0038-如何在CDH集群安装Anaconda&搭建Python私有源

Posted 2021-01-22

温馨提示：要看高清无码套图，请使用手机打开并单击图片放大查看。

1.文档编写目的

---

Anaconda是一个用于科学计算的Python发行版，支持 Linux, Mac, Windows系统，提供了包管理与环境管理的功能，可以很方便地解决多版本python并存、切换以及各种第三方包安装问题。

在企业集群无外网环境下，通过搭建企业私有的Python源，可以使用pip工具方便的安装，以解决Anaconda环境中没有的依赖包的问题。

本文档讲述如何在CDH集群离线安装Anaconda、搭建Python私有源及pip工具使用。

• 内容概述

1.如何在CDH集群上安装Anaconda

2.如何搭建python私有源

3.pip工具的使用

• 测试环境

1.操作系统：CentOS6.5

2.CM和CDH版本为5.12.1

3.采用root用户操作

• 前置条件

1.CDH集群运行正常

2.Apache服务正常运行

2.CDH集群安装Anaconda

---

使用Parcel包的方式在CDH集群安装Anaconda，下载地址：

https://repo.continuum.io/pkgs/misc/parcels/

根据自己操作系统版本选择Parcel包，本文档是在CentOS6.5版本下操作

2.1部署Anaconda的Parcel包http访问

---

1.下载Anaconda的Parcel包，下载地址如下

https://repo.continuum.io/pkgs/misc/parcels/Anaconda-4.2.0-el6.parcel
https://repo.continuum.io/pkgs/misc/parcels/Anaconda-4.2.0-el6.parcel.sha
https://repo.continuum.io/pkgs/misc/parcels/manifest.json

2.将上述3个文件下载至HTTP所在服务器的/var/×××w/html/anaconda4.2.0目录下

[[email protected] html]# mkdir -p anaconda4.2.0/
[[email protected] html]# cd anaconda4.2.0/
[[email protected] anaconda4.2.0]# ll
total 541416
-rw-r--r-- 1 root root 554396033 Jun 19 17:47 Anaconda-4.2.0-el6.parcel
-rw-r--r-- 1 root root        41 Jul  7 19:43 Anaconda-4.2.0-el6.parcel.sha
-rw-r--r-- 1 root root      3890 Jul  7 19:43 manifest.json
[[email protected] anaconda4.2.0]# 

3.验证http是否正常访问

2.2安装Anaconda

---

1.配置Anaconda的parcels包访问地址

2.点击下载、分配并激活

2.3配置Anaconda环境变量

---

Anaconda默认安装在/opt/cloudera/parcels/Anaconda目录，在集群所有节点配置Anaconda的环境变量，操作如下：

1. 使用root用户或有sudo权限用户编辑/etc/profile文件，在文件末尾增加如下配置

export ANACONDA_HOME=/opt/cloudera/parcels/Anaconda
export PATH=$ANACONDA_HOME/bin:$PATH

2.执行source命令，使环境变量立即生效

[[email protected] anaconda4.2.0]# source /etc/profile

3.验证环境变量是否配置成功

[[email protected] anaconda4.2.0]# python
Python 2.7.12 |Anaconda 4.2.0 (64-bit)| (default, Jul  2 2016, 17:42:40) 
[GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
Anaconda is brought to you by Continuum Analytics.
Please check out: http://continuum.io/thanks and https://anaconda.org
>>> 

3.搭建Pyton私有源

---

• 前置准备：

1.下载pip2pi安装包，下载地址：

https://codeload.github.com/wolever/pip2pi/zip/master

2.下载Python源码依赖包（tar.gz/zip/whl等格式安装包）

3.1安装pip2pi工具

---

1. 将下载的pip2pi源码包上传至服务器，并解压

[[email protected] ~]# unzip master 

2.使用Python命令安装pip2pi

[[email protected] python]# cd pip2pi-master/
[[email protected] pip2pi-master]# ll
total 44
-rw-r--r-- 1 root root 2797 Mar 19 15:51 CHANGELOG.txt
drwxr-xr-x 2 root root 4096 Mar 19 15:51 libpip2pi
-rw-r--r-- 1 root root 1649 Mar 19 15:51 LICENSE.txt
-rw-r--r-- 1 root root   19 Mar 19 15:51 MANIFEST.in
-rw-r--r-- 1 root root 5080 Mar 19 15:51 README.rst
-rwxr-xr-x 1 root root   84 Mar 19 15:51 run-tests
-rw-r--r-- 1 root root   20 Mar 19 15:51 setup.cfg
-rw-r--r-- 1 root root 1422 Mar 19 15:51 setup.py
drwxr-xr-x 5 root root 4096 Mar 19 15:51 tests
-rw-r--r-- 1 root root  194 Mar 19 15:51 tox.ini
[[email protected] pip2pi-master]# python setup.py install

3.测试pip2pi和dir2pi命令是否安装成功

3.2部署依赖包

---

1. 在http服务器的/var/×××w/html目录下新建python-packages目录

[[email protected] html]# mkdir python-packages
[[email protected] html]# pwd
/var/×××w/html
[[email protected] html]# ll
total 32
drwxr-xr-x 2 root root 4096 Sep 11 15:52 python-packages
[[email protected] html]# 

2.将本地下载的Python依赖包上传至服务器的/var/×××w/html/python-packages目录

[[email protected] python-packages]# ll
total 1780
-rw-r--r-- 1 root root   2421 Sep 24  2016 bit_array-0.1.0.tar.gz
-rw-r--r-- 1 root root   1254 Sep 12 07:18 enum-compat-0.0.2.tar.gz
-rw-r--r-- 1 root root 339017 Sep 12 07:27 pymongo-3.3.0-cp27-cp27mu-manylinux1_x86_64.whl
-rw-r--r-- 1 root root 368294 Sep 12 07:25 pymongo-3.5.1-cp27-cp27mu-manylinux1_x86_64.whl
-rw-r--r-- 1 root root 799601 Sep 12 07:04 Routes-1.12.3.tar.gz
-rw-r--r-- 1 root root  30409 Apr 23  2016 sasl-0.2.1.tar.gz
-rw-r--r-- 1 root root  29630 Apr 21  2016 six-1.10.0.tar.gz
-rw-r--r-- 1 root root  87832 Jan 14  2017 thrift-0.10.0.zip
-rw-r--r-- 1 root root   3885 Jun  9  2016 thrift_sasl-0.2.1.tar.gz
-rw-r--r-- 1 root root 140861 Dec  3  2016 ym-impyla-0.14.0.tar.gz

3.使用dir2pi命令建立包索引

[[email protected] python-packages]# dir2pi /var/×××w/html/python-packages/
[[email protected] python-packages]# ll
total 1784
-rw-r--r--  1 root root   2421 Sep 24  2016 bit_array-0.1.0.tar.gz
-rw-r--r--  1 root root   1254 Sep 12 07:18 enum-compat-0.0.2.tar.gz
-rw-r--r--  1 root root 339017 Sep 12 07:27 pymongo-3.3.0-cp27-cp27mu-manylinux1_x86_64.whl
-rw-r--r--  1 root root 368294 Sep 12 07:25 pymongo-3.5.1-cp27-cp27mu-manylinux1_x86_64.whl
-rw-r--r--  1 root root 799601 Sep 12 07:04 Routes-1.12.3.tar.gz
-rw-r--r--  1 root root  30409 Apr 23  2016 sasl-0.2.1.tar.gz
drwxr-xr-x 11 root root   4096 Sep 12 08:06 simple
-rw-r--r--  1 root root  29630 Apr 21  2016 six-1.10.0.tar.gz
-rw-r--r--  1 root root  87832 Jan 14  2017 thrift-0.10.0.zip
-rw-r--r--  1 root root   3885 Jun  9  2016 thrift_sasl-0.2.1.tar.gz
-rw-r--r--  1 root root 140861 Dec  3  2016 ym-impyla-0.14.0.tar.gz
[[email protected] python-packages]# 

4.验证Http是否正常访问，访问地址

http://ip-172-31-6-148.fayson.com/python-packages/simple/

至此Python的私有源服务器搭建完成。

3.3配置pip命令使用私有源

---

1.在当前用户的home目录（即“~”）目录，创建.pip目录

[[email protected] ~]# mkdir -p .pip
[[email protected] ~]# pwd
/root
[[email protected] ~]# 

2.在~/.pip目录下新建pip.conf文件，文件内容如下

[[email protected] ~]# cd .pip/
[[email protected] .pip]# pwd
/root/.pip
[[email protected] .pip]# vim pip.conf 
[global]
index-url = http://ip-172-31-6-148.fayson.com/python-packages/simple
[install]
trusted-host = ip-172-31-6-148.fayson.com

3.测试pip命令是否从私有源下载依赖包

[[email protected] .pip]# pip install bit-array
Collecting bit-array
  Downloading http://ip-172-31-6-148.fayson.com/python-packages/simple/bit-array/bit-array-0.1.0.tar.gz
Building wheels for collected packages: bit-array
  Running setup.py bdist_wheel for bit-array ... done
  Stored in directory: /root/.cache/pip/wheels/7a/7f/04/3bdc776730b06ce5bedd46f9023d263a32f82941e1365a054f
Successfully built bit-array
Installing collected packages: bit-array
Successfully installed bit-array-0.1.0
[[email protected] .pip]# 

可以看到上图标注部分显示从刚搭建好的源下载依赖包。

4.pip使用

---

1.指定版本安装package

 [[email protected] ~] pip install SomePackage            # 最新版本 
[[email protected] ~] pip install SomePackage==1.0.4     # 具体的版本
[[email protected] ~] pip install ‘SomePackage>=1.0.4‘     # 最低版本

2.批量安装packages

创建文本文件requirements.txt，内容如下

[[email protected] ~]# vim requirements.txt
six
bit_array
thrift==0.10.0

使用pip批量安装

[[email protected] ~]# pip install -r requirements.txt 
Collecting six (from -r requirements.txt (line 1))
Collecting bit_array (from -r requirements.txt (line 2))
Collecting thrift==0.10.0 (from -r requirements.txt (line 3))
  Downloading http://ip-172-31-6-148.fayson.com/python-packages/simple/thrift/thrift-0.10.0.zip (87kB)
    100% |████████████████████████████████| 92kB 58.1MB/s 
Building wheels for collected packages: thrift
  Running setup.py bdist_wheel for thrift ... done
  Stored in directory: /root/.cache/pip/wheels/21/53/2d/121a953df877103ba4f2fe9041be0d28d6979438c6dd7b2b15
Successfully built thrift
Installing collected packages: six, bit-array, thrift
Successfully installed bit-array-0.1.0 six-1.10.0 thrift-0.10.0
[[email protected] ~]# 

3.输出当前python环境安装的所有package及版本到指定文件

[[email protected] ~]# pip freeze > requirements.txt

4.卸载package

[[email protected] ~]# pip uninstall six

5.列出当前Python环境安装的所有package

[[email protected] ~]# pip list --format=columns
Package                            Version                
---------------------------------- -----------------------
alabaster                          0.7.9                  
anaconda-clean                     1.0                    
anaconda-client                    1.5.1                  
anaconda-navigator                 1.3.1                  
argcomplete                        1.0.0                  
astroid                            1.4.7                  
astropy                            1.2.1                  
Babel                              2.3.4                  
backports-abc                      0.4                    

6.查看安装的package详细信息

[[email protected] ~]# pip show six
Name: six
Version: 1.10.0
Summary: Python 2 and 3 compatibility utilities
Home-page: http://pypi.python.org/pypi/six/
Author: Benjamin Peterson
Author-email: [email protected]
License: MIT
Location: /opt/cloudera/parcels/Anaconda-4.2.0/lib/python2.7/site-packages
Requires: 
[[email protected] ~]# 

5.备注

---

package下载方式

• 使用pip2pi工具进行批量下载
• 通过官网手动下载，官网下载地址：https://pypi.python.org/pypi

在本地安装pip2pi工具，可以通过pip2tgz命令批量下载package

1. 在本地创建requirements.txt文件，内容如下

six
bit_array
thrift==0.10.0

由于pip2tgz是通过将参数传递给pip调用来完成包的下载，所以可以指定任何pip可以识别的参数格式。

2.使用pip2tgz命令进行批量下载

[[email protected] .pip]# pip2tgz /root/python/ -r /root/requirements.txt

注意：pip命令默认使用的是国外的源下载package，如果国外源不稳定则可以参考3.3章节配置国内源。

6.常见问题

---

1. 搭建完Python私有源，pip安装package时报错

[[email protected] .pip]# pip install six
Collecting six
  The repository located at ip-172-31-6-148.fayson.com is not a trusted or secure host and is being ignored. If this repository is available via HTTPS it is recommended to use HTTPS instead, otherwise you may silence this warning and allow it anyways with ‘--trusted-host ip-172-31-6-148.fayson.com‘.
  Could not find a version that satisfies the requirement six (from versions: )
No matching distribution found for six
[[email protected] .pip]# 

解决方法：

在pip.conf文件中增加如下配置：

trusted-host = ip-172-31-6-148.fayson.com

醉酒鞭名马，少年多浮夸！ 岭南浣溪沙，呕吐酒肆下！挚友不肯放，数据玩的花！
温馨提示：要看高清无码套图，请使用手机打开并单击图片放大查看