使用JRuby 9k构建可扩展Web应用： JVM部署实战 - CSDN文库

需积分: 10 129 浏览量更新于2024-07-20 收藏 6.62MB PDF 举报

"Deploying with JRuby 9k - Deliver Scalable Web Apps Using JVM" 本书是一本关于如何利用JRuby 9k在JVM上部署可扩展Web应用的专业指南。JRuby是一种在JVM上运行的快速、可扩展且功能强大的语言，它保留了传统Ruby环境的所有优点，同时提供了更少的组件和更低的内存消耗，从而实现高性能和可扩展性。书中介绍了三种JRuby部署策略，旨在帮助开发者在保持高效开发速度的同时，确保应用的性能和稳定性。书中的章节涵盖了从基础到高级的各种主题： 1. **第1章：JRuby入门** - 介绍JRuby的基本概念，解释为什么选择JRuby以及它的优势，包括与MRI（Matz的Ruby实现）的对比。 2. **第2章：创建部署环境** - 展示如何构建一个基于JRuby的微服务，利用JVM的原生并发能力，并讨论使用Docker构建稳定、可复现的虚拟生产环境。 3. **第3章：部署Rails应用** - 指导读者将现有Rails应用迁移到JRuby，详细说明如何准备应用以充分利用JVM平台，同时保持对Ruby开发者的友好性。还包括使用多线程Puma服务器将Rails应用部署到Heroku或其他私有云的步骤。 4. **第4章：使用JRuby消费后端服务** - 讨论如何集成JRuby来利用强大的Java库，提高应用的功能和性能。 5. **第5章：在企业环境中部署JRuby** - 介绍TorqueBox，这是一个一体化的JRuby环境，内置消息传递、调度和守护进程支持，适合处理大规模任务，展示JRuby在企业场景中的应用。 6. **第6章：管理JRuby应用** - 讨论JRuby应用的生命周期管理，包括版本控制、更新和维护。 7. **第7章：JRuby应用调优** - 提供深入的JVM性能优化建议，帮助提升JRuby应用的运行效率。 8. **第8章：监控生产环境中的JRuby** - 阐述如何有效地监控JRuby应用，确保在生产环境中能够及时发现问题并进行故障排除。 9. **第9章：使用持续集成服务器** - 教授如何配置Travis CI进行持续部署，实现自动化测试和构建流程。该书由JRuby社区的专家编写，得到了包括JRuby项目负责人在内的多位专家的高度评价，适合既有Ruby背景又有Java背景的开发者，以及希望深入了解JRuby部署的读者。通过阅读这本书，读者将能够掌握JRuby 9k的实际应用，实现高效、可扩展的Web应用部署。为了运行书中的示例，你需要一台可以运行Java虚拟机的计算机和互联网连接。

Conventions

The examples in this book can be run on Linux, Mac, Windows, and many

other operating systems. But some small changes to the command-line

statements may be required for certain platforms.

We’ll use notation from

bash

, which is the default shell on Mac OS X and many

Linux distributions. The

$

prompt will be used for all command-line examples.

Windows command prompts typically use something like

C:\>

instead, so when

you see a command like this

$ bundle install

you’ll know not to type the dollar sign and to read it like this:

C:\> bundle install

Most commands will be compatible between Windows and

bash

systems (such

as

cd

and

mkdir

). In the cases where they’re not compatible, the appropriate

commands for both systems will be spelled out. One case in particular is the

rm

command, which will look like this:

$ rm temp.txt

$ rm -rf tmp/

On Windows this should be translated to these two commands, respectively:

C:\> del temp.txt

C:\> rd /s /q tmp/

Another Unix notation that’s used in this book is the

~

(tilde) to represent a

user’s home directory. When you see a command like this

$ cd ~/code/twitalytics

you can translate it to Windows 10 as this command:

C:\> cd C:\Users\yourname\code\twitalytics

On earlier versions of Windows, the user’s home directory can be found in

the

Documents and Settings

directory. You can also use the

%USERPROFILE%

environ-

ment variable. Its value is the location of the current user’s profile directory.

Other than these minor notation changes, the examples in this book are

compatible with Windows by virtue of the Java Virtual Machine.

Getting the Source Code

You’re ready to set up the Twitalytics application. Start by downloading the

source code from

http://pragprog.com/titles/jkdepj2/source_code

. Unpack the downloaded

Preface • xvi

report erratum • discuss

There are many problems with this kind of architecture, and those problems

have been realized by Twitter, GitHub, and countless others. They include

the following:

Stuck processes Sometimes the processes will get into a stuck state and

need to be killed by an external tool like god or Monit.

Slow restarts There’s a lot of overhead in starting a new process. Several

instances may end up fighting each other for resources if they’re restarted

at the same time.

Memory growth Each of the processes keeps its own copy of an application,

along with Rails and any supporting gems, in memory. Each new instance

means you’ll also need more memory for the server.

Several frameworks, such as Passenger and Unicorn, have improved on this

model. But they all suffer from the same underlying constraint. A single MRI

process has a scalability ceiling because Ruby code cannot execute in parallel.

Other operations, such as I/O, can be done in parallel, and frameworks such

as EventMachine and Celluloid leverage this to create event-based environ-

ments. But these frameworks also have an upper limit, and that’s why some

of their biggest implementers, including Logstash

3

and Venntro,

4

still choose

to run them on JRuby.

JRuby enables a similar model to the one used by MRI but with only one JVM

process. Inside this JVM process is a single application instance capable of

handling all of a website’s traffic. This works by allowing the platform to create

many threads that run against the same application instance in parallel. You

can create far more JVM threads than MRI processes because they’re much

lighter in weight. This model is illustrated in the figure on page 4.

Apache is included in the architecture diagram, but its role for a single

instance is greatly reduced. It may be used to serve up static content and

load balance a distributed cluster, but it won’t need to distribute requests

across multiple processes on a single machine.

In the coming chapters, you’ll build an architecture like the one just described

with each of the technologies you use. You’ll start by using Warbler to package

a simple Rack application, which will get you up and running quickly. But

first, you’ll need to install JRuby and a few other pieces of software.

3.

https://www.elastic.co/products/logstash

4.

http://dev.venntro.com/2013/07/euruko-2013-summary/

report erratum • discuss

What Makes JRuby So Great? • 3

剩余172页未读，继续阅读

ramissue

粉丝: 354
资源: 1487

最新资源