When organizing things helps to simplify life.

In the previous article, we start to explore dynamic privileges and the interaction with static ones. We also saw how to remove SUPER privilege from a DBA account. 

What we did was go by subtraction. But in real life, we should act differently. We should ADD only what is really needed for the account to work correctly.

Adding privilege one by one, and for each user is problematic given the level of interaction they may have, and also prone to mistakes. 

Instead, we can use ROLES to group, assign, and revoke the correct privileges in a much easier way.

This is becoming even more important in MySQL with the advent of dynamic privileges.

What should we do to correctly use ROLES? Well first of all design.   …

When trying to make things better, make it very complicated.

I was working on a Security Threat Tool script when I had to learn more about the interaction between static and dynamic privileges in MySQL 8.

Dynamic privileges is a “new” thing added in MySQL 8 to easily extend the privileges definition, and at the same time to provide more granularity. For instance, the FLUSH operation now has dedicated Privileges and by scope. 

Dynamic privileges are assigned at runtime. Most of them are active when the server starts. But they can also change with respect to the components or plugin when activated. ( …

In the previous release of our Percona Distribution for MySQL Operator, we implemented one interesting feature, which can be seen as “self-healing”: https://jira.percona.com/browse/K8SPXC-564.

I do not think it got enough attention, so I want to write more about this.

As it is well known, a 3-node cluster can survive a crash of one node (or pod, in Kubernetes terminology), and this case is very well handled by itself. However, if there is a problem with 2 nodes at the same time, this scenario is problematic for Percona XtraDB Cluster. Let’s see why this is a problem.

First, let’s review if the first node goes offline:

In this case, the cluster can continue work, because Node 1 and Node 2 figure out …

In my previous post Back From a Long Sleep, MyDumper Lives!, I told you that Fast Index Creation was in the queue and (after fixing several bugs) it will package in release 0.10.7 next month. But why am I so excited about it? Well, this feature opens new opportunities, what I’m going to call Restore Strategies.

In the Past…

In the past, the only option was to first restore the table definition and then insert the data rows in two simple steps. On tables with millions of rows, we already know why it takes more time as it inserts in the clustered index and in the secondary index, instead of building the secondary indexes after the data has been inserted, as myloader is able to do now. 

Nowadays, we have the option to do it, in …

The combination of the information obtained from the “pt-summaries” discussed in the previous posts of this series (Part 1: The Percona Support Way, Part 2: Knowing the Server, Part 3: What MySQL?) helps us come up with the first impression of a MySQL server. However, apart from the quick glance we get at two samples of a selective group of MySQL status variables, they provide what I call a “static” view of the server, akin to looking at a still picture of an engine. We get the chance to spot some major discrepancies in the MySQL configuration in view of the available resources in the …

In the previous post of this series, we looked at the hardware specifications and operating system settings of the host server through the lenses of a pt-summary report. Now that we know the foundation on which the database is running, we can turn our focus to MySQL itself. The second of our triad of tools from the Percona Toolkit will help us with that:

pt-mysql-summary conveniently summarizes the status and configuration of a MySQL database server so that you can learn about it at a glance. 

The goal for this part is to find what MySQL distribution and version is being used if the server …

The new MyDumper 0.10.5 version, which includes many new features and bug fixes, is now available.  You can download the code from here.

For this release, we focused on fixing some old issues and testing old pull requests to get higher quality code. On releases 0.10.1, 0.10.3, and 0.10.5, we released the packages compiled against MySQL 5.7 libraries, but from now on, we are also compiling against MySQL 8 libraries for testing purposes, not releasing, as we think that more people in the community will start compiling against the latest version, and we should be prepared.

New Features:

• Password obfuscation #312
• Using …

There is a common pattern in life, you often discover or understand things by accident. Many scientific discoveries fit such a description. In our database world, I was looking to see how BLOB/TEXT columns are allocated using overlay pages and I stumbled upon something interesting and unexpected. Let me present to you my findings, along with my attempt at explaining what is happening.

InnoDB Tablespaces

The first oddity I found is a bunch of free pages in each tablespace it is skipping. Here’s an example from a simple table with only an integer primary key and a char(32) column:

root@LabPS8_1:~/1btr# innodb_space -f /var/lib/mysql/test/t1.ibd space-extents
start_page  page_used_bitmap
0               #####################################........................... <--- free pages
64              ################################################################
128 …

This blog post will show how to compile the Percona XtraBackup (PXB) tool for ARM. For this, we are going to use an AWS EC2 ARM instance with Ubuntu 20.04(Focal Fossa).

The motivation for this was born in my interest in the new generation of ARM processors and if this is a viable option for the future. Ideally, I do not recommend installing all the necessary packages to compile Xtrabackup in a production environment for security reasons. Still, you can have a “compiling” server for this purpose and then move the binaries around.

Machine Configuration

For this blog post, I picked a c6g.2xlarge instance. The machine has the following hardware configuration:

Multi-writer replication has been a challenge in the MySQL ecosystem for years before truly dedicated solutions were introduced – first Galera (and so Percona XtradDB Cluster (PXC)) replication (around 2011), and then Group Replication (first GA in 2016).

Now, with both multi-writer technologies available, do we still need traditional asynchronous replication, set up in active-active topology? Apparently yes, there are still valid use cases. And you may need it not only when for some reason Galera/PXC or GR are not suitable, but also when you actually use them. Of course, the most typical case is to have a second cluster in a different …