I just wanted to collect the links of the Google and MariaDB relationship. These were all items in September 2013.

The Register: Google swaps out MySQL, moves to MariaDB – this was largely based on a presentation at SLAC by Jeremy Cole (slides). Same news was picked up by ReadWrite (with a different angle naturally), Google Waves Goodbye To MySQL In Favor Of MariaDB.

If you’re more audio inclined, there was great discussion on the This Week in Google podcast: checkout episode …

Education is everything. Education can pull you out of poverty, it can propel you into upper echelons. Education can make you see what’s truly important in the life. Education can enable you to appreciate the people. Blogging is all about education. This Log Buffer brings education about database arena from the blogs.

Oracle:

Large CLOBs and attention to detail needed while reading Oracle Support notes.

Here is a simple way to get SQL_ID of large and similar queries in Oracle.

Robin Moffatt is sharing tips adn tricks for the OBIEE …

Percona Server provides users with a variety of powerful benefits including enterprise-grade features, high performance in demanding use cases, rich diagnostics, and access to unique Percona XtraBackup capabilities, all for free and open source. In addition, Percona Server is a drop-in MySQL replacement. Compatibility with MySQL has important ramifications (as well as some risks) as Giuseppe Maxia highlighted in a recent blog post about MySQL alternatives. Our commitment to drop-in compatibility with MySQL has important benefits for Percona Server users and drives our development process as I’ll describe.

Drop-in …

This article is providing the details to the 500K QPS with MySQL 5.7 benchmark results published earlier and explained during my talk at MySQL Connect..

Looking back in the history of MySQL / InnoDB improvement, you may easily realize that there was no any so big speed-up(s) between releases in read-only (RO) performance as they came with MySQL 5.6. And it's easy to understand as well that without having a good level of scalability on read-only, it's hard to expect to reach a good level on read+write (specially when reads are dominating ;-))

However, we were so happy with RO performance in MySQL 5.6, that with 5.7 we turned our main focus on the read+write (RW), because on big data volumes things are not yet going as we wish.. But the RW dependency on RO jumped up once more again, and then …

MySQL Enterprise Monitor (MEM) 3.0 is a huge improvement over MEM 2.x and I really hope you'll take a look at it.  My goal here is to tell you about a new feature that is near-and-dear to my heart--looking at a server in the context of a group of servers.
Why is this important?
As many of you know it's important that each server in a replication topology have a unique server_id (each slave, really).  It's not hard to give each server a server_id, but it's also very easy to forget this step, especially if you're cloning a slave from another slave, or if you're under the gun to do something quickly.  It's surprising how often our Support engineers and consultants tell us they see this in the field.
The problem is that if more than one server has the same server_id as another, problems can occur that may be difficult …

As a follow-up to my earlier post describing the process to create custom Advisors for MySQL Enterprise Monitor, this post will demonstrate how to create custom graphs to track metrics over time. The password policy scripts I introduced earlier will again be the basis for the data used in this post.

Collecting new data

Similar to the custom Advisor created in the earlier post, creating a custom graph starts with custom data collection.  Why can’t I just use the data collections I defined in the earlier example?  Because there’s a restriction on using multiple custom data types (defined as namespace+classname combinations) in a single graph, and in the earlier example, I used …

We are very pleased and excited to announce that MySQL Enterprise Monitor 3.0.2 is now available for download on the My Oracle Support (MOS) web site. It will also be available via the Oracle Software Delivery Cloud with the October update in about 1 week.

Based on feedback from our customers, MySQL Enterprise Monitor (MEM) 3.0 offers many significant improvements over previous releases. Highlights include:

• Policy-based automatic scheduling of rules and event handling (including email notifications) make administration of scale-out easier and automatic
• Enhancements such as automatic discovery of MySQL instances, centralized agent configuration and multi-instance monitoring further improve ease of configuration and management
• The new cloud and virtualization-friendly, "agent-less" design allows remote monitoring of MySQL databases without the need for any remote agents
• Trends, projections …

One of our customers reported that ‘create table select from’ statements stall for a period of time equal to the TokuDB lock timeout.  This indicated a lock conflict between multiple transactions.  In addition, other MySQL clients that were opening unrelated tables were also stalled.  This indicated that some shared mutex is held too long.  We discuss details about this bug and how it was fixed.  The bug fix will be distributed in TokuDB 7.1.0.

Example
Suppose that we set the tokudb lock timeout to 60 seconds just to exaggerate the stall.

mysql> set global tokudb_lock_timeout=60000;
Query OK, 0 rows affected (0.00 sec)


We then create a simple table.

mysql> create table s (id int primary key);
Query OK, 0 rows affected (0.02 sec)


When we create new table (t) and populate it from some …

One of the major announcements made at the MySQL Connect conference just over a week ago was the labs release of MySQL Fabric, which supports management of MySQL Server farms in a sharded deployment.  It’s available on labs.mysql.com, which means it is just an early release with some rough edges.  One of those rough edges that I’d like to see resolved soon is that it’s difficult to install on Windows, as there is no installer package provided.  There is some documentation on how to install Fabric (you’ll find it in section 15.8.2 of the PDF that makes up the downloadable documentation), but it’s focused on Linux platforms, so I thought I would post the process that …

Sreekar,
Yes, in general you can specify multiple output threads.
And you can definitely play with that.
See what works and what does not.
There are sometimes worst performance numbers when you have multiple
write threads (even in non-raid disks).
Regards,
Sanjay