Neuigkeiten

Galera features

Galera provides the following features:

• Synchronous replication
• Active/active multi-master topology
• Read and write to any cluster node
• Automatic membership control, failed nodes drop from the cluster
• Automatic node joining
• True parallel row level replication
• Direct client connections
• Drop-in replacement for native MySQL

Galera benefits

Benefits using Galera Replication:

• High Availability
• No slave lag
• No lost transactions
• No more data inconsistency
• Smaller client latencies
• Read scalability …

Table of Contents

• Reasons to Upgrade MariaDB or MySQL
• How to Upgrade MariaDB or MySQL
• MariaDB and MySQL Reserved Keywords
• MariaDB and MySQL Upgrade Problems we hit in real life
  • Upgrade MySQL 4.1 to MySQL 5.1
  • Upgrade MySQL 4.1 to MySQL 5.6
  • Upgrade MySQL 5.0 to MySQL 5.1
  • Upgrade MySQL 5.0 to MySQL 5.6
  • Upgrade MySQL 5.0 to MariaDB 10.3
  • Upgrade MySQL 5.1 to MySQL 5.5
  • Upgrade MySQL 5.5 to MySQL 5.6
  • Upgrade MySQL 5.6 to MySQL 5.7
  • Upgrade MySQL 5.7.12 to MySQL 5.7.21
  • Upgrade MySQL 5.7 to MySQL 8.0
  • Upgrade …

With a partner we are planning and Advanced MySQL DBA workshop. As the name says it should be a workshop. Its planned duration is 2 days. So within 2 days we can cover 4 to 8 topics more deeply. Requirements: VirtualBox, VMware, own Laptop?

Possible exercises during the workshop

• Set-up a Master-Master replication with 2 Slaves
• Load balance on master with MySQL Proxy and on Slaves with LVS.
• Design a little schema and load with data from foodmart
• Do a backup with XtraDB and LVM
• Do a PITR and and InnoDB crash …

Workshop topics

Duration: 3 days, Exercises

• Partitioning
  • Partitioning by time (temporal data)
  • Optimizer and Partitioning
  • Limitations
  • New 5.5 and 5.6 features of Partitioning
• InnoDB Architecture
  • Clustered PK
  • InnoDB Internals
  • InnoDB Online features
  • New Performance Features in 5.5 and 5.6
  • Operations of InnoDB
• Performance Tuning and Benchmarking
  • Basics, Latency and Throughput
  • Profiling
  • Benchmarking
  • Benchmarking Tools
• Memcached
  • Caching strategies
  • Operations
  • ding
• HA and replication solutions …

Hi Partha Dutta,

What you are describing is mentioned in Literature reference [3].

Hi Baron,

90% of the MySQL installations have 1 or 2 concurrent running queries. And most of the MySQL users do not have the time or the capabilities to tune queries, applications or add indexes.

I know that you, Domas and others think the Query Cache should NOT be used at all. However from what we see on the market it is for more MySQL users useful l than it harms.

Regards, Oli

After further investigations with MariaDB developers we found out, that the Problem can be solved when an explicit BEGIN or START TRANSACTION is set in front of the SELECT.

For example:

SET autocommit=0; BEGIN; SELECT * FROM test; COMMIT;

This is similar to other problems we have seen earlier this year where a customers got back several rows on a PK lookup (which IMHO is a Bug)!

As a result: Always use BEGIN or START TRANSACTION when you set autocommit=0!

It looks really evil as you said. I could reproduce it with the following sequence:

SHOW GLOBAL STATUS
WHERE variable_name = 'Qcache_hits'
   OR variable_name = 'Qcache_inserts'
   OR variable_name = 'Qcache_not_cached'
   OR variable_name = 'Qcache_queries_in_cache'
   OR variable_name = 'Com_select';

SET autocommit=0; SELECT * FROM test; COMMIT;
SHOW GLOBAL STATUS
WHERE variable_name = 'Qcache_hits'
   OR variable_name = 'Qcache_inserts'
   OR variable_name = 'Qcache_not_cached'
   OR …

Hi luke,

Great! Thank you for the hint. I somehow missed this bug… :(

Oli

It was just a complex Join Query as far as I can remember:

SELECT * FROM a
JOIN b ON ...
JOIN c ON ...
...
WHERE ...

Possibly it had some GROUP BY or DISTINCT in it.

Regards, Oli

MySQL 5.6

The focus of this release seems to be on Replication, Partitioning and InnoDB/Performance.

• New InnoDB-related INFORMATION_SCHEMA tables and more information in the PERFORMANCE_SCHEMA.
• Persistent Optimizer statistics for InnoDB tables.
• InnoDB table information can be limited when many tables are opened.
• Several internal performance enhancements for InnoDB.
• InnoDB deadlocks can be written to the error log.
• Delayed Replication
• ALTER TABLE ... EXCHANGE PARTITION.
• Binlog shipping with the mysqlbinlog …

In our opinion remote Consulting is non optimal for the customer because tight interaction is difficult remotely.

Please visit Remote-DBA or Consulting…

Workshop topics

Duration: 3 days, register here.

• Partitioning
  • Partitioning by time (temporal data)
  • Optimizer and Partitioning
  • Limitations
  • New 5.5 and 5.6 features of Partitioning
• InnoDB Architecture
  • Clustered PK
  • InnoDB Internals
  • InnoDB Online features
  • New Performance Features in 5.5 and 5.6
  • Operations of InnoDB
• Performance Tuning and Benchmarking
  • Basics, Latency and Throughput
  • Profiling
  • Benchmarking
  • Benchmarking Tools
• Memcached
  • Caching strategies
  • Operations
  • ding
• HA and replication solutions …

Dear MySQL and MariaDB User,

With this Newsletter you receive the News about FromDual for Spring 2011:

Topics

• Technical Information
• Support and Remote-DBA services for MySQL and MariaDB
• Open dates
• News about FromDual
• News about MySQL

Technical Information

We have compiled the following technical information for you:

• Configuration of MySQL for Shared Hosting: Operating MySQL in a Shared Hosting environment is one of the trickiest things you can do with MySQL. Some hints and information about it you can …

We do not currently operate any newsletter.

Old Newsletter

• Summer 2019
• Spring 2015
• Summer 2014
• Winter 2013
• Spring 2013
• Spring 2012
• Winter 2012
• Autumn 2011
• Summer 2011
• Spring 2011

Hi vishnuraj

Simple examples are always good to explain or to try to understand. So I created the following table:

In my database I have 3 fields as id,medicinename,quantity.
In medicinename I have many different names. Many names comes twice or thrice.

+------+----------------------+---------+
| id   | medicinename         | quantiy |
+------+----------------------+---------+
|    1 | Trifolium arvense L. |      25 |
|    2 | Agave americana      |     120 |
|    3 | Hypericum perforatum |      12 | …

The FromDual Performance Monitor for MySQL is now available on our download page.

An other customer was asking the following question related to this blog post:

Does this mean that the other node cannot even respond to read only queries?

Yes this is exactly what it means. It has nothing to do with Master-Slave Replication where you have for example a read-only Slave. In simple set-ups on the passive node there runs nothing except the Heartbeat and the DRBD process.

More details you can find on Wikipedia.

Some people who have their set-ups really under control run on the …

A friend made some private comments and allowed me to post them. IMHO they are worth considering. translate.google.com may help:

We configured Heartbeat to use the version 1 mode because of 3 reasons:

HA ist über Verfügbarkeit. Einfachheit der Konfiguration und eine geringe Anzahl zu testender Systemzustände sind ein wichtiger Punkt. CRM, und alles was danach kommt lösen keine Probleme sondern schaffen sie. Die Konfiguration wird so komplex, daß man sie nicht mehr ganz testen kann. Es geht darum, die …

Hi Henrik,

Yes great idea!

Basically the differences are: Zabbix is an integrated Monitoring and Alerting solution where Nagios is an Alerting Solution with Monitoring modules. Zabbix is technologically newer than Nagios. I have heard from some customers that they are not happy with how Fixes and Changes are included into Nagios. Incinga is a fork of Nagios and seems to address these problems.

More information you can find in the discussion: Can some one recommend some good monitoring tools for …