You are here

Agrégateur de flux

Impact of column types on MySQL JOIN performance

Shinguz - Wed, 2013-12-11 20:12
Taxonomy upgrade extras: sqlquerytuningmysql

In our MySQL trainings and consulting engagements we tell our customers always to use the smallest possible data type to get better query performance. Especially for the JOIN columns. This advice is supported as well by the MySQL documentation in the chapter Optimizing Data Types:

Use the most efficient (smallest) data types possible. MySQL has many specialized types that save disk space and memory. For example, use the smaller integer types if possible to get smaller tables. MEDIUMINT is often a better choice than INT because a MEDIUMINT column uses 25% less space.

I remember somewhere the JOIN columns where explicitly mentioned but I cannot find it any more.

Test set-up

To get numbers we have created a little test set-up:

CREATE TABLE `a` ( `id` int(10) unsigned NOT NULL AUTO_INCREMENT , `data` varchar(64) DEFAULT NULL , `ts` timestamp NOT NULL , PRIMARY KEY (`id`) ) ENGINE=InnoDB CHARSET=latin1  
CREATE TABLE `b` ( `id` int(10) unsigned NOT NULL AUTO_INCREMENT , `data` varchar(64) DEFAULT NULL , `ts` timestamp NOT NULL , `a_id` int(10) unsigned DEFAULT NULL , PRIMARY KEY (`id`) ) ENGINE=InnoDB DEFAULT CHARSET=latin1
1048576 rows 16777216 rows

The following query was used for the test:

EXPLAIN SELECT * FROM a JOIN b ON b.a_id = a.id WHERE a.id BETWEEN 10000 AND 15000; +----+-------------+-------+--------+---------------+---------+---------+-------------+----------+-------------+ | id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra | +----+-------------+-------+--------+---------------+---------+---------+-------------+----------+-------------+ | 1 | SIMPLE | b | ALL | NULL | NULL | NULL | NULL | 16322446 | Using where | | 1 | SIMPLE | a | eq_ref | PRIMARY | PRIMARY | 4 | test.b.a_id | 1 | NULL | +----+-------------+-------+--------+---------------+---------+---------+-------------+----------+-------------+

And yes: I know this query could be more optimal by setting an index on b.a_id.

Results

The whole workload was executed completely in memory and thus CPU bound (we did not want to measure the speed of our I/O system).

SEJOIN columnbytesquery timeGainSpaceCharacter setInnoDBMEDIUMINT35.28 s96%4% faster75%InnoDBINT45.48 s100%100%100%InnoDBBIGINT85.65 s107%7% slower200%InnoDBNUMERIC(7, 2)~46.77 s124%24% slower~100%InnoDBVARCHAR(7)7-86.44 s118%18% slower~200%latin1InnoDBVARCHAR(16)7-86.44 s118%18% slower~200%latin1InnoDBVARCHAR(32)7-86.42 s118%18% slower~200%latin1InnoDBVARCHAR(128)7-86.46 s118%18% slower~200%latin1InnoDBVARCHAR(256)8-96.17 s114%14% slower~225%latin1InnoDBVARCHAR(16)7-86.96 s127%27% slower~200%utf8InnoDBVARCHAR(128)7-86.82 s124%24% slower~200%utf8InnoDBCHAR(16)166.85 s125%25% slower400%latin1InnoDBCHAR(128)1289.68 s177%77% slower3200%latin1InnoDBTEXT8-910.7 s195%95% slower~225%latin1MyISAMINT43.16 s58%42% fasterTokuDBINT44.52 s82%18% faster

Some comments to the tests:

  • MySQL 5.6.13 was used for most of the tests.
  • TokuDB v7.1.0 was tested with MySQL 5.5.30.
  • As results the optimistic cases were taken. In reality the results can be slightly worse.
  • We did not take into consideration that bigger data types will eventually cause more I/O which is very slow!
Commands ALTER TABLE a CONVERT TO CHARACTER SET latin1; ALTER TABLE b CONVERT TO CHARACTER SET latin1; ALTER TABLE a MODIFY COLUMN id INT UNSIGNED NOT NULL; ALTER TABLE b MODIFY COLUMN a_id INT UNSIGNED NOT NULL;

MySQL Environment MyEnv 1.0.2 has been released

FromDual.en - Wed, 2013-12-11 15:58

FromDual has the pleasure to announce the release of the new version 1.0.2 of its popular multi-instance MySQL Environment MyEnv.

You can download MyEnv from here.

In the inconceivable case that you find a bug in MyEnv please report it to our Bugtracker.

Any feedback, statements and testimonials are welcome as well! Please send them to feedback@fromdual.com.

If you have questions about MyEnv we will answer them in the MyEnv forum.

Upgrade from 1.0.x to 1.0.2 # cd /home/mysql/product # tar xf /tmp/myenv-1.0.2.tar.gz # rm -f myenv # ln -s myenv-1.0.2 myenv

Add the following line to your ~/.bash_profile:

# BEGIN MyEnv # Written by the MyEnv installMyEnv.php script. . /etc/myenv/MYENV_BASE MYENV_PWD=`pwd` cd $MYENV_BASE/bin . myenv.profile up cd $MYENV_PWD # END MyEnv Changes in MyEnv 1.0.2 MyEnv
  • Missing [client] section in my.cnf should not lead to error any more (Bug #95).
  • profile.template removed because it is redundant to installer information.
  • up removed from myenv.profile.
  • Also root should be allowed to start MySQL (Bug #99).
  • Status for OEM agent implemented (oem_agent.php). To activate this feature copy it from utl/oem_agent.php to plg/showMyEnvStatus/.
  • Plug-in interface for showMyEnvStatus.php implemented.
  • Debug information improved.
  • Empty line after environment switch removed.
  • One missing error line in myenv_start_stop.php was added.
  • Return code in myenv_start_stop.php is nicer now.
  • Difference between my.cnf and myenv.conf should be warning and not an error.
  • Error messages made more precise.
  • Made it more clear that MyEnv is distributed under GPL v2.
  • Deleting the actual instance with MyEnv Installer should no longer confuse MyEnv (Bug #104).
  • RPM packages for CentOS 6, Fedora 18 and 19, RHEL 5 and 6 are available.
MyEnv Installer
  • User experience during deleting an instance improved (Bug #96).
  • Deleting a running instance is not allowed. Much more verbose information about (Bug #89).
  • Instance name black list added.
  • Instances should be listed when change was chosen (#93).
  • Replace multi-line readline output to fixed crippled output in menu selection (Bug #103).
MyEnv Utilities
  • alter_engine.pl should now work for socket and port, local and remote.
  • Password is not exposed any more in alter_engine.pl.
  • Query bench (query_bench.phpx) added for micro bench marks.
MySQL Backup Manager
  • Automated mysql_bman tests should all pass again.
Taxonomy upgrade extras: myenvoperationMySQL Operationsmulti instanceconsolidationrelease

MySQL Environment MyEnv 1.0.2 has been released

FromDual.en - Wed, 2013-12-11 15:58

FromDual has the pleasure to announce the release of the new version 1.0.2 of its popular multi-instance MySQL Environment MyEnv.

You can download MyEnv from here.

In the inconceivable case that you find a bug in MyEnv please report it to our Bugtracker.

Any feedback, statements and testimonials are welcome as well! Please send them to feedback@fromdual.com.

If you have questions about MyEnv we will answer them in the MyEnv forum.

Upgrade from 1.0.x to 1.0.2 # cd /home/mysql/product # tar xf /tmp/myenv-1.0.2.tar.gz # rm -f myenv # ln -s myenv-1.0.2 myenv

Add the following line to your ~/.bash_profile:

# BEGIN MyEnv # Written by the MyEnv installMyEnv.php script. . /etc/myenv/MYENV_BASE MYENV_PWD=`pwd` cd $MYENV_BASE/bin . myenv.profile up cd $MYENV_PWD # END MyEnv Changes in MyEnv 1.0.2 MyEnv
  • Missing [client] section in my.cnf should not lead to error any more (Bug #95).
  • profile.template removed because it is redundant to installer information.
  • up removed from myenv.profile.
  • Also root should be allowed to start MySQL (Bug #99).
  • Status for OEM agent implemented (oem_agent.php). To activate this feature copy it from utl/oem_agent.php to plg/showMyEnvStatus/.
  • Plug-in interface for showMyEnvStatus.php implemented.
  • Debug information improved.
  • Empty line after environment switch removed.
  • One missing error line in myenv_start_stop.php was added.
  • Return code in myenv_start_stop.php is nicer now.
  • Difference between my.cnf and myenv.conf should be warning and not an error.
  • Error messages made more precise.
  • Made it more clear that MyEnv is distributed under GPL v2.
  • Deleting the actual instance with MyEnv Installer should no longer confuse MyEnv (Bug #104).
  • RPM packages for CentOS 6, Fedora 18 and 19, RHEL 5 and 6 are available.
MyEnv Installer
  • User experience during deleting an instance improved (Bug #96).
  • Deleting a running instance is not allowed. Much more verbose information about (Bug #89).
  • Instance name black list added.
  • Instances should be listed when change was chosen (#93).
  • Replace multi-line readline output to fixed crippled output in menu selection (Bug #103).
MyEnv Utilities
  • alter_engine.pl should now work for socket and port, local and remote.
  • Password is not exposed any more in alter_engine.pl.
  • Query bench (query_bench.phpx) added for micro bench marks.
MySQL Backup Manager
  • Automated mysql_bman tests should all pass again.
Taxonomy upgrade extras: myenvoperationMySQL Operationsmulti instanceconsolidationrelease

MySQL Environment MyEnv 1.0.2 has been released

FromDual.en - Wed, 2013-12-11 15:58
Taxonomy upgrade extras: myenvoperationMySQL Operationsmulti instanceconsolidation

FromDual has the pleasure to announce the release of the new version 1.0.2 of its popular multi-instance MySQL Environment MyEnv.

You can download MyEnv from here.

In the inconceivable case that you find a bug in MyEnv please report it to our Bugtracker.

Any feedback, statements and testimonials are welcome as well! Please send them to feedback@fromdual.com.

If you have questions about MyEnv we will answer them in the MyEnv forum.

Upgrade from 1.0.x to 1.0.2 # cd /home/mysql/product # tar xf /tmp/myenv-1.0.2.tar.gz # rm -f myenv # ln -s myenv-1.0.2 myenv

Add the following line to your ~/.bash_profile:

# BEGIN MyEnv # Written by the MyEnv installMyEnv.php script. . /etc/myenv/MYENV_BASE MYENV_PWD=`pwd` cd $MYENV_BASE/bin . myenv.profile up cd $MYENV_PWD # END MyEnv Changes in MyEnv 1.0.2 MyEnv
  • Missing [client] section in my.cnf should not lead to error any more (Bug #95).
  • profile.template removed because it is redundant to installer information.
  • up removed from myenv.profile.
  • Also root should be allowed to start MySQL (Bug #99).
  • Status for OEM agent implemented (oem_agent.php). To activate this feature copy it from utl/oem_agent.php to plg/showMyEnvStatus/.
  • Plug-in interface for showMyEnvStatus.php implemented.
  • Debug information improved.
  • Empty line after environment switch removed.
  • One missing error line in myenv_start_stop.php was added.
  • Return code in myenv_start_stop.php is nicer now.
  • Difference between my.cnf and myenv.conf should be warning and not an error.
  • Error messages made more precise.
  • Made it more clear that MyEnv is distributed under GPL v2.
  • Deleting the actual instance with MyEnv Installer should no longer confuse MyEnv (Bug #104).
  • RPM packages for CentOS 6, Fedora 18 and 19, RHEL 5 and 6 are available.
MyEnv Installer
  • User experience during deleting an instance improved (Bug #96).
  • Deleting a running instance is not allowed. Much more verbose information about (Bug #89).
  • Instance name black list added.
  • Instances should be listed when change was chosen (#93).
  • Replace multi-line readline output to fixed crippled output in menu selection (Bug #103).
MyEnv Utilities
  • alter_engine.pl should now work for socket and port, local and remote.
  • Password is not exposed any more in alter_engine.pl.
  • Query bench (query_bench.phpx) added for micro bench marks.
MySQL Backup Manager
  • Automated mysql_bman tests should all pass again.

Wir suchen Dich: MySQL DBA für FromDual Support

Oli Sennhauser - Wed, 2013-12-04 16:23

FromDual sucht enthusiastische Mitarbeiter die:

  • Kenntnisse über MySQL, Percona Server oder MariaDB aufweisen oder sich aneignen wollen
  • mit dem Open-Source Ökosystem vertraut sind und Beiträge dazu geleistet haben
  • als DBA oder DevOps wissen, wie man Datenbank-Systeme betreibt
  • verstehen, was beim Betrieb von Datenbanken falsch gemacht werden kann
  • gerne selbständig remote arbeiten und über IRC, Skype, Mail und Telefon zu kommunizieren gewohnt sind
  • sich auf Linux Systemen gut auskennen und wohl fühlen
  • gute Team-Player sind und zum Wachstum der Firma beitragen wollen
  • gerne den direkten Kontakt mit Kunden haben und
  • auf der Suche nach einer neuen Herausforderung sind

Stellenbeschreibung

Wir suchen deutschsprachige vollzeit MySQL Support DBA's (Sie oder Ihn), welche primär für unsere MySQL Support Dienstleistungen zuständig sind und unseren Kunden helfen, ihre MySQL Datenbanken zu betreiben (Support und remote-DBA).

Du bist fit in MySQL und:

  • hast Erfahrung im Betrieb kritischer und hoch verfügbarer produktiver MySQL Datenbanken hauptsächlich auf Linux.
  • Deine tägliche Arbeit ist MySQL-Replikation in allen Variationen.
  • weisst, wie die meist verbreitetsten MySQL HA Setups funktionieren und wie man sie wieder effizient repariert, wenn ein Problem auftritt. (Wenn Du bereits Erfahrungen mit Galera Cluster gesammelt hast, ist das ein Vorteil!)
  • kennst die gängigen Open-Source Technologien (LAMP Stack, etc.)
  • kannst Bash skripten und einfache Programme in mindestens einer verbreiteten Programmier-/Skripting-Sprache (Perl, PHP, ...) erstellen.

Du wirst im direkten Kontakt mit Kunden stehen. Du hast ein gutes Gespür für deren Probleme und kannst zuhören, weisst wie antworten und findest die eigentlichen Probleme. Du wirst proaktiv handeln, bevor etwas passiert und den Kunden wieder auf den richtigen Pfad führen.
Du bist ein guter Kommunikator und ein aktiver Team Player.

Um Deine Arbeit erledigen zu können, arbeitest Du in einer europäischen Zeitzone. Deine Arbeitszeit kannst Du in bestimmten Grenzen flexibel gestalten. Wir erwarten, dass Du Deinen Beitrag zum Bereitschaftsdienst leistest. FromDual hat voraussichtlich keine Büroräumlichkeiten in Deinem Wohnort, aber ein Umzug ist nicht notwendig: Wir ermöglichen das Arbeiten von zu hause oder unterstützen bei der Suche einer geeigneten Arbeitsräumlichkeit. Gute schriftliche und mündliche Englischkenntnisse sind zwingend.

Neben Deiner Tätigkeit als Support DBA erwarten wir, dass Du Dir laufend neue Kenntnisse aneignest und Deine Fähigkeiten verbesserst sowie dazu beiträgst, unsere Monitoring-Lösung, unsere Datenbank-Steuerung und unseren weiteren Tools zu verbessern. Im weiteren würden wir es sehr schätzen, wenn Du regelmässig zur Verfassung technischer Artikel (Blog oder Zeitschriften) beiträgst und überall mit hilfst, wo Hilfe nötig ist...

Du solltest in der Lage sein, die meiste Zeit selbständig zu arbeiten, denken und zu handeln und Dir neues Wissen selbständig anzueignen (durch Google, die MySQL Dokumentation, Ausprobieren, etc.). Wenn Du mal nicht weiterkommst, werden Dir Deine Kollegen von FromDual gerne helfen.

Wenn Du jemanden brauchst, der Dir die ganze Zeit Dein Händchen hält, ist FromDual nicht die richtige Wahl.

Wer ist FromDual?

FromDual ist die führende unabhängige MySQL Beratungs- und Dienstleistungs-Firma in Europa mit ihrem Hauptsitz in der Schweiz.

Unsere Kunden befinden sich hauptsächlich in Europa und reichen vom kleinen Start-Up bis zur europäischen Top-500 Firma.

Du wirst in einer spannenden Zeit zu uns stossen. Wir sind am wachsen und brauchen die entsprechenden Leute, welche selbst und mit uns wachsen wollen. In dem Mass, wie wir uns weiter entwickeln, muss auch unser Team wachsen uns seine Fähigkeiten erweitern.

Wie geht es weiter

Wenn Du an dieser Chance interessiert bist und Du denkst, dass Du die passende Kandidatin oder der passende Kandidat bist (wir wissen, dass es niemanden gibt, der 100% auf diese Stellenbeschreibung passt!), würden wir uns freuen, von Dir zu hören.

Bitte schicke Deinen ungeschönten Lebenslauf mit Deinen Lohnvorstellungen und einer Liste Deiner Open-Source Beiträgen, Blog-Artikel, Vorträgen, Tweets etc. an jobs@fromdual.com. Wenn Du mehr über diese Stelle erfahren oder wenn Du mit mir persönlich sprechen möchtest, ruf mich bitte an unter +41 79 830 09 33 (Oli Sennhauser, CTO). Bitte nur Bewerber, KEINE Headhunter!

Nachdem wir Deinen Lebenslauf erhalten und geprüft haben, laden wir Dich ein, Deine technischen Fähigkeiten in einem kleinen MySQL-Test unter Beweis zu stellen. Wenn Du den Test bestanden hast, laden wir Dich für die finalen Interviews ein.

Dieses Stellenangebot ist offen bis 31. Januar 2014

Wir suchen Dich: MySQL DBA für FromDual Support

Oli Sennhauser - Wed, 2013-12-04 16:23

FromDual sucht enthusiastische Mitarbeiter die:

  • Kenntnisse über MySQL, Percona Server oder MariaDB aufweisen oder sich aneignen wollen
  • mit dem Open-Source Ökosystem vertraut sind und Beiträge dazu geleistet haben
  • als DBA oder DevOps wissen, wie man Datenbank-Systeme betreibt
  • verstehen, was beim Betrieb von Datenbanken falsch gemacht werden kann
  • gerne selbständig remote arbeiten und über IRC, Skype, Mail und Telefon zu kommunizieren gewohnt sind
  • sich auf Linux Systemen gut auskennen und wohl fühlen
  • gute Team-Player sind und zum Wachstum der Firma beitragen wollen
  • gerne den direkten Kontakt mit Kunden haben und
  • auf der Suche nach einer neuen Herausforderung sind

Stellenbeschreibung

Wir suchen deutschsprachige vollzeit MySQL Support DBA's (Sie oder Ihn), welche primär für unsere MySQL Support Dienstleistungen zuständig sind und unseren Kunden helfen, ihre MySQL Datenbanken zu betreiben (Support und remote-DBA).

Du bist fit in MySQL und:

  • hast Erfahrung im Betrieb kritischer und hoch verfügbarer produktiver MySQL Datenbanken hauptsächlich auf Linux.
  • Deine tägliche Arbeit ist MySQL-Replikation in allen Variationen.
  • weisst, wie die meist verbreitetsten MySQL HA Setups funktionieren und wie man sie wieder effizient repariert, wenn ein Problem auftritt. (Wenn Du bereits Erfahrungen mit Galera Cluster gesammelt hast, ist das ein Vorteil!)
  • kennst die gängigen Open-Source Technologien (LAMP Stack, etc.)
  • kannst Bash skripten und einfache Programme in mindestens einer verbreiteten Programmier-/Skripting-Sprache (Perl, PHP, ...) erstellen.

Du wirst im direkten Kontakt mit Kunden stehen. Du hast ein gutes Gespür für deren Probleme und kannst zuhören, weisst wie antworten und findest die eigentlichen Probleme. Du wirst proaktiv handeln, bevor etwas passiert und den Kunden wieder auf den richtigen Pfad führen.
Du bist ein guter Kommunikator und ein aktiver Team Player.

Um Deine Arbeit erledigen zu können, arbeitest Du in einer europäischen Zeitzone. Deine Arbeitszeit kannst Du in bestimmten Grenzen flexibel gestalten. Wir erwarten, dass Du Deinen Beitrag zum Bereitschaftsdienst leistest. FromDual ist eine vollständig virtuelle Firma. Ein Umzug ist daher nicht notwendig (Home-Office). Gute schriftliche und mündliche Englischkenntnisse sind zwingend.

Neben Deiner Tätigkeit als Support DBA erwarten wir, dass Du Dir laufend neue Kenntnisse aneignest und Deine Fähigkeiten verbesserst sowie dazu beiträgst, unsere Monitoring-Lösung, unsere Datenbank-Steuerung und unseren weiteren Tools zu verbessern. Im weiteren würden wir es sehr schätzen, wenn Du regelmässig zur Verfassung technischer Artikel (Blog oder Zeitschriften) beiträgst und überall mit hilfst, wo Hilfe nötig ist...

Du solltest in der Lage sein, die meiste Zeit selbständig zu arbeiten, denken und zu handeln und Dir neues Wissen selbständig anzueignen (durch Google, die MySQL Dokumentation, Ausprobieren, etc.). Wenn Du mal nicht weiterkommst, werden Dir Deine Kollegen von FromDual gerne helfen.

Wenn Du jemanden brauchst, der Dir die ganze Zeit Dein Händchen hält, ist FromDual nicht die richtige Wahl.

Wer ist FromDual?

FromDual ist die führende unabhängige MySQL Beratungs- und Dienstleistungs-Firma in Europa mit ihrem Hauptsitz in der Schweiz.

Unsere Kunden befinden sich hauptsächlich in Europa und reichen vom kleinen Start-Up bis zur europäischen Top-500 Firma.

Du wirst in einer spannenden Zeit zu uns stossen. Wir sind am wachsen und brauchen die entsprechenden Leute, welche selbst und mit uns wachsen wollen. In dem Mass, wie wir uns weiter entwickeln, muss auch unser Team wachsen uns seine Fähigkeiten erweitern.

Wie geht es weiter

Wenn Du an dieser Chance interessiert bist und Du denkst, dass Du die passende Kandidatin oder der passende Kandidat bist (wir wissen, dass es niemanden gibt, der 100% auf diese Stellenbeschreibung passt!), würden wir uns freuen, von Dir zu hören.

Bitte schicke Deinen ungeschönten Lebenslauf mit Deinen Lohnvorstellungen und einer Liste Deiner Open-Source Beiträgen, Blog-Artikel, Vorträgen, Tweets etc. an jobs@fromdual.com. Wenn Du mehr über diese Stelle erfahren oder wenn Du mit mir persönlich sprechen möchtest, ruf mich bitte an unter +41 79 830 09 33 (Oli Sennhauser, CTO). Bitte nur Bewerber, KEINE Headhunter!

Nachdem wir Deinen Lebenslauf erhalten und geprüft haben, laden wir Dich ein, Deine technischen Fähigkeiten in einem kleinen MySQL-Test unter Beweis zu stellen. Wenn Du den Test bestanden hast, laden wir Dich für die finalen Interviews ein.

Dieses Stellenangebot ist offen bis 31. Januar 2014

Workbench starting/stopping multiple instance set-ups with myenv

Abdel-Mawla Gharieb - Sat, 2013-11-23 16:03
Table of Content
Introduction

MySQL Workbench is a very good and free GUI tool provided by Oracle to manage MySQL administration and development tasks. Opening many MySQL connections (same or different instances, remote or local MySQL servers) at the same time is one of its main features. While it's working fine to perform SQL statements on the different connections opened for multiple instances, but some people are asking if it is available as well to start and stop multiple MySQL instances using MySQL Workbench? if yes, how to configure it to perform such task? and also does that make any conflict with MyEnv tool - if it's installed - or not?
Yes, MySQL Workbench could be configured to start and stop multiple MySQL instances (local or remote) and it does not make any conflict with MyEnv tool.

In this article, I will describe how to configure MySQL Workbench to start and stop multiple MySQL instances and getting benefits from MyEnv scripts in this purpose.

Prerequisites

System information and installed packages:

  • Operating System: Ubuntu 12.04 (64 bit) .
  • MySQL Server: Any mysql version (I used MySQL 5.5 tarballs).
  • Number of MySQL Instances: Two instances are installed (mysql1 & mysql2).
  • MySQL Workbench: Version 6.0 .
  • MyEnv: Version 1.0.1.
What is MyEnv?

MyEnv is a set of scripts to run comfortably multiple MySQL, Percona Server or MariaDB database instances on the same server. You can even run multiple database instances with different binary versions. If you have MySQL multiple instance setups, you really should try out MyEnv.

I will not talk more about MyEnv features and its benefits rather, I'd like to mention that if you're using MyEnv and want to use MySQL Workbench at the same time, you will not face any conflict between them both and you can manage your MySQL instances by either MyEnv or MySQL Workbench. More over, you can use MyEnv scripts to configure MySQL Workbench starting/stopping multiple instances in an easy way!

For more information about MyEnv tool , you can check it out on our website myenv.


MySQL Workbench configuration
Add MySQL connections to MySQL Workbench
  • Choose a connection name for the 1st instance "mysql1" and specify the connection string:

    If you didn't adjust the "Configure Server Management" in this step - at the left bottom of the previous screen - you can open MySQL connections and perform SQL queries normally to this instance but you can neither edit the instance configuration parameters nor start/stop it.
    BTW, you can adjust it at anytime later and that what I did already in this example.

  • Add another connection for the 2nd instance "mysql2" the same like "mysql1".
Start/Stop instance configurations

To configure MySQL Workbench to start/stop instance, we need to have relevant start and stop commands because it just execute them as they would be execute in the system shell. In this case, we may get benefit of MyEnv scripts for that purpose using the following command:

$MYENV_BASE/bin/database.php $MYENV_DATABASE start|stop
Where $MYENV_BASE is the MyEnv basedir ("/opt/myenv" in this ex.) and $MYENV_DATABASE is the instance name in MyEnv (mysqld1 & mysqld2 for mysql1 & mysql2 respectively in this ex.)

Now, we can use the following window to modify System type,Configuration file path, start , stop and status commands to match each instance configurations:



Now MySQL Workbench should be able to start and stop the configured MySQL instances.


Notes:

  • MyEnv doesn't allow any user to start a mysql instance except mysql user (even if it's the root user), so that mysql OS user should be used to execute those commands and you might need to assign it SUDO permissions.
  • Start/Stop Server button in MySQL Workbench depends on the output of the status command used ("cat /opt/mysql1/data/*.pid 2>/dev/null" for checking mysql1 instance status), and the button label will be changed to execute the appropriate command accordingly (for ex. if the instance is not running, the label should be changed to "Start Server" and the start command will be executed if the button clicked and vise verse), so you should make sure that this command is returning the expected results, otherwise, Workbench wont be able to manage the instance.
  • It doesn't matter how MySQL was installed (RPM, tarballs or from source), it's the same concept, you just need to provide MyEnv start and stop commands along with the status command and then MySQL Workbench will work.
  • If MyEnv is not installed, you can also configure MySQL Workbench to start/stop multiple instances by providing normal start|stop instance commands in the "Manage Server Connections" window but those need to be prepared first.
Taxonomy upgrade extras: MyEnv

Workbench starting/stopping multiple instance set-ups with myenv

Abdel-Mawla Gharieb - Sat, 2013-11-23 16:03
Table of Content
Introduction

MySQL Workbench is a very good and free GUI tool provided by Oracle to manage MySQL administration and development tasks. Opening many MySQL connections (same or different instances, remote or local MySQL servers) at the same time is one of its main features. While it's working fine to perform SQL statements on the different connections opened for multiple instances, but some people are asking if it is available as well to start and stop multiple MySQL instances using MySQL Workbench? if yes, how to configure it to perform such task? and also does that make any conflict with MyEnv tool - if it's installed - or not?
Yes, MySQL Workbench could be configured to start and stop multiple MySQL instances (local or remote) and it does not make any conflict with MyEnv tool.

In this article, I will describe how to configure MySQL Workbench to start and stop multiple MySQL instances and getting benefits from MyEnv scripts in this purpose.

Prerequisites

System information and installed packages:

  • Operating System: Ubuntu 12.04 (64 bit) .
  • MySQL Server: Any mysql version (I used MySQL 5.5 tarballs).
  • Number of MySQL Instances: Two instances are installed (mysql1 & mysql2).
  • MySQL Workbench: Version 6.0 .
  • MyEnv: Version 1.0.1.
What is MyEnv?

MyEnv is a set of scripts to run comfortably multiple MySQL, Percona Server or MariaDB database instances on the same server. You can even run multiple database instances with different binary versions. If you have MySQL multiple instance setups, you really should try out MyEnv.

I will not talk more about MyEnv features and its benefits rather, I'd like to mention that if you're using MyEnv and want to use MySQL Workbench at the same time, you will not face any conflict between them both and you can manage your MySQL instances by either MyEnv or MySQL Workbench. More over, you can use MyEnv scripts to configure MySQL Workbench starting/stopping multiple instances in an easy way!

For more information about MyEnv tool , you can check it out on our website myenv.


MySQL Workbench configuration
Add MySQL connections to MySQL Workbench
  • Choose a connection name for the 1st instance "mysql1" and specify the connection string:

    If you didn't adjust the "Configure Server Management" in this step - at the left bottom of the previous screen - you can open MySQL connections and perform SQL queries normally to this instance but you can neither edit the instance configuration parameters nor start/stop it.
    BTW, you can adjust it at anytime later and that what I did already in this example.

  • Add another connection for the 2nd instance "mysql2" the same like "mysql1".
Start/Stop instance configurations

To configure MySQL Workbench to start/stop instance, we need to have relevant start and stop commands because it just execute them as they would be execute in the system shell. In this case, we may get benefit of MyEnv scripts for that purpose using the following command:

$MYENV_BASE/bin/database.php $MYENV_DATABASE start|stop
Where $MYENV_BASE is the MyEnv basedir ("/opt/myenv" in this ex.) and $MYENV_DATABASE is the instance name in MyEnv (mysqld1 & mysqld2 for mysql1 & mysql2 respectively in this ex.)

Now, we can use the following window to modify System type,Configuration file path, start , stop and status commands to match each instance configurations:



Now MySQL Workbench should be able to start and stop the configured MySQL instances.


Notes:

  • MyEnv doesn't allow any user to start a mysql instance except mysql user (even if it's the root user), so that mysql OS user should be used to execute those commands and you might need to assign it SUDO permissions.
  • Start/Stop Server button in MySQL Workbench depends on the output of the status command used ("cat /opt/mysql1/data/*.pid 2>/dev/null" for checking mysql1 instance status), and the button label will be changed to execute the appropriate command accordingly (for ex. if the instance is not running, the label should be changed to "Start Server" and the start command will be executed if the button clicked and vise verse), so you should make sure that this command is returning the expected results, otherwise, Workbench wont be able to manage the instance.
  • It doesn't matter how MySQL was installed (RPM, tarballs or from source), it's the same concept, you just need to provide MyEnv start and stop commands along with the status command and then MySQL Workbench will work.
  • If MyEnv is not installed, you can also configure MySQL Workbench to start/stop multiple instances by providing normal start|stop instance commands in the "Manage Server Connections" window but those need to be prepared first.
Taxonomy upgrade extras: MyEnv

Workbench starting/stopping multiple instance set-ups with myenv

Abdel-Mawla Gharieb - Sat, 2013-11-23 16:03
Table of Content
Introduction

MySQL Workbench is a very good and free GUI tool provided by Oracle to manage MySQL administration and development tasks. Opening many MySQL connections (same or different instances, remote or local MySQL servers) at the same time is one of its main features. While it's working fine to perform SQL statements on the different connections opened for multiple instances, but some people are asking if it is available as well to start and stop multiple MySQL instances using MySQL Workbench? if yes, how to configure it to perform such task? and also does that make any conflict with MyEnv tool - if it's installed - or not?
Yes, MySQL Workbench could be configured to start and stop multiple MySQL instances (local or remote) and it does not make any conflict with MyEnv tool.

In this article, I will describe how to configure MySQL Workbench to start and stop multiple MySQL instances and getting benefits from MyEnv scripts in this purpose.

Prerequisites

System information and installed packages:

  • Operating System: Ubuntu 12.04 (64 bit) .
  • MySQL Server: Any mysql version (I used MySQL 5.5 tarballs).
  • Number of MySQL Instances: Two instances are installed (mysql1 & mysql2).
  • MySQL Workbench: Version 6.0 .
  • MyEnv: Version 1.0.1.
What is MyEnv?

MyEnv is a set of scripts to run comfortably multiple MySQL, Percona Server or MariaDB database instances on the same server. You can even run multiple database instances with different binary versions. If you have MySQL multiple instance setups, you really should try out MyEnv.

I will not talk more about MyEnv features and its benefits rather, I'd like to mention that if you're using MyEnv and want to use MySQL Workbench at the same time, you will not face any conflict between them both and you can manage your MySQL instances by either MyEnv or MySQL Workbench. More over, you can use MyEnv scripts to configure MySQL Workbench starting/stopping multiple instances in an easy way!

For more information about MyEnv tool , you can check it out on our website myenv.


MySQL Workbench configuration
Add MySQL connections to MySQL Workbench
  • Choose a connection name for the 1st instance "mysql1" and specify the connection string:

    If you didn't adjust the "Configure Server Management" in this step - at the left bottom of the previous screen - you can open MySQL connections and perform SQL queries normally to this instance but you can neither edit the instance configuration parameters nor start/stop it.
    BTW, you can adjust it at anytime later and that what I did already in this example.

  • Add another connection for the 2nd instance "mysql2" the same like "mysql1".
Start/Stop instance configurations

To configure MySQL Workbench to start/stop instance, we need to have relevant start and stop commands because it just execute them as they would be execute in the system shell. In this case, we may get benefit of MyEnv scripts for that purpose using the following command:

$MYENV_BASE/bin/database.php $MYENV_DATABASE start|stop
Where $MYENV_BASE is the MyEnv basedir ("/opt/myenv" in this ex.) and $MYENV_DATABASE is the instance name in MyEnv (mysqld1 & mysqld2 for mysql1 & mysql2 respectively in this ex.)

Now, we can use the following window to modify System type,Configuration file path, start , stop and status commands to match each instance configurations:



Now MySQL Workbench should be able to start and stop the configured MySQL instances.


Notes:

  • MyEnv doesn't allow any user to start a mysql instance except mysql user (even if it's the root user), so that mysql OS user should be used to execute those commands and you might need to assign it SUDO permissions.
  • Start/Stop Server button in MySQL Workbench depends on the output of the status command used ("cat /opt/mysql1/data/*.pid 2>/dev/null" for checking mysql1 instance status), and the button label will be changed to execute the appropriate command accordingly (for ex. if the instance is not running, the label should be changed to "Start Server" and the start command will be executed if the button clicked and vise verse), so you should make sure that this command is returning the expected results, otherwise, Workbench wont be able to manage the instance.
  • It doesn't matter how MySQL was installed (RPM, tarballs or from source), it's the same concept, you just need to provide MyEnv start and stop commands along with the status command and then MySQL Workbench will work.
  • If MyEnv is not installed, you can also configure MySQL Workbench to start/stop multiple instances by providing normal start|stop instance commands in the "Manage Server Connections" window but those need to be prepared first.

Virident Flash Speicher und MySQL Datenbanken - 100 x mehr I/O Durchsatz?

FromDual.de - Sat, 2013-11-23 15:57

Sehr geehrte Damen und Herren,

Am Donnerstag, 5. Dezember kommt Virident nach Zürich um ihre Enterprise Flash Speicher für den Datenbank-Einsatz vorzustellen.

Ryan Taylor und Mark Lucas werden diese fortschrittliche Technologie vorstellen und aufzeigen, wie führende Unternehmen wie LinkedIn Virident Produkte wirksam einsetzten um ihre Datenbankanwendungen zu beschleunigen. Danach zeigt Oli Sennhauser (CTO von FromDual) auf, wo und wie Flash Speicher mit MySQL (und anderen Datenbanken) am besten zum Einsatz gebracht wird und wie er die Datenbanken beschleunigt.

Nehmen sie teil an der Frage- und Antwort-Runde sowie dem Erfahrungsaustausch über Flash Speicher im allgemeinen und dem Datenbankbetrieb im speziellen.

Die Runde ist kostenlos und erfolgt vorwiegend in englischer Sprache. Virident wird die Produkte mitbringen, damit sich die Teilnehmer ein Bild davon machen können. Sollten sie diese fortschrittliche Technologie selber bei sich testen wollen, unterstützt Virident sie dabei kostenlos und ohne weitere Verpflichtungen. Die perfekte Art selber praktische und unabhängige Benchmarks mit ihren Daten und Anwendungen durchzuführen.

Der Anlass findet am Donnerstag, 5. Dezember von 15:00 bis 17:30 im Crowne Plaza Hotel an der Badenerstrasse 420 in 8048 Zürich statt.

Bitte kündigen sie uns ihre Teilnahme vorher per Mail an, damit wir ihnen einen Platz reservieren und die benötigten Ressourcen zur Verfügung stellen können.

Mit freundlichen Grüssen,
Ihr FromDual Team

Virident Flash Speicher und MySQL Datenbanken - 100 x mehr I/O Durchsatz?

FromDual.de - Sat, 2013-11-23 15:57

Sehr geehrte Damen und Herren,

Am Donnerstag, 5. Dezember kommt Virident nach Zürich um ihre Enterprise Flash Speicher für den Datenbank-Einsatz vorzustellen.

Ryan Taylor und Mark Lucas werden diese fortschrittliche Technologie vorstellen und aufzeigen, wie führende Unternehmen wie LinkedIn Virident Produkte wirksam einsetzten um ihre Datenbankanwendungen zu beschleunigen. Danach zeigt Oli Sennhauser (CTO von FromDual) auf, wo und wie Flash Speicher mit MySQL (und anderen Datenbanken) am besten zum Einsatz gebracht wird und wie er die Datenbanken beschleunigt.

Nehmen sie teil an der Frage- und Antwort-Runde sowie dem Erfahrungsaustausch über Flash Speicher im allgemeinen und dem Datenbankbetrieb im speziellen.

Die Runde ist kostenlos und erfolgt vorwiegend in englischer Sprache. Virident wird die Produkte mitbringen, damit sich die Teilnehmer ein Bild davon machen können. Sollten sie diese fortschrittliche Technologie selber bei sich testen wollen, unterstützt Virident sie dabei kostenlos und ohne weitere Verpflichtungen. Die perfekte Art selber praktische und unabhängige Benchmarks mit ihren Daten und Anwendungen durchzuführen.

Der Anlass findet am Donnerstag, 5. Dezember von 15:00 bis 17:30 im Crowne Plaza Hotel an der Badenerstrasse 420 in 8048 Zürich statt.

Bitte kündigen sie uns ihre Teilnahme vorher per Mail an, damit wir ihnen einen Platz reservieren und die benötigten Ressourcen zur Verfügung stellen können.

Mit freundlichen Grüssen,
Ihr FromDual Team

Virident Flash Speicher und MySQL Datenbanken - 100 x mehr I/O Durchsatz?

FromDual.de - Sat, 2013-11-23 15:57

Sehr geehrte Damen und Herren,

Am Donnerstag, 5. Dezember kommt Virident nach Zürich um ihre Enterprise Flash Speicher für den Datenbank-Einsatz vorzustellen.

Ryan Taylor und Mark Lucas werden diese fortschrittliche Technologie vorstellen und aufzeigen, wie führende Unternehmen wie LinkedIn Virident Produkte wirksam einsetzten um ihre Datenbankanwendungen zu beschleunigen. Danach zeigt Oli Sennhauser (CTO von FromDual) auf, wo und wie Flash Speicher mit MySQL (und anderen Datenbanken) am besten zum Einsatz gebracht wird und wie er die Datenbanken beschleunigt.

Nehmen sie teil an der Frage- und Antwort-Runde sowie dem Erfahrungsaustausch über Flash Speicher im allgemeinen und dem Datenbankbetrieb im speziellen.

Die Runde ist kostenlos und erfolgt vorwiegend in englischer Sprache. Virident wird die Produkte mitbringen, damit sich die Teilnehmer ein Bild davon machen können. Sollten sie diese fortschrittliche Technologie selber bei sich testen wollen, unterstützt Virident sie dabei kostenlos und ohne weitere Verpflichtungen. Die perfekte Art selber praktische und unabhängige Benchmarks mit ihren Daten und Anwendungen durchzuführen.

Der Anlass findet am Donnerstag, 5. Dezember von 15:00 bis 17:30 im Crowne Plaza Hotel an der Badenerstrasse 420 in 8048 Zürich statt.

Bitte kündigen sie uns ihre Teilnahme vorher per Mail an, damit wir ihnen einen Platz reservieren und die benötigten Ressourcen zur Verfügung stellen können.

Mit freundlichen Grüssen,
Ihr FromDual Team

Galera Cluster 3.1 GA is out!

Shinguz - Wed, 2013-11-13 11:14

Great News: Galera Cluster v3.1 GA for MySQL 5.6 was released at Percona Live London (PLUK) 2013. The information is still a bit hidden...

You can find it here:


Or directly on our download page.

Careful: Online-Upgrade from 5.5 to 5.6 will not work yet. We have to find a work-around...

Xtrabackup in a nutshell

Abdel-Mawla Gharieb - Wed, 2013-11-06 14:23
Table of Content
Introduction

No one can deny that one of the most important and daily tasks for DBAs is performing backup and restore operations, we're not required to perform backup and restore operations only when we want to add new replication slave, when we want to implement disaster recovery procedures or when we want to prepare testing or staging server for the running production system, but even if we're going to make any changes to the database schema in order to enhance the database performance, it's recommended to have fresh backup copy before making any live changes, so if backup and restore operations cannot be handled smoothly, we're going to face many troubles in our daily work. If we're going to talk about backup and restore operations, Xtrabackup tool will be strongly appeared.

Xtrabackup tool is a free open source tool developed by Percona to perform physical backup and restore operations which is much faster than performing logical backup and restore using the MySQL utilities (mysqldump and mysql), and many other advantages.

Xtrabackup tool has many options and features which are very useful, but in this article, I'll go through only on how to use this tool to perform simple full, incremental and partial backups and restores, advantages and disadvantages of each method and some important tips.

For more information about Xtrabackup tool, you can browse the manual document from here.

Prerequisites Full Backup

If you want to have a full backup from your entire database system with the shortest/fastest backup and restore time, this method will be very useful for you. As compared to the full logical database backup using mysqldump and mysql utilities (very long time to backup and more than the doubled time to restore), taking a full physical backup using Xtrabackup tool will make your life much easier.

Below is the needed steps to make a full physical database backup using XtraBackup tool:

Create Backup

A simple Xtrabackup command to backup the full databases should be something like:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp /backup/dir/path/full-backup . . . innobackupex: completed OK!

A timestamped folder (for ex. "2013-11-06_00-00-00") would be created to contain all backup files if we didn't use the option "--no-timestamp" in the above command (I didn't use the timestamped folders here to just simplify the names for the reader, otherwise, it's very useful when using automated backup scripts).

Xtrabackup tool now created the backup files under the folder "full-backup" plus some extra files like "xtrabackup-checkpoints" file which contains some information (useful in the incremental backups) like:

  • backup_type = full-backuped : which indicates the backup type "full backup".
  • from_lsn = 0 : which indicates the log sequence number where the backup process started from (0 means from the beginning).
  • to_lsn = 3768762 : which indicates the log sequence number where the backup process ended at.

Another important file is "xtrabackup_binlog_info" which is very useful in replication setups:

[root@ ~]# cat xtrabackup_binlog_info mysql-bin.000027 191

WHERE:

  • mysql-bin.000027: is the binary log file name of the master when the backup created.
  • 191: is the binary log position of the backup.
Prepare Full Backup

The backed up files are not ready at the moment to be restored, we must prepare the backup files first as follows:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log /backup/dir/path/full-backup . . . innobackupex: completed OK!

Now, the full backup is ready to be restored ...

Restore Full Backup

To get the full backup restored, the MySQL instance should be stopped first and then one of the following two procedures should be done:

  • Using the copy back option: [root@ ~]# innobackupex --user=db_user –-password='db_password' --copy-back /backup/dir/path/full-backup . . . innobackupex: completed OK!

    Xtrabackup tool - in this method - will copy all files under the backup folder (full-backup) to the MySQL datadir which must be indicated in the my.cnf file, otherwise, it wouldn't know where the datadir should be placed.

  • Using the operating system copy or move commands:

    If you don't want to keep the backup files on your local system (you have another copy in an external server), the move command will be very fast to get your backup restored:

    [root@ ~]# mv /backup/dir/path/full-backup /var/lib/mysql
  • As the user who moves/copies the files into MySQL datadir is not "mysql" user, you should make sure that mysql user has the right permissions on its datadir and also the path "/var/lib/mysql" should be replaced with the MySQL datadir if it's set to a different path.

[root@ ~]# chown -R mysql:mysql /var/lib/mysql

After moving/copying the backup files into MySQL datadir, you are free to start the MySQL instance again.

Prepare slave from full backup

Preparing a slave using Xtrabackup is pretty easy and a straight forward process:

  • Restore the full backup as explained above.
  • Check the binary logs information of the backup:
    [root@ ~]# cat xtrabackup_binlog_info mysql-bin.000027 191
  • Execute the CHANGE MASTER TO command using the above info and start the slave:
    SQL> CHANGE MASTER TO -> MASTER_HOST='master_ip', -> MASTER_PORT=master_port, -> MASTER_USER='slave_user_name', -> MASTER_PASSWORD='slave_password', -> MASTER_LOG_FILE='mysql-bin.000027', ## taken from xtrabackup_binlog_info -> MASTER_LOG_POS=191; ## taken from xtrabackup_binlog_info SQL> START SLAVE;

For more information on how to set up MySQL Replication, check out this nice manual link.

Prepare GTID slave from full backup

GTID is supported in Xtrabackup starting from version 2.1.0. To restore a GTID slave server, the GTID_MODE should be enabled on the master server before creating its backup, otherwise, no GTID values will be included in the "xtrabackup_binlog_info" file under the backup directory. However, the following steps should be done:

  • Restore the full backup normally as explained above.
  • Check the GTID value of the backup:
    [root@ ~]# cat xtrabackup_binlog_info mysql-bin.000027 191 b9b4712a-df64-11e3-b391-60672090eb04:1-12
  • Set the variable GTID_PURGED with the GTID value of the backup:
    SQL> SET GLOBAL GTID_PURGED="b9b4712a-df64-11e3-b391-60672090eb04:1-12";
  • Execute the auto position CHANGE MASTER TO command and start the slave:
    SQL> CHANGE MASTER TO -> MASTER_HOST='master_ip', -> MASTER_PORT=master_port, -> MASTER_USER='slave_user_name', -> MASTER_PASSWORD='slave_password', -> MASTER_AUTO_POSITION = 1; SQL> START SLAVE;

For more information on how to set up Transaction-based Replication in MySQL, check out this link.

Advantages / Disadvantages
  • Advantages:
    • Fast, simple and easy way to get your full database backed up and restored.
    • All Xtrabackup tool features (like streaming: move the backed up files directly to a remote server) are supported in the full backup method.
    • Simple way to introduce a new slave to the master.
  • Disadvantages:
    • We have to replace the entire MySQL datadir with the new one (In other words, the datadir folder should be empty/removed before the restore process).
    • We can't extract one single database or single table from the whole backup (Unless it's MyISAM table), which means that you have to take it all or leave it all.
Important Hints
  • The message innobackupex: completed OK! should be printed at the end of every xtrabackup command , otherwise, it would be failed to make a successful command (backup, prepare or restore).
  • The ib_logfile files size should be the same in both source and destination servers, if not, you have to either remove them from the backup folder (which will be restored) before starting the MySQL instance and MySQL will create new ones for you OR change those files size in the destination server's configuration file to match the same size in the backup before starting the MySQL instance
  • The MySQL user used in the Xtrabackup tool, should have at least the following privileges (RELOAD, LOCK TABLES and REPLICATION CLIENT).
  • To prepare a new slave from another slave, just add the two options (“--slave-info" & --safe-slave-backup”) to the backup command and use the information in the file "xtrabackup_slave_info" under the backup folder to issue the "CHANGE MASTER TO" command in the new slave after finishing the restore.
  • To Accelerate the preparation process of your backup, just add the option "--use-memory" in the prepare command in order to allocate more used memory (Xtrabackup will use the specified memory as an internal innodb_buffer_pool_size for the prepare process), for ex: [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --use-memory=512M /backup/dir/path/full-backup . . . innobackupex: completed OK!
  • The preparation process consists of two steps, replaying the committed transactions and rolling back the uncommitted transactions, using the --apply-log option only in the preparation command will do both steps for you.
  • The backup folder "/backup/dir/path/full-backup" SHOULD NOT be created before executing the backup command, because Xtrabackup will create that folder for you, and it will fail to continue processing if that folder was already exist.
Incremental Backup

When you have a very large database system, you will need large enough storage to store your database backups in, and if you want to perform a daily backup then your mission will be more difficult. In such cases, the incremental database backup method will be very useful. It allows you to have only the database changes (delta) - after the physical full backup – with the minimum storage space required in a fast way, and hence, you can perform the daily backup operations to your database system without the need to having large storage available.

The following steps describe a simple way to perform your physical incremental database backup using XtraBackup tool:

Create Incremental Backup

To perform an incremental backup, we should first perform a full backup - the same like we did in the previous section - to be the base backup of the upcoming incremental backups.

Creating the full backup (Base Backup):

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp /backup/dir/path/full-backup . . . innobackupex: completed OK!

The "xtrabackup-checkpoints" file contents will be something like:

  • backup_type = full-backuped : which indicates the backup type "full backup".
  • from_lsn = 0 : which indicates the log sequence number where the backup process started from (0 means from the beginning).
  • to_lsn = 3768762 : which indicates the log sequence number where the backup process ended at.

Creating the first incremental backup:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp --incremental /backup/dir/path/inc1 --incremental-basedir=/backup/dir/path/full-backup . . . innobackupex: completed OK!

We informed the Xtrabackup tool to perform an incremental backup by adding the command "--incremental", and by specifying the full-backup path as the basedir, we informed it from which backup it should start tracking the database changes.

The "xtrabackup-checkpoints" file contents will be something like:

  • backup_type = incremental : which indicates the backup type "incremental backup".
  • from_lsn = 3768762 : which indicates the log sequence number where the backup process started from (the same LSN as the previous full backup ended at).
  • to_lsn = 4908762 : which indicates the log sequence number where the backup process ended at.

Creating the second incremental backup:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp --incremental /backup/dir/path/inc2 --incremental-basedir=/backup/dir/path/inc1 . . . innobackupex: completed OK!

We informed the Xtrabackup tool to perform an incremental backup by adding the command "--incremental", and by specifying the 1st incremental backup path as the basedir, we informed it to start tracking the database changes since the last incremental (not the full backup).

The "xtrabackup-checkpoints" file contents will be something like:

  • backup_type = incremental : which indicates the backup type "incremental backup".
  • from_lsn = 4908762 : which indicates the log sequence number where the backup process started from (the same LSN as the 1st incremental backup ended at).
  • to_lsn = 6508762 : which indicates the log sequence number where the backup process ended at.

Note: We can create as many incremental backups as we want by using the same procedure above.

Prepare Incremental Backup

As mentioned earlier in the article, the preparation process consists of two steps (replaying the committed transactions and rolling back the uncommitted transactions) and using the --apply-log option only will do both of them (like we did in the full backup) but in the incremental backups, we MUST do them separately as follows:

  • Replay the committed transactions on the base backup (by adding the option "--redo-only"): [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --redo-only /backup/dir/path/full-backup . . . innobackupex: completed OK!
  • Replay the committed transactions on the 1st incremental backup: [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --redo-only /backup/dir/path/full-backup --incremental-dir=/backup/dir/path/inc1 . . . innobackupex: completed OK!

    Note: we specified the full backup folder here, because replaying the committed transactions steps, appends all changes from the incremental backup to the full backup.

  • Replay the committed transactions on the 2nd incremental backup: [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --redo-only /backup/dir/path/full-backup --incremental-dir=/backup/dir/path/inc2 . . . innobackupex: completed OK!

    Note: here, we didn't use the 1st incremental backup folder, because all changes in the 1st incremental was already appended to the full backup in the previous step.

  • Finally, roll back all uncommitted transactions: [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log /backup/dir/path/full-backup . . . innobackupex: completed OK!

    Note: as the full backup folder contains all data now (full + 1st & 2nd incremental), there's no need to do this step on the incremental backup folders.

Now, the incremental backup is ready to be restored ...

Restore Incremental Backup

The full backup folder will be the only folder to be restored (there's no need to the incremental backup folders) as it contains all data after appending the changes from all incremental backup. We can restore it the same way we did in the full backup restore.

Advantages / Disadvantages
  • Advantages:
    • Less storage resources needed.
    • Faster than the full backup.
  • Disadvantages:

    In addition to the disadvantages of the full backup, there are other ones:

    • Complicate and hard process to implement as compared to the full backup.
    • The incremental backup strategy, based on Log Sequence Number which affects only XtraDB and InnoDB storage engines while the others (like MyISAM) will be backed up completely (changed + unchanged data) in each incremental backup process.
    • If we have many incremental backups, appending all of them might consume time and might be confusing as well.
    • If one of the incremental backups become corrupted or not available for any reason, we will not be able to add all incremental backups after that to the full backup.
Important Hints
  • The backup preparation sequence steps above, MUST be followed using the same order.
  • If the "--redo-only" option was not be used in any of the preparation steps (except the final step), all up coming incremental backups will be useless as we won't be able to add them to the base backup anymore.
  • Replaying the committed transactions steps bring all incremental data and append it to the full backup, so that, the rolling back of the uncommitted transactions step should be execute only on the full backup (as it contains already the whole data).).
  • In the incremental backups, Xtrabackup generates two files for every table ".delta" & ".meta"(for ex. test.ibd.delta & test.ibd.meta), the delta file size reflects the changes which was applied on that table since the last incremental backup.
  • The preparation time of the individual incremental backup will depend on how much data changed there since the last incremental.
  • The preparation time for the full backup - in most cases - is really small as compared to the incremental ones because full backups apply the redo logs only while the incremental backups apply the deltas plus the redo logs. So if the delta files are big, the preparation process will take longer.
  • Full backups is recommended against Incremental backups if there are many changes applied on the DB, while the incremental backups are recommended when there are few changes applied on the DB.
Differential backup

We can use the incremental backup strategy in order to perform differential backups, but we should consider the following points:

  • We always specify the full backup folder as the base backup (in the incremental we specify the previous incremental folder as a base backup)
  • All incremental backups between differential and full backups MUST BE ignored when preparing the backup files because the differential backup contains already all changes since the last full backup.
  • In the backup preparation process, we should consider the last differential backup as the first incremental backup and all incremental backups after that could be applied normally.

Note: Having differnetial backups in the middle of incremental backups will be useful for many reasons, such as:

  • Differential backups reduce the backup preparation steps/time needed because differential backp will replace all its previous incremental backups.
  • Differential backups reduce the chances of loosing the incremental backups if we have corrupted incremental backup in the middle, because in this case, differential backup will act as a backup of the previous incremental backups.
Partial Backup

Unlike MyISAM, having physical database backup for a single database or table is not possible if the table engine type is InnoDB. But by using the partial database backup method in the XtraBackup tool, it will be possible to have physical InnoDB tables backup the same like MyISAM ones (but with some restrictions).

The following steps describing how to perform partial database backup using XtraBackup tool:

Create Partial Backup

A simple Xtrabackup command to backup some databases (or tables) should be something like:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp --databases=”db1 db2 db3.tbl1” /backup/dir/path/partial-backup . . . innobackupex: completed OK!
Prepare Partial Backup

The same like the other backup methods, the backed up files are not ready until we get them prepared by adding the "--export" option as follows:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --export /backup/dir/path/partial-backup . . . innobackupex: completed OK!

Some errors regarding those not included InnoDB tables from the backup may be appeared, but that's fine. Also there will be a notification of creating the ".exp" file for each table which will be used in the import (restore) process.


Now, the partial backup is ready to be restored ...

Restore Partial Backup

The restore process of the partial backup is quite different than the full and incremental backups.

To restore a partial backup, the following steps should be made:

  • Unlike the other methods (Full and Incremental backups), MySQL instance on the destination server shouldn't be stopped because we will have to execute some SQL commands.
  • On the destination server, we should create new tables (as many as we have in the partial backup or as we will restore) with the same structure like the one in the partial backup and then discard its table space: mysql> CREATE TABLE db.tbl1 (...)ENGINE=INNODB; mysql> ALTER TABLE db.tbl1 DISCARD TABLESPACE;
  • Copy “.ibd” and “.exp” files for each table into the corresponding DB directory then assign the right permissions to mysql user: [root@ ~]# cp /backup/dir/path/partial-backup/db/tbl1.ibd /var/lib/mysql/db [root@ ~]# cp /backup/dir/path/partial-backup/db/tbl1.exp /var/lib/mysql/db [root@ ~]# chown -R mysql:mysql /var/lib/mysql/db
  • Now we should tell MySQL to use the new table spaces: mysql> ALTER TABLE db.tbl1 IMPORT TABLESPACE;
Advantages / Disadvantages
  • Advantages:
    • Although it's a complicated process, but it allows us to backup and restore individual InnoDB tables the same like MyISAM.
    • Useful when having huge InnoDB tables and we want to backup/restore them only.
  • Disadvantages:
    • The streaming feature is not available in the partial backup.
    • Restoring/importing individual tables or databases from a partial backup is not applicable unless the destination server is Percona Server.
    • In addition to restoring the files(copy back), three SQL statements should be executed for each table (table creation + two ALTER statements) in order to get them ready for use, which means that we might do a very boring job (or we have to create a special script) to get the partial backup restored if it contains many tables.
Important Hints
  • Although we didn't remove the MySQL datadir before the restore process (like full and incremental backups) as well as having the MySQL instance running, but we can restore the partial backup using the same way (remove the datadir contents and copy/move the backup files to the datadir), but we should take into our consideration that we'll have only the backed up databases/tables and all other databases/tables (which are not included in the partial backup) will be missed.
  • “innodb_file_per_table” server option must be enabled (in both source and destination servers).
  • "innodb_expand_import" option must be enabled in the destination server which is available only in Percona server (and that explain why we can restore partial backup on Percona server only).
  • Beside the "--databases" option, other two alternative options to perform the same needs could be used but we must provide each table with the fully qualified naming format:
    • --include='db.tbl'.
    • --tables-file=/path/to/file.txt ==> in that file, we can add multiple tables one per line in the fully qualified naming format.

Now, you can use the Xtrabackup tool to perform full, incremental and partial database backups, you can decide which method(s) of them are suitable for you according to the advantages and disadvantages of each one, and by considering the important hints for each method you can perform your backup efficiently.

I hope you found this article useful for you and to be familiar with such wonderful tool.

Taxonomy upgrade extras: GTIDxtrabackupBackup

Xtrabackup in a nutshell

Abdel-Mawla Gharieb - Wed, 2013-11-06 14:23
Table of Content
Introduction

No one can deny that one of the most important and daily tasks for DBAs is performing backup and restore operations, we're not required to perform backup and restore operations only when we want to add new replication slave, when we want to implement disaster recovery procedures or when we want to prepare testing or staging server for the running production system, but even if we're going to make any changes to the database schema in order to enhance the database performance, it's recommended to have fresh backup copy before making any live changes, so if backup and restore operations cannot be handled smoothly, we're going to face many troubles in our daily work. If we're going to talk about backup and restore operations, Xtrabackup tool will be strongly appeared.

Xtrabackup tool is a free open source tool developed by Percona to perform physical backup and restore operations which is much faster than performing logical backup and restore using the MySQL utilities (mysqldump and mysql), and many other advantages.

Xtrabackup tool has many options and features which are very useful, but in this article, I'll go through only on how to use this tool to perform simple full, incremental and partial backups and restores, advantages and disadvantages of each method and some important tips.

For more information about Xtrabackup tool, you can browse the manual document from here.

Prerequisites Full Backup

If you want to have a full backup from your entire database system with the shortest/fastest backup and restore time, this method will be very useful for you. As compared to the full logical database backup using mysqldump and mysql utilities (very long time to backup and more than the doubled time to restore), taking a full physical backup using Xtrabackup tool will make your life much easier.

Below is the needed steps to make a full physical database backup using XtraBackup tool:

Create Backup

A simple Xtrabackup command to backup the full databases should be something like:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp /backup/dir/path/full-backup . . . innobackupex: completed OK!

A timestamped folder (for ex. "2013-11-06_00-00-00") would be created to contain all backup files if we didn't use the option "--no-timestamp" in the above command (I didn't use the timestamped folders here to just simplify the names for the reader, otherwise, it's very useful when using automated backup scripts).

Xtrabackup tool now created the backup files under the folder "full-backup" plus some extra files like "xtrabackup-checkpoints" file which contains some information (useful in the incremental backups) like:

  • backup_type = full-backuped : which indicates the backup type "full backup".
  • from_lsn = 0 : which indicates the log sequence number where the backup process started from (0 means from the beginning).
  • to_lsn = 3768762 : which indicates the log sequence number where the backup process ended at.

Another important file is "xtrabackup_binlog_info" which is very useful in replication setups:

[root@ ~]# cat xtrabackup_binlog_info mysql-bin.000027 191

WHERE:

  • mysql-bin.000027: is the binary log file name of the master when the backup created.
  • 191: is the binary log position of the backup.
Prepare Full Backup

The backed up files are not ready at the moment to be restored, we must prepare the backup files first as follows:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log /backup/dir/path/full-backup . . . innobackupex: completed OK!

Now, the full backup is ready to be restored ...

Restore Full Backup

To get the full backup restored, the MySQL instance should be stopped first and then one of the following two procedures should be done:

  • Using the copy back option: [root@ ~]# innobackupex --user=db_user –-password='db_password' --copy-back /backup/dir/path/full-backup . . . innobackupex: completed OK!

    Xtrabackup tool - in this method - will copy all files under the backup folder (full-backup) to the MySQL datadir which must be indicated in the my.cnf file, otherwise, it wouldn't know where the datadir should be placed.

  • Using the operating system copy or move commands:

    If you don't want to keep the backup files on your local system (you have another copy in an external server), the move command will be very fast to get your backup restored:

    [root@ ~]# mv /backup/dir/path/full-backup /var/lib/mysql
  • As the user who moves/copies the files into MySQL datadir is not "mysql" user, you should make sure that mysql user has the right permissions on its datadir and also the path "/var/lib/mysql" should be replaced with the MySQL datadir if it's set to a different path.

[root@ ~]# chown -R mysql:mysql /var/lib/mysql

After moving/copying the backup files into MySQL datadir, you are free to start the MySQL instance again.

Prepare slave from full backup

Preparing a slave using Xtrabackup is pretty easy and a straight forward process:

  • Restore the full backup as explained above.
  • Check the binary logs information of the backup:
    [root@ ~]# cat xtrabackup_binlog_info mysql-bin.000027 191
  • Execute the CHANGE MASTER TO command using the above info and start the slave:
    SQL> CHANGE MASTER TO -> MASTER_HOST='master_ip', -> MASTER_PORT=master_port, -> MASTER_USER='slave_user_name', -> MASTER_PASSWORD='slave_password', -> MASTER_LOG_FILE='mysql-bin.000027', ## taken from xtrabackup_binlog_info -> MASTER_LOG_POS=191; ## taken from xtrabackup_binlog_info SQL> START SLAVE;

For more information on how to set up MySQL Replication, check out this nice manual link.

Prepare GTID slave from full backup

GTID is supported in Xtrabackup starting from version 2.1.0. To restore a GTID slave server, the GTID_MODE should be enabled on the master server before creating its backup, otherwise, no GTID values will be included in the "xtrabackup_binlog_info" file under the backup directory. However, the following steps should be done:

  • Restore the full backup normally as explained above.
  • Check the GTID value of the backup:
    [root@ ~]# cat xtrabackup_binlog_info mysql-bin.000027 191 b9b4712a-df64-11e3-b391-60672090eb04:1-12
  • Set the variable GTID_PURGED with the GTID value of the backup:
    SQL> SET GLOBAL GTID_PURGED="b9b4712a-df64-11e3-b391-60672090eb04:1-12";
  • Execute the auto position CHANGE MASTER TO command and start the slave:
    SQL> CHANGE MASTER TO -> MASTER_HOST='master_ip', -> MASTER_PORT=master_port, -> MASTER_USER='slave_user_name', -> MASTER_PASSWORD='slave_password', -> MASTER_AUTO_POSITION = 1; SQL> START SLAVE;

For more information on how to set up Transaction-based Replication in MySQL, check out this link.

Advantages / Disadvantages
  • Advantages:
    • Fast, simple and easy way to get your full database backed up and restored.
    • All Xtrabackup tool features (like streaming: move the backed up files directly to a remote server) are supported in the full backup method.
    • Simple way to introduce a new slave to the master.
  • Disadvantages:
    • We have to replace the entire MySQL datadir with the new one (In other words, the datadir folder should be empty/removed before the restore process).
    • We can't extract one single database or single table from the whole backup (Unless it's MyISAM table), which means that you have to take it all or leave it all.
Important Hints
  • The message innobackupex: completed OK! should be printed at the end of every xtrabackup command , otherwise, it would be failed to make a successful command (backup, prepare or restore).
  • The ib_logfile files size should be the same in both source and destination servers, if not, you have to either remove them from the backup folder (which will be restored) before starting the MySQL instance and MySQL will create new ones for you OR change those files size in the destination server's configuration file to match the same size in the backup before starting the MySQL instance
  • The MySQL user used in the Xtrabackup tool, should have at least the following privileges (RELOAD, LOCK TABLES and REPLICATION CLIENT).
  • To prepare a new slave from another slave, just add the two options (“--slave-info" & --safe-slave-backup”) to the backup command and use the information in the file "xtrabackup_slave_info" under the backup folder to issue the "CHANGE MASTER TO" command in the new slave after finishing the restore.
  • To Accelerate the preparation process of your backup, just add the option "--use-memory" in the prepare command in order to allocate more used memory (Xtrabackup will use the specified memory as an internal innodb_buffer_pool_size for the prepare process), for ex: [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --use-memory=512M /backup/dir/path/full-backup . . . innobackupex: completed OK!
  • The preparation process consists of two steps, replaying the committed transactions and rolling back the uncommitted transactions, using the --apply-log option only in the preparation command will do both steps for you.
  • The backup folder "/backup/dir/path/full-backup" SHOULD NOT be created before executing the backup command, because Xtrabackup will create that folder for you, and it will fail to continue processing if that folder was already exist.
Incremental Backup

When you have a very large database system, you will need large enough storage to store your database backups in, and if you want to perform a daily backup then your mission will be more difficult. In such cases, the incremental database backup method will be very useful. It allows you to have only the database changes (delta) - after the physical full backup – with the minimum storage space required in a fast way, and hence, you can perform the daily backup operations to your database system without the need to having large storage available.

The following steps describe a simple way to perform your physical incremental database backup using XtraBackup tool:

Create Incremental Backup

To perform an incremental backup, we should first perform a full backup - the same like we did in the previous section - to be the base backup of the upcoming incremental backups.

Creating the full backup (Base Backup):

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp /backup/dir/path/full-backup . . . innobackupex: completed OK!

The "xtrabackup-checkpoints" file contents will be something like:

  • backup_type = full-backuped : which indicates the backup type "full backup".
  • from_lsn = 0 : which indicates the log sequence number where the backup process started from (0 means from the beginning).
  • to_lsn = 3768762 : which indicates the log sequence number where the backup process ended at.

Creating the first incremental backup:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp --incremental /backup/dir/path/inc1 --incremental-basedir=/backup/dir/path/full-backup . . . innobackupex: completed OK!

We informed the Xtrabackup tool to perform an incremental backup by adding the command "--incremental", and by specifying the full-backup path as the basedir, we informed it from which backup it should start tracking the database changes.

The "xtrabackup-checkpoints" file contents will be something like:

  • backup_type = incremental : which indicates the backup type "incremental backup".
  • from_lsn = 3768762 : which indicates the log sequence number where the backup process started from (the same LSN as the previous full backup ended at).
  • to_lsn = 4908762 : which indicates the log sequence number where the backup process ended at.

Creating the second incremental backup:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp --incremental /backup/dir/path/inc2 --incremental-basedir=/backup/dir/path/inc1 . . . innobackupex: completed OK!

We informed the Xtrabackup tool to perform an incremental backup by adding the command "--incremental", and by specifying the 1st incremental backup path as the basedir, we informed it to start tracking the database changes since the last incremental (not the full backup).

The "xtrabackup-checkpoints" file contents will be something like:

  • backup_type = incremental : which indicates the backup type "incremental backup".
  • from_lsn = 4908762 : which indicates the log sequence number where the backup process started from (the same LSN as the 1st incremental backup ended at).
  • to_lsn = 6508762 : which indicates the log sequence number where the backup process ended at.

Note: We can create as many incremental backups as we want by using the same procedure above.

Prepare Incremental Backup

As mentioned earlier in the article, the preparation process consists of two steps (replaying the committed transactions and rolling back the uncommitted transactions) and using the --apply-log option only will do both of them (like we did in the full backup) but in the incremental backups, we MUST do them separately as follows:

  • Replay the committed transactions on the base backup (by adding the option "--redo-only"): [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --redo-only /backup/dir/path/full-backup . . . innobackupex: completed OK!
  • Replay the committed transactions on the 1st incremental backup: [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --redo-only /backup/dir/path/full-backup --incremental-dir=/backup/dir/path/inc1 . . . innobackupex: completed OK!

    Note: we specified the full backup folder here, because replaying the committed transactions steps, appends all changes from the incremental backup to the full backup.

  • Replay the committed transactions on the 2nd incremental backup: [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --redo-only /backup/dir/path/full-backup --incremental-dir=/backup/dir/path/inc2 . . . innobackupex: completed OK!

    Note: here, we didn't use the 1st incremental backup folder, because all changes in the 1st incremental was already appended to the full backup in the previous step.

  • Finally, roll back all uncommitted transactions: [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log /backup/dir/path/full-backup . . . innobackupex: completed OK!

    Note: as the full backup folder contains all data now (full + 1st & 2nd incremental), there's no need to do this step on the incremental backup folders.

Now, the incremental backup is ready to be restored ...

Restore Incremental Backup

The full backup folder will be the only folder to be restored (there's no need to the incremental backup folders) as it contains all data after appending the changes from all incremental backup. We can restore it the same way we did in the full backup restore.

Advantages / Disadvantages
  • Advantages:
    • Less storage resources needed.
    • Faster than the full backup.
  • Disadvantages:

    In addition to the disadvantages of the full backup, there are other ones:

    • Complicate and hard process to implement as compared to the full backup.
    • The incremental backup strategy, based on Log Sequence Number which affects only XtraDB and InnoDB storage engines while the others (like MyISAM) will be backed up completely (changed + unchanged data) in each incremental backup process.
    • If we have many incremental backups, appending all of them might consume time and might be confusing as well.
    • If one of the incremental backups become corrupted or not available for any reason, we will not be able to add all incremental backups after that to the full backup.
Important Hints
  • The backup preparation sequence steps above, MUST be followed using the same order.
  • If the "--redo-only" option was not be used in any of the preparation steps (except the final step), all up coming incremental backups will be useless as we won't be able to add them to the base backup anymore.
  • Replaying the committed transactions steps bring all incremental data and append it to the full backup, so that, the rolling back of the uncommitted transactions step should be execute only on the full backup (as it contains already the whole data).).
  • In the incremental backups, Xtrabackup generates two files for every table ".delta" & ".meta"(for ex. test.ibd.delta & test.ibd.meta), the delta file size reflects the changes which was applied on that table since the last incremental backup.
  • The preparation time of the individual incremental backup will depend on how much data changed there since the last incremental.
  • The preparation time for the full backup - in most cases - is really small as compared to the incremental ones because full backups apply the redo logs only while the incremental backups apply the deltas plus the redo logs. So if the delta files are big, the preparation process will take longer.
  • Full backups is recommended against Incremental backups if there are many changes applied on the DB, while the incremental backups are recommended when there are few changes applied on the DB.
Differential backup

We can use the incremental backup strategy in order to perform differential backups, but we should consider the following points:

  • We always specify the full backup folder as the base backup (in the incremental we specify the previous incremental folder as a base backup)
  • All incremental backups between differential and full backups MUST BE ignored when preparing the backup files because the differential backup contains already all changes since the last full backup.
  • In the backup preparation process, we should consider the last differential backup as the first incremental backup and all incremental backups after that could be applied normally.

Note: Having differnetial backups in the middle of incremental backups will be useful for many reasons, such as:

  • Differential backups reduce the backup preparation steps/time needed because differential backp will replace all its previous incremental backups.
  • Differential backups reduce the chances of loosing the incremental backups if we have corrupted incremental backup in the middle, because in this case, differential backup will act as a backup of the previous incremental backups.
Partial Backup

Unlike MyISAM, having physical database backup for a single database or table is not possible if the table engine type is InnoDB. But by using the partial database backup method in the XtraBackup tool, it will be possible to have physical InnoDB tables backup the same like MyISAM ones (but with some restrictions).

The following steps describing how to perform partial database backup using XtraBackup tool:

Create Partial Backup

A simple Xtrabackup command to backup some databases (or tables) should be something like:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp --databases=”db1 db2 db3.tbl1” /backup/dir/path/partial-backup . . . innobackupex: completed OK!
Prepare Partial Backup

The same like the other backup methods, the backed up files are not ready until we get them prepared by adding the "--export" option as follows:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --export /backup/dir/path/partial-backup . . . innobackupex: completed OK!

Some errors regarding those not included InnoDB tables from the backup may be appeared, but that's fine. Also there will be a notification of creating the ".exp" file for each table which will be used in the import (restore) process.


Now, the partial backup is ready to be restored ...

Restore Partial Backup

The restore process of the partial backup is quite different than the full and incremental backups.

To restore a partial backup, the following steps should be made:

  • Unlike the other methods (Full and Incremental backups), MySQL instance on the destination server shouldn't be stopped because we will have to execute some SQL commands.
  • On the destination server, we should create new tables (as many as we have in the partial backup or as we will restore) with the same structure like the one in the partial backup and then discard its table space: mysql> CREATE TABLE db.tbl1 (...)ENGINE=INNODB; mysql> ALTER TABLE db.tbl1 DISCARD TABLESPACE;
  • Copy “.ibd” and “.exp” files for each table into the corresponding DB directory then assign the right permissions to mysql user: [root@ ~]# cp /backup/dir/path/partial-backup/db/tbl1.ibd /var/lib/mysql/db [root@ ~]# cp /backup/dir/path/partial-backup/db/tbl1.exp /var/lib/mysql/db [root@ ~]# chown -R mysql:mysql /var/lib/mysql/db
  • Now we should tell MySQL to use the new table spaces: mysql> ALTER TABLE db.tbl1 IMPORT TABLESPACE;
Advantages / Disadvantages
  • Advantages:
    • Although it's a complicated process, but it allows us to backup and restore individual InnoDB tables the same like MyISAM.
    • Useful when having huge InnoDB tables and we want to backup/restore them only.
  • Disadvantages:
    • The streaming feature is not available in the partial backup.
    • Restoring/importing individual tables or databases from a partial backup is not applicable unless the destination server is Percona Server.
    • In addition to restoring the files(copy back), three SQL statements should be executed for each table (table creation + two ALTER statements) in order to get them ready for use, which means that we might do a very boring job (or we have to create a special script) to get the partial backup restored if it contains many tables.
Important Hints
  • Although we didn't remove the MySQL datadir before the restore process (like full and incremental backups) as well as having the MySQL instance running, but we can restore the partial backup using the same way (remove the datadir contents and copy/move the backup files to the datadir), but we should take into our consideration that we'll have only the backed up databases/tables and all other databases/tables (which are not included in the partial backup) will be missed.
  • “innodb_file_per_table” server option must be enabled (in both source and destination servers).
  • "innodb_expand_import" option must be enabled in the destination server which is available only in Percona server (and that explain why we can restore partial backup on Percona server only).
  • Beside the "--databases" option, other two alternative options to perform the same needs could be used but we must provide each table with the fully qualified naming format:
    • --include='db.tbl'.
    • --tables-file=/path/to/file.txt ==> in that file, we can add multiple tables one per line in the fully qualified naming format.

Now, you can use the Xtrabackup tool to perform full, incremental and partial database backups, you can decide which method(s) of them are suitable for you according to the advantages and disadvantages of each one, and by considering the important hints for each method you can perform your backup efficiently.

I hope you found this article useful for you and to be familiar with such wonderful tool.

Taxonomy upgrade extras: GTIDxtrabackupBackup

Xtrabackup in a nutshell

Abdel-Mawla Gharieb - Wed, 2013-11-06 14:23
Table of Content
Introduction

No one can deny that one of the most important and daily tasks for DBAs is performing backup and restore operations, we're not required to perform backup and restore operations only when we want to add new replication slave, when we want to implement disaster recovery procedures or when we want to prepare testing or staging server for the running production system, but even if we're going to make any changes to the database schema in order to enhance the database performance, it's recommended to have fresh backup copy before making any live changes, so if backup and restore operations cannot be handled smoothly, we're going to face many troubles in our daily work. If we're going to talk about backup and restore operations, Xtrabackup tool will be strongly appeared.

Xtrabackup tool is a free open source tool developed by Percona to perform physical backup and restore operations which is much faster than performing logical backup and restore using the MySQL utilities (mysqldump and mysql), and many other advantages.

Xtrabackup tool has many options and features which are very useful, but in this article, I'll go through only on how to use this tool to perform simple full, incremental and partial backups and restores, advantages and disadvantages of each method and some important tips.

For more information about Xtrabackup tool, you can browse the manual document from here.

Prerequisites Full Backup

If you want to have a full backup from your entire database system with the shortest/fastest backup and restore time, this method will be very useful for you. As compared to the full logical database backup using mysqldump and mysql utilities (very long time to backup and more than the doubled time to restore), taking a full physical backup using Xtrabackup tool will make your life much easier.

Below is the needed steps to make a full physical database backup using XtraBackup tool:

Create Backup

A simple Xtrabackup command to backup the full databases should be something like:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp /backup/dir/path/full-backup . . . innobackupex: completed OK!

A timestamped folder (for ex. "2013-11-06_00-00-00") would be created to contain all backup files if we didn't use the option "--no-timestamp" in the above command (I didn't use the timestamped folders here to just simplify the names for the reader, otherwise, it's very useful when using automated backup scripts).

Xtrabackup tool now created the backup files under the folder "full-backup" plus some extra files like "xtrabackup-checkpoints" file which contains some information (useful in the incremental backups) like:

  • backup_type = full-backuped : which indicates the backup type "full backup".
  • from_lsn = 0 : which indicates the log sequence number where the backup process started from (0 means from the beginning).
  • to_lsn = 3768762 : which indicates the log sequence number where the backup process ended at.

Another important file is "xtrabackup_binlog_info" which is very useful in replication setups:

[root@ ~]# cat xtrabackup_binlog_info mysql-bin.000027 191

WHERE:

  • mysql-bin.000027: is the binary log file name of the master when the backup created.
  • 191: is the binary log position of the backup.
Prepare Full Backup

The backed up files are not ready at the moment to be restored, we must prepare the backup files first as follows:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log /backup/dir/path/full-backup . . . innobackupex: completed OK!

Now, the full backup is ready to be restored ...

Restore Full Backup

To get the full backup restored, the MySQL instance should be stopped first and then one of the following two procedures should be done:

  • Using the copy back option: [root@ ~]# innobackupex --user=db_user –-password='db_password' --copy-back /backup/dir/path/full-backup . . . innobackupex: completed OK!

    Xtrabackup tool - in this method - will copy all files under the backup folder (full-backup) to the MySQL datadir which must be indicated in the my.cnf file, otherwise, it wouldn't know where the datadir should be placed.

  • Using the operating system copy or move commands:

    If you don't want to keep the backup files on your local system (you have another copy in an external server), the move command will be very fast to get your backup restored:

    [root@ ~]# mv /backup/dir/path/full-backup /var/lib/mysql
  • As the user who moves/copies the files into MySQL datadir is not "mysql" user, you should make sure that mysql user has the right permissions on its datadir and also the path "/var/lib/mysql" should be replaced with the MySQL datadir if it's set to a different path.

[root@ ~]# chown -R mysql:mysql /var/lib/mysql

After moving/copying the backup files into MySQL datadir, you are free to start the MySQL instance again.

Prepare slave from full backup

Preparing a slave using Xtrabackup is pretty easy and a straight forward process:

  • Restore the full backup as explained above.
  • Check the binary logs information of the backup:
    [root@ ~]# cat xtrabackup_binlog_info mysql-bin.000027 191
  • Execute the CHANGE MASTER TO command using the above info and start the slave:
    SQL> CHANGE MASTER TO -> MASTER_HOST='master_ip', -> MASTER_PORT=master_port, -> MASTER_USER='slave_user_name', -> MASTER_PASSWORD='slave_password', -> MASTER_LOG_FILE='mysql-bin.000027', ## taken from xtrabackup_binlog_info -> MASTER_LOG_POS=191; ## taken from xtrabackup_binlog_info SQL> START SLAVE;

For more information on how to set up MySQL Replication, check out this nice manual link.

Prepare GTID slave from full backup

GTID is supported in Xtrabackup starting from version 2.1.0. To restore a GTID slave server, the GTID_MODE should be enabled on the master server before creating its backup, otherwise, no GTID values will be included in the "xtrabackup_binlog_info" file under the backup directory. However, the following steps should be done:

  • Restore the full backup normally as explained above.
  • Check the GTID value of the backup:
    [root@ ~]# cat xtrabackup_binlog_info mysql-bin.000027 191 b9b4712a-df64-11e3-b391-60672090eb04:1-12
  • Set the variable GTID_PURGED with the GTID value of the backup:
    SQL> SET GLOBAL GTID_PURGED="b9b4712a-df64-11e3-b391-60672090eb04:1-12";
  • Execute the auto position CHANGE MASTER TO command and start the slave:
    SQL> CHANGE MASTER TO -> MASTER_HOST='master_ip', -> MASTER_PORT=master_port, -> MASTER_USER='slave_user_name', -> MASTER_PASSWORD='slave_password', -> MASTER_AUTO_POSITION = 1; SQL> START SLAVE;

For more information on how to set up Transaction-based Replication in MySQL, check out this link.

Advantages / Disadvantages
  • Advantages:
    • Fast, simple and easy way to get your full database backed up and restored.
    • All Xtrabackup tool features (like streaming: move the backed up files directly to a remote server) are supported in the full backup method.
    • Simple way to introduce a new slave to the master.
  • Disadvantages:
    • We have to replace the entire MySQL datadir with the new one (In other words, the datadir folder should be empty/removed before the restore process).
    • We can't extract one single database or single table from the whole backup (Unless it's MyISAM table), which means that you have to take it all or leave it all.
Important Hints
  • The message innobackupex: completed OK! should be printed at the end of every xtrabackup command , otherwise, it would be failed to make a successful command (backup, prepare or restore).
  • The ib_logfile files size should be the same in both source and destination servers, if not, you have to either remove them from the backup folder (which will be restored) before starting the MySQL instance and MySQL will create new ones for you OR change those files size in the destination server's configuration file to match the same size in the backup before starting the MySQL instance
  • The MySQL user used in the Xtrabackup tool, should have at least the following privileges (RELOAD, LOCK TABLES and REPLICATION CLIENT).
  • To prepare a new slave from another slave, just add the two options (“--slave-info" & --safe-slave-backup”) to the backup command and use the information in the file "xtrabackup_slave_info" under the backup folder to issue the "CHANGE MASTER TO" command in the new slave after finishing the restore.
  • To Accelerate the preparation process of your backup, just add the option "--use-memory" in the prepare command in order to allocate more used memory (Xtrabackup will use the specified memory as an internal innodb_buffer_pool_size for the prepare process), for ex: [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --use-memory=512M /backup/dir/path/full-backup . . . innobackupex: completed OK!
  • The preparation process consists of two steps, replaying the committed transactions and rolling back the uncommitted transactions, using the --apply-log option only in the preparation command will do both steps for you.
  • The backup folder "/backup/dir/path/full-backup" SHOULD NOT be created before executing the backup command, because Xtrabackup will create that folder for you, and it will fail to continue processing if that folder was already exist.
Incremental Backup

When you have a very large database system, you will need large enough storage to store your database backups in, and if you want to perform a daily backup then your mission will be more difficult. In such cases, the incremental database backup method will be very useful. It allows you to have only the database changes (delta) - after the physical full backup – with the minimum storage space required in a fast way, and hence, you can perform the daily backup operations to your database system without the need to having large storage available.

The following steps describe a simple way to perform your physical incremental database backup using XtraBackup tool:

Create Incremental Backup

To perform an incremental backup, we should first perform a full backup - the same like we did in the previous section - to be the base backup of the upcoming incremental backups.

Creating the full backup (Base Backup):

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp /backup/dir/path/full-backup . . . innobackupex: completed OK!

The "xtrabackup-checkpoints" file contents will be something like:

  • backup_type = full-backuped : which indicates the backup type "full backup".
  • from_lsn = 0 : which indicates the log sequence number where the backup process started from (0 means from the beginning).
  • to_lsn = 3768762 : which indicates the log sequence number where the backup process ended at.

Creating the first incremental backup:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp --incremental /backup/dir/path/inc1 --incremental-basedir=/backup/dir/path/full-backup . . . innobackupex: completed OK!

We informed the Xtrabackup tool to perform an incremental backup by adding the command "--incremental", and by specifying the full-backup path as the basedir, we informed it from which backup it should start tracking the database changes.

The "xtrabackup-checkpoints" file contents will be something like:

  • backup_type = incremental : which indicates the backup type "incremental backup".
  • from_lsn = 3768762 : which indicates the log sequence number where the backup process started from (the same LSN as the previous full backup ended at).
  • to_lsn = 4908762 : which indicates the log sequence number where the backup process ended at.

Creating the second incremental backup:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp --incremental /backup/dir/path/inc2 --incremental-basedir=/backup/dir/path/inc1 . . . innobackupex: completed OK!

We informed the Xtrabackup tool to perform an incremental backup by adding the command "--incremental", and by specifying the 1st incremental backup path as the basedir, we informed it to start tracking the database changes since the last incremental (not the full backup).

The "xtrabackup-checkpoints" file contents will be something like:

  • backup_type = incremental : which indicates the backup type "incremental backup".
  • from_lsn = 4908762 : which indicates the log sequence number where the backup process started from (the same LSN as the 1st incremental backup ended at).
  • to_lsn = 6508762 : which indicates the log sequence number where the backup process ended at.

Note: We can create as many incremental backups as we want by using the same procedure above.

Prepare Incremental Backup

As mentioned earlier in the article, the preparation process consists of two steps (replaying the committed transactions and rolling back the uncommitted transactions) and using the --apply-log option only will do both of them (like we did in the full backup) but in the incremental backups, we MUST do them separately as follows:

  • Replay the committed transactions on the base backup (by adding the option "--redo-only"): [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --redo-only /backup/dir/path/full-backup . . . innobackupex: completed OK!
  • Replay the committed transactions on the 1st incremental backup: [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --redo-only /backup/dir/path/full-backup --incremental-dir=/backup/dir/path/inc1 . . . innobackupex: completed OK!

    Note: we specified the full backup folder here, because replaying the committed transactions steps, appends all changes from the incremental backup to the full backup.

  • Replay the committed transactions on the 2nd incremental backup: [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --redo-only /backup/dir/path/full-backup --incremental-dir=/backup/dir/path/inc2 . . . innobackupex: completed OK!

    Note: here, we didn't use the 1st incremental backup folder, because all changes in the 1st incremental was already appended to the full backup in the previous step.

  • Finally, roll back all uncommitted transactions: [root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log /backup/dir/path/full-backup . . . innobackupex: completed OK!

    Note: as the full backup folder contains all data now (full + 1st & 2nd incremental), there's no need to do this step on the incremental backup folders.

Now, the incremental backup is ready to be restored ...

Restore Incremental Backup

The full backup folder will be the only folder to be restored (there's no need to the incremental backup folders) as it contains all data after appending the changes from all incremental backup. We can restore it the same way we did in the full backup restore.

Advantages / Disadvantages
  • Advantages:
    • Less storage resources needed.
    • Faster than the full backup.
  • Disadvantages:

    In addition to the disadvantages of the full backup, there are other ones:

    • Complicate and hard process to implement as compared to the full backup.
    • The incremental backup strategy, based on Log Sequence Number which affects only XtraDB and InnoDB storage engines while the others (like MyISAM) will be backed up completely (changed + unchanged data) in each incremental backup process.
    • If we have many incremental backups, appending all of them might consume time and might be confusing as well.
    • If one of the incremental backups become corrupted or not available for any reason, we will not be able to add all incremental backups after that to the full backup.
Important Hints
  • The backup preparation sequence steps above, MUST be followed using the same order.
  • If the "--redo-only" option was not be used in any of the preparation steps (except the final step), all up coming incremental backups will be useless as we won't be able to add them to the base backup anymore.
  • Replaying the committed transactions steps bring all incremental data and append it to the full backup, so that, the rolling back of the uncommitted transactions step should be execute only on the full backup (as it contains already the whole data).).
  • In the incremental backups, Xtrabackup generates two files for every table ".delta" & ".meta"(for ex. test.ibd.delta & test.ibd.meta), the delta file size reflects the changes which was applied on that table since the last incremental backup.
  • The preparation time of the individual incremental backup will depend on how much data changed there since the last incremental.
  • The preparation time for the full backup - in most cases - is really small as compared to the incremental ones because full backups apply the redo logs only while the incremental backups apply the deltas plus the redo logs. So if the delta files are big, the preparation process will take longer.
  • Full backups is recommended against Incremental backups if there are many changes applied on the DB, while the incremental backups are recommended when there are few changes applied on the DB.
Differential backup

We can use the incremental backup strategy in order to perform differential backups, but we should consider the following points:

  • We always specify the full backup folder as the base backup (in the incremental we specify the previous incremental folder as a base backup)
  • All incremental backups between differential and full backups MUST BE ignored when preparing the backup files because the differential backup contains already all changes since the last full backup.
  • In the backup preparation process, we should consider the last differential backup as the first incremental backup and all incremental backups after that could be applied normally.

Note: Having differnetial backups in the middle of incremental backups will be useful for many reasons, such as:

  • Differential backups reduce the backup preparation steps/time needed because differential backp will replace all its previous incremental backups.
  • Differential backups reduce the chances of loosing the incremental backups if we have corrupted incremental backup in the middle, because in this case, differential backup will act as a backup of the previous incremental backups.
Partial Backup

Unlike MyISAM, having physical database backup for a single database or table is not possible if the table engine type is InnoDB. But by using the partial database backup method in the XtraBackup tool, it will be possible to have physical InnoDB tables backup the same like MyISAM ones (but with some restrictions).

The following steps describing how to perform partial database backup using XtraBackup tool:

Create Partial Backup

A simple Xtrabackup command to backup some databases (or tables) should be something like:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --no-timestamp --databases=”db1 db2 db3.tbl1” /backup/dir/path/partial-backup . . . innobackupex: completed OK!
Prepare Partial Backup

The same like the other backup methods, the backed up files are not ready until we get them prepared by adding the "--export" option as follows:

[root@ ~]# innobackupex --user=db_user –-password='db_password' --apply-log --export /backup/dir/path/partial-backup . . . innobackupex: completed OK!

Some errors regarding those not included InnoDB tables from the backup may be appeared, but that's fine. Also there will be a notification of creating the ".exp" file for each table which will be used in the import (restore) process.


Now, the partial backup is ready to be restored ...

Restore Partial Backup

The restore process of the partial backup is quite different than the full and incremental backups.

To restore a partial backup, the following steps should be made:

  • Unlike the other methods (Full and Incremental backups), MySQL instance on the destination server shouldn't be stopped because we will have to execute some SQL commands.
  • On the destination server, we should create new tables (as many as we have in the partial backup or as we will restore) with the same structure like the one in the partial backup and then discard its table space: mysql> CREATE TABLE db.tbl1 (...)ENGINE=INNODB; mysql> ALTER TABLE db.tbl1 DISCARD TABLESPACE;
  • Copy “.ibd” and “.exp” files for each table into the corresponding DB directory then assign the right permissions to mysql user: [root@ ~]# cp /backup/dir/path/partial-backup/db/tbl1.ibd /var/lib/mysql/db [root@ ~]# cp /backup/dir/path/partial-backup/db/tbl1.exp /var/lib/mysql/db [root@ ~]# chown -R mysql:mysql /var/lib/mysql/db
  • Now we should tell MySQL to use the new table spaces: mysql> ALTER TABLE db.tbl1 IMPORT TABLESPACE;
Advantages / Disadvantages
  • Advantages:
    • Although it's a complicated process, but it allows us to backup and restore individual InnoDB tables the same like MyISAM.
    • Useful when having huge InnoDB tables and we want to backup/restore them only.
  • Disadvantages:
    • The streaming feature is not available in the partial backup.
    • Restoring/importing individual tables or databases from a partial backup is not applicable unless the destination server is Percona Server.
    • In addition to restoring the files(copy back), three SQL statements should be executed for each table (table creation + two ALTER statements) in order to get them ready for use, which means that we might do a very boring job (or we have to create a special script) to get the partial backup restored if it contains many tables.
Important Hints
  • Although we didn't remove the MySQL datadir before the restore process (like full and incremental backups) as well as having the MySQL instance running, but we can restore the partial backup using the same way (remove the datadir contents and copy/move the backup files to the datadir), but we should take into our consideration that we'll have only the backed up databases/tables and all other databases/tables (which are not included in the partial backup) will be missed.
  • “innodb_file_per_table” server option must be enabled (in both source and destination servers).
  • "innodb_expand_import" option must be enabled in the destination server which is available only in Percona server (and that explain why we can restore partial backup on Percona server only).
  • Beside the "--databases" option, other two alternative options to perform the same needs could be used but we must provide each table with the fully qualified naming format:
    • --include='db.tbl'.
    • --tables-file=/path/to/file.txt ==> in that file, we can add multiple tables one per line in the fully qualified naming format.

Now, you can use the Xtrabackup tool to perform full, incremental and partial database backups, you can decide which method(s) of them are suitable for you according to the advantages and disadvantages of each one, and by considering the important hints for each method you can perform your backup efficiently.

I hope you found this article useful for you and to be familiar with such wonderful tool.

MySQL Environment MyEnv 1.0.1 has been released

FromDual.en - Tue, 2013-11-05 16:03
Taxonomy upgrade extras: myenvoperationMySQL Operationsmulti instanceconsolidation

FromDual has the pleasure to announce the release of the new version 1.0.1 of its popular multi-instance MySQL Environment MyEnv.

You can download MyEnv from here.

In the inconceivable case that you find a bug in MyEnv please report it to our Bugtracker.

Any feedback, statements and testimonials are welcome as well! Please send them to feedback@fromdual.com.

If you have questions about MyEnv we will answer them in the MyEnv forum.

Upgrade from 1.0 or 1.0.x to 1.0.y # cd /home/mysql/product # tar xf /tmp/myenv-1.0.y.tar.gz # rm -f myenv # ln -s myenv-1.0.y myenv
Changes in MyEnv 1.0.1 MyEnv
  • New alias v introduce for echo $MYENV_VERSION.
  • Function start was extended by --option. This allows you to start a Galera cluster node as follows: start --wsrep-cluster-address="gcomm://" (Bug #58).
  • Requirements for Ubuntu in README added.
  • Mismatch between my.cnf and myenv.conf or [mysqld] section and [client] section in my.cnf is checked and reported (Bug #71).
  • Before an instance is started MyEnv will check if file mysql/user.frm is created just in case installation was missing (Bug #82).
  • Non default PID file name is handled properly (Bug #64).
  • Starting an instance will also report SUCCESS or ERROR.
  • Instance check mysqladmin ping was improved by adding a timeout (Bug #56).
  • User to start MySQL instance is checked (Bug #69).
  • The scripts start and stop are much more robust (Bug #66).
  • Internal sequence of checks during start and stop has been changed.
  • Package recommendation for SLES/OpenSuSE completed.
  • Fixes around the POSIX functions for SLES/OpenSuSE.
  • Fake around missing signal constants for SLES/OpenSuSE and more verbose error messages.
  • Check for missing POSIX features added (Bug #59).
  • MyEnv start-up and check sequence changed.
MyEnv Installer
  • Deleting a non existing instance is handled properly.
  • Empty basedir is handled properly.
  • Installation on /etc/my.cnf is handled properly (Bug #62).
  • Installation of my.cnf is redesigned, prepared for future features and error handling is improved.
  • mysql_install_db is also searched in bin path for NDB (MySQL Cluster) set-ups.
  • Installer is checking at the right location to decide if an instance is already installed or not (#Bug #83).
  • Installer should not be started as root user. Throwing warning (Bug #61).
  • Installer has additional suggestion for socket name: mysql-instance.sock, whole installation process is broken up in more pieces for future improvements, test cases were added (Bug #60).
  • Delete instance error messages made more clear (Bug #89).
MyEnv Utilities

No changes in this release.

MySQL Backup Manager

No changes in this release.

Upgrade from Galera Cluster 2.x to 3.0

Abdel-Mawla Gharieb - Thu, 2013-10-24 21:25
Table of Content
Introduction

Codership announced from weeks ago introducing the Galera Cluster new release 3.0 having many bug fixes, performance enhancements plus the main purpose which is working with MySQL 5.6. In this article, I'll go through the upgrade steps from Galera 2.x to the new release 3.0, but at the time of writing this article - as mentioned in the Codership release notes - THIS IS A BETA QUALITY RELEASE FOR TESTING PURPOSES. NOT RECOMMENDED FOR PRODUCTION YET.

Important note: a new Galera version (3.1) will be available soon for production and it will be INCOMPATIBLE with this beta version (3.0) but still compatible with 2.x, so again DO NOT go for production using 3.0 and postpone the production upgrade process until 3.1 become available.

Prerequisites System information

The following are the cluster system information:

  • Operating System: CentOS release 6.4 (64 bit)
  • Cluster system consists of 3 cluster nodes (192.168.1.251 "gcservera",192.168.1.252 "gcserverb" & 192.168.1.253 "gcserverc")
Installed packages

The following are the packages installed on the three cluster nodes:

  • MySQL version: mysql-5.5.33_wsrep_23.7.6 (RPM)
  • Galera provider version:galera-23.2.7 (RPM)
mysql> show global variables like'%version%'; +-------------------------+--------------------------------------------------+ | Variable_name | Value | +-------------------------+--------------------------------------------------+ | innodb_version | 5.5.33 | | protocol_version | 10 | | slave_type_conversions | | | version | 5.5.33-log | | version_comment | MySQL Community Server (GPL), wsrep_23.7.6.r3919 | | version_compile_machine | x86_64 | | version_compile_os | Linux | +-------------------------+--------------------------------------------------+ 7 rows in set (0.03 sec) Required packages

The following are the needed packages to be installed:

  • MySQL 5.6 + Galera plugin: Could be downloaded from here.
  • Galera provider 3.0: Could be downloaded from here
Upgrade the first node Upgrade the installed binaries

To upgrade the installed binaries, you have to stop the mysqld first.

Important: If you are using a load balancer in your cluster system, you should bring the node in question out from the load balancer before stopping the mysqld.

[root@gcservera ~]# /etc/init.d/mysql stop

And then, upgrade with the binaries:

[root@gcservera ~]# rpm -qa|grep MySQL MySQL-server-5.5.33_wsrep_23.7.6-1.rhel6.x86_64 MySQL-client-5.5.34-1.el6.x86_64 [root@gcservera ~]# rpm -e MySQL-server-5.5.33_wsrep_23.7.6-1.rhel6.x86_64 [root@gcservera ~]# rpm -ivh /downloads/MySQL-server-5.6.13_wsrep_24.0-1.rhel6.x86_64.rpm Preparing... ########################################### [100%] 1:MySQL-server ########################################### [100%] [root@gcservera ~]# rpm -U /downloads/galera-24.3.0-1.rhel6.x86_64.rpm Upgrade mysql schema

It's recommended to perform the mysql schema upgrade before joining the cluster, so that, we need to start the node first as a standalone instance by disabling the provider option:

#my.cnf [mysqld] . . #wsrep_provider=/usr/lib64/galera/libgalera_smm.so

Then start the instance and perform the upgrade using the mysql_upgrade utility:

[root@gcservera ~]# /etc/init.d/mysql start Starting MySQL........... SUCCESS! [root@gcservera ~]# mysql_upgrade Prepare the node to join the cluster

All running nodes (second and third nodes) are using the old galera version (2.7) at the moment, so that a backward compatibility option (wsrep_provider_options="socket.checksum=1") MUST be added in the node's configuration in order to join the cluser, otherwise, it will fail to join it back.

Also don't forget to enable again the provider option:

#my.cnf [mysqld] . . wsrep_provider=/usr/lib64/galera/libgalera_smm.so wsrep_provider_options="socket.checksum=1" Joining the cluster

Now we are ready to join the cluster by restarting the node

[root@gcservera ~]# /etc/init.d/mysql restart Stopping MySQL........... SUCCESS! Starting MySQL........... SUCCESS!

We can check the new version as follows:

mysql> show global variables like'%version%'; +-------------------------+------------------------------------------------+ | Variable_name | Value | +-------------------------+------------------------------------------------+ | innodb_version | 5.6.13 | | protocol_version | 10 | | slave_type_conversions | | | version | 5.6.13-log | | version_comment | MySQL Community Server (GPL), wsrep_24.0.r3937 | | version_compile_machine | x86_64 | | version_compile_os | Linux | +-------------------------+------------------------------------------------+ 7 rows in set (0.05 sec)

And the cluster status as well:

mysql> show global status like'wsrep%'; +----------------------------+--------------------------------------+ | Variable_name | Value | +----------------------------+--------------------------------------+ | wsrep_local_state_uuid | ac53dc1e-3aff-11e3-b970-eb7044f6dc77 | . . | wsrep_local_state_comment | Synced | . . | wsrep_cluster_size | 3 | | wsrep_cluster_state_uuid | ac53dc1e-3aff-11e3-b970-eb7044f6dc77 | | wsrep_cluster_status | Primary | | wsrep_connected | ON | | wsrep_local_index | 2 | | wsrep_provider_name | Galera | | wsrep_provider_vendor | Codership Oy | | wsrep_provider_version | 24.3.0(r159) | | wsrep_ready | ON | +----------------------------+--------------------------------------+ 43 rows in set (0.13 sec)

Important: if you are using a load balancer in the cluster, it's now the time to add this node back again to it.


Rolling upgrade the other nodes

You can start doing the rolling upgrade to the other nodes the same like the first one but after making sure that all nodes state (wsrep_local_state_comment) is Synced. If you stopped a node while it's in the Donor state, then the donor and the joiner nodes might be crashed, so make sure of that first.


Get rid of the old release option

After making the upgrade on all the cluster nodes, the backward compatibility option (wsrep_provider_options="socket.checksum=1") is not needed anymore, so removing it from the configuration files and doing a rolling restart on all nodes will do the mission.


Have fun with the new MySQL and Galera releases :)


Taxonomy upgrade extras: upgradegalera

Upgrade from Galera Cluster 2.x to 3.0

Abdel-Mawla Gharieb - Thu, 2013-10-24 21:25
Table of Content
Introduction

Codership announced from weeks ago introducing the Galera Cluster new release 3.0 having many bug fixes, performance enhancements plus the main purpose which is working with MySQL 5.6. In this article, I'll go through the upgrade steps from Galera 2.x to the new release 3.0, but at the time of writing this article - as mentioned in the Codership release notes - THIS IS A BETA QUALITY RELEASE FOR TESTING PURPOSES. NOT RECOMMENDED FOR PRODUCTION YET.

Important note: a new Galera version (3.1) will be available soon for production and it will be INCOMPATIBLE with this beta version (3.0) but still compatible with 2.x, so again DO NOT go for production using 3.0 and postpone the production upgrade process until 3.1 become available.

Prerequisites System information

The following are the cluster system information:

  • Operating System: CentOS release 6.4 (64 bit)
  • Cluster system consists of 3 cluster nodes (192.168.1.251 "gcservera",192.168.1.252 "gcserverb" & 192.168.1.253 "gcserverc")
Installed packages

The following are the packages installed on the three cluster nodes:

  • MySQL version: mysql-5.5.33_wsrep_23.7.6 (RPM)
  • Galera provider version:galera-23.2.7 (RPM)
mysql> show global variables like'%version%'; +-------------------------+--------------------------------------------------+ | Variable_name | Value | +-------------------------+--------------------------------------------------+ | innodb_version | 5.5.33 | | protocol_version | 10 | | slave_type_conversions | | | version | 5.5.33-log | | version_comment | MySQL Community Server (GPL), wsrep_23.7.6.r3919 | | version_compile_machine | x86_64 | | version_compile_os | Linux | +-------------------------+--------------------------------------------------+ 7 rows in set (0.03 sec) Required packages

The following are the needed packages to be installed:

  • MySQL 5.6 + Galera plugin: Could be downloaded from here.
  • Galera provider 3.0: Could be downloaded from here
Upgrade the first node Upgrade the installed binaries

To upgrade the installed binaries, you have to stop the mysqld first.

Important: If you are using a load balancer in your cluster system, you should bring the node in question out from the load balancer before stopping the mysqld.

[root@gcservera ~]# /etc/init.d/mysql stop

And then, upgrade with the binaries:

[root@gcservera ~]# rpm -qa|grep MySQL MySQL-server-5.5.33_wsrep_23.7.6-1.rhel6.x86_64 MySQL-client-5.5.34-1.el6.x86_64 [root@gcservera ~]# rpm -e MySQL-server-5.5.33_wsrep_23.7.6-1.rhel6.x86_64 [root@gcservera ~]# rpm -ivh /downloads/MySQL-server-5.6.13_wsrep_24.0-1.rhel6.x86_64.rpm Preparing... ########################################### [100%] 1:MySQL-server ########################################### [100%] [root@gcservera ~]# rpm -U /downloads/galera-24.3.0-1.rhel6.x86_64.rpm Upgrade mysql schema

It's recommended to perform the mysql schema upgrade before joining the cluster, so that, we need to start the node first as a standalone instance by disabling the provider option:

#my.cnf [mysqld] . . #wsrep_provider=/usr/lib64/galera/libgalera_smm.so

Then start the instance and perform the upgrade using the mysql_upgrade utility:

[root@gcservera ~]# /etc/init.d/mysql start Starting MySQL........... SUCCESS! [root@gcservera ~]# mysql_upgrade Prepare the node to join the cluster

All running nodes (second and third nodes) are using the old galera version (2.7) at the moment, so that a backward compatibility option (wsrep_provider_options="socket.checksum=1") MUST be added in the node's configuration in order to join the cluser, otherwise, it will fail to join it back.

Also don't forget to enable again the provider option:

#my.cnf [mysqld] . . wsrep_provider=/usr/lib64/galera/libgalera_smm.so wsrep_provider_options="socket.checksum=1" Joining the cluster

Now we are ready to join the cluster by restarting the node

[root@gcservera ~]# /etc/init.d/mysql restart Stopping MySQL........... SUCCESS! Starting MySQL........... SUCCESS!

We can check the new version as follows:

mysql> show global variables like'%version%'; +-------------------------+------------------------------------------------+ | Variable_name | Value | +-------------------------+------------------------------------------------+ | innodb_version | 5.6.13 | | protocol_version | 10 | | slave_type_conversions | | | version | 5.6.13-log | | version_comment | MySQL Community Server (GPL), wsrep_24.0.r3937 | | version_compile_machine | x86_64 | | version_compile_os | Linux | +-------------------------+------------------------------------------------+ 7 rows in set (0.05 sec)

And the cluster status as well:

mysql> show global status like'wsrep%'; +----------------------------+--------------------------------------+ | Variable_name | Value | +----------------------------+--------------------------------------+ | wsrep_local_state_uuid | ac53dc1e-3aff-11e3-b970-eb7044f6dc77 | . . | wsrep_local_state_comment | Synced | . . | wsrep_cluster_size | 3 | | wsrep_cluster_state_uuid | ac53dc1e-3aff-11e3-b970-eb7044f6dc77 | | wsrep_cluster_status | Primary | | wsrep_connected | ON | | wsrep_local_index | 2 | | wsrep_provider_name | Galera | | wsrep_provider_vendor | Codership Oy | | wsrep_provider_version | 24.3.0(r159) | | wsrep_ready | ON | +----------------------------+--------------------------------------+ 43 rows in set (0.13 sec)

Important: if you are using a load balancer in the cluster, it's now the time to add this node back again to it.


Rolling upgrade the other nodes

You can start doing the rolling upgrade to the other nodes the same like the first one but after making sure that all nodes state (wsrep_local_state_comment) is Synced. If you stopped a node while it's in the Donor state, then the donor and the joiner nodes might be crashed, so make sure of that first.


Get rid of the old release option

After making the upgrade on all the cluster nodes, the backward compatibility option (wsrep_provider_options="socket.checksum=1") is not needed anymore, so removing it from the configuration files and doing a rolling restart on all nodes will do the mission.


Have fun with the new MySQL and Galera releases :)


Upgrade from Galera Cluster 2.x to 3.0

Abdel-Mawla Gharieb - Thu, 2013-10-24 21:25
Table of Content
Introduction

Codership announced from weeks ago introducing the Galera Cluster new release 3.0 having many bug fixes, performance enhancements plus the main purpose which is working with MySQL 5.6. In this article, I'll go through the upgrade steps from Galera 2.x to the new release 3.0, but at the time of writing this article - as mentioned in the Codership release notes - THIS IS A BETA QUALITY RELEASE FOR TESTING PURPOSES. NOT RECOMMENDED FOR PRODUCTION YET.

Important note: a new Galera version (3.1) will be available soon for production and it will be INCOMPATIBLE with this beta version (3.0) but still compatible with 2.x, so again DO NOT go for production using 3.0 and postpone the production upgrade process until 3.1 become available.

Prerequisites System information

The following are the cluster system information:

  • Operating System: CentOS release 6.4 (64 bit)
  • Cluster system consists of 3 cluster nodes (192.168.1.251 "gcservera",192.168.1.252 "gcserverb" & 192.168.1.253 "gcserverc")
Installed packages

The following are the packages installed on the three cluster nodes:

  • MySQL version: mysql-5.5.33_wsrep_23.7.6 (RPM)
  • Galera provider version:galera-23.2.7 (RPM)
mysql> show global variables like'%version%'; +-------------------------+--------------------------------------------------+ | Variable_name | Value | +-------------------------+--------------------------------------------------+ | innodb_version | 5.5.33 | | protocol_version | 10 | | slave_type_conversions | | | version | 5.5.33-log | | version_comment | MySQL Community Server (GPL), wsrep_23.7.6.r3919 | | version_compile_machine | x86_64 | | version_compile_os | Linux | +-------------------------+--------------------------------------------------+ 7 rows in set (0.03 sec) Required packages

The following are the needed packages to be installed:

  • MySQL 5.6 + Galera plugin: Could be downloaded from here.
  • Galera provider 3.0: Could be downloaded from here
Upgrade the first node Upgrade the installed binaries

To upgrade the installed binaries, you have to stop the mysqld first.

Important: If you are using a load balancer in your cluster system, you should bring the node in question out from the load balancer before stopping the mysqld.

[root@gcservera ~]# /etc/init.d/mysql stop

And then, upgrade with the binaries:

[root@gcservera ~]# rpm -qa|grep MySQL MySQL-server-5.5.33_wsrep_23.7.6-1.rhel6.x86_64 MySQL-client-5.5.34-1.el6.x86_64 [root@gcservera ~]# rpm -e MySQL-server-5.5.33_wsrep_23.7.6-1.rhel6.x86_64 [root@gcservera ~]# rpm -ivh /downloads/MySQL-server-5.6.13_wsrep_24.0-1.rhel6.x86_64.rpm Preparing... ########################################### [100%] 1:MySQL-server ########################################### [100%] [root@gcservera ~]# rpm -U /downloads/galera-24.3.0-1.rhel6.x86_64.rpm Upgrade mysql schema

It's recommended to perform the mysql schema upgrade before joining the cluster, so that, we need to start the node first as a standalone instance by disabling the provider option:

#my.cnf [mysqld] . . #wsrep_provider=/usr/lib64/galera/libgalera_smm.so

Then start the instance and perform the upgrade using the mysql_upgrade utility:

[root@gcservera ~]# /etc/init.d/mysql start Starting MySQL........... SUCCESS! [root@gcservera ~]# mysql_upgrade Prepare the node to join the cluster

All running nodes (second and third nodes) are using the old galera version (2.7) at the moment, so that a backward compatibility option (wsrep_provider_options="socket.checksum=1") MUST be added in the node's configuration in order to join the cluser, otherwise, it will fail to join it back.

Also don't forget to enable again the provider option:

#my.cnf [mysqld] . . wsrep_provider=/usr/lib64/galera/libgalera_smm.so wsrep_provider_options="socket.checksum=1" Joining the cluster

Now we are ready to join the cluster by restarting the node

[root@gcservera ~]# /etc/init.d/mysql restart Stopping MySQL........... SUCCESS! Starting MySQL........... SUCCESS!

We can check the new version as follows:

mysql> show global variables like'%version%'; +-------------------------+------------------------------------------------+ | Variable_name | Value | +-------------------------+------------------------------------------------+ | innodb_version | 5.6.13 | | protocol_version | 10 | | slave_type_conversions | | | version | 5.6.13-log | | version_comment | MySQL Community Server (GPL), wsrep_24.0.r3937 | | version_compile_machine | x86_64 | | version_compile_os | Linux | +-------------------------+------------------------------------------------+ 7 rows in set (0.05 sec)

And the cluster status as well:

mysql> show global status like'wsrep%'; +----------------------------+--------------------------------------+ | Variable_name | Value | +----------------------------+--------------------------------------+ | wsrep_local_state_uuid | ac53dc1e-3aff-11e3-b970-eb7044f6dc77 | . . | wsrep_local_state_comment | Synced | . . | wsrep_cluster_size | 3 | | wsrep_cluster_state_uuid | ac53dc1e-3aff-11e3-b970-eb7044f6dc77 | | wsrep_cluster_status | Primary | | wsrep_connected | ON | | wsrep_local_index | 2 | | wsrep_provider_name | Galera | | wsrep_provider_vendor | Codership Oy | | wsrep_provider_version | 24.3.0(r159) | | wsrep_ready | ON | +----------------------------+--------------------------------------+ 43 rows in set (0.13 sec)

Important: if you are using a load balancer in the cluster, it's now the time to add this node back again to it.


Rolling upgrade the other nodes

You can start doing the rolling upgrade to the other nodes the same like the first one but after making sure that all nodes state (wsrep_local_state_comment) is Synced. If you stopped a node while it's in the Donor state, then the donor and the joiner nodes might be crashed, so make sure of that first.


Get rid of the old release option

After making the upgrade on all the cluster nodes, the backward compatibility option (wsrep_provider_options="socket.checksum=1") is not needed anymore, so removing it from the configuration files and doing a rolling restart on all nodes will do the mission.


Have fun with the new MySQL and Galera releases :)


Pages

Subscribe to FromDual aggregator