基于Linux环境实现Oracle goldengate远程抽取MySQL同步数据到MySQL
场景说明:
先有项目需要读取生产库数据,但是不能直接读取生产库数据,需要把生产数据同步到一个中间库,下游系统从中间库读取数据。
生产库mysql - OGG - 中间库 mysql
备注:中间库也可以是其Oracle或者其他数据库类型,为了方便处理,采用同类型数据库
ogg版本:ogg21.5
前提条件
Ogg18.1 remote capture要求mysql为5.7版本,只能从linux远程捕获mysql on windows or linux,且不支持DDL捕获。支持远程mysql为community/commerical版本。
OGG抽取配置
远程捕获不支持DDL同步。
远程捕获参考网址:
https://docs.oracle.com/en/middleware/goldengate/core/19.1/oggmp/oracle-goldengate-classic-mysql.html#GUID-A70C0798-CE2A-4938-A54E-16BD9BA415AE
https://docs.oracle.com/en/middleware/goldengate/core/21.3/gghdb/preparing-and-configuring-system-oracle-goldengate.html#GUID-DCEF7BEE-CCF0-4531-8266-36F15105EDFA
通常,从站应用的事务会记录到中继日志中,而不是从站的二进制日志中。
要使从属服务器在其二进制日志中写入从属服务器从属服务器接收的事务,
必须将 my.cnf 中的 log-slave-updates 选项作为 1 与 Oracle GoldenGate
的其他二进制日志记录参数一起启动复制从属服务器。在主节点的事务进入从节点的二进制日志后,
您可以在节点上设置常规的 Oracle GoldenGate 捕获,以捕获和处理节点的二进制日志。
准备工作 源端创建OGG应用用户
--源端创建OGG同步用户(测试环境)
CREATE USER 'ogguser_s'@'10.10.xx.xx' IDENTIFIED BY 'Sjztdb#2023';
GRANT ALL PRIVILEGES ON *.* TO 'ogguser_s'@'10.10.xx.xx' WITH GRANT OPTION;
FLUSH PRIVILEGES;
--源端创建OGG同步用户(生产环境)
CREATE USER 'ogguser_s'@'10.128.xx.xx' IDENTIFIED BY 'Sjztdb#2023';
GRANT ALL PRIVILEGES ON *.* TO 'ogguser_s'@'10.128.xx.xx' WITH GRANT OPTION;
FLUSH PRIVILEGES;
创建测试表:
create table mtab (mid int primary key,mname varchar(100),vdate date);
insert into mtab values(1001,'oggtomysql','2024-01-01');
准备工作 目标端创建ogg同步用户
-- 语法
dbhostname是OGG服务器地址
mysql > CREATE USER 'targetuser'@'dbhostname' IDENTIFIED BY 'password';
mysql > GRANT ALL PRIVILEGES ON *.* TO 'targetuser'@'dbhostname';
mysql > FLUSH PRIVILEGES;
--创建用户并授权(测试环境)
CREATE USER 'ogguser_t'@'10.10.xx.xx' IDENTIFIED BY 'Sjztdb#2023';
GRANT ALL PRIVILEGES ON *.* TO 'ogguser_t'@'10.10.xx.xx';
FLUSH PRIVILEGES;
ALTER USER 'ogguser_t'@'10.10.194.7' IDENTIFIED WITH mysql_native_password BY 'Sjztdb#2023';
flush privileges;
--目标端数据库创建用户并授权(生产环境)
CREATE USER 'ogguser_t'@'10.128.xx.xx' IDENTIFIED BY 'Sjztdb#2023';
GRANT ALL PRIVILEGES ON *.* TO 'ogguser_t'@'10.128.xx.xx';
FLUSH PRIVILEGES;
ALTER USER 'ogguser_t'@'10.128.xx.xx' IDENTIFIED WITH mysql_native_password BY 'Sjztdb#2023';
flush privileges;
--测试验证登录目标端数据库
dblogin sourcedb [email protected]:3306, userid ogguser_t, password Sjztdb#2023
安装OGG
--创建OGG目录
mkdir /data/ogg181_mysql
--解压OGG
unzip -d /data/ogg181_mysql/ ogg181mysql.zip
--配置别名
alias mggsci='/data/ogg181_mysql/ggsci'
--创建OGG目录
create subdirs
MGR 进程配置
GGSCI (oggser) 2> view param mgr
port 7920
dynamicportlist 7920-7925
AUTORESTART EXTRACT *,RETRIES 5,WAITMINUTES 3
PURGEOLDEXTRACTS ./dirdat/*, USECHECKPOINTS,MINKEEPDAYS 3
ACCESSRULE, PROG *, IPADDR 10.10.*.*, ALLOW
LAGREPORTHOURS 1
LAGINFOMINUTES 30
LAGCRITICALMINUTES 45
--确认OGG能远程连接到mysql
--测试验证登录源端数据库
dblogin sourcedb [email protected]:3306, userid ogguser_s, password Sjztdb#2023
GGSCI (oggsend-194-7) 2> dblogin sourcedb [email protected]:3306, userid ogguser_s, password Sjztdb#2023
Successfully logged into database.
创建OGG登录别名
配置用户凭证
add credentialstore
alter credentialstore add user [email protected]:3306/apvll, password Sjztdb#2023 alias s_mysql1 ##源端
alter credentialstore add user [email protected]:3306/apvll, password Sjztdb#2023 alias t_mysql2 ##目标端
测试别名登录
DBLOGIN SOURCEDB [email protected]:3306 USERIDALIAS s_mysql1
DBLOGIN SOURCEDB [email protected]:3306 USERIDALIAS t_mysql2
配置抽取进程
--添加进程
ADD EXTRACT eap, TRANLOG, BEGIN NOW
ADD EXTTRAIL ./dirdat/eb, EXTRACT eap
--编辑参数 (生产配置)
EDIT PARAMS eap
EXTRACT eap
sourcedb [email protected]:3306, userid ogguser_s, password Sjztdb#2023
DISCARDFILE ./dirrpt/eap.dsc, APPEND, MEGABYTES 1024
DISCARDROLLOVER AT 05:30 ON MONDAY
TRANLOGOPTIONS ALTLOGDEST REMOTE
EXTTRAIL ./dirdat/eb
TABLE ap.arrival_;
--删除抽取进程,如果需要
delete EXTRACT eap
配置应用进程
--添加应用进程
--1.目标端添加检查表(生产环境)
dblogin sourcedb [email protected]:3306, userid ogguser_t, password Sjztdb#2023
ADD CHECKPOINTTABLE ap.ggcheck2
--2.编辑参数(生产环境)
EDIT PARAMS rap
REPLICAT rap
TARGETDB [email protected]:3306, userid ogguser_t, password Sjztdb#2023
DISCARDFILE ./dirrpt/rap.dsc, APPEND, MEGABYTES 1024
DISCARDROLLOVER AT 05:30 ON MONDAY
MAP ap.arrival_, TARGET ap.arrival_;
--3.添加应用进程
ADD REPLICAT rap, EXTTRAIL ./dirdat/eb, CHECKPOINTTABLE ap.ggcheck2
--4.删除应用进程
DELETE REPLICAT rap
数据初始化
-- 批量导出dbname数据库中多张表结构
mysqldump -uroot -pAa123456 -d --databases ap --tables arrival_ > db0408.sql
//方式1:通过mysqldump导出导入的方式进行初始化
--从源端导出数据
--导出某个表 (需要加入--master-data 记录导出时候的binlog日志文件和pos位置)
mysqldump -uroot -pAa123456 --master-data --databases apvll --tables apv_common_relation >/tmp/apv_common_relation.sql
mysqldump -uroot -pAa123456 --master-data --databases apvll --tables arrival_area_enter_out apv_result_common arrival_area_hour_info arrival_area_total_info > /tmp/newdata0409.sql
--目标端导入数据
//方式2:
--通过OGG初始化进程进行初始化
--编辑初始化抽取进程参数
edit param einit1
extract einit1
sourcedb [email protected]:3306, userid ogguser_s, password Sjztdb#2023
TRANLOGOPTIONS ALTLOGDEST REMOTE
rmthost 10.128.255.146, mgrport 7920
rmttask replicat, group rinit1
TABLE apvll.arrival_area_enter_out;
--添加初始化抽取进程
add extract einit1, sourceistable
--OGG服务器添加初始化应用进程
--编辑初始化应用进程参数
edit param rinit1
replicat rinit1
TARGETDB [email protected]:3306, userid ogguser_t, password Sjztdb#2023
discardfile ./dirrpt/rinit1.dsc, purge
MAP apvll.arrival_area_enter_out, TARGET apvll.arrival_area_enter_out;
--添加初始化应用进程
add replicat rinit1,specialrun
--删除初始化进程
delete replicat rinit1
启动同步进程
--查看mysqldump导出文件的binlog位置和pos位置
--CHANGE MASTER TO MASTER_LOG_FILE='mysql-bin.002414', MASTER_LOG_POS=767854389;
-- 修改抽取进程开始抽取的日志文件和pos位置
ALTER EXTRACT eap,lognum 2414,logpos 767854389
--启动抽取进程
start eap
--启动应用进程
start rap
相关问题
问题1:启动抽取进程报错
2024-03-21 15:35:06 INFO OGG-01052 No recovery is required for target file ./dirdat/et000000000, at RBA 0 (file
not opened).
2024-03-21 15:35:06 INFO OGG-01478 Output file ./dirdat/et is using format RELEASE 12.3/18.1.
2024-03-21 15:35:06 INFO OGG-00182 VAM API running in single-threaded mode.
2024-03-21 15:35:06 INFO OGG-01515 Positioning to begin time 2024 M03 21 15:34:26.
terminate called after throwing an instance of 'std::invalid_argument'
what(): stol
更换OGG版本为19.1,重新启动抽取进程出现如下提示
2024-03-21 15:45:57 ERROR OGG-00146 Call to VAMInitialize returned with error status 600: VAM Client Report <CAU
SE OF FAILURE : Invalid argument: stol for log file: mysql-bin.000001�. Y logFileNameLen: 20, CheckSumLength: 0. Log
file name contains non numeric characters
WHEN FAILED : While reading log event from binary log
WHERE FAILED : MySQLBinLog Reader Module
CONTEXT OF FAILURE : No Information Available!>.
更换版本为21.3版本之后正常启动。
问题2:启动应用进程报错
--问题
Process Abending : 2024-04-09 14:42:02.536013
Oracle GoldenGate Delivery for MySQL process started, group rap discard file opened: 2024-04-09 14:51:53.977658
Current time: 2024-04-09 14:51:54
Discarded record from action ABEND on error 1403
No data found
Aborting transaction on /OSdata/ogg213_mysql/dirdat/ed beginning at seqno 0 rba 2,171.
error at seqno 0 rba 2171
Problem replicating apvll.arrival_area_enter_out to ap.arrival_
Record not found
Mapping problem with delete record (target format) SCN:000000000000000002414:000000616156738...
*
id = 0368072643414db9929565eff2aa2364
000000: 30 33 36 38 30 37 32 36 34 33 34 31 34 64 62 39 |0368072643414db9|
000010: 39 32 39 35 36 35 65 66 66 32 61 61 32 33 36 34 |929565eff2aa2364|
*
Process Abending : 2024-04-09 14:51:54.563740
--检查处理 ,源端和目标端都未查询到该记录
mysql> select id from ap.arrival_ where id = '0368072643414db9929565eff2aa2364';
Empty set (0.01 sec)
mysql> select id from ap.arrival_ where id = '0368072643414db9929565eff2aa2364';
Empty set (0.00 sec)
--继续分析队列文件
GGSCI (oggser- DBLOGIN as ogguser_t) 37> info rap
Replicat rapLast Started 2024-04-09 14:51 Status ABENDED
Checkpoint Lag 00:00:00 (updated 00:26:29 ago)
Log Read Checkpoint File ./dirdat/ed000000000
First Record RBA 1341
./logdump
open ./dirdat/ed000000000
Logdump 3 >pos 1341
----pos + RBA号,直接跳到这个记录上,这里为什么跳到1259呢?因为 info rap命令查询出来的结果告诉我们,1341 是他当前正在处理的事务。
Logdump 4 >sfh
2024/04/09 14:42:00.877.974 Metadata Len 84 RBA 1341
Database Name:
3000 5000 0100 0200 0100 0200 4600 0100 0400 0000 | 0.P.........F.......
0000 0200 0400 0200 0000 0300 0700 0500 656e 5f55 | ................en_U
5304 000f 000d 0041 7369 612f 5368 616e 6768 6169 | S......Asia/Shanghai
0500 1400 1000 0000 1112 1122 2211 1112 1111 1111 | ..........."".......
1111 1111 | ....
Logdump 5 >N
2024/04/09 14:42:00.877.975 Metadata Len 616 RBA 1476
Table Name: ap.arrival_
3040 6202 0000 0100 0200 0b00 0200 0a00 0100 0600 | 0@b.................
0100 0200 0100 0340 4802 0000 0a00 3200 3000 0200 | [email protected]...
6964 4000 6000 0000 6000 0000 6000 0000 0000 0000 | id@.`...`...`.......
0000 0000 0000 0000 0000 0000 fd00 0000 0000 0000 | ....................
0000 0000 2300 3700 3500 0700 7061 726b 5f69 6440 | ....#.7.5...park_id@
0060 0000 0060 0000 0060 0000 0000 0000 0000 0000 | .`...`...`..........
0065 0000 0000 0000 00fd 0000 0000 0000 0000 0000 | .e..................
Logdump 6 >
--分析可能的原因:
是因为没有指定抽取进程开始抽取的pos号,所以抽取进程把之前的操作记录也都抽取了,所以之前的记录都已经没有了