ibr - Innobase Reader

Install

1. Download latest package from github release page | here
   • Windows user choose innobase_reader_xxx_x86_64-pc-windows-gnu.zip
   • Linux user choose innobase_reader_xxx_x86_64-unknown-linux-musl.tar.gz
2. Uncompress package to get executable binary file (ibr or ibr.exe), THAT IT
   • The executable binary has no dependencies
   • You can put it on any directory in your system PATH

Introduction

Usage

ibr -h

The innobase datafile(*.ibd) reader

Usage: ibr <INPUT> <COMMAND>

Commands:
  info  Print basic information
  list  List all page. page_type, page_number and more
  desc  Describe datafile information by sdi page
  sdi   Print SDI json
  view  View page data with given page_no
  dump  Dump index page user records
  help  Print this message or the help of the given subcommand(s)

Arguments:
  <INPUT>  Input innodb datafile. for example departments.ibd

Options:
  -h, --help     Print help
  -V, --version  Print version

Commands

info - list basic information

ibr ./data/departments.ibd info

Meta Information:
     version => server(80037), space(1)
    space_id => 3
  page_count => 8
   file_size => 131072
PageTypes Statistics:
   ALLOCATED => 2
       INODE => 1
 IBUF_BITMAP => 1
     FSP_HDR => 1
         SDI => 1
       INDEX => 2

list - list all page basic information

ibr ./data/departments.ibd list

page_no=0, page_type=FSP_HDR, space_id=3, lsn=20329829, offset=0x0(0)
page_no=1, page_type=IBUF_BITMAP, space_id=3, lsn=20328997, offset=0x4000(16384)
page_no=2, page_type=INODE, space_id=3, lsn=20329829, offset=0x8000(32768)
page_no=3, page_type=SDI, space_id=3, lsn=20338300, offset=0xc000(49152)
page_no=4, page_type=INDEX, space_id=3, lsn=20449347, offset=0x10000(65536)
page_no=5, page_type=INDEX, space_id=3, lsn=20449393, offset=0x14000(81920)
page_no=6, page_type=ALLOCATED, space_id=0, lsn=0, offset=0x18000(98304)
page_no=7, page_type=ALLOCATED, space_id=0, lsn=0, offset=0x1c000(114688)

desc - like desc on mysql client, it prints table structure

ibr ./data/departments.ibd desc

COL1: name=dept_no, type=STRING, nullable=false, data_len=16, utf8_def=char(4)
COL2: name=dept_name, type=VARCHAR, nullable=false, data_len=161, utf8_def=varchar(40)
COL3: name=DB_TRX_ID, type=INT24, nullable=false, data_len=6, utf8_def=
COL4: name=DB_ROLL_PTR, type=LONGLONG, nullable=false, data_len=7, utf8_def=
IDX1: name=PRIMARY, type=IT_PRIMARY, id=155, root=4, algorithm=IA_BTREE
 (1-1): column_opx=0, col_name=dept_no, order=ORDER_ASC, ele_len=16, hidden=false, isnil=false, isvar=true
 (1-2): column_opx=2, col_name=DB_TRX_ID, order=ORDER_ASC, ele_len=-1, hidden=true, isnil=false, isvar=false
 (1-3): column_opx=3, col_name=DB_ROLL_PTR, order=ORDER_ASC, ele_len=-1, hidden=true, isnil=false, isvar=false
 (1-4): column_opx=1, col_name=dept_name, order=ORDER_ASC, ele_len=-1, hidden=true, isnil=false, isvar=true
IDX2: name=dept_name, type=IT_UNIQUE, id=156, root=5, algorithm=IA_BTREE
 (2-1): column_opx=1, col_name=dept_name, order=ORDER_ASC, ele_len=160, hidden=false, isnil=false, isvar=true
 (2-2): column_opx=0, col_name=dept_no, order=ORDER_ASC, ele_len=-1, hidden=true, isnil=false, isvar=true

sdi - like ibd2sdi, it prints SDI data on SDI page

ibr ./data/departments.ibd sdi | grep -A 40 "columns"

"columns": [
  {
    "name": "dept_no",
    "type": 29,
    "is_nullable": false,
    "is_zerofill": false,
    "is_unsigned": false,
    "is_auto_increment": false,
    "is_virtual": false,
    "hidden": 1,
    "ordinal_position": 1,
    "char_length": 16,
    "numeric_precision": 0,
    "numeric_scale": 0,
    "numeric_scale_null": true,
    "datetime_precision": 0,
    "datetime_precision_null": 1,
    "has_no_default": true,
    "default_value_null": false,
    "srs_id_null": true,
    "srs_id": 0,
    "default_value": "ICAgICAgICAgICAgICAgIA==",
    "default_value_utf8_null": true,
    "default_value_utf8": "",
    "default_option": "",
    "update_option": "",
    "comment": "",
    "generation_expression": "",
    "generation_expression_utf8": "",
    "options": "interval_count=0;",
    "se_private_data": "table_id=1065;",
    "engine_attribute": "",
    "secondary_engine_attribute": "",
    "column_key": 2,
    "column_type_utf8": "char(4)",
    "elements": [],
    "collation_id": 255,
    "is_explicit_collation": false
  },
  {
    "name": "dept_name",

view - prints page structure

ibr ./data/departments.ibd view 0 | grep -A 20 BasePage

BasePage {
    addr: 0x0000@(0),
    fil_hdr: FilePageHeader {
        addr: 0x0000@(0),
        check_sum: 0xbfb1ff11(3216113425),
        page_no: 0,
        prev_page: 0x000138a5(80037),
        next_page: 0x00000001(1),
        lsn: 0x0000000001363565(20329829),
        page_type: FSP_HDR,
        flush_lsn: 0x0000000000000000(0),
        space_id: 3,
    },
    page_body: FileSpaceHeaderPageBody {
        addr: 0x0026@(38),
        fsp_hdr: FileSpaceHeader {
            addr: 0x0026@(38),
            space_id: 3,
            notused: 0,
            fsp_size: 8,
            free_limit: 64,

dump - dump row data

ibr ./data/departments.ibd dump 4 -l 3

**************************************** Row 1 ****************************************
     dept_no => Str("d001")
   DB_TRX_ID => TrxId(0x00000000053d)
 DB_ROLL_PTR => RollPtr(0x820000008d0110)
   dept_name => Str("Marketing")
**************************************** Row 2 ****************************************
     dept_no => Str("d002")
   DB_TRX_ID => TrxId(0x00000000053d)
 DB_ROLL_PTR => RollPtr(0x820000008d011d)
   dept_name => Str("Finance")
**************************************** Row 3 ****************************************
     dept_no => Str("d003")
   DB_TRX_ID => TrxId(0x00000000053d)
 DB_ROLL_PTR => RollPtr(0x820000008d012a)
   dept_name => Str("Human Resources")
ONLY dump 3 of 9 rows, use `--limit num' to dump more

Build

Install rust toolchain and build with cargo

cargo build

Limitation

1. only support ibd in MySQL 8.0.x
   • cannot parse frm table structure now
2. only support limited data type
   • char/varchar
   • int/bigint
   • date/datetime/timestamp
3. don’t support overflow page now