Codership Oy http://www.codership.com DISCLAIMER THIS SOFTWARE PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT SHALL CODERSHIP OY BE HELD LIABLE TO ANY PARTY FOR ANY DAMAGES RESULTING DIRECTLY OR INDIRECTLY FROM THE USE OF THIS SOFTWARE. Trademark Information. MySQL is a trademark or registered trademark of Oracle and/or its affiliates. Other trademarks are the property of their respective owners. Licensing Information. Please see file COPYING that came with this distribution Source code can be found at wsrep API: https://launchpad.net/wsrep MySQL patch: https://launchpad.net/codership-mysql ABOUT THIS DOCUMENT This document covers installation and configuration issues specific to this wsrep-patched MySQL distribution by Codership. It does not cover the use or administration of MySQL server per se. The reader is assumed to know how to install, configure, administer and use standard MySQL server version 5.1.xx. MYSQL-5.5.x/wsrep-23.x CONTENTS: ========= 1. WHAT IS WSREP PATCH FOR MYSQL 2. INSTALLATION 3. FIRST TIME SETUP 3.1 CONFIGURATION FILES 3.2 DATABASE PRIVILEGES 3.3 CHECK AND CORRECT FIREWALL SETTINGS 3.4 SELINUX 3.5 APPARMOR 3.6 CONNECT TO CLUSTER 4. UPGRADING FROM MySQL 5.1.x 5. CONFIGURATION OPTIONS 5.1 MANDATORY MYSQL OPTIONS 5.2 WSREP OPTIONS 6. ONLINE SCHEMA UPGRADE 6.1 TOTAL ORDER ISOLATION (TOI) 6.2 ROLLING SCHEMA UPGRADE (RSU) 7. LIMITATIONS 1. WHAT IS WSREP PATCH FOR MYSQL/INNODB Wsrep API developed by Codership Oy is a modern generic (database-agnostic) replication API for transactional databases with a goal to make database replication/logging subsystem completely modular and pluggable. It is developed with flexibility and completeness in mind to satisfy broad range of modern replication scenarios. It is equally suitable for synchronous and asynchronous, master-slave and multi-master replication. wsrep stands for Write Set REPlication. Wsrep patch for MySQL/InnoDB allows MySQL server to load and use various wsrep API implementations ("wsrep providers") with different qualities of service. Without wsrep provider MySQL-wsrep server will function like a regular standalone server. 2. INSTALLATION In the examples below mysql authentication options are omitted for brevity. 2.1 Download and install mysql-wsrep package. Download binary package for your Linux distribution from https://launchpad.net/codership-mysql/ 2.1.1 On Debian and Debian-derived distributions. Upgrade from mysql-server-5.0 to mysql-wsrep is not supported yet, please upgrade to mysql-server-5.1 first. If you're installing over an existing mysql installation, mysql-server-wsrep will conflict with mysql-server-5.1 package, so remove it first: $ sudo apt-get remove mysql-server-5.1 mysql-server-core-5.1 mysql-server-wsrep requires psmisc and mysql-client-5.1.47 (or later). MySQL 5.1 packages can be found from backports repositories. For further information about configuring and using Debian or Ubuntu backports, see: * http://backports.debian.org * https://help.ubuntu.com/community/UbuntuBackports For example, installation of required packages on Debian Lenny: $ sudo apt-get install psmisc $ sudo apt-get -t lenny-backports install mysql-client-5.1 Now you should be able to install mysql-wsrep package: $ sudo dpkg -i 2.1.2 On CentOS and similar RPM-based distributions. If you're migrating from existing MySQL installation, there are two variants: a) If you're already using official MySQL-server-community 5.1.x RPM from Oracle: # rpm -e mysql-server b) If you're upgrading from the stock mysql-5.0.77 on CentOS: 1) Make sure that the following packages are not installed: # rpm --nodeps --allmatches -e mysql-server mysql-test mysql-bench 2) Install *official* MySQL-shared-compat-5.1.x from http://dev.mysql.com/downloads/mysql/5.1.html Actual installation: # rpm -Uvh If this fails due to unsatisfied dependencies, install missing packages (e.g. yum install perl-DBI) and retry. Additional packages to consider (if not yet installed): * galera (multi-master replication provider, https://launchpad.net/galera) * MySQL-client-community (for connecting to server and mysqldump-based SST) * rsync (for rsync-based SST) * xtrabackup and nc (for xtrabackup-based SST) 2.2 Upgrade system tables. If you're upgrading a previous MySQL installation, it might be advisable to upgrade system tables. To do that start mysqld and run mysql_upgrade command. Consult MySQL documentation in case of errors. Normally they are not critical and can be ignored unless specific functionality is needed. 3. FIRST TIME SETUP Unless you're upgrading an already installed mysql-wsrep package, you will need to set up a few things to prepare server for operation. 3.1 CONFIGURATION FILES * Make sure system-wide my.cnf does not bind mysqld to 127.0.0.1. That is, if you have the following line in [mysqld] section, comment it out: #bind-address = 127.0.0.1 * Make sure system-wide my.cnf contains "!includedir /etc/mysql/conf.d/" line. * Edit /etc/mysql/conf.d/wsrep.cnf and set wsrep_provider option by specifying a path to provider library. If you don't have a provider, leave it as it is. * When a new node joins the cluster it'll have to receive a state snapshot from one of the peers. This requires a privileged MySQL account with access from the rest of the cluster. Edit /etc/mysql/conf.d/wsrep.cnf and set mysql login/password pair for SST, for example: wsrep_sst_auth=wsrep_sst:wspass * See CONFIGURATION section below about other configuration parameters that you might want to change at this point. 3.2 DATABASE PRIVILEGES Restart MySQL server and connect to it as root to grant privileges to SST account (empty users confuse MySQL authentication matching rules, we need to delete them too): $ mysql -e "SET wsrep_on=OFF; DELETE FROM mysql.user WHERE user='';" $ mysql -e "SET wsrep_on=OFF; GRANT ALL ON *.* TO wsrep_sst@'%' IDENTIFIED BY 'wspass'"; 3.3 CHECK AND CORRECT FIREWALL SETTINGS. MySQL-wsrep server needs to be accessible from other cluster members through its client listening socket and through wsrep provider socket. See your distribution and wsrep provider documentation for details. For example on CentOS you might need to do something along these lines: # iptables --insert RH-Firewall-1-INPUT 1 --proto tcp --source /24 --destination /32 --dport 3306 -j ACCEPT # iptables --insert RH-Firewall-1-INPUT 1 --proto tcp --source /24 --destination /32 --dport 4567 -j ACCEPT If there is a NAT firewall between the nodes, it must be configured to allow direct connections between the nodes (e.g. via port forwarding). 3.4 SELINUX If you have SELinux enabled, it may block mysqld from doing required operations. You'll need to either disable it or configure to allow mysqld to run external programs and open listen sockets at unprivileged ports (i.e. things that an unprivileged user can do). See SELinux documentation about it. To quickly disable SELinux: 1) run 'setenforce 0' as root. 2) set 'SELINUX=permissive' in /etc/selinux/config 3.5 APPARMOR AppArmor automatically comes with Ubuntu and may also prevent mysqld to from opening additional ports or run scripts. See AppArmor documentation about its configuration. To disable AppArmor for mysqld: $ cd /etc/apparmor.d/disable/ $ sudo ln -s /etc/apparmor.d/usr.sbin.mysqld $ sudo service apparmor restart 3.6 CONNECT TO CLUSTER Now you're ready to connect to cluster by setting wsrep_cluster_address variable and monitor status of wsrep provider: mysql> SET GLOBAL wsrep_cluster_address=''; mysql> SHOW STATUS LIKE 'wsrep%'; 4 UPGRADING FROM MySQL 5.1.x !!! THESE INSTRUCTIONS ARE PRELIMINARY AND INCOMPLETE !!! 1) BEFORE UPGRADE (while running 5.1.x): - comment out 'wsrep_provider' setting from configuration files (my.cnf and/or wsrep.cnf) - If performing a rolling upgrade on a running cluster, set wsrep_sst_method=mysqldump. You might also need to configure wsrep_sst_receive_address and wsrep_sst_auth appropriately. mysqldump is the only way to transfer data from 5.1.x to 5.5.x reliably. - remove innodb_plugin settings from configuration files. 2) Perform upgrade as usual: http://dev.mysql.com/doc/refman/5.5/en/upgrading-from-previous-series.html Don't forget to run 'mysql_upgrade' command. 3) AFTER UPGRADING individual node: - uncomment 'wsrep_provider' line in configuration file. - restart the server and join the cluster. 4) AFTER UPGRADING the whole cluster: - revert to usual wsrep SST settings if not 'mysqldump'. 5. CONFIGURATION OPTIONS 5.1 MANDATORY MYSQL OPTIONS binlog_format=ROW This option is required to use row-level replication as opposed to statement-level. For performance and consistency considerations don't change that. As a side effect, binlog, if turned on, can be ROW only. In future this option won't have special meaning. innodb_autoinc_lock_mode=2 This is a required parameter. Without it INSERTs into tables with AUTO_INCREMENT column may fail. autoinc lock modes 0 and 1 can cause unresolved deadlock, and make system unresponsive. innodb_locks_unsafe_for_binlog=1 This option is required for parallel applying. 5.2 WSREP OPTIONS All options are optional except for wsrep_provider, wsrep_cluster_address, and wsrep_sst_auth. wsrep_provider=none A full path to the library that implements WSREP interface. If none is specified, the server behaves like a regular mysqld. wsrep_provider_options= Provider-specific option string. Check wsrep provider documentation or http://www.codership.com/wiki wsrep_cluster_address= Provider-specific cluster address string. This is used to connect a node to the desired cluster. This option can be given either on mysqld startup or set during runtime. See wsrep provider documentation for possible values. wsrep_cluster_name="my_wsrep_cluster" Logical cluster name, must be the same for all nodes of the cluster. wsrep_node_address= An option to explicitly specify the network address of the node in the form
[:port] if autoguessing for some reason does not produce desirable results (multiple network interfaces, NAT, etc.) If not explicitly overridden by wsrep_sst_receive_address, the
part will be used to listen for SST (see below). And the whole
[:port] will be passed to wsrep provider to be used as a base address in its communications. wsrep_node_name= Human readable node name (for easier log reading only). Defaults to hostname. wsrep_slave_threads=1 Number of threads dedicated to processing of writesets from other nodes. For best performance should be few per CPU core. wsrep_dbug_option Options for the built-in DBUG library (independent from what MySQL uses). Empty by default. Not currently in use. wsrep_debug=0 Enable debug-level logging. wsrep_convert_LOCK_to_trx=0 Implicitly convert locking sessions into transactions inside mysqld. By itself it does not mean support for locking sessions, but it prevents the database from going into logically inconsistent state. Note however, that loading large database dump with LOCK statements might result in abnormally large transactions and cause an out-of-memory condition wsrep_retry_autocommit=1 Retry autocommit queries and single statement transactions should they fail certification test. This is analogous to rescheduling an autocommit query should it go into deadlock with other transactions in the database lock manager. wsrep_auto_increment_control=1 Automatically adjust auto_increment_increment and auto_increment_offset variables based on the number of nodes in the cluster. Significantly reduces certification conflict rate for INSERTS. wsrep_drupal_282555_workaround=1 MySQL seems to have an obscure bug when INSERT into table with AUTO_INCREMENT column with NULL value for that column can fail with a duplicate key error. When this option is on, it retries such INSERTs. Required for stable Drupal operation. Documented at: http://bugs.mysql.com/bug.php?id=41984 http://drupal.org/node/282555 wsrep_causal_reads=0 Enforce strict READ COMMITTED semantics on reads and transactions. May result in additional latencies. It is a session variable. wsrep_OSU_method=TOI Online Schema Upgrade (OSU) can be performed with two alternative methods: Total Order Isolation (TOI) runs DDL statement in all cluster nodes in same total order sequence locking the affected table for the duration of the operation. This may result in the whole cluster being blocked for the duration of the operation. Rolling Schema Upgrade (RSU) executes the DDL statement only locally, thus blocking only one cluster node. During the DDL processing, the node is not replicating and may be unable to process replication events (due to table lock). Once DDL operation is complete, the node will catch up and sync with the cluster to become fully operational again. The DDL statement or its effects are not replicated, so it is user's responsibility to manually perform this operation on each of the nodes. wsrep_forced_binlog_format=none Force every transaction to use given binlog format. When this variable is set to something else than NONE, all transactions will use the given forced format, regardless of what the client session has specified in binlog_format. Valid choices for wsrep_forced_binlog_format are: ROW, STATEMENT, MIXED and special value NONE, meaning that there is no forced binlog format in effect. This variable was introduced to support STATEMENT format replication during rolling schema upgrade processing. However, in most cases ROW replication is valid for asymmetrict schema replication. State snapshot transfer options. When a new node joins the cluster it has to synchronize its initial state with the other cluster members by transferring state snapshot from one of them. The options below govern how this happens and should be set up before attempting to join or start a cluster. wsrep_sst_method=rsync What method to use to copy database state to a newly joined node. Supported methods: - mysqldump: slow (except for small datasets) but allows for upgrade between major MySQL versions or InnoDB features. - rsync: much faster on large datasets (default). - rsync_wan: same as rsync but with deltaxfer to minimize network traffic. - xtrabackup: very fast and practically non-blocking SST method based on Percona's xtrabackup tool. (for xtrabackup to work the following settings must be present in my.cnf on all nodes: [mysqld] wsrep_sst_auth=root: datadir= [client] socket= ) wsrep_sst_receive_address= Address (hostname:port) at which this node wants to receive state snapshot. Defaults to mysqld bind address, and if that is not specified (0.0.0.0) - to the first IP of eth0 + mysqld bind port. NOTE: check that your firewall allows connections to this address from other cluster nodes. wsrep_sst_auth= Authentication information needed for state transfer. Depends on the state transfer method. For mysqldump-based SST it is : and should be the same on all nodes - it is used to authenticate with both state snapshot receiver and state snapshot donor. wsrep_sst_donor= A name of the node which should serve as state snapshot donor. This allows to control which node will serve state snapshot request. By default the most suitable node is chosen by wsrep provider. This is the same as given in wsrep_node_name. 6. ONLINE SCHEMA UPGRADE Schema upgrades mean any data definition statements (DDL statemnents) run for the database. They change the database structure and are non- transactional. Release 22.3 brings a new method for performing schema upgrades. User can now choose whether to use the traditional total order isolation or new rolling schema upgrade method. The OSU method choice is done by global parameter: 'wsrep_OSU_method'. 6.1 Total Order Isolation (TOI) With earlier releases, DDL processing happened always by Total Order Isolation (TOI) method. With TOI, the DDL was scheduled to be processed in same transaction seqeuncing 'slot' in each cluster node. The processing is secured by locking the affected table from any other use. With TOI method, the whole cluster has part of the database locked for the duration of the DDL processing. 6.2 Rolling Schema Upgrade (RSU) Rolling schema upgrade is new DDL processing method, where DDL will be processed locally for the node. The node is disconnected of the replication for the duration of the DDL processing, so that there is only DDL statement processing in the node and it does not block the rest of the cluster. When the DDL processing is complete, the node applies delayed replication events and synchronizes back with the cluster. The DDL can then be executed cluster-wide by running the same DDL statement for each node in turn. When this rolling schema upgrade proceeds, part of the cluster will have old schema structure and part of the cluster will have new schema structure. 7. LIMITATIONS 1) Currently replication works only with InnoDB storage engine. Any writes to tables of other types, including system (mysql.*) tables are not replicated. However, DDL statements are replicated in statement level, and changes to mysql.* tables will get replicated that way. So, you can safely issue: CREATE USER..., but issuing: INSERT INTO mysql.user..., will not be replicated. 2) DELETE operation is unsupported on tables without primary key. Also rows in tables without primary key may appear in different order on different nodes. As a result SELECT...LIMIT... may return slightly different sets. 3) Unsupported queries: * LOCK/UNLOCK TABLES cannot be supported in multi-master setups. * lock functions (GET_LOCK(), RELEASE_LOCK()... ) 4) Query log cannot be directed to table. If you enable query logging, you must forward the log to a file: log_output = FILE Use general_log and general_log_file to choose query logging and the log file name 5) Maximum allowed transaction size is defined by wsrep_max_ws_rows and wsrep_max_ws_size. Anything bigger (e.g. huge LOAD DATA) will be rejected. 6) Due to cluster level optimistic concurrency control, transaction issuing COMMIT may still be aborted at that stage. There can be two transactions. writing to same rows and committing in separate cluster nodes, and only one of the them can successfully commit. The failing one will be aborted. For cluster level aborts, MySQL/galera cluster gives back deadlock error. code (Error: 1213 SQLSTATE: 40001 (ER_LOCK_DEADLOCK)). 7) XA transactions can not be supported due to possible rollback on commit.