You are here

Shinguz's blog

MySQL Multi-Master – Single-Slave – Replication (Episode 2)

Introduction

One of the features that make MySQL so great is its easy replication set-up. If you are experienced and know-how to do it, it takes you about 15 minutes to set-up a slave. What you have in the end is a replication from one master to one or several slaves. So you can build a top-down data stream pyramid and spread your data on many slaves.

MySQL Replication

Test application for MySQL high availability (HA) set-up

When I set-up a MySQL HA environment for customers I usually do some final fail over tests after configuring the whole beast.

To check if the application behaves like expected I always run my little test application (test.sh) from the server(s) where the customers application runs. It displays "graphically" how the application behaves and you can show to the customer immediately what is going on...

Make sure, that you point it to the virtual IP (VIP) or the load balancer (LB).

Citation of the week

Taxonomy upgrade extras: 

Das dreieckige Rad hat gegenüber dem viereckigen einen gewaltigen Vorteil: Ein Rumms weniger pro Umdrehung!

Translation:

The triangular wheel has one enormous advantage over the quadrangular: One knock less per revolution!

Maybe not new, but I have not heard it yet and I love it. It was about reinventing functionality in a well known product...

MySQL Cluster: No more room in index file

Taxonomy upgrade extras: 

Recently we were migrating an InnoDB/MyISAM schema to NDB. I was too lazy to calculate all the needed MySQL Cluster parameters (for example with ndb_size.pl) and just took my default config.ini template.

Because I am really lazy I have a little script doing this for me (alter_engine.pl).

But suddenly my euphoria was stopped abruptly by the following error:

Possible memory leak in NDB-API applications?

Taxonomy upgrade extras: 

A customer has recently experienced a possible memory leak in its NDB-API application. What he did was something like:

# ps aux | grep <pid>

over time and then he saw the RSS increasing. When he would have had a look a little longer he would have seen that the RSS consumption would increase up to a certain level and then becomes stable. Which is the expected behavior.

Possible Memory Leak

Active/active fail over cluster with MySQL Replication

Electing a slave as new master and aligning the other slaves to the new master

In a simple MySQL Replication set-up you have high-availability (HA) on the read side (r). But for the master which covers all the writes (w) and the time critical read (rt) there is no HA implemented. For some situations this can be OK. For example if you have rarely writes or if you can wait until a new Master is set up.

But in other cases you need a fast fail-over to a new master.

Typical automated MySQL maintenance jobs, query cache

The following maintenance jobs are typically run against a MySQL database:

  • Backup
  • Clean-up binary logs
  • Optimize tables
  • Purge query cache
  • Rotate binary logs

Backup

A backup is not a typical maintenance job. But it behaves more or less like one. The backup should be done regularly depending on the restore/PITR (Point in Time Recovery) requirements.

My thoughts about MySQL (Cluster) replication

Taxonomy upgrade extras: 

According to Johans wishes I write down my concerns about MySQL (Cluster) replication. These items are things I run again and again into it with customers:

Some more details about DiskSyncSize

The parameter DiskSyncSize is a MySQL Cluster parameter and was added in MySQL 5.1.23.

After the amount of stored bytes of data per file, the data node will fsync (flush) the LCP file to disk, even if a fsync is not needed for consistency.

This is done because the OS will otherwise buffer all the writes, and when a fsync is really needed, it can take a lot of time...

Originally this parameter was hard coded. Now it defaults to 4 Mbyte.

With MySQL-Enterprise Montior through firewalls

Taxonomy upgrade extras: 

Sometimes it is nice to show customers the functionality of MySQL-Enterprise Monitor (aka Merlin). I install the agents on the servers and the dashboard runs on my laptop. But very often only ssh is open to these servers.

So how to dig a whole through the firewall for MySQL-Enterprise Monitor?

# ssh -R 18080:localhost:18080 oli@where_the_agent_sits

Maybe trivial for you but for me its hard to remember...

Sparse files

Taxonomy upgrade extras: 

What is a sparse file?

"A sparse file is a file where space has been allocated but not actually filled with data. These space is not written to the file system. Instead, brief information about these empty regions is stored, which takes up much less disk space. These regions are only written to disk at their actual size when data is written to them. The file system transparently converts reads from empty sections into blocks filled with zero bytes at runtime." [ 1 ]

Ranking in MySQL results

Taxonomy upgrade extras: 

A friend of me asked me long time ago: "How can I have a ranking on a result with MySQL?". Now I found some time to write it down:

Lets do first some preparation for the example:

CREATE TABLE sales (
  id     INT UNSIGNED NOT NULL AUTO_INCREMENT PRIMARY KEY
, fruit  VARCHAR(32)
, amount DECIMAL
);

INSERT INTO sales
VALUES (NULL, 'apple', 12.75), (NULL, 'orange', 1.89), (NULL, 'pear', 19.23)
     , (NULL, 'banana', 4.25), (NULL, 'cherry', 123.75), (NULL, 'plum', 23.15)
;

Now lets query:

MySQL logon trigger

With MySQL 5.0 the database provides trigger functionality on INSERT, REPLACE, UPDATE and DELETE.

Those of you who know some other RDBMS know, that there are also some system events where one would like to have triggers.

Unfortunately MySQL does not (yet) provide such functionality. This is sad because as database administrator this would be sometimes very helpful.

But you can build your own LOGON and STARTUP trigger.

MySQL provides some hooks for these events...

MySQL Cluster restore

Recently the question came up if it is faster to restore a MySQL cluster when all nodes are up or only ONE node from each node group during restore.

The answer from our gurus was: All nodes up during restore! I wanted to find out why. So I set up the following cluster and started to measure:

MySQL Cluster set up

Cluster set-up

MyEnv (MySQL and MariaDB BasEnv)

What is MyEnv?

MyEnv is a tool to run several MySQL or MariaDB database instances on ONE host. You can even run multiple database instances with different binary versions. We call this multi-instance set-ups.

With MyEnv a multi-instance set-up is more comfortable to handle than with mysqld_multi (old) or Systemd services (new) and it provides more useful functionality.

MySQL Active - Active Clustering

It is possible to use an active - active shared-disk cluster in MySQL in some cases. For doing this you have to fulfill the following requirements:

Transaction performance

Transaction performance relates among other things from I/O performance. This means hard disk performance.

Hard disk performance

When you select a hard disk, an important feature to consider is the performance (speed) of the drive. Hard disks come in a wide range of performance capabilities. As is true of many things, one of the best indicators of a drive’s relative performance is its price. An old saying from the automobile-racing industry is appropriate here: "Speed costs money. How fast do you want to go?"

Round-Robin Database Storage Engine (RRD)

In a round-robin database (RRD) usually time-series data like network bandwidth, temperatures, CPU load etc. is stored. The data is stored in the way that system storage footprint remains constant over time. This avoids resource expensive purge jobs and reduces complexity.

RRD

MySQL does NOT yet provide this kind of storage engine. Although some people were thinking about and some prototypes exists.

SATA Flash Solid State Disk up to 160 Gbyte announced!

The price for a 160 Gbyte disk will be around USD 15'000. This is still a bit expensive. But the access time is around 0.5 ms (both for reading and writing) which is around 10 times faster than a normal 15'000 rpm SCSI disk! The disk has NO cache because it is a cache itself (according to the supplier. Maybe this will change in the future). And the lifetime of a cell is > 5 mio writes. For the same performance one needs usually an array of around 10 disks. If your database is heavily write-I/O bound you should consider this solution.

MySQL Multi-Master - Single-Slave - Replication

MySQL provides its replication for High Availability (HA) and for read Scale-out. Generally it is known that in a MySQL replication you can only replicate from one Master to many slaves. In this paper it is shown how a set-up can look like to replicate from two masters to one slave.

Caution: Handle this information with care!!!

Profiling MySQL with oprofile

Taxonomy upgrade extras: 

Probably the answer to this question is already known. But we want to prove it and by the way learn to deal with MySQL and oprofile.

Materialized Views (MV) with MySQL

Materialised View (MV) is the pre-calculated (materialised) result of a query. Unlike a simple VIEW the result of a Materialised View is stored somewhere, generally in a table. Materialised Views are used when immediate response is needed and the query where the Materialised View bases on would take to long to produce a result. Materialised Views have to be refreshed once in a while. It depends on the requirements how often a Materialised View is refreshed and how actual its content is.

Pittfalls with Federated Tables

A Federated Table is a table which points to a table in an other MySQL database instance (mostly on an other server). It can be seen as a view to this remote database table. Other RDBMS have similar concepts for example database links.

Read more about the pittfalls with MySQL Federated Tables...

Some more MySQL tools added

We have added some more tools to our MySQL consulting tools collection.

Some more benchmarks added

Taxonomy upgrade extras: 

We have added some more database benchmarks to our collection.

More details you can find on our Benchmarking page.

Pages

Subscribe to RSS - Shinguz's blog