You may have noticed that in many of my demos, I disable firewall and SELinux (I even use --initialize-insecure sometimes ). This is just to make things easier… But in fact enabling iptables and SELinux are not complicated.

Firewall

These examples are compatible with Oracle Linux, RedHat and CentOS. If you use another distro, the principle is the same.

For the firewall, we need first to allow incoming traffic to MySQL and MySQL X ports: 3306 and 33060:

# firewall-cmd --zone=public --add-port=3306/tcp --permanent
# firewall-cmd --zone=public --add-port=33060/tcp --permanent

If you don’t plan to restart the firewall, you just need to run the same commands without --permanent to make then immediately active.

Then we need to allow the Group Replication’s communication port. This is usually 33061 but it can be configured in …

You have seen in this previous post, that since 8.0.17, it’s now possible to query the MySQL Router using its REST API.

Additionally, we also saw in this post, that since 8.0.17, we are now able to write extensions to MySQL Shell using the Extension Framework.

Let’s combine both and see how we can integrate the MySQL Router’s REST API in the Shell.

I’ve created an extension in ext.router that creates a MySQL Router Object.

The new extension, as a method to create the object:

This is an example that illustrates how to create a MySQL Router Object, as you can see you can pass the password directly as parameter but it’s not recommended in interactive mode. It’s …

Since MySQL 8.0.16, the Router as the possibility to launch an internal webserver (see Jan’s blog post).

Even if this webserver could serve static files, it was the first piece of a much more interesting solution that is now available since 8.0.17.

It’s possible now to query the MySQL Router via its REST API and get a lot of useful information.

Setup

Let’s first configure our MySQL Router to take advantages of this new feature. In this example, I will add the following lines to /etc/mysqlrouter/mysqlrouter.conf that I created using the --bootsrapcommand line argument:

[http_server]
port=8080

[rest_api]

[rest_router]
require_realm=somerealm

[rest_routing]
require_realm=somerealm …

Create a MySQL InnoDB Cluster using MySQL 8.0 has always been very easy. Certainly thanks to MySQL Shell and server enhancements like SET PERSIST and RESTART statement (see this post).

The most complicated part to deal with was the existing and none existing data. In fact GTID sets must be compatible.

Let me explain that with some examples:

Example 1 – empty servers

If you have empty servers with GTID enabled, manually creating credentials to connect to each MySQL instances will generate GTIDs that will prevent nodes to …

Ok, so if you’re reading this, then I can guess you’ve got a MySQL InnoDB Cluster in an awkard shape, i.e. you need to restore a backup and add the instance back into the cluster, so we have all our instances again.

As it might be logical to think “ah, but I’ve only lost 1 instance, a read-only instance, so all I have to do is backup & restore the other read-only instance and I’m home free. Well I want to make it a little harder. So in this scenario, assume that we’ve lost both the READ-ONLY instances, so I need to backup my primary READ-WRITE instance.

I’ve got a 8.0.16 instance, on Oracle Linux 7.4. We’ll be looking at 2 hosts, ic1 & ic3.

We’ll be using the MySQL Enterprise Edition Server, that bundles MySQL Enterprise Backup with the rpm’s so we don’t need to install anything else.

I’ll assume you’ve got access to Oracle …

To operate more easily a MySQL Group Replication (InnoDB Cluster), the Group Replication plugins provides some UDFs.

If you have read the recent article from Tiago Vale about the Group Replication Communication Protocol, you may have heard about two new UDFs allowing to get or set  the communication protocol.

So what are all the UDFs provided with the Group Replication and what’s their purpose ?

SELECT UDF_NAME FROM performance_schema.user_defined_functions 
WHERE UDF_NAME LIKE 'group_repl%';
+-------------------------------------------------+
 | UDF_NAME                                        |
 +-------------------------------------------------+
 | group_replication_get_communication_protocol    |
 | group_replication_get_write_concurrency         |
 | group_replication_set_as_primary                | …

At the end of the month, some engineers of the MySQL Team will be present in Austin, TX !

We will attend the first edition of Percona Live USA in Texas.

During that show, you will have the chance to meet key engineers, product managers, as well as Dave and myself.

Let me present you the Team that will be present during the conference:

The week will start with the MySQL InnoDB Cluster full day tutorial by Kenny and myself. This tutorial is a full hands-on tutorial where we will start by migrating a classical asynchronous master-replicas topology to a new MySQL InnoDB Cluster. We will then experience …

As explained in this previous post, it’s now (since 8.0.16) possible to use the MySQL Shell Reporting Framework to monitor MySQL InnoDB Cluster.

Additionally, when a member of the MySQL InnoDB Cluster’s Group leaves the group for any reason, or when a new node is added from a backup, this member needs to sync up with the other nodes of the cluster. This process is called the Distributed Recovery.

During the Distributed Recovery, the joiner receives from a donor all the missing transactions using asynchronous replication on a dedicated channel.

It’s of course also possible to monitor the progress of this recovery process by calculating how many transactions have …

With MySQL Shell 8.0.16, a new very interesting feature was released: the Reporting Framework.

Jesper already blogged about it and I recommend you to read his articles if you are interested in writing your own report:

• https://mysql.wisborg.dk/2019/04/26/mysql-shell-8-0-16-built-in-reports/
• https://mysql.wisborg.dk/2019/04/27/mysql-shell-8-0-16-user-defined-reports/

I this post, I will show you one user-defined report that can be used to monitor your MySQL InnoDB Cluster / Group Replication.

Preparation

Before being able to use the report, you need to download 2 files. The first one is the …

We saw yesterday that when an issue (like network splitting), it’s possible to remain with a partitioned cluster where none of the partition have quorum (majority of members). For more info read how to manage a split-brain situation.

If your read the previous article you notice the red warning about forcing the quorum. As an advice is never too much, let me write it down again here : “Be careful that the best practice is to shutdown the other nodes to avoid any kind of conflicts if they reappear during the process of forcing quorum“.

But if some network problem is happening it might not be possible to shutdown those other nodes. Would it be really bad ?

YES !

Split-Brain

Remember, we were in this situation:

We …