There’s an instructive post on the HipChat blog regarding how their product’s message count has risen from 110 million to over a billion since they were acquired by Atlassian. Central to their scaling strategy to meet data growth was their migration from Lucene to Elasticsearch.
will continue to work…bugs will be fixed and important features will be added.
In my opinion, for the time being, Tire is still superior to the elasticsearch-ruby alternative in terms of features and its elegant DSL.
Single branch development is easy. But this strategy’s easiness blows up once you release code to the general public. If you need to hot-fix an instance of deployed code, while in the midst of a development cycle, single branch development gets in your way by requiring you think. On the other hand, having more than one branch at least allows you to jump back in time via an alternate branch to perform a patch, while not disturbing an unfinished developmental branch. And you can do this without having to think much.
Are you looking to get going with Elasticsearch as quickly as possible without having to worry about installing Java or Elasticsearch itself? Are you looking for a repeatable and automated mechanism for bringing up Elasticsearch instances for developmental and or testing purposes? While there’s certainly a number of Elasticsearch-as-a-platform service providers out there, there’s one other option: use Elasticsearch-in-a-box.
Elasticsearch-in-a-box is a freely available Vagrant base box. What that means is that you can quickly fire up and tear down an Elasticsearch environment with simple commands like
vagrant up and
Recently, the good folks over at Packt Publishing gave me a copy of their newly published Instant Mockito, by Marcin Grzejszczak. Packt’s Instant series are really enjoyable. The premise of these books is that they’re short and sweet. They’re slightly more than a tutorial; they get you up and running quickly while throwing in a few more facets that go beyond the typical tutorial.
As I’ve written about before, Vagrant is handy tool for creating localized VMs. It’s a lot like firing up EC2 images, but, for the most part, things are localized (you can, by the way, use Vagrant to fire up EC2 images). If you’ve ever used VMWare before, its the same thing, except Vagrant is free. You can create VMs of various operating systems, fire them up, and tear them down all with ease.
Vagrant plays nicely with hip DevOps frameworks like Chef and Puppet and if your installations require a number of components, then these tools are defiantly the way to go. Sometimes, however, a simple Bash script is good enough as in the case for auto-installing some base component, like Java, Node.js or Ruby.
Using Vagrant’s configuration file, aptly dubbed
Vagrantfile, you can instruct a VM instance to run a series of steps – these steps can be simple shell scripts, Chef cookbooks, or the Puppet equivalent.
I tend to favor Ubuntu as my preferred flavor of linux; consequently, all production EC2 instances use a customized Ubuntu AMI. Testing various aspects of this system with various software libraries, however, is initially tested locally using Vagrant VMs. What’s more, you can install localized VMs of other operating systems ranging from Debian to OpenSuse to Heroku’s Cedalon.
When you fire up an AWS AMI, you are given a small partition of disk space that survives reboots. For example, the base Ubuntu AMI I tend to favor comes with an 8GB primary partition; however, 8GB is often not enough, especially if you’re running a database or something that requires a lot of disk space.
If you poke around on an AMI instance, you’ll notice some AMI instances will have additional partitions and in many cases, these partitions will be huge; nevertheless, they’re transient and any data on those disks will disappear after a reboot.
Accordingly, if you need to gain some more permanent space on an AMI instance, you’ll need to leverage an Elastic Block Store (or EBS), which is basically a permanent hard disk that you can attach to a running AMI instance. The data on an EBS will survive a reboot.
When it comes to queues, whether they’re implemented as JMS, database tables (i.e. what Ruby’s Delayed::Job uses for a queue), or even Amazon’s SQS, the most common metric used to evaluate the state of a queue is its length. In essence, one derives an efficiency metric based upon how many messages are residing in a queue at any given time. If there are just a few messages, the queue is operating efficiently. If there are numerous messages, things are inefficient and alarms must be sounded.
HTML5 is important for three reasons. And its importance starts with the end of browser plugins. That’s right. With HTML5, rich media aspects that were formally handled by plugins (think Flash), are now built-in. That’s why there are new media tags like
Think about it for a second – when’s the last time you visited a site on your tablet that asked you to install a plugin? Never.
That’s also why some older plugin laden sites do not work on your mobile device. Remember the whole ”Steve Jobs no flash” kerfuffle years ago? Yep, Mr. Jobs was adamant that the iPhone would support HTML5 and not fall into the plugin trap. Incidentally, Google and other major vendors have since followed suit. HTML5 has the support of all major browser vendors now – Apple, Google, Firefox, Opera, and yes, even Microsoft.