MySQL Replication

From Kolmisoft Wiki

Jump to: navigation, search

MySQL Replication works only on servers where MySQL 5.x is installed




THIS TUTORIAL WILL GUIDE YOU THROUGH HOW TO IMPLEMENT MASTER<->MASTER REPLICATION

Before you start, make a backup of your database.



Contents

[edit] Create user on Server 1

  • Username: rep2
  • Password: rep2
  • Host: 222.222.222.222
  • Grant all privileges to this user (on database mor)

Login as root and issue such commands:

CREATE USER 'rep2'@'222.222.222.222' IDENTIFIED BY 'rep2';
GRANT REPLICATION SLAVE , REPLICATION CLIENT ON *.* TO 'rep2'@'222.222.222.222' IDENTIFIED BY 'rep2' WITH MAX_QUERIES_PER_HOUR 0 MAX_CONNECTIONS_PER_HOUR 0 MAX_UPDATES_PER_HOUR 0 MAX_USER_CONNECTIONS 0 ;
GRANT ALL PRIVILEGES ON *.* TO 'rep2'@'222.222.222.222' WITH GRANT OPTION ;

Go to Server 2 and test the connection:

mysql -h 111.111.111.111 -u rep2 -prep2



[edit] Create user on Server 2

  • Username: rep1
  • Password: rep1
  • Host: 111.111.111.111
  • Grant all privileges to this user (on database mor)

Login as root and issue such commands:

CREATE USER 'rep1'@'111.111.111.111' IDENTIFIED BY 'rep1';
GRANT REPLICATION SLAVE , REPLICATION CLIENT ON *.* TO 'rep1'@'111.111.111.111' IDENTIFIED BY 'rep1' WITH MAX_QUERIES_PER_HOUR 0 MAX_CONNECTIONS_PER_HOUR 0 MAX_UPDATES_PER_HOUR 0 MAX_USER_CONNECTIONS 0 ;
GRANT ALL PRIVILEGES ON *.* TO 'rep1'@'111.111.111.111' WITH GRANT OPTION ;

Go to Server 1 and test the connection:

mysql -h 222.222.222.222 -u rep1 -prep1


Note: Make sure you have applied REPLICATION CLIENT and REPLICATION SLAVE on both users.



[edit] Server 1 my.cnf

[mysqld]
innodb_file_per_table=1
bind-address = 0.0.0.0
binlog-do-db = mor
# binlog-do-db = mor_mnp          #uncomment if there is mor_mnp database
# binlog-do-db = opensips         #uncomment if there is opensips database
server-id = 10
auto_increment_increment = 10
auto_increment_offset = 1
replicate-do-db = mor
# replicate-do-db = mor_mnp       #uncomment if there is mon_mnp database
# replicate-do-db = opensips      #uncomment if there is opensips database
binlog_format=MIXED
log-bin = /var/lib/mysql/mysql-bin.log
log-bin-index = /var/lib/mysql/master-log-bin.index
# sync_binlog = 1               #increases reliability, but decreases performance
relay-log = /var/lib/mysql/slave-relay.log
relay-log-index = /var/lib/mysql/slave-relay-log.index

slave-skip-errors=1061,1062,1213,1060,1032
slave_transaction_retries=3600

expire_logs_days = 3
# innodb_flush_log_at_trx_commit = 1
# innodb_support_xa

max_connections = 300
slave_transaction_retries = 3600

datadir=/var/lib/mysql
socket=/var/lib/mysql/mysql.sock

[mysqld_safe]
log-error=/var/log/mysqld.log
pid-file=/var/run/mysqld/mysqld.pid



Stop services:

service elasticsearch stop
service httpd stop
service asterisk stop
service mor_server_loadstats stop
service crond stop
service mor_alerts stop
service mor_aggregates stop

[edit] Server 2 my.cnf

[mysqld]
innodb_file_per_table=1
bind-address = 0.0.0.0
binlog-do-db = mor
# binlog-do-db = mor_mnp          #uncomment if there is mor_mnp database
# binlog-do-db = opensips         #uncomment if there is opensips database
server-id = 20
auto_increment_increment = 10
auto_increment_offset = 2
replicate-do-db = mor
# replicate-do-db = mor_mnp       #uncomment if there is mon_mnp database
# replicate-do-db = opensips      #uncomment if there is opensips database
binlog_format=MIXED
log-bin = /var/lib/mysql/mysql-bin.log
log-bin-index = /var/lib/mysql/master-log-bin.index
# sync_binlog = 1               #increases reliability, but decreases performance
relay-log = /var/lib/mysql/slave-relay.log
relay-log-index = /var/lib/mysql/slave-relay-log.index

slave-skip-errors=1061,1062,1213,1060,1032
slave_transaction_retries=3600

expire_logs_days = 3
# innodb_flush_log_at_trx_commit = 1
# innodb_support_xa

max_connections = 300
slave_transaction_retries = 3600

datadir=/var/lib/mysql
socket=/var/lib/mysql/mysql.sock

[mysqld_safe]
log-error=/var/log/mysqld.log
pid-file=/var/run/mysqld/mysqld.pid



Stop services:

service elasticsearch stop
service httpd stop
service asterisk stop
service mor_server_loadstats stop
service crond stop
service mor_alerts stop
service mor_aggregates stop

[edit] Server 2

Make sure data on Server 2 = data on Server 1. E.g. MySQL table structure and data are identical on both machines.




[edit] Server 1

Replace x13 with your MOR version.

/usr/src/mor/x13/mysql/configure_mycnf.sh
/usr/src/mor/x13/maintenance/test_fix_scripts/mysql/add_skip_name_resolve.sh 
/usr/src/mor/x13/maintenance/test_fix_scripts/mysql/mysql_log_slow.sh
/usr/src/mor/x13/db/disable_mysql_secure_file_priv.sh 
service mysqld restart
mysql -u root -p 
USE mor;
FLUSH TABLES WITH READ LOCK; 
SHOW MASTER STATUS; 

Remember File and Position.

UNLOCK TABLES;
quit;




[edit] Server2

Check if you can connect to first server:

mysql -h 111.111.111.111 -u rep2 -p

Proceed if okay:

Replace x13 with your MOR version.

/usr/src/mor/x13/mysql/configure_mycnf.sh
/usr/src/mor/x13/maintenance/test_fix_scripts/mysql/add_skip_name_resolve.sh 
/usr/src/mor/x13/maintenance/test_fix_scripts/mysql/mysql_log_slow.sh
/usr/src/mor/x13/db/disable_mysql_secure_file_priv.sh 
service mysqld restart
mysql -u root -p 
SLAVE STOP; 
CHANGE MASTER TO MASTER_HOST='111.111.111.111', MASTER_USER='rep2', MASTER_PASSWORD='rep2', MASTER_LOG_FILE='mysql-bin.000001', MASTER_LOG_POS=205; 

NOTE: Change MASTER_HOST, MASTER_LOG_FILE and MASTER_LOG_POS accordingly.

START SLAVE;
USE mor;
FLUSH TABLES WITH READ LOCK; 
SHOW MASTER STATUS; 

Remember File and Position.

UNLOCK TABLES;
quit;




[edit] Server 1

Check if you can connect to first server:

mysql -h 222.222.222.222 -u rep1 -p

Proceed if okay:

mysql -u root -p 
SLAVE STOP; 
CHANGE MASTER TO MASTER_HOST='222.222.222.222', MASTER_USER='rep1', MASTER_PASSWORD='rep1', MASTER_LOG_FILE='mysql-bin.000001', MASTER_LOG_POS=205; 

NOTE: Change MASTER_HOST, MASTER_LOG_FILE and MASTER_LOG_POS accordingly.

START SLAVE;
quit;




Start services on both servers Server 1 and Server 2:

service elasticsearch start
service httpd start
service asterisk start
service mor_server_loadstats start
service crond start
service mor_alerts start
service mor_aggregates start

[edit] Log files

In order for MySQL Replication to function properly, you need to check log files in /var/lib/mysql

Open master-log-bin.index. It should contain names of the files which really exist in same folder. If there's some mismatch, edit this file: enter missing names for files or delete file names which do not represent real files in this folder.

Do the same with the file slave-relay-log.index.




[edit] Testing

Login to MySQL console on both servers:

mysql -u root -p 

Then:

SHOW SLAVE STATUS\G

If everything is okay, you should see lines between others on both servers:

Slave_IO_State: Waiting for master to send event
Slave_IO_Running: Yes
Slave_SQL_Running: Yes


If Replication is off, you will see:

Slave_IO_State:
Slave_IO_Running: No


The quick way to do this is:

mysql -u root -e 'SHOW SLAVE STATUS\G' | grep 'Slave'

If root has psw:

mysql -u root -pPASSWORD -e 'SHOW SLAVE STATUS\G' | grep 'Slave'



Never use phpmyadmin to check if data is same on both servers. Use mysql console and a command like count. For example: SELECT COUNT(*) FROM destinations;



[edit] Troubleshooting

Check file: /var/lib/mysql/mysqld.log
Check GUI production logs and look for error messages beginning with: "Replication exception:".



[edit] 'Could not find first log file name in binary log index file' from master when reading data from binary log

We go to the master and check /var/lib/mysql/master-log-bin.index file.

For example, it can show:

/var/lib/mysql/mysql-bin.000001
/var/lib/mysql/mysql-bin.000002
/var/lib/mysql/mysql-bin.000003
/var/lib/mysql/mysql-bin.000005

In /var/lib/mysql/ we also have file mysql-bin.000004, so its' name is missing in index file, lets add it, so master-log-bin.index looks like:

/var/lib/mysql/mysql-bin.000001
/var/lib/mysql/mysql-bin.000002
/var/lib/mysql/mysql-bin.000003
/var/lib/mysql/mysql-bin.000004
/var/lib/mysql/mysql-bin.000005

Restart mysql on this server and try with the slave again.

This is just an example - some other file's records can be missing. Make sure the file has all related files' records, which are in same folder.



[edit] Failed to open the relay log

  • stop mysql on the slave server
  • find the directory that contains the relay logs (usually /var/lib/mysql)
  • delete all relay logs called hostname-relay-bin.XXX, hostname-relay-bin.index and relay-log.info
  • restart mysql

Your mysql will start with fresh relay logs and is getting the missing data directly from the master server, as there will probably still be binlogs!



[edit] Couldn't find MySQL manager or server

Comment out basedir in my.cnf




  • Default mysql.cnf - can be useful to go back if something goes wrong and no backup is done.
  • If server is rebooted - it takes ~45s for the database to start synchronizing after MySQL is up.



[edit] Could not initialize master info structure, more error messages can be found in the MySQL error log

Please check this link: http://forums.mysql.com/read.php?26,163227,163241#msg-163241



[edit] Broken replication fix

http://rackerhacker.com/2008/01/09/mysql-replication-breakdown/



[edit] What to do if you have errors with binlogs

Delete all binlog files from both servers (but leave database structures and data!), then reapply replication. You have to test GUI on both servers after you apply MySQL replication.



[edit] No more MySQL Crash Safe Replication in 5.0?

http://www.mysqlperformanceblog.com/2008/01/29/no-more-mysql-crash-safe-replication-in-50/



[edit] Can't connect to MySQL server on 'xxx.xxx.xxx.xxx'

Error looks like this:

# mysql -h 222.222.222.222 -u rep1 -prep1
ERROR 2003 (HY000): Can't connect to MySQL server on '222.222.222.222' (113)

Please check firewalls and make sure that port TCP 3306 is not blocked.



[edit] TRUNCATE usage

Do not use TRUNCATE statement because it can ruin Replication.




[edit] See also

Personal tools

Try MOR

Try M2