|
|
|
|
|
Guidelines for Performance Testing the Sun Storage 7410 Unified Storage System Three-part Series from Fishworks Engineer Brendan Gregg June 12, 2009, Volume 136, Issue 2
Performance testing and developing best practices for the Sun Storage 7000 series is the subject of a three-part blog by Fishworks Engineer Brendan Gregg. The complexities of the hybrid storage pool make performance testing a more complicated but no less necessary process, he notes in part 1. As a first order of business, Gregg recommends doing a reality check on performance stats and doing that check in more than one manner just to double check one's own results. He cautions against reading client caching results as if they pertained to the file server, which they do not. He also recommends distribution client load over at least 10 clients. And then, "drive CPUs to saturation," he advises, unless you are willing to settle for results measuring less than peak performance. He then points out that testing must involve a minimum of two fully-populated JBODs if a proper sense of the performance of the 7410 is to be learned. A further consideration that Gregg recommends is to check storage profile by evaluating the desired redundancy profile against mirroring, since the default RAID-Z2 storage profile on the 7000 series provides double-parity, but can also deliver lower performance than mirroring, particularly with random reads. Another requirement that Gregg stipulates is the use of read-biased SSDs (Readzillas) -- as many as possible for concurrent I/O -- and letting them warm up before testing begins. Here he stresses the importance of adjusting database record size before creating files. Gregg also recommends using Logzillas to accelerate the pace of synchronous write workloads. Finally, document the test system you are using: If you have not tested a max configuration, let readers know that, Gregg writes. His comments on this aspect are unambiguous: "Every performance result should be accompanied by details on the target system, and a comparison to a maximum configuration, to give an accurate representation of a product's true capabilities." In part 2 of his series, Gregg provides a tuning checklist for achieving maximum performance on the Sun Storage 7410, particularly for finding performance limits. He notes that the list was originally written as an internal Sun document and has been released in the interests of transparency, adding that tuning of the sort outlined is used during product development to drive systems as fast as possible to identify and solve bottlenecks. In this blog, Gregg addresses Sun Storage 7410 hardware, settings, network considerations, and client number and workload. He advises against trusting benchmark tools and urges checking benchmark results with the analytics on the 7410. He cautions that using a benchmark tool without tuning the file size is almost certainly going to result in erroneous findings. Finally, in Part 3 of his serious, Gregg posts load generation scripts he has used to measure the limits of the Sun Storage 7410. About these scripts he writes, the scripts are very simple, as all they need to do is generate load. They don't generate a specific level of load or measure the performance of that load but just apply load as fast as they can. Measurements are taken on the target - which can be a more reliable way of measuring what actually happened (past the client caches). For anything more complex, he advises using a real benchmarking tool. Gregg includes scripts for random reads, sequential reads, client management, and scripts for running the workload. He includes as well scripts for stepping the workload. For all the detail in Gregg's method -- and it is considerable -- in the end he observes that "This type of testing can be useful to determine how fast a storage server can go - it's top speed. But that's all. For a better test of application performance, reach for a real benchmarking tool, or setup a test environment and run the application with a simulated workload." More Information
Sun Storage 7410 Unified Storage System Performance Testing the 7000 series, part 1 of 3 - Brendan Gregg blog
Keywords:
|
|
||||||||||||||||||||||||||||||
This type of testing can be useful to determine how fast a storage server can go - it's top speed. But that's all. 
Print
Email
Guidelines for Performance Testing the Sun Storage 7410 Unified Storage System
(this article)
|
News and Solutions for Users of Solaris, Java and Oracle's Sun hardware products
Just the news you need, none of what you don't – 42,000+ Members – 24,000+ Articles Published since 1998 |
Copyright © 1998-2011 System News, Inc.
