White Papers from Percona
A Checklist for Preventing Common But Deadly MySQL® Problems, Part 3
by Andrew Moore, Benjamin Mildren, David Busby, Max Bubenick, November, 2014
This is the third in a series of three white papers for busy DBAs and their management. The series provides a checklist of some of the most common MySQL implementation, performance, and recovery problems that can be minimized or prevented by proactive MySQL administration best practices. - Read More
A Checklist for Preventing Common But Deadly MySQL® Problems, Part 2
by Andrew Moore, Benjamin Mildren, David Busby, Max Bubenick, October, 2014
This is the second in a series of three white papers for busy DBAs and their management. The series provides a checklist of some of the most common MySQL implementation, performance, and recovery problems that can be minimized or prevented by proactive MySQL administration best practices. - Read More
A Checklist for Preventing Common But Deadly MySQL® Problems, Part 1
by Andrew Moore, Benjamin Mildren, David Busby, Max Bubenick, October, 2014
We created a checklist of common but deadly MySQL problems to make your life easier. If handled reactively, these issues can lead to unacceptable downtime, unnecessary costs, malicious attacks, frustrated users, and lost business. - Read More
MySQL Backup and Recovery Best Practices
by Ryan Lowe, June, 2014
Backup and recovery processes are a foundational piece of any application infrastructure. A well-tested backup and recovery system can be the difference between a minor outage and the end of a business. - Read More
Percona Report Measuring ScaleArc with Sysbench
by Peter Boros, March, 2014
ScaleArc hired Percona to perform various tests on its database traffic management product. This case study is the outcome of the benchmarks carried out by Uday Sawant (ScaleArc) and Peter Boros (Percona). - Read More
High Level Multi-Datacenter MySQL High Availability
by Jay Janssen, September, 2013
This document represents current MySQL high availability best practices. These form the core of what Percona consultants and engineers will typically discuss with customers exploring this space. - Read More
Intel 520 SSD - Sysbench Report
by Vadim Tkachenko, May, 2012
The goal of this benchmark is to show the performance impact of using an Intel 520 SSD drive under a MySQL workload. - Read More
MySQL Performance Analysis with Percona Toolkit and TCP/IP Network Traffic
by Baron Schwartz, February, 2012
The TCP network traffic conversation between a client and a MySQL server is a rich source of data for many types of analysis. In this paper, we show how the simplest type of traffic packet headers with no payload can provide a surprising variety of information about the system’s performance. The techniques that we demonstrate are quick, require o - Read More
Virident FlashMAX M1400 MySQL - tpcc-mysql report
by Vadim Tkachenko, November, 2011
The goal of this research is to compare the performance (throughput) of Virident FlashMAX M1400 card with Fusion-io ioDrive Duo 1.28TB card under a tpcc-mysql workload. - Read More
Virident FlashMAX M1400 MySQL - sysbench report
by Vadim Tkachenko, November, 2011
The goal of this research is to compare the performance (throughput) of Virident FlashMAX M1400 card with Fusion-io ioDrive Duo 1.28TB card under a sysbench oltp workload. The results and conclusions are shown on this paper. - Read More
Clustrix tpcc-mysql Benchmark
by Vadim Tkachenko, October, 2011
To evaluate the performance of Clustrix, we ran the tpcc-mysql benchmark-ing utility with a scale factor of 5000 warehouses. The purpose of running tpcc-mysql on each system was to determine: The median throughput of each system The 90% response time provided by each system - Read More
Fusion-io MySQL multi-instances report
by Vadim Tkachenko, October, 2011
Performance research on running MySQL multi-instance deployment on Fusion-io ioDrive - Read More
Improving Percona Server performance with Flashcache on the Virident tachIOn Drive
by Baron Schwartz, Vadim Tkachenko, August, 2011
Using flash storage as a second-level cache for MySQL, between disk and main memory, reduces I/O latency while avoiding the size and cost limitations of storing the data on flash alone. This white paper presents the results of our benchmarks with Facebook's Flashcache module and the Virident tachIOn PCI-E storage device. We explain when it is cost- - Read More
Scaling MySQL With Virident Flash Drives and Multiple Instances of Percona Server
by Baron Schwartz, Shridar Subramanian, Vadim Tkachenko, May, 2011
Hardware continues to outpace software, and with the large amounts of memory, many CPU cores, and enormously powerful flash drives readily available today, a hybrid strategy of multiple database servers per physical server is necessary to better utilize the hardware's capacity. Such a configuration can enable a 5x to 15x consolidation factor, and i - Read More
Causes of Downtime in Production MySQL Servers
by Baron Schwartz, March, 2011
Everyone wants to prevent database downtime by being proactive, but are measures such as inspecting logs and analyzing SQL really proactive? Although these are worthwhile, they almost always identify problems that already exist. To be truly proactive, one must prevent problems, which requires studying and understanding the reasons for downtime. T - Read More
Scaling MySQL Deployments With Percona Server and Virident tachIOn Drives
by Baron Schwartz, Shridar Subramanian, Vadim Tkachenko, February, 2011
Until recently, MySQL and InnoDB were unable to take advantage of high-performance hardware, particularly storage systems. Large-scale deployments made heavy use of so-called horizontal scaling with a technique called sharding — partitioning data across many small-to-midsize servers. Today, MySQL — especially Percon - Read More
Preventing MySQL Emergencies
by Baron Schwartz, February, 2011
Preventing downtime-causing emergencies in MySQL can be difficult because they are caused by complex combinations of several things going wrong. Efforts to be proactive may be sincere, but without knowledge of the causes of emergencies, they often fail to prevent further problems. This white paper explains dozens of ways that real emergencies coul - Read More
Configuring Percona Server with XtraDB for Fast Restarts
by Baron Schwartz, Vadim Tkachenko, January, 2011
InnoDB requires a long time to restart on servers with large amounts of memory. It can take many hours to warm InnoDB up enough to serve queries quickly. The database's "working set" of data is not in memory after a restart, so queries must perform random disk I/O to read pages, and are thus too slow for production use. Percona Server with XtraDB g - Read More
Forecasting MySQL Scalability with the Universal Scalability Law
by Baron Schwartz, November, 2010
Forecasting a system's scalability limitations can help answer questions such as “will my server handle ten times the existing load?” and “at what point will I need to upgrade my hardware?” Timely answers to these questions have more business value than exact predictions. Mathematical models can help reduce guesswork while - Read More
Percona Server with XtraDB for Software-as-a-Service Application Databases
by Baron Schwartz, Vadim Tkachenko, October, 2010
The Software as a Service (SaaS) deployment model is popular with customers, but SaaS databases are often difficult for providers to manage, due to the unique stresses placed upon them. The open-source MySQL database server can suffer from performance and operational limitations in SaaS deployments. Percona Server with XtraDB is an enhanced versio - Read More
Goal-Driven Performance Optimization
by Baron Schwartz, Peter Zaitsev, April, 2010
Goal-Driven Performance Optimization is the method we use at Percona to deliver results quickly for our clients. It is a simple, universal process that anyone can use. Benefits include time savings, a defined termination condition, avoidance of distractions, measurable results, and proof when goals are unachievable. - Read More
