In the first installment of the Virtual SAN Performance Testing series, we reviewed benchmarking performance using synthetic I/O generation tool Iometer, automated by the VMware I/O Analyzer appliance. Using Iometer, or other synthetic I/O generation tools is frequently the first option for benchmarking selected, as it is an operationally light weight method to benchmark storage performance. But what if you want to be able to simulate real world workloads within your Virtual SAN cluster, without the burden of building out applications. That is where I/O trace files can come into play.
Category Archives: Performance
As people begin to assess, design, build, and deploy VMware Virtual SAN based solutions for the first time, there is great curiosity in understanding the performance expectations to have, and results one can achieve when utilizing Virtual SAN in specific configurations. Most customers are running some type of benchmark in proof-of-concept environments in order to gauge the performance of VMware Virtual SAN in their environment. In working with customers and partners, we have seen a variety of methods used in attempting to benchmark and analyze Virtual SAN performance. In order to ease this process, we are developing guidance on how best to perform performance testing on Virtual SAN. This guidance will be presented in a four part series as follows:
- Virtual SAN Performance Testing Part I – Utilizing I/O Analyzer with Iometer
- Virtual SAN Performance Testing Part II – Utilizing I/O Analyzer with Application Trace Files
- Virtual SAN Performance Testing Part III – Utilizing Custom Application Trace files
- Virtual SAN Performance Testing Part IV – Analyzing Performance Results
Today VMware released Update 2 of its vSphere management solution, vCenter Server. In this release there are updates to the supported database versions and many resolved known issues.
- vCenter Server database support: vCenter Server now supports the following external databases:
- Oracle 12c. Important: For pre-requisite requirements, see KB 2079443.
- Microsoft SQL Server 2012 Service Pack 1
- Microsoft SQL Server 2014
- vCloud Hybrid Service: The vCloud Hybrid Service (vCHS) introduces a new container, Hybrid Cloud Service, on the vSphere Web Client home page. The Hybrid Cloud Service container contains the vCHS installer and the new vCloud Connector installer.
- Customer Experience Improvement Program: The vSphere customer experience improvement program is introduced to collect configuration data for vSphere and transmit weekly to VMware for analysis in understanding the usage and improving the product. For more details, see the vSphere Documentation Center.
Microsoft Exchange is a very common critical application managed by enterprises. Exchange 2013 has brought forth many changes to the performance and IO profile of its infrastructure.
Exchange Server 2013 is an even greater candidate for virtualization than its predecessors. Architectural changes and improvements to the core of Exchange Server, along with advancements in server hardware, make vSphere the default choice for Exchange 2013.
There are some really very interesting and useful virtualized Big Data-related talks and demos at VMworld 2014 during the week of the 24th-27th August. Among them is a customer panel at 4-5pm on Monday discussing vSphere, vCAC and Big Data Extensions for managing your virtualized Hadoop workloads as well as a talk from a VMware customer on providing Hadoop-as-a-Service (1-2pm on Monday and again at 2:30pm on Wednesday).
Here is a quick summary of the various Big Data events at the show. Come by and meet us at the various demo stations where we will show these technologies, as well as at the Hands-On Lab, HOL-SDDC-1409. Hope to see you there!
|VAPP1428 – Hadoop as a Service: Utilizing VMware vCloud Automation Center and Big Data Extensions at Adobe||Mon, Aug 25, 1-2pm (first occurrence)Wed, Aug 27, 2:30-3:30pm (second occurrence)|
|VAPP1807 – Best Practices of Virtualizing Hadoop on vSphere – Customer Panel||Mon, Aug 25, 4-5pm|
|VAPP1314 – Scaling Your Storage Architecture for Big Data – How the Isilon Server Fits Well with Virtual Hadoop Workloads||Tue, Aug 26, 4-5pm|
|VAPP1859 – Extreme Performance Series: How to Build a High Performance Platform for Big Data||Wed, Aug 27, 8:30-9:30am|
|HOL-SDDC-1409 – Big Data Extensions Hands on Lab :||All Week|
|Demonstrations on the VMware booth and at the Data Lake area of the EMC/EVP booth||All Week|
|Customer Meetings||Contact email@example.com|
|VMware Big Data Special Interest Group (SIG)||Contact firstname.lastname@example.org|
TPC-C Benchmark is an on-line transaction processing (OLTP). (TPCC Main site) TPC-C uses a mix of five concurrent transactions of different types and complexity. The database is comprised of nine types of tables with a wide range of record and population sizes. TPC-C is measured as transactions per minute (TPM).
The goal of this exercise was to see if 1 million TPM can be achieved on virtualized SQL 2014 backed by an all Flash storage array for a TPC-C like test. The TPC-C testing would be compared between two VM sizes (Within NUMA & Exceeding NUMA boundaries)
As many of you know, I work very closely with VMware Product Management and Engineering and we are always looking for data points to support concepts and aid the development of future generations with our products. We have created a 3 page survey that covers Performance and Availability within vCenter Server environments.
Help us make our products better by spending a few minutes to complete, your participation is much appreciated.
Now that I’ve scared you, lets take a look at these use cases.
I’ve been asked a few times recently how to determine what virtual topology vNUMA recommended and created for us and besides looking at the obvious guest OS for the final result, you can also check the vmware.log file for more detailed information.
Additional background here.
1) “Wide and “Flat” virtual machine – default configuration
This virtual machine was configured with 20 vCPUs (20 sockets and 1 corespersocket) on a 4 socket, 10 core, hyper-threading enabled, host:
numa: Exposing multicore topology with cpuid.coresPerSocket = 10 is suggested for best performance
numaHost: 2 virtual nodes, 20 virtual sockets, 2 physical domains
Here we see vNUMA has automatically set corespersocket = 10, which matches the physical topology, and presented 2 “virtual nodes” aka NUMA nodes.
2) Spanning pNUMA nodes – manually configured
This virtual machine was configured with 20 vCPUs (1 socket and 20 corespersocket) on a 4 socket, 10 core, hyper-threading enabled, host:
numa: Setting.vcpu.maxPerVirtualNode=20 to match cpuid.coresPerSocket
numaHost: 1 virtual nodes, 1 virtual sockets, 2 physical domains
Here we see vNUMA has respected the manual configuration and set the vNUMA advanced setting maxPerVirtualNode = 20 which doesn’t match the physical topology. 1 “virtual nodes” aka NUMA node is presented which spans 2 “physical domains,” aka pNUMA nodes.
So searching vmware.log for ‘numa’ and ‘numaHost’ will provide these details and again a reminder to let vNUMA provide the optimal configuration when possible.
Eighteen months ago at SAP SAPPHIRE Madrid 2012, VMware presented impressive performance data showcasing SAP HANA running on vSphere. At that time, vSphere 5.1 received support from SAP to run SAP HANA for non-production scenarios.
After an extended and successful joint testing program with SAP, we’re proud to announce that vSphere 5.5 has achieved production support for SAP HANA. Today, joint customers can run and scale single node SAP HANA databases up to 1TB in a virtual environment while taking advantage of all of the vSphere features they know and rely on to achieve high availability and improved Quality of Service for their mission critical workloads.