[转帖]py_innodb_page_info.py工具使用

目录

一、Linux安装Python3

1. 解压包

2. 安装环境

3. 生成编译脚本

4. 检查python3.10的编译器

5. 建立Python3和pip3的软链

6. 添加到PATH

7. 验证Python3和pip3是否正常

二、py_innodb_page_info.py工具

1. py_innodb_page_info.py

2. mylib.py

3. include.py

三、使用py_innodb_page_info.py工具

四、参考资料

---

一、Linux安装Python3

1. 解压包

tar -zxvf Python-3.10.5.tgz

2. 安装环境

yum install gcc

yum install libffi-devel -y

3. 生成编译脚本

进入解压目录下，使用下述命令来配置编译目录。注意--prefix路径是当前解压目录。

./configure --prefix=/home/python3/Python-3.10.5

编译并安装，若下图所示：

make & make install

4. 检查python3.10的编译器

进入bin目录下执行以下命令：

./python3.10

5. 建立Python3和pip3的软链

建立Python3和pip3的软链，若以下命令：

ln -s /home/python3/Python-3.10.5/bin/python3 /usr/bin/python3

ln -s /home/python3/Python-3.10.5/bin/pip3 /usr/bin/pip3

6. 添加到PATH

1. 进入：vim /etc/profile
2. python的bin加入到PATH中：export PATH=$PATH:$JAVA_HOME/bin:/home/python3/Python-3.10.5/bin
3. 重启profile：source ~/.bash_profile

7. 验证Python3和pip3是否正常

python3 -V

pip3 -V

二、py_innodb_page_info.py工具

此工具是《MySQL技术内幕 InnoDB存储引擎》作者姜承尧写的用来分析表空间中的各页的类型和信息，用python编写。

此工具由三个文件组成：py_innodb_page_info.py、mylib.py、include.py。注意：需要放在同一个目录下。

1. py_innodb_page_info.py

#! /usr/bin/env python





#encoding=utf-8





import mylib





from sys import argv





from mylib import myargv





 





if __name__ == '__main__':





    myargv = myargv(argv)





    if myargv.parse_cmdline() == 0:





        pass





    else:





        mylib.get_innodb_page_type(myargv)

2. mylib.py

#encoding=utf-8





import os





import include





from include import *





 





TABLESPACE_NAME='/home/MySQL5.7/mysql-5.7.35/data/test_mysql/page_structure.ibd'





VARIABLE_FIELD_COUNT = 1





NULL_FIELD_COUNT = 0





 





class myargv(object):





    def __init__(self, argv):





        self.argv = argv





        self.parms = {}





        self.tablespace = ''





   





    def parse_cmdline(self):





        argv = self.argv





        if len(argv) == 1:





            print 'Usage: python py_innodb_page_info.py [OPTIONS] tablespace_file'





            print 'For more options, use python py_innodb_page_info.py -h'





            return 0





        while argv:





            if argv[0][0] == '-':





                if argv[0][1] == 'h':





                    self.parms[argv[0]] = ''





                    argv = argv[1:]





                    break





                if argv[0][1] == 'v':





                    self.parms[argv[0]] = ''





                    argv = argv[1:]           





                else:





                    self.parms[argv[0]] = argv[1]





                    argv = argv[2:]





            else:





                self.tablespace = argv[0]





                argv = argv[1:]





        if self.parms.has_key('-h'):





            print 'Get InnoDB Page Info'





            print 'Usage: python py_innodb_page_info.py [OPTIONS] tablespace_file\n'





            print 'The following options may be given as the first argument:'





            print '-h        help '





            print '-o output put the result to file'





            print '-t number thread to anayle the tablespace file'





            print '-v        verbose mode'





            return 0





        return 1





       





def mach_read_from_n(page,start_offset,length):





    ret = page[start_offset:start_offset+length]





    return ret.encode('hex')





   





def get_innodb_page_type(myargv):





    f=file(myargv.tablespace,'rb')





    fsize = os.path.getsize(f.name)/INNODB_PAGE_SIZE





    ret = {}





    for i in range(fsize):





        page = f.read(INNODB_PAGE_SIZE)





        page_offset = mach_read_from_n(page,FIL_PAGE_OFFSET,4)





        page_type = mach_read_from_n(page,FIL_PAGE_TYPE,2)





        if myargv.parms.has_key('-v'):





            if page_type == '45bf':





                page_level = mach_read_from_n(page,FIL_PAGE_DATA+PAGE_LEVEL,2)





                print "page offset %s, page type <%s>, page level <%s>"%(page_offset,innodb_page_type[page_type],page_level)





            else:





                print "page offset %s, page type <%s>"%(page_offset,innodb_page_type[page_type])





        if not ret.has_key(page_type):





            ret[page_type] = 1





        else:





            ret[page_type] = ret[page_type] + 1





    print "Total number of page: %d:"%fsize





    for type in ret:





        print "%s: %s"%(innodb_page_type[type],ret[type])

3. include.py

#encoding=utf-8





INNODB_PAGE_SIZE = 16*1024*1024





 





# Start of the data on the page





FIL_PAGE_DATA = 38





FIL_PAGE_OFFSET = 4 # page offset inside space





FIL_PAGE_TYPE = 24 # File page type





 





# Types of an undo log segment */





TRX_UNDO_INSERT = 1





TRX_UNDO_UPDATE = 2





 





# On a page of any file segment, data may be put starting from this offset





FSEG_PAGE_DATA = FIL_PAGE_DATA





 





# The offset of the undo log page header on pages of the undo log





TRX_UNDO_PAGE_HDR = FSEG_PAGE_DATA





 





PAGE_LEVEL = 26    #level of the node in an index tree; the leaf level is the level 0 */         





 





innodb_page_type={





    '0000':u'Freshly Allocated Page',





    '0002':u'Undo Log Page',





    '0003':u'File Segment inode',





    '0004':u'Insert Buffer Free List',





    '0005':u'Insert Buffer Bitmap',





    '0006':u'System Page',





    '0007':u'Transaction system Page',





    '0008':u'File Space Header',





    '0009':u'扩展描述页',





    '000a':u'Uncompressed BLOB Page',





    '000b':u'1st compressed BLOB Page',





    '000c':u'Subsequent compressed BLOB Page',





    '45bf':u'B-tree Node'





}





 





innodb_page_direction={





    '0000': 'Unknown(0x0000)',





    '0001': 'Page Left',





    '0002': 'Page Right',





    '0003': 'Page Same Rec',





    '0004': 'Page Same Page',





    '0005': 'Page No Direction',





    'ffff': 'Unkown2(0xffff)'





}





INNODB_PAGE_SIZE=1024*16

三、使用py_innodb_page_info.py工具

进入py_innodb_page_info.py所在的目录下，使用以下命令：

python py_innodb_page_info.py -v /home/MySQL5.7/mysql-5.7.35/data/test_mysql/test_init_table.ibd

[root@488c1daa7967 py_innodb_page_info]# 





[root@488c1daa7967 py_innodb_page_info]# pwd





/home/py_innodb_page_info





[root@488c1daa7967 py_innodb_page_info]# 





[root@488c1daa7967 py_innodb_page_info]# 





[root@488c1daa7967 py_innodb_page_info]# python py_innodb_page_info.py -v /home/MySQL5.7/mysql-5.7.35/data/test_mysql/test_init_table.ibd





page offset 00000000, page type <File Space Header>





page offset 00000001, page type <Insert Buffer Bitmap>





page offset 00000002, page type <File Segment inode>





page offset 00000003, page type <B-tree Node>, page level <0000>





page offset 00000000, page type <Freshly Allocated Page>





page offset 00000000, page type <Freshly Allocated Page>





Total number of page: 6:





Freshly Allocated Page: 2





Insert Buffer Bitmap: 1





File Space Header: 1





B-tree Node: 1





File Segment inode: 1

内容	描述
page offset	1. 该页在表空间的偏移量； 2. 00000003页：3 * 16KB = 49152 = 0xc000，     打开.ibd文件，找到0xc000，就是00000003页的起始位置。
page type	页的类型，其中： 1. B-tree Node：B+树节点页，page level <0000>是叶子节点； 其他是非叶子节点； Freshly Allocated Page：最新分配的页（空闲页）； Insert Buffer Bitmap：插入缓冲Bitmap页；
page level	1. B+树的高度； 2. <0000>是叶子节点；非0是非叶子节点
Total number of page	表空间的页总数
Freshly Allocated Page	表空间的空闲页总数
Insert Buffer Bitmap	插入缓冲Bitmap的页数
File Space Header	File Space Header的页数
B-tree Node	B+树节点的页数量（叶子节点 + 非叶子节点）
File Segment inode	File Segment inode的页数

四、参考资料

Linux 安装Python3_Alex_Sheng_Sea的博客-CSDN博客_linux安装python

Download Python | Python.org

py_innodb_page_info工具使用_Hehuyi_In的博客-CSDN博客_py_innodb_page_info

文章知识点与官方知识档案匹配，可进一步学习相关知识

Python入门技能树基本技能数据库操作304107 人正在系统学习中