MySQL 5.6 upgrades are in full swing these days and knowing how to safely upgrade from MySQL 5.5 to 5.6 is important. When upgrading a replication environment, it’s important that you can build a migration plan that safely allows for your upgrade with minimal risk — rollback is often a very important component to this.
For many people this means upgrading slaves first and then the master. The strategy of an older master replicating to a newer slave is well known and has been supported in MySQL replication for a very long time. To be specific: you can have a MySQL 5.6 slave of a 5.5 master and this should work fine until you upgrade your master and/or promote one of the slaves to be the master.
However, there are those of us who like to live on the edge and do unsupported things. Suppose that when you cut over to that MySQL 5.6 master your application completely breaks. What would your rollback plan be? In such a case, leaving a 5.5 slave of the new 5.6 master (or perhaps a dual-master setup with 5.5 and 5.6) would be useful to allow you to rollback to but still have the data written on the 5.6 master.
With Statement-based replication (SBR), you are generally ok with this type of setup, provided you aren’t doing any MySQL 5.6 syntax-specific things until you don’t have any more 5.5 slaves. However, with Row-based replication (RBR), things are a bit trickier, particularly when column formats change.
Now, one nice new feature of MySQL 5.6 is the improvement of thestorage requirements forDATETIME fields as well as the addition of fractional second support for TIME, DATETIME, and TIMESTAMP. This is great, but unfortunately this is a new column format that 5.5 clearly would not understand. Does this put our 5.6 to 5.5 replication in jeopardy? The answer is, if we’re careful, NO.
Quite simply, MySQL 5.6 supports both old and new types and mysql_upgrade does not make such a conversion on existing tables. Only NEW tables or REBUILT tables in 5.6 will use the new format. Any tables from 5.5 with a simple mysql_upgrade to 5.6 will still be using the old types. For more information on how to find columns in 5.6 that are using the old format, seeIke Walker’s excellent blog post on the topic. (Thanks Ike!)
To test this out, I created a simple experiment. I have a master and slave using RBR, both on 5.5, and I setuppt-heartbeatto update the master. I realized that pt-heartbeat actually uses a varchar for the timestamp field — I suspect this makes multiple database support easier. However, since pt-heartbeat’s update uses a NOW() to populate that field, I can convert it to a DATETIME:
1 |
|
[root@master~]# pt-heartbeat --update --database percona --create-table CREATETABLE`heartbeat`( `ts`varchar(26)NOTNULL, `server_id`int(10)unsignedNOTNULL, `file`varchar(255)DEFAULTNULL, `position`bigint(20)unsignedDEFAULTNULL, `relay_master_log_file`varchar(255)DEFAULTNULL, `exec_master_log_pos`bigint(20)unsignedDEFAULTNULL, PRIMARYKEY(`server_id`) )ENGINE=InnoDBDEFAULTCHARSET=latin1 mastermysql>altertableheartbeatdropcolumnts,addcolumntsDATETIME; slavemysql>select*fromheartbeat/G ***************************1.row*************************** server_id:1 file:master-bin.000002 position:5107583 relay_master_log_file:NULL exec_master_log_pos:NULL ts:2014-05-0217:03:59 1rowinset(0.00sec) CREATETABLE`heartbeat`( `server_id`int(10)unsignedNOTNULL, `file`varchar(255)DEFAULTNULL, `position`bigint(20)unsignedDEFAULTNULL, `relay_master_log_file`varchar(255)DEFAULTNULL, `exec_master_log_pos`bigint(20)unsignedDEFAULTNULL, `ts`datetimeDEFAULTNULL, PRIMARYKEY(`server_id`) )ENGINE=InnoDBDEFAULTCHARSET=latin1 |
So my heartbeat table now has a 5.5 DATETIME, pt-heartbeat is working properly, and the heartbeat is replicating to the slave. Now I will upgrade my master to MySQL 5.6:
1 |
|
[root@master~]# rpm -e Percona-Server-devel-55-5.5.36-rel34.2.el6.x86_64 Percona-Server-shared-55-5.5.36-rel34.2.el6.x86_64 Percona-Server-client-55-5.5.36-rel34.2.el6.x86_64 Percona-Server-server-55-5.5.36-rel34.2.el6.x86_64 --nodeps [root@master~]# yum install Percona-Server-server-56.x86_64 ============================================================================================================== Package Arch Version Repository Size ============================================================================================================== Installing:Percona-Server-server-56 x86_64 5.6.16-rel64.2.el6 Percona 19M Installingfordependencies: Percona-Server-client-56 x86_64 5.6.16-rel64.2.el6 Percona 6.8M Percona-Server-shared-56 x86_64 5.6.16-rel64.2.el6 Percona 712k TransactionSummary ============================================================================================================== Install 3Package(s) ...[root@master~]# service mysql start StartingMySQL(PerconaServer).......SUCCESS! [root@master~]# mysql WelcometotheMySQLmonitor. Commandsendwith;or/g. YourMySQLconnectionidis1 Serverversion:5.6.16-64.2-56-logPerconaServer(GPL),Release64.2,Revision569 [root@master~]# mysql_upgrade Lookingfor'mysql'as:mysql Lookingfor'mysqlcheck'as:mysqlcheck Running'mysqlcheckwithdefaultconnectionarguments Running'mysqlcheckwithdefaultconnectionarguments mysql.columns_priv OK mysql.db OK mysql.event OK mysql.func OK mysql.general_log OK mysql.help_category OK mysql.help_keyword OK mysql.help_relation OK mysql.help_topic OK mysql.host OK mysql.ndb_binlog_index OK mysql.plugin OK mysql.proc OK mysql.procs_priv OK mysql.proxies_priv OK mysql.servers OK mysql.slow_log OK mysql.tables_priv OK mysql.time_zone OK mysql.time_zone_leap_second OK mysql.time_zone_name OK mysql.time_zone_transition OK mysql.time_zone_transition_type OK mysql.user OK Running'mysql_fix_privilege_tables'... Running'mysqlcheckwithdefaultconnectionarguments Running'mysqlcheckwithdefaultconnectionarguments percona.heartbeat OK OK |
I can now verify that Ike’s INFORMATION_SCHEMA queries correctly detect the ‘heartbeat.ts’ column as the old format:
1 |
|
mastermysql>selectt.table_schema,t.engine,t.table_name,c.column_name,c.column_type frominformation_schema.tablest innerjoininformation_schema.columnsconc.table_schema=t.table_schemaandc.table_name=t.table_name leftouterjoininformation_schema.innodb_sys_tablesistonist.name=concat(t.table_schema,'/',t.table_name) leftouterjoininformation_schema.innodb_sys_columnsisconisc.table_id=ist.table_idandisc.name=c.column_name wherec.column_typein('time','timestamp','datetime') andt.table_schemanotin('mysql','information_schema','performance_schema') andt.table_type='base table' and(t.engine!='innodb'or(t.engine='innodb'andisc.mtype=6)) orderbyt.table_schema,t.table_name,c.column_name; +--------------+--------+------------+-------------+-------------+ |table_schema|engine|table_name|column_name|column_type| +--------------+--------+------------+-------------+-------------+ |percona |InnoDB|heartbeat |ts |datetime | +--------------+--------+------------+-------------+-------------+ 1rowinset(0.04sec) |
To make replication work from MySQL 5.6 to 5.5, I also had to add a few backwards compatibility options on the master:
1 |
|
log_bin_use_v1_row_events=ON binlog_checksum=NONE |
Once I fixed that up, I can verify my slave is still working after this and receiving heartbeats. Clearly the new formats are not a show-stopper for backwards replication compatibility.
1 |
|
slavemysql>showslavestatus/G ***************************1.row*************************** Slave_IO_State:Waitingformastertosendevent Master_Host:192.168.70.2 Master_User:repl Master_Port:3306 Connect_Retry:60 Master_Log_File:master-bin.000005 Read_Master_Log_Pos:120 Relay_Log_File:slave-relay-bin.000002 Relay_Log_Pos:267 Relay_Master_Log_File:master-bin.000005 Slave_IO_Running:Yes Slave_SQL_Running:Yes mastermysql>select*fromheartbeat; +-----------+-------------------+----------+-----------------------+---------------------+---------------------+ |server_id|file |position|relay_master_log_file|exec_master_log_pos|ts | +-----------+-------------------+----------+-----------------------+---------------------+---------------------+ | 1|master-bin.000002| 5115935|NULL | NULL|2014-05-0217:04:23| +-----------+-------------------+----------+-----------------------+---------------------+---------------------+ 1rowinset(0.01sec) slavemysql>select*fromheartbeat; +-----------+-------------------+----------+-----------------------+---------------------+---------------------+ |server_id|file |position|relay_master_log_file|exec_master_log_pos|ts | +-----------+-------------------+----------+-----------------------+---------------------+---------------------+ | 1|master-bin.000002| 5115935|NULL | NULL|2014-05-0217:04:23| +-----------+-------------------+----------+-----------------------+---------------------+---------------------+ 1rowinset(0.00sec) |
But, if I’m not careful on MySQL 5.6, and rebuild the table, the new format does clearly bite me:
1 |
|
mastermysql>setsql_log_bin=0; QueryOK,0rowsaffected(0.00sec) mastermysql>altertablepercona.heartbeatforce; QueryOK,1rowaffected,1warning(0.18sec) Records:1 Duplicates:0 Warnings:1 mastermysql>showwarnings; +-------+------+-------------------------------------------------------------------------------------+ |Level|Code|Message | +-------+------+-------------------------------------------------------------------------------------+ |Note |1880|TIME/TIMESTAMP/DATETIMEcolumnsofoldformathavebeenupgradedtothenewformat.| +-------+------+-------------------------------------------------------------------------------------+ 1rowinset(0.00sec) slavemysql>showslavestatus/G ***************************1.row*************************** ...Slave_IO_Running:Yes Slave_SQL_Running:No ...Last_Errno:1677 Last_Error:Column5oftable'percona.heartbeat'cannotbeconvertedfromtype''totype'datetime' ...Last_SQL_Errno:1677 Last_SQL_Error:Column5oftable'percona.heartbeat'cannotbeconvertedfromtype''totype'datetime' Replicate_Ignore_Server_Ids: Master_Server_Id:1 1rowinset(0.00sec) |
What does all this teach us?
While the MySQL version is important, for RBR what matters most is the actual current format for each column. Your master and slave(s) MUST have the same column formats for RBR to work right.
So, the new temporal formats do not necessarily break RBR replication back to 5.5, provided:
However, I want to make it clear that MySQL 5.6 to 5.5 replication is technically unsupported. I have not exhausted all possibilities for problems with 5.6 to 5.5 RBR replication, just this specific one. If you choose to make an upgrade strategy that relies on backwards replication in this way, be prepared for it to not work and test it thoroughly in advance. The purpose of this post is to simply point out that data type formats, in and of themselves, do not necessarily break RBR backwards compatibility.