Without further ado, here’s what you need to do to deploy your applications from the command line:

1. Email beta@engineyard.com to be added to the CLI beta program.
2. $ gem install engineyard
3. $ ey deploy

And you are off to the races! Bundler works out of the box, and you can deploy to Solo and clustered environments on Engine Yard Cloud.

For more details, read on.

Command Line Beta Details

Before getting started, make sure that your environment has been created and instances have been launched.

First, create a new environment for the application. If you are using Rails 3, please set the application type to rack 1.1.0. All rails 3 apps work correctly as just a rack app. Also note that you will need to select Ruby 1.8.7 to use Rails 3.

Note: The git url that you set on EY Cloud must match one of the git remotes in your local repo.

Set your deploy key if it is a private repo. Then setup your gems. If you are using bundler, Bundler 0.9 will already be selected for you.

Setup your SSH key. You must be able to ssh to the instance as your deploy user to use the CLI deploy beta.

Boot your desired cluster. The deployments work fine with either a Solo instance or a Cluster.

While the cluster is booting we need to do a few things on the command line.

You need to be in the directory for the app you want to deploy. One of the git remotes for the current directory must match the git url you entered on the Cloud Web Ui.

Next install the engineyard gem

blogin15(master): gem install engineyard
      Building native extensions.  This could take a while...
      Successfully installed mime-types-1.16
      Successfully installed rest-client-1.4.2
      Successfully installed highline-1.5.2
      Successfully installed json-1.2.4
      Successfully installed engineyard-0.2.6
      5 gems installed
      

Then run ey environments to see what environments are configured for this app

blogin15(master): ey environments
      We need to fetch your API token, please login
      Email:  adelcambre@engineyard.com
      Password: *********
      Cloud environments for blogin15:
        blogin15_production  1 instance  blogin15
      

You will need to enter your Engine Yard Cloud credentials to fetch your API token. You will only need to do this once.

Your cluster should now be booted. Once it is, you can do the deploy.

blogin15(master): ey deploy

      Connecting to the server...
      Server does not have ey-deploy gem installed
      Installing ey-deploy gem...
      ssh -q deploy@184.73.238.74 "sudo gem install ey-deploy -v '~>0.2.4.pre1'"
      Successfully installed ey-deploy-0.2.4.pre2
      1 gem installed
      Installing ri documentation for ey-deploy-0.2.4.pre2...
      Installing RDoc documentation for ey-deploy-0.2.4.pre2...
      Running deploy on server...
      ssh -q deploy@184.73.238.74 "eysd deploy --app blogin15 --branch master"
      Warning: Permanently added 'github.com,207.97.227.239' (RSA) to the list of known hosts.
      ~> Pushing code to all servers
      ~> Starting full deploy
      ~> copying to /data/blogin15/releases/20100415212737
      ~> ensuring proper ownership
      ~> Gemfile detected, bundling gems
      Fetching source index from http://gemcutter.org/
      Installing rake (0.8.7) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing ZenTest (4.3.1) from rubygems repository at http://gemcutter.org/
      Installing RubyInline (3.8.4) from rubygems repository at http://gemcutter.org/
      Installing sexp_processor (3.0.4) from rubygems repository at http://gemcutter.org/
      Installing ParseTree (3.0.5) from rubygems repository at http://gemcutter.org/
      Using abstract (1.0.0) from system gems
      Using builder (2.1.2) from system gems
      Installing i18n (0.3.6) from rubygems repository at http://gemcutter.org/
      Installing memcache-client (1.7.8) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing tzinfo (0.3.18) from rubygems repository at http://gemcutter.org/
      Installing activesupport (3.0.0.beta) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing activemodel (3.0.0.beta) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing erubis (2.6.5) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Using rack (1.1.0) from system gems
      Installing rack-mount (0.4.7) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing rack-test (0.5.3) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing actionpack (3.0.0.beta) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Using mime-types (1.16) from system gems
      Installing polyglot (0.3.1) from rubygems repository at http://gemcutter.org/
      Installing treetop (1.4.5) from rubygems repository at http://gemcutter.org/
      Installing mail (2.1.5.3) from rubygems repository at http://gemcutter.org/
      Installing text-hyphen (1.0.0) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing text-format (1.0.0) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing actionmailer (3.0.0.beta) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing arel (0.2.1) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing activerecord (3.0.0.beta) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing activeresource (3.0.0.beta) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing bundler (0.9.15) from rubygems repository at http://gemcutter.org/
      Installing mysql (2.8.1) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      with native extensions Installing thor (0.13.4) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing railties (3.0.0.beta) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing rails (3.0.0.beta) from .gem files at /data/blogin15/releases/20100415212737/vendor/cache
      Installing rails-action-args (0.1.0) from rubygems repository at http://gemcutter.org/
      Installing ruby_parser (2.0.4) from rubygems repository at http://gemcutter.org/
      Installing ruby2ruby (1.2.4) from rubygems repository at http://gemcutter.org/
      Your bundle is complete! Use `bundle show gemname` to see where a bundled gem is installed.
      Copying .gem files into vendor/cache
        * rake-0.8.7.gem
        * ZenTest-4.3.1.gem
        * RubyInline-3.8.4.gem
        * sexp_processor-3.0.4.gem
        * ParseTree-3.0.5.gem
        * abstract-1.0.0.gem
        * builder-2.1.2.gem
        * i18n-0.3.6.gem
        * memcache-client-1.7.8.gem
        * tzinfo-0.3.18.gem
        * activesupport-3.0.0.beta.gem
        * activemodel-3.0.0.beta.gem
        * erubis-2.6.5.gem
        * rack-1.1.0.gem
        * rack-mount-0.4.7.gem
        * rack-test-0.5.3.gem
        * actionpack-3.0.0.beta.gem
        * mime-types-1.16.gem
        * polyglot-0.3.1.gem
        * treetop-1.4.5.gem
        * mail-2.1.5.3.gem
        * text-hyphen-1.0.0.gem
        * text-format-1.0.0.gem
        * actionmailer-3.0.0.beta.gem
        * arel-0.2.1.gem
        * activerecord-3.0.0.beta.gem
        * activeresource-3.0.0.beta.gem
        * bundler-0.9.15.gem
        * mysql-2.8.1.gem
        * thor-0.13.4.gem
        * railties-3.0.0.beta.gem
        * rails-3.0.0.beta.gem
        * rails-action-args-0.1.0.gem
        * ruby_parser-2.0.4.gem
        * ruby2ruby-1.2.4.gem
      ~> Symlinking configs
      ~> running deploy hook: deploy/before_migrate.rb
      ~> migrating
      ~> Migrating: cd /data/blogin15/releases/20100415212737 && RAILS_ENV=production RACK_ENV=production MERB_ENV=production rake db:migrate
      (in /data/blogin15/releases/20100415212737)
      ==  CreatePosts: migrating ====================================================
      -- create_table(:posts)
         -> 0.0221s
      ==  CreatePosts: migrated (0.0222s) ===========================================

      ~> running deploy hook: deploy/before_symlink.rb
      ~> symlinking code
      ~> Restarting app servers
      ~> restarting app: /data/blogin15/releases/20100415212737
      ~> running deploy hook: deploy/after_restart.rb
      ~> cleaning up old releases
      ~> finalizing deploy
      Deploy complete
      

Now your site should be deployed! Please send feedback on this new system, or problem reports, to beta@engineyard.com.

Caveats:

1. If your environment has more than one application with a private git repository, CLI deploys are not yet functional. This will be fixed in a future beta release.

Today we are excited to announce a new Beta Program for Engine Yard Cloud. This is an opt-in program for users to use beta-only features before we open them up for everyone. If you like staying on the cutting edge, have a favorite feature that you just have to try today, or would like to help us by providing feedback on upcoming features, this is the program for you.

What features are you making available in the beta program today?

We have an exciting lineup ahead, but for today we’d like to announce a couple of features that we know a lot of you are waiting for.

The first one is support for Ruby 1.8.7. You can now update individual environments to use either Ruby 1.8.6 or 1.8.7. To do so, simple go to the edit environment page for an existing application and change the Ruby version there, or create a new environment. As part of this change, we have also isolated the system Ruby used to run the Engine Yard management stack from the version that your applications would use. This change has resulted in greater stability on the platform, and will also allow us to add newer Ruby versions easily to the platform.

As a result of this, we are now also ready for Rails 3 when it releases. If you like to stay on the bleeding edge and do not require support for non-production applications, you can use the Rails 3 pre-release gems today (just make sure to switch your environment to use Ruby 1.8.7).

How does the Beta Program work?

Its simple – just email beta@engineyard.com with your Engine Yard Cloud login email address. We will be rolling out access to Ruby 1.8.7 over the coming few days to all beta customers who sign up.

Thanks for helping us test out our newest set of features, and watch this space for more product-related announcements soon!

Unicorn is a very interesting ruby application server. It brings the robust HTTP parser from Mongrel into a Unix centric preforking webserver configuration. This means that when you start Unicorn, it starts a master process that opens the socket to listen on, preloads your application code and then forks worker children to do the actual application serving. The linux kernel takes care of load balancing between child processes, since all the children are calling accept on the same socket fd inherited from the master process. This means you get global queueing for free, because a process will never get more then one request at once and the kernel will efficiently queue requests until a child worker is ready to accept a new request.

Unicorn ends up being the fastest option available to run your ruby applications on. In my benchmarking Unicorn handily beats Mongrel and Passenger in throughput tests as well as memory consumption tests. So you can think of Unicorn as what Mongrel should have been. It has the ease of management of Passenger but without some of the downsides (dealing with amqp, threading weirdness). I highly recommend Unicorn as the highest performance stack we have for you to run on, but please do test this for your self with your own applications as all apps are different. Also, Unicorn is currently an experimental stack option (for more details, see our technology stack page).

To get started right now with Unicorn on our cloud all you need to do is to create a new environment, or terminate the instances in an existing environment and click the pencil icon to edit the environment. On the environment details page you will be able to select the stack you want to run, choose Nginx+Unicorn and then go ahead and boot up some new servers to try it out.

Go ahead, try it out, and as always, send us any feedback you have on this latest feature.

So go ahead, create that 100 node clone of production to run your massive scalability tests on. Be up and running in  minutes, and tear down your environment when you are done.

Cloning your applications with Engine Yard Cloud from Engine Yard on Vimeo.

New Pricing
Effective November 1st, we will be reducing the prices that we charge for instance hours in Engine Yard Cloud, passing through the announced price reduction in Amazon EC2 pricing. These prices will be in effect automatically for the November billing cycle. New and existing customers will not have to do anything to activate the new pricing, you will get the price reduction automatically starting with your November usage.

High Memory Instances
Amazon also introduced a new class of high-memory instances, the High Mem 2XL and 4XL. They have 34.2GB and 68.4GB of RAM respectively, making them ideal for memory intensive, high throughput applications, such as database servers and caching applications. Starting today, you can add each of these instances to your environments – simply pick the type of instance that is appropriate for your needs and start your environments with them. They are priced at $1.32/hour for the 2XL instance, and $2.55/hour for the 4XL instance.

Bug Fixes
We also resolved a number of customer-facing issues with this deploy. They include:

• Deleting SSH keys no longer blocks additional instances from joining a cluster (LH 684)
• Utility instance alerts now show up in the summary section of the detailed alerts view
• Installing support SSH keys on customer instances no longer sets (and leaves) the cluster status as “deploying configuration changes”
• Volumes now mount consistenly after being cached
• Snapshots on non-master instances no longer report a chef failure in the UI
• The billing page now works correctly for charges very near zero (LH 704)

High-memory instances, reduced pricing and more. Try them out and send us your feedback – we would love to hear from you!

As with all “Hello World” applications, this one focuses on the first workflow most new users see: deploying a new application on Engine Yard Cloud. This 2 minute screencast shows you how to get your application up and running on a load-balanced application cluster with a dedicated database tier, using Engine Yard Cloud.

Engine Yard Cloud includes three broad classes of monitors: resource monitors, service monitors and URL monitors. Let’s look at each of them in detail.

1. Resource Monitors

Each instance in your environment includes a set of resource monitors enabled by default. Behind the scenes, we use collectd to collect statistics and trigger alerts for host-based resource usage. Currently we include 5 types of resource monitors enabled by default: load, free memory, free disk space, swap and cpu. Each has associated limits for warning and failure triggers, which generate alerts into an environment’s alert stream.

2. Process Monitors

While resource monitors work on the instances, it is often desirable to monitor services that are externally visible. Engine Yard Cloud uses an extensible infrastructure to monitor processes that are essential to the health of your application. Currently, we check for availability of sshd, the database tier (MySQL), and core services (nginx, haproxy and internal Engine Yard services). A change in state of any of these processes triggers alerts on that environment.

3. URL Availability

Today we are adding the third in our set of monitoring tools: URL availability. While most of our users have used 3rd party services to monitor their applications, one of the issues we have heard users talk about is integrating those alerts with their Engine Yard dashboard. Starting today, all Flex users can specify one URL for each of their environments to monitor for uptime. Simply look for the “Monitored URLs” tab for your environment on the dashboard.

Our default implementation will run checks every 2 minutes to ensure uptime, and trigger alerts every time there is a state transition. With the upcoming premium support and proactive monitoring, our support staff will also be notified immediately when a downtime does occur, and will initiate action to ensure that your services are back up as soon as possible.

So there you have it. Process, resource and URL monitoring. Just three of the ways in which we continue to strive to give you increased visibility into your cloud environments.

1. Engine Yard Cloud is now out of beta and in general release!
2. We’re introducing Premium Support for Engine Yard Cloud, with 24×7 proactive monitoring and trouble-shooting for production applications
3. Solo and Flex pricing plans have been unified into a single Engine Yard Cloud plan
   • Pricing for resources is at the old “Solo” level for everyone
   • Paid support is now optional and available to all users, with a three month minimum commitment

For more details, jump over to the post on the Engine Yard Blog.

Today we are excited to announce out-of-the-box support for master-slave replication on the Engine Yard Cloud. When you add additional nodes to your database tier, we will configure them as slave nodes. All writes will continue to go to your master node, but you can read from any node in the tier. This is a great way to get read-scalability. For example, your backup and reporting applications can now run exclusively on the slave tier without impacting master performance. Moreover, you get enhanced write performance as well because your master can focus on performing only write queries.

Adding slave replicas is easy. While deploying your application, simply add more than one node for the database tier. The first node will be configured as the master, and all remaining nodes as slaves. You can increase and decrease that number at any time, simply by adding or removing database nodes to a running cluster.

There are several ways to get your rails application to work with database replication. When you add slave databases, we will automatically add entries to your database.yml files, which will contain information about your new replica (you will need to restart your application to re-read this information). While you can use this information manually to configure your own database access scheme, we recommend looking at MySQL Proxy. This makes the configuration of your database tier transparent to your application, and allows you to scale up and scale down without affecting application configuration. You can also use one of the several Rails plugins available for this (such as Masochism, master_slave_adapter, and many others).

Enjoy the latest horsepower-boost to your database tier, and keep us posted on your experience with this feature. We’d love to get your feedback!

Database Slaves

We’re excited to announce out-of-the-box support for database master-slave configurations. We have been working hard over the last few weeks to ensure that all corner cases are handled correctly and that database clusters scale well under heavy loads. You can add a database slave to a running application, or create a database tier when you first boot the application. Watch this space for more details on this feature.

Quick Start workflow updates

New customers will be dropped directly into the quick start workflow rather than on the new environment page.

Bug fixes

We also fixed the following bugs:

- SSH Key deletion was broken #634

- Add instance to cluster page did not have a default size for the utility server volume

- Add instance to cluster page did not remember instance type selected if there were validation errors

- The dashboard and alerts layouts were broken #639 #632

- Chef more accurately reports its progress back to the app, if it gets stuck the message should reflect what was happening #631

- The support dashboard had a really big query that was bloating the mongrels and swapping the boxes.

- The FRAMEWORK_ENV was not set in the chef-deploy hooks #619

Thanks, and as always keep us posted on your experiences with the product.

– The Engine Yard Cloud team