Skip to content

Latest commit

 

History

History
105 lines (68 loc) · 3.53 KB

README.md

File metadata and controls

105 lines (68 loc) · 3.53 KB
Raw

Zookeeper Demo

A simple tutorial to run a replicated zookeeper service in EC2.

The configuration of the EC2 instances is left to be done manually. The deployment of zookeeper is automated with fabric.

The practical exercises of reading and writing to zk work on the command line and with the python zk client.

After completing this tutorial you'll have practical experience of writing with the zk data model and the kazoo API. You'll also be familiar how to use fabric to automate configuration of VMs

Configure AWS EC2

Launch n (>2) Ubuntu 20 instances in EC2. (ami-0dba2cb6798deb6d8). This has been tested with micro instances and m5 instances.

Configure a security group to allow SSH and TCP 2181 from your dev machine.

Also add a TCP rule 0-65535 for 172.31.0.0/16

Configure Instances

Prerequisites

  • python >3.6
  • AWS ssh key configured

Setup Python Dependencies

  • create new virtual environment
  • run pip install -r requirements.txt

Configure EC2 instances

  • copy config.yaml.template to config.yaml
  • edit config.yaml and insert public and private IP addresses of your EC2 instances
  • configure path to your AWS ssh key (chosen when starting EC2 instances)

Use boto_tutorial

A quick way to start the VMs is via the boto_tutorial To prepare machines for the zookeeper tutorial first create a security group as described above. Note the id of the security group. For illustration, I have used sg-0f8c026df0032d5c1 - you need to change that in your own code. You can use any name you like.

fab create --name zk --securitygroup sg-0f8c026df0032d5c1 
fab create --name zk --securitygroup sg-0f8c026df0032d5c1 
fab create --name zk --securitygroup sg-0f8c026df0032d5c1

then run fab instancedetails --name zk

And copy the ip address pairs at the end of the output (use rectangle selection to make that easier for you)

INFO:fabfile:Instance up and running at ec2-52-55-249-15.compute-1.amazonaws.com with internal ip 172.31.45.4: 52.55.249.15: 172.31.45.4
INFO:fabfile:Instance up and running at ec2-18-232-78-191.compute-1.amazonaws.com with internal ip 172.31.39.236: 18.232.78.191: 172.31.39.236
INFO:fabfile:Instance up and running at ec2-54-224-224-79.compute-1.amazonaws.com with internal ip 172.31.32.31: 54.224.224.79: 172.31.32.31

into the config.yaml file

  52.55.249.15: 172.31.45.4
  18.232.78.191: 172.31.39.236
  54.224.224.79: 172.31.32.31
  • [] run start_zk_cluster.py to deploy ZK on the machines
  • [] run query_zk_cluster.py to check that everything worked.

You should see something like the below.

Congratulations You now have a running ZK cluster.

Play with ZK command line client

ssh into two of your hosts (host A and B) and run sudo /usr/local/zookeeper/bin/zkCli.sh -server 127.0.0.1:2181

Ephemeral nodes

One node A enter create -e /app 'COMSM0072' to create an ephemeral node One node B enter get -w /app to watch the node

Now close the session on host A You should see a message in the session on host B WatchedEvent state:SyncConnected type:NodeDeleted path:/app

Play with the ZK Kazoo Python client

The client registers a watcher with callback handler on node app. In order to see that in action,

  1. create a node /app (as above)
  2. Then, run client.py,
  3. then change the node through the command line set /app "Happy Days"