I’m happy to announce that a new version of the MariaDB Audit Plugin is available. Version 1.1.5 can be downloaded here. As you can see the Audit Plugin is available from SkySQL, who has developed the plugin.

However, now with the Audit Plugin being GA for a couple of months since 7th of November last year and customers using it in production, SkySQL has decided to contribute the Audit Plugin to the MariaDB project and I’m happy to tell you that starting from MariaDB versions 5.5.37 and 10.0.9 the Audit Plugin will be included by default. Notice that these versions of MariaDB aren’t yet released.

The MariaDB Audit Plugin introduces the capabilities of tracking user access to data. By having the Audit Plugin available by default in MariaDB, all users can easily set up tracking in their own systems and follow in real time who’s doing what in …

I’m frequently tracking semaphores waits, and if you’ve examined them before, it can be a little matching up the threads listed in the SEMAPHORES section with the transactions in the TRANSACTIONS section.

Semaphore waits are related to internal synchronization between threads in mysqld, and not directly to row locks or other items associated with user queries, so that’s why the SEMAPHORES section only reports the OS thread id.

Fortunately, the TRANSACTIONS sections also reports the OS thread handle, but in hex format.

Here is an example semaphore wait:

--Thread 1079654736 has waited at ibuf0ibuf.c line 3549
for 943.00 seconds the semaphore:
X-lock (wait_ex) on RW-latch at 0x7f2a48830bf8 '&block->lock'
a writer (thread id 1079654736) has reserved it in mode wait exclusive
number of readers 1, waiters flag 1, lock_word: ffffffffffffffff
Last time read locked in file buf0flu.c line 1318

So how do …

If you’re like me, you’ve probably heard of the Spider storage engine, but not used it yet.

While it has been available for some years now, I just simply haven’t used it before until now.

I suspect that has to do with ease of installation. Previously, one had to compile it with MySQL in order to use it, which excludes a lot of people. However, in MariaDB 10.0 (as of 10.0.4), it is very easy to add and use.

And with MariaDB 10.0.8 being declared RC, combined with Spider’s sheer usefulness, I only suspect its usage will become more and more widespread.

What is the Spider storage engine, and why will it be useful?

“The Spider storage engine is a storage engine with built-in sharding features. It supports partitioning and xa transactions, and allows tables of different MariaDB instances to be handled as if they were on the same instance. It refers to one possible implementation of ISO/IEC …

MySQL 5.6.16 was recently released (it is the latest MySQL 5.6, is GA), and is available for download here:

http://dev.mysql.com/downloads/mysql/5.6.html

As opposed to the latest 5.5 release, this latest 5.6 release has quite a few more bug fixes, but that’s expected since 5.5 has been GA for much longer.

There were 2 minor functionality changes:

• Previously, ALTER TABLE in MySQL 5.6 could alter a table such that the result had temporal columns in both 5.5 and 5.6 format. Now ALTER TABLE upgrades old temporal columns to 5.6 format for ADD COLUMN, CHANGE COLUMN, MODIFY COLUMN, ADD INDEX, and FORCE operations. This conversion cannot be done using the INPLACE algorithm, so specifying ALGORITHM=INPLACE in these cases results in an error. (Bug #17246318)
• CMake now supports a -DTMPDIR=dir_name option to specify the default …

MySQL 5.5.36 was recently released (it is the latest MySQL 5.5, is GA), and is available for download here:

http://dev.mysql.com/downloads/mysql/5.5.html

I was reading through the changelogs to review the changes and fixes, and to summarize, I must say this release is mostly uneventful.

There was one new feature added (for building, so not really applicable to everyone), and only 17 bugs fixed.

The new feature is this:

• CMake now supports a -DTMPDIR=dir_name option to specify the default tmpdir value. If unspecified, the value defaults to P_tmpdir in . Thanks to Honza Horak for the patch. (Bug #68338, Bug #16316074)

Out of the 17 bugs, there was only 1 I thought worth mentioning (because it is a wrong results bug):

• COUNT(DISTINCT) sometimes produced an incorrect result when the last read row …

I recently wrote a post on tracking metadata locks (MDL) in MySQL 5.7, and I wanted to take a moment to expand on it by explaining a couple of the associated variables in more detail.

First off, once you have enabled the performance_schema *and* the metadata lock instrumentation, you can verify it with:

mysql> SELECT * FROM performance_schema.setup_instruments
    -> WHERE NAME = 'wait/lock/metadata/sql/mdl';
+----------------------------+---------+-------+
| NAME                       | ENABLED | TIMED |
+----------------------------+---------+-------+
| wait/lock/metadata/sql/mdl | YES     | YES   |
+----------------------------+---------+-------+

“ENABLED” will report “YES” if it is enabled properly, and “NO” if not.

“TIMED” (referring to event timing) reports “YES” if it was enabled via the my.cnf or my.ini file, and it reports …

I recently blogged about tracking metadata locks in the latest MySQL, and now I want to discuss how to track these metadata locks in MariaDB.

In MySQL 5.7, there is a table named `metadata_locks` added to the performance_schema (performance_schema must be enabled *and* the metadata_locks instrument must be specifically enabled as well.

In the MariaDB 10.0 implementation (as of 10.0.7), there is a table named METADATA_LOCK_INFO added to the *information_schema*. This is a new plugin, so the plugin must be installed, but that is very simple with:

INSTALL SONAME 'metadata_lock_info';

Then, you will have the table.

To see it in action:

Connection #1:

mysql> create table t (id int) engine=myisam;
mysql> begin;
mysql> select * from t;

Connection #2:

mysql> alter table t add index …

If you’re running MySQL 5.7, you might have noticed the ibtmp1 file located in the datadir, and you might be wondering exactly what this file is.

In 5.7, InnoDB added a separate tablespace for all non-compressed InnoDB temporary tables. This new tablespace is named ibtmp1 and is located in the datadir by default.

“The new tablespace is always recreated on server startup. … A newly added configuration file option, innodb_temp_data_file_path, allows for a user-defined temporary data file path. For related information, see InnoDB Temporary Table Undo Logs.”

You can read that, and the full changelog entry, here:

http://dev.mysql.com/doc/relnotes/mysql/5.7/en/news-5-7-1.html

And the full variable description is here:

…

I’ve blogged about metadata locks (MDL) in the past (1 2 3) and in particular discussed how best to track them down and troubleshoot threads stuck waiting on metadata locks.

If you’ve had any experience with these, you’ll know finding them isn’t always the most straight-forward task.

So I was glad to see metadata lock instrumentation added to MySQL 5.7.3 as part of performance_schema, which makes tracking these down a breeze! (Note this is only in 5.7.3 currently, and therefore is some time from being GA as of today)!

To use these, performance_schema must be enabled (i.e., performance_schema=1 in your config file).

But, also, the metadata_locks instrument is disabled by default, so even if you enable the …

I recently upgraded an instance to 5.7.3 the other day, and ran into an error, so I wanted to share the resolution for it here.

In my case, I was upgrading 5.7.1 to 5.7.3. However, this will apply to anyone wanting to upgrade from pre-5.7.2 (including 5.6/5.5) to 5.7.2+.

I performed the upgrade, in-place, and restarted mysqld. This was fine. However, then I attempted to connect via the command-line, and received the following error:

shell> mysql -uroot -ppass -P3310
ERROR 1130 (HY000): Host 'localhost' is not allowed
to connect to this MySQL server

Searching the net, you’ll mostly find RTM replies, which were all accurate as far as I could tell. In all of those prior reported cases, the issues were expected behavior and the issues were ultimately user error.

Of course I double-checked my config and data files. I knew I didn’t change anything in the user table, or any system table, for that matter. …