2019独角兽企业重金招聘Python工程师标准>>> 
Deployment 部署
===========
#由于pyspider有不同的组件，所以您可以运行“pyspider”来启动一个独立的和第三方服务的免费实例。或者使用MySQL或MongoDB和RabbitMQ来部署分布式抓取集群
Since pyspider has various components, you can just run `pyspider` to start a standalone and third service free instance. Or using MySQL or MongoDB and RabbitMQ to deploy a distributed crawl cluster.
#在生产环境中部署pyspider，在每个流程中运行组件，在数据库服务中存储数据更加可靠和灵活。
To deploy pyspider in product environment, running component in each process and store data in database service is more reliable and flexible.
#安装
Installation
------------
#要在每个单独进程中部署pyspider组件，至少需要一个数据库服务，pyspider现在支持[MySQL](http://www.mysql.com/), [MongoDB](http://www.mongodb.org/) and [PostgreSQL](http://www.postgresql.org/),你可以选择其中一个
To deploy pyspider components in each single processes, you need at least one database service. pyspider now supports [MySQL](http://www.mysql.com/), [MongoDB](http://www.mongodb.org/) and [PostgreSQL](http://www.postgresql.org/). You can choose one of them.
#并且你必须有一个消息队列服务让组件连接到一起，你可以使用[RabbitMQ](http://www.rabbitmq.com/), [Beanstalk](http://kr.github.io/beanstalkd/) or [Redis](http://redis.io/)作为消息队列
And you need a message queue service to connect the components together. You can use [RabbitMQ](http://www.rabbitmq.com/), [Beanstalk](http://kr.github.io/beanstalkd/) or [Redis](http://redis.io/) as message queue.
#
`pip install --allow-all-external pyspider[all]`
#即使你之前已经使用“pip”安装了pyspider，也必须带上‘pyspider[all]’安装，以满足安装MySQL/MongoDB/RabbitMQ的需求
> Even if you had install pyspider using `pip` before. Install with `pyspider[all]` is necessary to install the requirements for MySQL/MongoDB/RabbitMQ.
#如果你使用Ubuntu系统，尝试
if you are using Ubuntu, try:
```
apt-get install python python-dev python-distribute python-pip libcurl4-openssl-dev libxml2-dev libxslt1-dev python-lxml
```
安装二进制包
to install binary packages.
#部署
Deployment
----------
#以下文档是以MySQL + RabbitMQ为例
**This document is based on MySQL + RabbitMQ**
### config.json #json配置文件
#即使你有能力使用 命令行去指定参数，配置文件也是一个更好的选择
Although you can use command-line to specify the parameters. A config file is a better choice.
```
{
"taskdb": "mysql+taskdb://username:password@host:port/taskdb",
"projectdb": "mysql+projectdb://username:password@host:port/projectdb",
"resultdb": "mysql+resultdb://username:password@host:port/resultdb",
"message_queue": "amqp://username:password@host:port/%2F",
"webui": {
"username": "some_name",
"password": "some_passwd",
"need-auth": true
}
}
```
#你可以通过运行“pyspider - help”和“pyspider webui - help”来获得完整的子命令选项。“webui”是JSON配置文件的子命令。您可以用这种方式添加更多其它组件
you can get complete options by running `pyspider --help` and `pyspider webui --help` for subcommands. `"webui"` in JSON is configs for subcommands. You can add parameters for other components similar to this one.
#数据库连接url
#### Database Connection URI
#`"taskdb"`, `"projectdb”`, `"resultdb"` 使用下面的格式连接数据库
`"taskdb"`, `"projectdb”`, `"resultdb"` is using database connection URI with format below:
```
mysql:
mysql+type://user:passwd@host:port/database
sqlite:
# relative path
sqlite+type:///path/to/database.db
# absolute path
sqlite+type:////path/to/database.db
# memory database
sqlite+type://
mongodb:
mongodb+type://[username:password@]host1[:port1][,host2[:port2],...[,hostN[:portN]]][/[database][?options]]
more: http://docs.mongodb.org/manual/reference/connection-string/
sqlalchemy:
sqlalchemy+postgresql+type://user:passwd@host:port/database
sqlalchemy+mysql+mysqlconnector+type://user:passwd@host:port/database
more: http://docs.sqlalchemy.org/en/rel_0_9/core/engines.html
local:
local+projectdb://filepath,filepath
type:
#可以是`taskdb`, `projectdb`, `resultdb`中的一个
should be one of `taskdb`, `projectdb`, `resultdb`.
```
#### Message Queue URL 消息队列地址
#你可以使用连接地址指定消息队列
You can use connection URL to specify the message queue:
```
rabbitmq:
amqp://username:password@host:5672/%2F
Refer: https://www.rabbitmq.com/uri-spec.html
beanstalk:
beanstalk://host:11300/
redis:
redis://host:6379/db
redis://host1:port1,host2:port2,...,hostn:portn (for redis 3.x in cluster mode)
builtin:
None
```
#提示postgresql:您需要用自己创建utf8编码数据库。pyspider不会为您创建数据库。
> Hint for postgresql: you need to create database with encoding utf8 by your own. pyspider will not create database for you.
#运行
running
-------
#您应该使用子命令独自运行组件。你可以“&”命令后添加[screen](http://linux.die.net/man/1/screen)或[nohup](http://linux.die.net/man/1/nohup)让程序在后台运行,以防止您的ssh会话结束后退出。* *建议使用[Supervisor]管理组件
You should run components alone with subcommands. You may add `&` after command to make it running in background and use [screen](http://linux.die.net/man/1/screen) or [nohup](http://linux.die.net/man/1/nohup) to prevent exit after your ssh session ends. **It's recommended to manage components with [Supervisor](http://supervisord.org/).**
```
# start **only one** scheduler instance 启动唯一的调度器实例
pyspider -c config.json scheduler
# phantomjs
pyspider -c config.json phantomjs
# start fetcher / processor / result_worker instances as many as your needs 启动你必须使用的实例
pyspider -c config.json --phantomjs-proxy="localhost:25555" fetcher
pyspider -c config.json processor
pyspider -c config.json result_worker
#启动webui，设置“--scheduler-rpc”，如果调度程序不在与webui相同的主机上运行
# start webui, set `--scheduler-rpc` if scheduler is not running on the same host as webui
pyspider -c config.json webui
```
#在docker中运行
Running with Docker
-------------------
[Running pyspider with Docker](Running-pyspider-with-Docker)
Deployment of demo.pyspider.org
-------------------------------
[Deployment of demo.pyspider.org](Deployment-demo.pyspider.org)