root_squash)
(changing network/mask by your local values)
On the other nodes:
sudo aptitude install nfs-common
And add to /etc/fstab
:
your-main-host:/export/volumes /media/volumes nfs auto,user 0 0
http://jermsmit.com/my-raspberry-pi-needs-a-swap/
dd if=/dev/zero of=/media/volumes/swap bs=1M count=2048
chmod 600 /media/volumes/swap
mkswap /media/volumes/swap
swapon /media/volumes/swap
Add to /etc/fstab:
/media/volumes/swap swap swap defaults 0 0
Repeat for worker nodes (changing name of swap file)
When the dockers are running, some service users (e.g. dovecot
or mysqld
) can have conflicting ids with the one of avahi, making it fail. To avoid that, we can just increase its uid
, e.g.:
sudo systemctl stop avahi-daemon
sudo usermod -u 205 avahi
sudo systemctl restart dbus
sudo systemctl start avahi-daemon
Install missing libnss-mdns
package (see explanation here):
sudo aptitude install libnss-mdns
Also make sure avahi-daemon
works, and otherwise restart it. See this issue.
Login to the main RPI and start the swarm:
docker swarm init --listen-addr eth0
And join from the other ones, just copy-paste command provided by the master from them:
docker swarm join --token your-token your-main_rpi:2377
If you have existing data, create folders (otherwise setup script will do it) and copy it data:
sudo mkdir -p /media/volumes/mail/data
sudo mkdir -p /media/volumes/mail/state
sudo mkdir -p /media/volumes/nextcloud
sudo chown -R pirate:pirate /media/volumes/*
sudo mkdir -p /media/volumes/openldap/data
sudo mkdir -p /media/volumes/openldap/config
sudo mkdir -p /media/volumes/openldap/certs
sudo chown -R 999 /media/volumes/openldap*
From your current installation:
rsync -auv --delete -e "ssh -i ~/.ssh/your-key_rsa" /var/www/nextcloud/data your-main-host:/media/volumes/nextcloud/
mysqldump --lock-tables -u nextcloud -p -h localhost nextcloud > /var/www/nextcloud/nextcloud_db_backup.sql
rsync -auv --delete -e "ssh -i ~/.ssh/your-key_rsa" /srv/vmail/ your-main-host:/media/volumes/mail/data
First download the repos:
git clone https://github.com/bingen/rpi_docker_home_server.git
cd rpi_docker_home_server
Set up your preferences:
./setup.sh
(Optional, can be downloaded from registry, unless you changed them) Build aux images:
cd images/rpi-nginx
docker build . -t bingen/rpi-nginx
cd ../../
cd images/rpi-nginx-php
docker build . -t bingen/rpi-nginx-php
cd ../../
(Optional, can be downloaded from registry, unless you changed them) Build images:
docker-compose build
Deploy docker stack (it will also rebuild components)
./deploy.sh your-stack-name
If you add or modify a service, you can update it running:
docker-compose build && docker push your-container && env $(cat .env | grep "^[A-Z]" | xargs) docker stack deploy --compose-file docker-compose.yml your-stack
docker node ls
docker stack ls
docker stack ps your-stack-name
To see logs of a docker swarm/stack service reference:
docker logs $(docker inspect --format "{{.Status.ContainerStatus.ContainerID}}" `docker stack ps your-stack-name | grep your-service-name | cut -f1 -d' '`)
To shutdown the stack:
docker stack rm your-stack-name
To get into containers:
docker ps # in the swarm node containing it
docker exec -ti 5105b27d9cf0 bash
To view swarm token:
docker swarm join-token worker
I was experience the issue described and fixed [here](Docker swarm nodes down after reboot! https://forums.docker.com/t/docker-worker-nodes-shown-as-down-after-re-start/22329/8?u=bingen):
To avoid swarm nodes showing up as Down on reboot, you can do:
sudo crontab -e
then add a line like this
@reboot docker ps
ldapsearch -x -w your-admin-ldap-password -D cn=admin,dc=your-domain,dc=com -b dc=your-domain,dc=com -LLL
To reset a user's password:
Copy this into a file, user_pwd.ldif
:
dn: uniqueIdentifier=your-user,ou=people,dc=your-domain,dc=com
changetype: modify
replace: userPassword
userPassword: {SSHA}Djpd2d+kbQm4ftHupSaS65wl8l8EbDot
And the run:
ldapadd -W -D "cn=admin,dc=your-domain,dc=com" -f user_pwd.ldif
You can generate the password with:
slappasswd -s your-password
You can use the following script to add users if you have previously created ldif
files:
./add_users <your-stack-name>
If you have existing data, make sure root password matches and access from outside ('%') is allowed.
After first run, set DATA_CHOWN=0. Otherwise every time you deploy the whole folder with all your data will be recursed to change ownership, and it can take long when it's only needed for the first time.
Need to log in as admin for the first time and enable Apps manually.
Run the following script to enable Let's Encrypt for Nextcloud:
./letsencrypt.sh <your-stack-name>
Follow the instructions here to set up your own registry:
docker service create --name registry --publish published=5000,target=5000 registry:2
Check your domain registration provider
Install fail2ban in you docker swarm master node if you want to allow ssh connections from outside.
sudo aptitude install fail2ban
Have a look at the documentation for configuration.
Get into your router admin page and redirect ports:
80
,443
for Web (Nextcloud and eventually other through HaProxy)25
,143
,587
,993
for mail server22
for ssh
to your docker swarm master node IP.
- Install and enable Nextcloud apps automatically
- XMPP
- Wordpress
- VPN
- Open social networks (GNU social, Diaspora)
- Transmission
- Sia storage
- Alternative: run your own registry for images.