Scott Drummonds on Virtualization
Just a month ago I was in Hong Kong for customer meetings. One customer explained that their Oracle databases were still un-virtualized because “Oracle does not support VMware”. Sigh.
I have for years been correcting this misunderstanding among customers. Despite Oracle’s clear statement of support we often see conflicting messages from Oracle’s sales teams. I was pleased to see a video from Oracle’s Richard Garstagen at VMworld 2012 on the unambiguous support of VMware environments:
Happily, Richard even mentions the effectiveness of host affinity in limiting the license costs. He also addresses the fear that reproduction on physical is a serious concern for virtualized Oracle databases.
I first read about this on License Consulting. Thanks for your article!
The Singapore Solutions Center builds live demonstrations for our customers. We use these demonstrations for training, experimentation, and sometimes to create videos that show cool technologies in action. This week Osmund Amoroso in the SSC released a canned demo (a video) showing work they have done on VFCache.
In this demo Osmund shows an Oracle database running on vSphere. The database is tested by an OLTP workload. Osmund shows results before and after enabling VFCache. Note the 43% increase in transaction rate. And be reminded that applications that spend less time waiting for data will use more of the CPU, as the video shows.
A couple weeks ago I joined a discussion between engineers and customer-facing technologists within EMC and VMware. There was some confusion around a claim by EMC with respect to Transparent Page Sharing (TPS). There exists an EMC paper that hints at disabling TPS. The astute Michael Webster thought this contradicted best practices I provided when leading VMware’s performance technical marketing team. Michael was correct so I decided to jump in and see what I could learn.
Fixed recommendations for consolidation ratios are cancerous. Whether we are talking about vCPUs per core, virtual machines per host, or VMDKs per LUN, there is no single number the represents the “right” ratio. Accurate guidance requires workload characterization and fine tuning using vSphere’s performance counters. Today I want to highlight one experiment that shows application choice impacting VMDK-to-LUN consolidation. The inescapable conclusion is that sequential access data must be separated from random access files!
I love VMware’s performance blog, VROOM! It is our most popular performance communication vehicle and its content is backed by a stellar engineering team with unmatched integrity. Each article details the nuances of VMware performance and educates on application and platform best practices. I love all the articles but am always surprised as to which our readers find most popular. Here is a countdown of the five entries most read in 2009.
[Taken from my communities blog, this article shows you why you should “Love Your Balloon Driver”.]
Earlier this month we finally published one of my favorite papers from ongoing vSphere launch activities. This paper on ESX memory management, written by Fei Guo of performance engineering, has three graphs that are absolute gems. They show balloon driver memory savings next to throughput numbers for three common benchmarks. The conclusion is inescapable: the balloon driver reclaims memory from over-provisioned VMs with virtually no impact to performance. This is true on every workload save one: Java.
Read the rest of this entry »