This tutorial will guide you through setting up a Redis stack and a RhoConnect application stack in the Amazon Web Service (AWS) cloud.
The RhoConnect application stack is scalable; it uses load balancing to automatically distribute and balance the incoming application traffic among all the instances you are running. You will set the maximum number of instances that the load balancer can create for your RhoConnect application.
Go to the Amazon Web Service (AWS) Console and click on the Sign in to the AWS Console button. You can sign into an existing AWS account, or create a new one.
Or you can open the URL for your account directly if you have the account number:
Enter your User Name and Password for your AWS account, then click the Sign in button.
In the AWS Management Console, change the Region to US West (N. California).
NOTE: The stacks created in this document only work in the N. California region.
We provide a few public preconfigured images that contains a blank RhoConnect app for
Us West N. California region:
ami-cbc59e8e(RhoConnect-3.1.2 with Nginx and phusion passenger)
ami-0dfca748(RhoConnect-3.2.0 with Nginx and cluster of thin servers)
To create your own RhoConnect Image, launch an EC2 Image:
Click the EC2 tab, click EC2 Dashboard, then click the Launch Instance button.
Click the Launch Classic Wizard radio button. Click Continue.
In the Choose an AMI section, click the Community AMIs tab.
Then paste in the AMI ID, i. e.
ami-cbc59e8e. When the preconfigured RhoConnect AMI appears, click Select.
In the Instance Details section, select the Instance Type (such as Large).
In the Instance Details section, Advanced Instance Options screen, change nothing. Click Continue.
In the Instance Details section, Add Tags screen, enter a name for your custom RhoConnect EC2 instance in the Value Column of the Name key, then click Continue.
In the Create Key Pair section, select your existing key pair or create a new keypair.
In the Configure Firewall section, click the Create New Security Group radio button.
Click the Add Rule button to add each port.
You may use a preexisting security group if it has ports 22, 80, and 6379 open.
Enter a firewall name and a group description, then click Continue.
Make sure all the instance information is correct as shown on the Review page and then click Launch.
In the window titled “Launch Instance Wizard”, click Close and wait for your instance to be Running.
In the EC2 tab, click on Instances.
For Viewing, select Running Instances. Select your newly created RhoConnect instance and scroll down to see Private DNS in the lower section of the My Instances screen.
You will need:
From the command line, use ssh to login to the running instance, using the keypair, DNS name, and user name. For example:
$ ssh -i ~/.ssh/devkey.pem firstname.lastname@example.org
Replace the blank ‘rhoapp’ application in the /opt/nginx/html folder with your app, and set its owner as ‘nginx’ user.
$ cd /opt/nginx/html # ... copy your application here under name 'rhoapp' $ sudo chown -R nginx:nginx rhoapp
You can test the application.
$ sudo /etc/init.d/redis start $ sudo /etc/init.d/nginx start
If your instance based on
ami-0dfca748 (RhoConnect-3.2.0) image, then you need to do one more step:
$ sudo /etc/init.d/thin start
If everything is configured properly, then you can login to the RhoConnect web console in your browser by using the instance’s public DNS name as the URL in your browser. When you are done, you can stop the application.
$ sudo /etc/init.d/nginx stop $ sudo /etc/init.d/redis stop
Do not forget also stop cluster of thin servers if your instance based on
$ sudo /etc/init.d/thin stop
In the EC2 tab, right-click the RhoConnect instance and select Create Image (EBS AMI).
In the Create Image window, enter the image name and description for your custom image. Then click Create This Image.
In the Create Image window showing that the request is received, click on “View pending image: ami-XXXXXXXX to go to your Amazon Machine Image.
Copy the AMI number; you will use it when you create the scaleable RhoConnect stack.
Click the Cloud Formation tab.
Ensure that the Region selected is US West (N. California) and Click the Create New Stack button.
In the Select Template section, you must select the template file for your redis stack. This is the template to run the AMI image with the Redis server.
First, enter the name that you want for your stack into the Stack Name field.
Now get the template by either providing the URL for the template file, or by uploading the template file.
Click the Provide a Template URL radio button.
If your custom image based on
ami-cbc59e8e(RhoConnect-3.1.2) image, then copy the following URL and paste it into the text field below the Provide a Template URL radio button:
If your custom image based on
ami-0dfca748 (RhoConnect-3.2.0), then use this URL:
Click the Upload a Template File radio button.
In another browser window, go to https://s3-us-west-1.amazonaws.com/rhoconnect-ca-deploy/rhoconnect_template.txt
if your custom image based on
If your custom image based on
ami-0dfca748 (RhoConnect-3.2.0), then to go
Save that text file to your computer.
Click the Choose File button, and navigate to and select your template file.
In the Create Stack: Specify Parameters window:
Click Continue to see a window entitled \“Create Stack\” showing that the stack is being created.
In the AWS Management Console, Cloud Formation tab, click the checkbox for your stack. To see the progress of the creation of your stack, click the Refresh button; the status should become CREATE_COMPLETE soon.
To see the progress of the components in your stack, you can click the Events tab. You can click the refresh button to see the progress.
In the AWS Management Console, Cloud Formation tab, click the Outputs tab. You will see a list of the Stack Outputs.
At the top of the output list is the URL of the RhoConnect application. You can click on it to see the RhoConnect Console for the running RhoConnect server.
To use this scaleable RhoConnect server in your Rhodes application, copy the URL for the RhoConnect application into the rhoconfig.txt file in your Rhodes application as the syncserver variable. For example:
syncserver = http://rhoconnec-RhoLoadB-OX64XR1IKNT2-1364705573.us-west-1.elb.amazonaws.com
You are charged for the stack that you have created, by the hour. When you no longer want to use the stack, you can delete it by right-clicking on the stack name in the AWS Management Console, Cloud Formation tab, and clicking Delete Stack.
The described architecture has a weak spot: a single point of failure (SPOF). If the Redis server is down or stops responding, then the entire Rhoconnect stack stops functioning. To address this issue, we need to implement failover — the continuation of a service after the failure of one or more of its components. Redis provides a very simple solution to
configure master-slave replication that allows slave Redis servers to be exact copies of master servers.
So if the master server goes down, we can issue a
SLAVEOF NO ONE command to the slave and then switch RhoConnect apps to use the slave that is now the master.
The following is a one possible scenario of how it might be done on the Amazon cloud.
Redis High Availability proposal:
/home/ubuntu/failoverif there is an failure to get a valid response upon performing a PING command three times in a row.
simpleBDservice to store current redis servers configuration (master/slave/backup). So, if another instance of monitor server detects failure and calls the failover script, it will know the current status of the redis servers and will wait until the recovery process that was initiated by another Monit service is over. Finally, the failover script reconfigures and restarts the Monit service to track a new pair of master/slave redis servers.
The URL for the CloudFormation template that implements the described proposal is available here https://s3-us-west-1.amazonaws.com/rhoconnect-ca-deploy/rhoconnect-ha.txt and can be used for evaluation purposes.
The template will create a sample scalable rhoconnect stack with 2 elastic load balancers and 5 EC2 instances.
Under heavy load Redis server might be a bottleneck in RhoConnect stack performance due to hight network traffic, large memory footprints of stored data in server instance. To address redis server performance issues RhoConnect (>= v4.0.0) supports multiple Redis instances per app, and evenly distributing load across available Redis server instances.
Edit your app
settings/settings.yml file and replace :redis one liner by array of redis servers you are going to use:
# ... :production: :licensefile: settings/license.key :syncserver: http://localhost:9292/api/application/ :push_server: http://appname:secret@localhost:8675/ # Typical case: one redis per app #:redis: localhost:6379 # Two or more redis servers per app :redis: - localhost:6379 - localhost:6380 # ...
The following recommendations are based on Ubuntu Amazon image, where RhoConnect Redis package installed and you wanted host 2 redis instances on ports 6379 and 6380 respectively.
/opt/rhoconnect/etc/redis.conf configuration file to
/opt/rhoconnect/etc/redis-6379.conf and make sure that it has settings for the 1st instance:
... daemonize yes pidfile /var/run/redis/redis6379.pid port 6379 # DO NOT FORGET comment persistence cmds # save ... ...
/opt/rhoconnect/etc/redis-6380.conf file and do similar setting for another redis instance:
... daemonize yes pidfile /var/run/redis/redis6380.pid port 6380 # DO NOT FORGET comment persistence cmds # save ... ...
/etc/init/rhoconnect-redis.conf file if it’s there.
Create upstart script
/etc/init/rhoconnect-redis-6379.conf for the 1-st instance:
# /etc/init/rhoconnect-redis-6379.conf description "Redis Datastore Server" # Redis is a key value in memory persistent datastore start on (local-filesystems and runlevel ) stop on runlevel  expect fork respawn pre-start script mkdir -p /var/run/redis chown redis:redis /var/run/redis end script exec start-stop-daemon --start --chuid redis:redis --pidfile /var/run/redis/redis6379.pid --umask 007 \ --exec /opt/rhoconnect/bin/redis-server -- /opt/rhoconnect/etc/redis-6379.conf
Define upstart script
/etc/init/rhoconnect-redis-6380.conf for the 2-d instance:
# /etc/init/rhoconnect-redis-6380.conf description "Redis Datastore Server" # Redis is a key value in memory persistent datastore start on (local-filesystems and runlevel ) stop on runlevel  expect fork respawn pre-start script mkdir -p /var/run/redis chown redis:redis /var/run/redis end script exec start-stop-daemon --start --chuid redis:redis --pidfile /var/run/redis/redis6380.pid --umask 007 \ --exec /opt/rhoconnect/bin/redis-server -- /opt/rhoconnect/etc/redis-6380.conf
We provide a public preconfigured image with above settings for Us
West N. California region:
ami-d8d0f29d(Redis-Server-2.6.7 with 2 instances on ports 6379 and 6380)
This Amazon CloudFormation template file shows how 2 redis servers with 2 instances per sever might be used for a typical rhoconncect application.