Admin Training VM Setup

This documents provides steps to create a VM image from scratch.

Instructors spin up VMs from the image.

Each VM provides:

6 RE nodes
A DNS server
A VNC desktop
Redis Insight.

as follows: https://drive.google.com/open?id=17tM53iHHTu-DQNPD48dYQEAias0qmqz30hpP2embUV4

Nodes run in containers, but they look like they run on VMs.

Users access the VM by VNC on port 80. All they need is the VM public IP and VNC password.

Setup is in 2 stages:

Start Docker/VNC and configure DNS
Configure VNC.

Stage 1

Create a VPC in GCP with subnet 172.18.0.0/16 in the region where you want to run VMs.

Requirement	Specification
Name	admin-training-vpc
Subnet Creation Mode	Custom
Subnet Name	admin-training-subnet
Subnet IP Address Range	172.18.0.0/16

Create a firewall rule that allows ingress on all ports from all sources (0.0.0.0/0) to all targets.
Create the base VM in the region and VPC where you want to run instances.

Requirement	Specification
Name	admin-training-stage-1
CPU	4
Memory	15 GB
OS	Ubuntu 18.04 LTS
Disk	30 GB
Networking	admin-training-vpc

SSH to the base VM from GCP console to finish setup.
Install vim and add 'trainee' user to the 'docker' group so it can start, stop, and SSH to containers.

sudo su
apt -y update
apt -y install vim

update-alternatives --config editor <<< 3

adduser --disabled-password --gecos "" trainee
groupadd docker
usermod -aG docker trainee

Install Docker.

apt-get install \
    apt-transport-https \
    ca-certificates \
    curl \
    software-properties-common <<< Y
    
curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -

apt-key fingerprint 0EBFCD88
   
add-apt-repository \
   "deb [arch=amd64] https://download.docker.com/linux/ubuntu \
   $(lsb_release -cs) \
   stable"

apt-get update
apt-get -y install docker-ce

Run 'sudo visudo' and add the following line so 'trainee' can start and stop containers without 'sudo'.

trainee ALL=(ALL) NOPASSWD:ALL

Switch to 'trainee' to create the Docker network, add scripts, and build/run the VNC image.

sudo su - trainee

Uncomment the line '#force_color_prompt' in the user's .bashrc file. This sets VM user prompt to 'green' so you can tell it apart from the VNC user.
Generate keys so students can 'silently' SSH from VNC user to base VM user and start/stop/SSH RE nodes.

mkdir .ssh
ssh-keygen -q -t rsa -N '' -f .ssh/id_rsa 2>/dev/null <<< y >/dev/null
cp -r .ssh/id_rsa.pub .ssh/authorized_keys 

mkdir vnc_docker
cp -r .ssh/ vnc_docker/ssh

Create bashrc with aliases to run necessary commands.

cat << EOF > vnc_docker/bashrc
source \$STARTUPDIR/generate_container_user

export PS1='\e[1;33m\u@\h\e[m:\e[1;34m\w\e[m\$ '

alias create_north_cluster="ssh -t [email protected] ./scripts/create_north_cluster.sh "
alias create_south_cluster="ssh -t [email protected] ./scripts/create_south_cluster.sh "

alias run_dnsutils="ssh -t [email protected] ./scripts/run_dnsutils.sh "

alias run_redis_start="ssh -t [email protected] docker run -it --name redis -h redis -w / redis bash"
alias run_redis_stop="ssh -t [email protected] docker container rm \$\(docker container ls -q -f '\''status=exited'\''\)"

alias ssh_base-vm="ssh -t [email protected]"        # only used by admins when building VMs

alias start_north_nodes="ssh -t [email protected] ./scripts/start_north_nodes.sh "
alias start_south_nodes="ssh -t [email protected] ./scripts/start_south_nodes.sh "

alias start_n1="ssh -t [email protected] docker start n1 "
alias start_n2="ssh -t [email protected] docker start n2 "
alias start_n3="ssh -t [email protected] docker start n3 "
alias start_s1="ssh -t [email protected] docker start s1 "
alias start_s2="ssh -t [email protected] docker start s2 "
alias start_s3="ssh -t [email protected] docker start s3 "

alias stop_n1="ssh -t [email protected] docker stop n1 "
alias stop_n2="ssh -t [email protected] docker stop n2 "
alias stop_n3="ssh -t [email protected] docker stop n3 "
alias stop_s1="ssh -t [email protected] docker stop s1 "
alias stop_s2="ssh -t [email protected] docker stop s2 "
alias stop_s3="ssh -t [email protected] docker stop s3 "
EOF

mkdir scripts

cat << EOF > scripts/start_north_nodes.sh
docker kill n1; docker rm n1;
docker kill n2; docker rm n2;
docker kill n3; docker rm n3;
docker run --name n1 -d --restart=always --cap-add=ALL --net rlabs --dns 172.18.0.20 --hostname n1.rlabs.org --ip 172.18.0.21 redislabs/redis
docker run --name n2 -d --restart=always --cap-add=ALL --net rlabs --dns 172.18.0.20 --hostname n2.rlabs.org --ip 172.18.0.22 redislabs/redis
docker run --name n3 -d --restart=always --cap-add=ALL --net rlabs --dns 172.18.0.20 --hostname n3.rlabs.org --ip 172.18.0.23 redislabs/redis
docker exec --user root n1 bash -c "iptables -t nat -I PREROUTING -p udp --dport 53 -j REDIRECT --to-ports 5300"
docker exec --user root n2 bash -c "iptables -t nat -I PREROUTING -p udp --dport 53 -j REDIRECT --to-ports 5300"
docker exec --user root n3 bash -c "iptables -t nat -I PREROUTING -p udp --dport 53 -j REDIRECT --to-ports 5300"
sleep 60
EOF

cat << EOF > scripts/start_south_nodes.sh
docker kill s1; docker rm s1;
docker kill s2; docker rm s2;
docker kill s3; docker rm s3;
docker run --name s1 -d --restart=always --cap-add=ALL --net rlabs --dns 172.18.0.20 --hostname s1.rlabs.org --ip 172.18.0.31 redislabs/redis
docker run --name s2 -d --restart=always --cap-add=ALL --net rlabs --dns 172.18.0.20 --hostname s2.rlabs.org --ip 172.18.0.32 redislabs/redis
docker run --name s3 -d --restart=always --cap-add=ALL --net rlabs --dns 172.18.0.20 --hostname s3.rlabs.org --ip 172.18.0.33 redislabs/redis
docker exec --user root s1 bash -c "iptables -t nat -I PREROUTING -p udp --dport 53 -j REDIRECT --to-ports 5300"
docker exec --user root s2 bash -c "iptables -t nat -I PREROUTING -p udp --dport 53 -j REDIRECT --to-ports 5300"
docker exec --user root s3 bash -c "iptables -t nat -I PREROUTING -p udp --dport 53 -j REDIRECT --to-ports 5300"
sleep 60
EOF

cat << EOF > scripts/create_north_cluster.sh
docker exec -it n1 bash -c "/opt/redislabs/bin/rladmin cluster create persistent_path \
        /var/opt/redislabs/persist ephemeral_path /var/opt/redislabs/tmp addr 172.18.0.21 \
        name north.rlabs.org username [email protected] password admin";

docker exec -it n2 bash -c "/opt/redislabs/bin/rladmin cluster join persistent_path \
        /var/opt/redislabs/persist ephemeral_path /var/opt/redislabs/tmp addr 172.18.0.22 \
        username [email protected] password admin nodes 172.18.0.21";

docker exec -it n3 bash -c "/opt/redislabs/bin/rladmin cluster join persistent_path \
        /var/opt/redislabs/persist ephemeral_path /var/opt/redislabs/tmp addr 172.18.0.23 \
        username [email protected] password admin nodes 172.18.0.21";
EOF

cat << EOF > scripts/create_south_cluster.sh
docker exec -it s1 bash -c "/opt/redislabs/bin/rladmin cluster create persistent_path \
        /var/opt/redislabs/persist ephemeral_path /var/opt/redislabs/tmp addr 172.18.0.31 \
        name south.rlabs.org username [email protected] password admin";
docker exec -it s2 bash -c "/opt/redislabs/bin/rladmin cluster join persistent_path \
        /var/opt/redislabs/persist ephemeral_path /var/opt/redislabs/tmp addr 172.18.0.32 \
        username [email protected] password admin nodes 172.18.0.31";
docker exec -it s3 bash -c "/opt/redislabs/bin/rladmin cluster join persistent_path \
        /var/opt/redislabs/persist ephemeral_path /var/opt/redislabs/tmp addr 172.18.0.33 \
        username [email protected] password admin nodes 172.18.0.31";
EOF

cat << EOF > scripts/run_dnsutils.sh
docker kill dnsutils; docker rm dnsutils
docker run --name dnsutils -it --net rlabs --dns 172.18.0.20 --hostname dnsutils.rlabs.org --ip 172.18.0.6 tutum/dnsutils
EOF

# make the scripts executable
chmod 755 scripts/start_north_nodes.sh
chmod 755 scripts/start_south_nodes.sh
chmod 755 scripts/create_north_cluster.sh
chmod 755 scripts/create_south_cluster.sh
chmod 755 scripts/run_dnsutils.sh

Create the Docker network.

#mkdir resolve
#echo 'nameserver 172.18.0.20' > resolve/resolv.conf

docker network create --subnet=172.18.0.0/16 rlabs

Build the VNC Docker image and run it as a container.

cat << EOF > vnc_docker/Dockerfile
## Custom Dockerfile
FROM consol/ubuntu-xfce-vnc

# Switch to root user to install additional software
USER 0

## Install SSH, DNS Utils, and the VNC user's bashrc and chromium-browser.init
RUN apt update; apt install -y ssh dnsutils;
RUN mkdir /headless/.ssh
COPY ./ssh /headless/.ssh
RUN chown -R 1000 /headless/.ssh/
COPY bashrc /headless/.bashrc
RUN chown -R 1000 /headless/.bashrc

## switch back to default user
USER 1000
EOF

cd vnc_docker
docker build -t vanilla-vnc .

docker run --name vanilla-vnc  -d -e VNC_PW=trainee! --restart=always --net rlabs --hostname vnc-terminal.rlabs.org --ip 172.18.0.2 -p 80:6901  vanilla-vnc

Run Redis Insight as a container so students can view database contents in a GUI.

docker run --name insight -d -v redisinsight:/db --restart=always --net rlabs --dns 172.18.0.20 --hostname insight.rlabs.org --ip 172.18.0.4  redislabs/redisinsight

Run Bind DNS as a container.

docker run --name vanilla-dns -d --restart=always --net rlabs --dns 172.18.0.20 --hostname ns.rlabs.org --ip 172.18.0.20 -p 10000:10000/tcp  sameersbn/bind

Someday, you may use CoreDNS with Corefile and rlabs.db.

docker run --name vanilla-dns -d -v /home/trainee/coredns/:/root/ --restart=always --net rlabs --dns 172.18.0.20 --hostname ns.rlabs.org --ip 172.18.0.20  coredns/coredns -conf /root/Corefile

Configure Bind DNS with its GUI on VNC desktop.

Point your browser to the VM public IP (it's in GCP console).
Sign in to VNC with password 'trainee!'.
Open Chrome browser on the VNC desktop.
Point it to https://172.18.0.20:10000 (this is Bind's GUI).
Sign in as 'root' with 'password'.
Configure the server using these steps.

https://drive.google.com/open?id=1QBFCyuU9uUikC5jEy1QwiEe4ln2S7MzekM0qke1Y9Jg

Test DNS config. If DNS doesn't work, remove the container and try again.

From a VNC shell terminal, run the following.

run_dnsutils
nslookup n1.rlabs.org
nslookup s1.rlabs.org
exit

Start Redis Enterprise nodes.

start_north_nodes
start_south_nodes

Build the cluster.

create_north_cluster

Resolve cluster names - they only resolve when you have a cluster.

run_dnsutils
dig @ns.rlabs.org north.rlabs.org
exit

Push DNS changes to a GCR image.

Return to SSH from GCP console so you're using your GCP account.

If you run these as 'trainee' you'll get 'config.json' errors later when running containers. If that happens, log in as 'root' and remove /home/trainee/.docker/.

# authenticate Docker to GCR
exit
exit
gsutil cp gs://admin-training-bucket/ru-gcr-write-key.json /tmp
cat /tmp/ru-gcr-write-key.json | sudo docker login -u _json_key --password-stdin https://gcr.io

# Commit DNS changes to image, tag it for upload, and push the image.
sudo docker commit vanilla-dns admin-training-dns
sudo docker tag admin-training-dns gcr.io/redislabs-university/admin-training-dns
sudo docker push gcr.io/redislabs-university/admin-training-dns

Reset your VM before saving your work.

Replace 'vanilla-dns' with 'configured-dns' from the GCR iamge.

sudo docker stop vanilla-dns
sudo docker rm vanilla-dns
sudo docker rmi sameersbn/bind
sudo docker rmi admin-training-dns

sudo docker run --name configured-dns -d --restart=always --net rlabs --dns 172.18.0.20 --hostname ns.rlabs.org --ip 172.18.0.20 -p 10000:10000/tcp  gcr.io/redislabs-university/admin-training-dns

Stop and remove node containers. Removing them forces manual restart on a new VM. Otherwise, cluster names do not resolve after you create a cluster.

sudo docker stop n1 n2 n3 s1 s2 s3
sudo docker rm n1 n2 n3 s1 s2 s3

Return to VNC shell. Remove 'known_hosts' and restart 'north' nodes.

'known_hosts' gives 'REMOTE HOST ID HAS CHANGED! Host key verification failed' errors. Clustered nodes would re-create a cluster on startup.

start_north_nodes
rm /headless/.ssh/known_hosts

Now you have:

vanilla VNC
configured DNS
Redis Insight
RE node image, containers stopped and removed.

Save your work.

Create a snapshot of the VM called 'admin-training-stage-1'.
Create an image from the snapshot called 'admin-training-stage-1'.

Stage 2

Now you have:

vanilla VNC
configured DNS
Redis Insight
RE node image, containers stopped and removed.

You'll configure VNC with:

A background image
2 workspaces
5 launchers (Chrome and 4 terminal shells).

Create a new VM called 'admin-training-stage-2' from image 'admin-training-stage-1'.

Configure VNC.

IMPORTANT: There is a dependency for VNC. It includes the private key for 'trainee' to silently authenticate to the base VM. If you configure a new VM that alters keys, you must create a new VNC as well.

Sign in to VNC desktop from your laptop browser with password 'trainee!'.
Open a terminal shell window.
Restart RE node containers.

start_north_nodes
start_south_nodes

Open Chrome browser in VNC and point it to admin consoles:

https://172.18.0.20:10000
http://insight:8001
https://n1:8443
https://n2:8443
https://n3:8443
https://s1:8443
https://s2:8443
https://s3:8443

Save tabs as bookmarks.
Set pages to open on startup.
Follow steps here to set up VNC with background, workspaces, and launchers.

https://docs.google.com/document/d/1X8K2jZTwBLr_jG9a01-u_dxRrTwb2URNCy0zXoQZUG4/edit#heading=h.ihyyp7do45r3

Push VNC to GCR, download and test.

SSH to the VM from GCP console so you're using your GCP account.
Download the service account key again and authenticate Docker to GCR.

gsutil cp gs://admin-training-bucket/ru-gcr-write-key.json /tmp/
cat /tmp/ru-gcr-write-key.json | sudo docker login -u _json_key --password-stdin https://gcr.io

Return to VNC terminal.

NOTE: Had to move this up before taking VNC Docker image.

rm /headless/.ssh/known_hosts

Commit changes and upload to GCR.

sudo docker commit vanilla-vnc admin-training-vnc
sudo docker tag admin-training-vnc gcr.io/redislabs-university/admin-training-vnc
sudo docker push gcr.io/redislabs-university/admin-training-vnc

Replace with GCR image and test.

sudo docker stop vanilla-vnc
sudo docker rm vanilla-vnc
sudo docker rmi vanilla-vnc
sudo docker rmi consol/ubuntu-xfce-vnc
sudo docker rmi admin-training-vnc

sudo docker run --name configured-vnc  -d -e VNC_PW=trainee! --restart=always --net rlabs --hostname vnc-terminal.rlabs.org --ip 172.18.0.2 -p 80:6901  gcr.io/redislabs-university/admin-training-vnc

Stop and remove node containers. Removing them forces manual restart on a new VM. Otherwise, cluster names do not resolve after you create a cluster.

sudo docker stop n1 n2 n3 s1 s2 s3
sudo docker rm n1 n2 n3 s1 s2 s3

Now you have:

configured DNS
configured VNC
Redis Insight
RE node image, containers stopped and removed.

You're ready to create user instances.

Save your work.

Create a snapshot of the VM called 'admin-training-stage-2'.
Create an image from the snapshot called 'admin-training-stage-2'.

Name		Name	Last commit message	Last commit date
Latest commit History 176 Commits
images		images
README.md		README.md

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

Admin Training VM Setup

Stage 1

Stage 2

About

Releases

Packages

barry-waldbaum-rl/reat-vm-setup

Folders and files

Latest commit

History

Repository files navigation

Admin Training VM Setup

Stage 1

Stage 2

About

Resources

Stars

Watchers

Forks

Releases

Packages 0

Packages