mysql binlog协议分析--具体event

这几天在修改canal, 连接mysql和maria接收到的event有所区别

拿一个简单的insert sql来举例

mysql 会有以下几个event写入到binlog里

1.ANONYMOUS_GTID_LOG_EVENT

2.QUERY_EVENT
header {
version: 1
logfileName: "20170105-162017-bin.000001"
logfileOffset: 1920
serverId: 1
serverenCode: "UTF-8"
executeTime: 1508809530000
sourceType: MYSQL
schemaName: ""
tableName: ""
eventLength: 72
}
entryType: TRANSACTIONBEGIN
storeValue: " \354\001"

3.TABLE_MAP_EVENT

4.WRITE_ROWS_EVENT
header {
version: 1
logfileName: "20170105-162017-bin.000001"
logfileOffset: 2040
serverId: 1
serverenCode: "UTF-8"
executeTime: 1508809530000
sourceType: MYSQL
schemaName: "test"
tableName: "test1"
eventLength: 40
eventType: INSERT
}
entryType: ROWDATA
storeValue: "\b\333\001\020\001P\000b\035\022\033\b\000\020\004\032\002id \000(\0010\000B\00215R\aint(11)"

5.XID_EVENT
header {
version: 1
logfileName: "20170105-162017-bin.000001"
logfileOffset: 2080
serverId: 1
serverenCode: "UTF-8"
executeTime: 1508809530000
sourceType: MYSQL
schemaName: ""
tableName: ""
eventLength: 31
}
entryType: TRANSACTIONEND
storeValue: "\022\003184"

如果一个事务里不仅一条sql, 4.WRITE_ROWS_EVENT就会有N条, 我们可以注意到一个完整事务TRANSACTIONBEGIN和TRANSACTIONEND必然会配对出现

mysql event部分格式说明

https://dev.mysql.com/doc/internals/en/event-data-for-specific-event-types.html

mariadb 的事件和mysql不一样

1.GTID_EVENT

2.TABLE_MAP_EVENT

3.WRITE_ROWS_EVENT
header {
version: 1
logfileName: "20170105-162017-bin.000001"
logfileOffset: 2040
serverId: 1
serverenCode: "UTF-8"
executeTime: 1508809530000
sourceType: MYSQL
schemaName: "test"
tableName: "test1"
eventLength: 40
eventType: INSERT
}
entryType: ROWDATA
storeValue: "\b\333\001\020\001P\000b\035\022\033\b\000\020\004\032\002id \000(\0010\000B\00215R\aint(11)"

4.XID_EVENT
header {
version: 1
logfileName: "20170105-162017-bin.000001"
logfileOffset: 2080
serverId: 1
serverenCode: "UTF-8"
executeTime: 1508809530000
sourceType: MYSQL
schemaName: ""
tableName: ""
eventLength: 31
}
entryType: TRANSACTIONEND
storeValue: "\022\003184"

gtid_event就是用来替代begin query event的

Binlog event的header格式如下：

4              timestamp

1              event type

4              server-id

4              event-size

4              log pos

2              flags

header的长度固定为19，event type用来标识这个event的类型，event size则是该event包括header的整体长度，而log pos则是下一个event所在的位置

binlog基础格式介绍可参考

http://www.jianshu.com/p/5e6b33d8945f

canal源码里并没有对gtid_event进行解析处理，查看mariadb官网

Event Header

Type[1] = 0xa2
Flags[2] = 08 00 => LOG_EVENT_SUPPRESS_USE_F

Fields

uint<8> GTID sequence
uint<4> Replication Domain ID
uint<1> Flags

if flag & FL_GROUP_COMMIT_ID

uint<8> commit_id

else

uint<6> 0

具体地址参考 https://mariadb.com/kb/en/library/gtid_event/

下面是截取到2个事务gtid_event的字节码

41, 0, 0, 34, 0, -17, 0, -71, 75, 51, 90, -94,    2, 0, 0, 0,    38, 0, 0, 0,   56, 20, 0, 0,    8, 0,    115, 25, 0, 0, 0, 0, 0, 0,    0, 0, 0, 0, 12, 0, 0, 0, 0, 0, 0, 31, 0, 0, 0, 0, 0, 0, 0, 11, 0, 0, 0, 0, 3, 8, 8, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 56, 13, 0, 8, 0, 18, 0, 4, 4, 4, 4, 18, 0, 0, -35, 0, 4, 26, 8, 0, 0, 0, 8, 8, 8, 2, 0, 0, 0, 10, 10, 10,

41, 0, 0, 38, 0, -17, 0, -30, 75, 51, 90, -94,    2, 0, 0, 0,    38, 0, 0, 0,   -28, 20,0, 0,    8, 0,    116, 25, 0, 0, 0, 0, 0, 0,    0, 0, 0, 0, 12, 0, 0, 0, 0, 0, 0, 32, 0, 0, 0, 0, 0, 0, 0, 11, 0, 0, 0, 0, 3, 8, 8, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 56, 13, 0, 8, 0, 18, 0, 4, 4, 4, 4, 18, 0, 0, -35, 0, 4, 26, 8, 0, 0, 0, 8, 8, 8, 2, 0, 0, 0, 10, 10, 10, 
红色部分是message header, 41, 0, 0是message长度, 是message sequence_id, -17, 0是semi标识位, -71或者-30开始就是具体event的内容了 115, 25, 0, 0, 0, 0, 0, 0 就是gtid了，用小序列标识=6516

Protocol::Packet

Data between client and server is exchanged in packets of max 16MByte size.

Payload Type    Name    Description

int<>     payload_length     Length of the payload. The number of bytes in the packet beyond the initial 4 bytes that make up the packet header.

int<>     sequence_id     Sequence ID

string<var>     payload     payload of the packet

Example:

01 00 00 00 01

    length: 1

    sequence_id: x00

    payload: 0x01