Frequently Asked Questions

RightGrid is a scalable batch computing system used for grid type applications. Whether you’re analyzing forms and statistics, transcoding videos, photos, or music, producing dynamic content, or even just doing raw calculations, just plug in your slave task and job producer into RightGrid and scale! RightGrid will monitor a job queue and launch servers as needed to handle the number of tasks based on parameters that you define. For instance, you may be transcoding videos and need to launch one instance for every 10 videos in the queue. RightScale will do that for you, so go ahead and queue up 100, or 1000, or more! When the workers are done with their task they will check back in with the queue and decide whether or not they are still needed before terminating themselves. With RightScale and EC2, it no longer take days for your servers to handle these tasks, get it done in an hour and pay less for it!

In order to upgrade from a RightScale Developer Account and unlock the premium features, follow the instructions below.

1. Log in to the RightScale Dashboard with your Developer Account information.
2. Navigate to the Settings tab and select Account.
3. Select the Plan tab and click Upgrade.
4. Input your RightScale Upgrade Code.
5. Verify your new plan details and the payment terms. If there are discrepancies, please call us at 1-866-720-0208.
6. Click on Subscribe.
7. Use Paypal to complete the transaction. (You do not need a Paypal Account.)
8. Click on Return to Merchant. You should now see that your plan has been upgraded and that new features are available.

Congratulations and enjoy the premium features of the Dashboard!

Load balancing is a two part process that involves first balancing incoming requests, and then balancing the load across your available application servers.

The first part uses round robin DNS from an external DNS service with a low TTL (time to live) setting. This handles the incoming requests and balances them across your web front ends. Generally, the number of web front ends is left static to handle incoming traffic. If one fails, then browsers will fail over to the other active server(s). You can then launch a new web front end and RightScale will automatically work it back into the round robin rotation. This process can also be automated using RightScale's system of alerts and escalations.

Each web front end also has a software load balancing utility called HAproxy. HAproxy will receive the incoming request and then dynamically distribute it to the application server that is the least busy, no matter what server it’s running on. It will also perform health checks on these application servers to make sure everything is configured to receive the request correctly before it sends the request out. It will also automatically detect new instances that you launch in an array during autoscaling and then dump traffic to these newly operational servers to decrease the load on your existing application servers.

Yes. Before you can take advantage of all the features in the RightScale Dashboard, you will first need to create an account with Amazon and sign-up for their EC2, S3, and SQS web services. When you login to the Dashboard, you will need to input your Amazon credentials.

If you need create an account with Amazon, please see How do I sign up for AWS?

To get signed up for AWS services and get them working with RightScale's dashboard requires a few steps.

1. Start by going to http://aws.amazon.com.
2. Near the top righthand side of the page you'll see the "Sign Up for AWS" section.
3. Create an account with Amazon.
4. Log in to your new account.
5. Click on the drop down menu "Your Web Services Account" > "Payment Method"
6. Enter your billing information.
7. On the left hand side of the page click on "Amazon Elastic Compute Cloud" (EC2)
8. On the right side click "Sign up for this Web Service"
9. Repeat this step for "Amazon Simple Storage Service" (S3) and "Amazon Simple Queue Service" (SQS) That's it, you're done! Now you're ready to input your credentials into RightScale!

RightScale helps you cluster your applications into Deployments and Arrays.

A Deployment is a clustered architecture that represents all the servers in your application. Using RightScale you can start and stop entire deployments on demand, clone deployment clusters to create identical staging or QA environments, and define and link variables that apply to scripts distributed across all of your servers. Creating a deployment is as easy as selecting and adding the servers you want, defining your variables, and clicking Launch.

An Array is a single server type used for horizontal scaling. It may be a single worker slave task plugged into our RightGrid system to scale based on the number of jobs in a queue. Or it may be your Deployment's application server that you want to scale dynamically based on the load or even schedule to scale based on known traffic patterns. Defining an array is simple. Select the type of server you would like to scale and then just define what deployment you're scaling and how you would like RightScale to automate the process for you.

Define your application server as the Server Template that you would like to scale horizontally to handle additional load. Incoming requests are handled by load balanced web front ends; these front ends also run a software application load balancer called HAproxy which will dynamically redirect incoming traffic to the available application servers.

RightScale uses a unique voting system combined with a system of alert conditions and escalations; actions that RightScale will take on your behalf. When one application server or web front end running an application server becomes too busy and a conditional alert is met it will escalate and take an action, placing a vote to grow that Array. RightScale will not launch additional servers until enough instances have voted that they are busy enough to scale up based on a decision threshold that you define. For instance, if you were running four application servers and defined a decision threshold of 51%, then 3 of those 4 servers would need to vote that they are busy before launching additional servers. The voting system is also used to scale the array back down then the load subsides.

Don't worry about it! Your load will increase and your RightScale deployment will sense that additional application servers are needed to handle the load and launch them for you. Incoming traffic will be dynamically balanced across these new servers and your site will continue to serve your visitors. No bandwidth overages, no stressed out CPUs, no headaches. Administrators, you are now free to take a vacation…without the PDA!

EC2 is based on the utility computing model where you only pay for what you comsume. No upfront costs. No long term contracts. No minimum charge.

The use of EC2 instances are charged by the hour. See Amazon EC2 for a complete pricing breakdown.

• Small Instance - $0.10/hr
• Large Instance - $0.40/hr
• X-Large Instance - $0.80/hr

To estimate your monthly EC2 charges, you can use Amazon's AWS Simple Monthly Calculator.

RightScale's MySQL Manager provides tools, templates, and scripts for replication and failover for your database running on EC2. A standard configuration consists of a Master database replicating in real time to a Slave. The Slave will take rolling backups every day and store them to S3 while the master will leave a "time machine" backup daily. In the event that your Master fails, it takes one click to promote your slave to the Master position to begin immediately accepting write requests. With a couple of more clicks you can bring up a new Slave and initialize it to begin replicating from the Master. This task can also be automated using our system of alerts and escalations if that's your preference. However, you are ultimately responsible for your own data. If both database servers were to ever go down, you would simply have to launch a new Master by clicking one button to pull up your latest backup from S3 and configure the database to be the Master.

RightScale and AWS recommend planning for failure. This is not to say that failure is to be expected, but it does mean that single points of failure are no longer a concern. RightScale recommends running at least two web front ends with a low TTL on the DNS and at least two database servers (using our MySQL Manager for redundancy and persistence).

When an instance is terminated, everything about it disappears, everything in memory, everything on the local disk, everything. The idea is to plan for an unexpected termination. Whenever you’re designing a system, you should always ask yourself, “What do I do when it fails?” Therefore, files that must be persisted should be stored to S3. We can also help you set up a synchronization link between S3 and local disks in your application array. RightScale also maintains your log files and persists them to the dashboard so you can view them on both operational and terminated servers. We also maintain an audit trail for RIghtScale-specific actions.

Based on RightScale’s personal experience, instances don’t just “disappear” arbitrarily. In fact, most of our experience with frantic emails about instances "disappearing" have been the result of operator error. Fortunately, you can use our alerting, log file persistence, and audit trail features to drill down on the problem and figure out the root cause of the problem. Basic troubleshooting steps include;

• Try to SSH into the instance, check the security group and make sure that port 22 is open for SSH to distinguish an SSH timeout from key issues (timeout vs. permission denied type of errors)
• Use ping to test connectivity (be sure to enable ICMP)
• Check the console output (use the convenient button on the RightScale dashboard)
• Look at the RightScale monitoring to see whether the instance is still sending monitoring data
• Use an instance in the same security group and try connecting from post details on the forum (instance ID, what you’ve tried, symptoms observed) and set the instance aside launch a new instance

You can also see this blog post for more information: Top Reasons Amazon EC2 Instances Disappear

AWS is currently the only company that provides utility computing resources to the masses at a generous scale. Unlike smaller “cloud vendors,” you can be sure that if you need the resources of a thousand CPUs, then you can get them delivered, fast. RightScale intends to support future cloud offerings at this level as they are provided.

There is currently no SLA for EC2. In a land of disposable computers, defining that SLA becomes difficult, especially when the best solution for most problems in EC2 often becomes, “launch a new instance.” So instead of 5-9s of uptime, 15 minute response times and 10,000% guarantees; just take matters into your own hands and fix the problem yourself in a couple of mouse clicks. Imagine no longer having to call your management host and wait for them to locate, diagnose, and troubleshoot your server. You no longer have to worry about unneccessary downtime and sacrificing potential business.

What is the RightScale solution? Simply login to the RightScale Dashboard and launch another instance. Problem solved.

S3 has a 99.9% uptime guarantee and AWS will credit your account for service unavailable. What is a “degraded instance”? Some users will get an email from Amazon stating that their “instance is running on degraded hardware.” This means that the underlying hardware contributing to your compute resources is having trouble. The email from Amazon will suggest steps to resolve the problem, but the easiest solution is to simply launch a new instance inside the Dashboard.

The initial consultation and setup fee includes 8 hours of hands-on assistance with setting up and deploying your application on RightScale and AWS. During this time you will be working closely with one of our integration engineers to learn about some of the more robust features of the RightScale management system, set up your environment, plug into RightScale's replication and failover features for your database, define the automation parameters you’d like to use, and deploy a production ready application on EC2.