Overview The Skinny

When it comes to zero downtime, proxies are the first line components of a cluster.

In order to achieve High Availability (HA) for MySQL, MariaDB and Percona Server, a commonly deployed setup consists of configuring load balancers (hardware or software) on top of those proxies.

A Strong Architecture How is Maintenance Made Possible?

With this proxy + load balancer architecture, server maintenance is made possible on any of the proxy hosts, as follows:

• the proxy is stopped
• the load balancer detects the dead proxy and removes it from the pool
• new connection requests go to live proxies

The Problem What Happens to Existing Sessions?

But wait… even though new connections are re-routed correctly, what happens …

The latest enhancement in ProxySQL v2.0.0 is the native support of Galera Cluster. No more need to use an external script within the scheduler like explained also in this post of ours.

This document will cover how to take an advantage of the new feature and integrate ProxySQL with Percona XtraDB Cluster to monitor galera node status and read-write split performed by ProxySQL. To illustrate we will use a cluster of 3 nodes, below are the details:

In this blog post, we demonstrate how to convert a single standalone cluster into a Composite Primary/DR topology running in two data centers.

Our example starting cluster has 5 nodes (1 master and 4 slaves) and uses service name alpha. Our target cluster will have 6 nodes (3 per cluster) in 2 member clusters alpha_east and alpha_west in composite service alpha.

This means that we will reuse the existing service name alpha as the name of the new composite service, and create two new service names, one for each cluster (alpha_east and alpha_west).

Below is an INI file extract example for our starting standalone cluster with 5 nodes:

[defaults]
...

[alpha]
connectors=db1,db2,db3,db4,db5
master=db1
members=db1,db2,db3,db4,db5
topology=clustered

To convert the above configuration to a Composite Primary/DR:

1. First you must stop all services on all existing nodes:

   shell> stopall
   

   …

In this blog post, we talk about what happened during an installation of the Tungsten Cluster into an environment with SELinux running and mis-configured.

An attempt to execute `tpm install` on v5.3.2 recently failed with the below error:

ERROR >> node3_production_customer_com >> Unable to run 'sudo systemctl status mysqld.service' or the database server is not running (DatasourceBootScriptCheck) 
Update the /etc/sudoers file or disable sudo by adding --enable-sudo-access=false 

Worse, this customer reported that this appeared as a WARNING only in Dev and Staging tests. So we checked, and it seemed we were able to access systemctl properly:

shell> sudo systemctl status mysqld.service
● mysqld.service - MySQL Percona Server
Loaded: loaded (/usr/lib/systemd/system/mysqld.service; enabled; vendor preset: disabled)
Active: activating (start-post) since Tue 2018-06-19 17:46:19 BST; 1min 15s ago …

In this blog post, we talk about how to configure automatic start at boot time for the Tungsten Clustering components.

By default, Tungsten Clustering does not start automatically on boot. To enable Tungsten Clustering to start at boot time, use the deployall script provided to create the necessary boot scripts:

shell> sudo /opt/continuent/tungsten/cluster-home/bin/deployall

To disable automatic startup at boot time, use the undeployall command:

shell> sudo /opt/continuent/tungsten/cluster-home/bin/undeployall

For Multisite/Multimaster deployments in specific, there are separate cross-site replicators running. In this case, a custom startup script must be created, otherwise the replicator will be unable to start as it has been configured in a different directory.

1. Create a link from the Tungsten Replicator service startup script in the operating system …

In this blog post, we talk about the Automatic Reconnect feature in the Tungsten Connector.

Automatic reconnect enables the Connector to re-establish a connection in the event of a transient failure. Under specific circumstances, the Connector will also retry the query.

Connector automatic reconnect is enabled by default in Proxy and Smartscale modes.

Use the following tpm command option on the command line (remove the leading hyphens inside INI files):

--connector-autoreconnect=false

to disable automatic reconnect.

This feature is not available while running in Bridge Mode. Use the tpm command option --connector-bridge-mode=false to disable Bridge mode.

Automatic reconnect enables retries of statements under the following circumstances:

• not in bridge mode
• not inside a transaction
• no temp table has been created
…

In this blog post, we talk about how existing client connections are handled by the Tungsten Connector when a manual master role switch is invoked and how to adjust that behavior.

When a graceful switch is invoked via cctrl or the Tungsten Dashboard, by default the Connector will wait for five (5) seconds to allow in-flight activities to complete before forcibly disconnecting all active connections from the application side, no matter what type of query was in use.

If connections still exist after the timeout interval, they are forcibly closed, and the application will get back an error.

This configuration setting ONLY applies to a manual switch. During a failover caused by loss of MySQL availability, there is no wait and all connections are force-closed immediately.

This timeout is adjusted via the tpm option …

In this blog post, we talk about how query connections are handled by the Tungsten Connector, especially read-only connections.

There are multiple ways to configure session handling in the Connector. The three main modes are Bridge, Proxy/Direct and Proxy/SmartScale.

In Bridge mode, the data source to connect to is chosen ONCE for the lifetime of the connection, which means that the selection of a different node will only happen if a NEW connection is opened through the Connector.

So if your application reuses its connections, all traffic sent through that session will continue to land on the selected read slave, i.e., when using connection pooling.

http://docs.continuent.com/tungsten-clustering-6.0/connector-bridgemode.html

The key difference is in how the slave latency checking is handled: …

The website or mobile app is the storefront for participating in the modern digital era. It’s your portal for inviting users to come and survey your products and services. Much attention focuses on front-end development; this is where the HMTL5, CSS, and JavaScript are coded to develop the landing page that everyone sees when they visit your site.

But the real magic happens on the backend. This is the ecosystem that really powers your website. One writer has articulated this point very nicely as follows:

The technology and programming that “power” a site—what your end user doesn’t see but what makes the site run—is called the back end. Consisting of the server, the database, and the server-side applications, it’s the behind-the-scenes functionality—the brain of a site. …

Welcome to another post in the series of Percona Live featured tutorial speakers blogs! In these blogs, we’ll highlight some of the tutorial speakers that will be at this year’s Percona Live conference. We’ll also discuss how these tutorials can help you improve your database environment. Make sure to read to the end to get a special Percona Live 2017 registration bonus!

In this Percona Live featured tutorial, we’ll meet Derek Downey (OSDB Practice Advocate, Pythian), David Turner (Storage SRE, Uber) and René Cannaò (MySQL SRE, Dropbox / ProxySQL). Their session is ProxySQL Tutorial. There is a stigma attached to database proxies when it comes to MySQL. This tutorial hopes to blow away that stigma by showing you what can be done with a proxy designed from the …