Jun 132017
 

PASS has just announced this year’s precon sessions for the PASS Summit in Seattle this October, and I am humbled to announce that my session, “Virtual SQL Servers – Right to Ludicrous Speed” has been selected for one of the Tuesday precon slots. This topic is near and dear to my heart, and I’m thrilled to be able to present to the group advanced SQL Server VM tuning techniques learned from over 16 years of performance tuning.

If you are less interested in SQL Server infrastructure than other topics, I urge you to check out the other precons being presented at this conference. The topics and presenters are world class, and if I was not presenting this precon on Tuesday, I’d be in one of the other precons!

I hope all of you can attend what I consider the best SQL Server ecosystem conference in the world. I’ll never miss it, and I look forward to meeting you all there!

Session Details:

Join this full-day introduction session focused on managing and boosting the performance of a virtualized SQL Server environment. The focus of the course is to help those new to virtualization, infrastructure, and cloud-based infrastructure-as-a-service concepts to become familiar with the concept and purposes of virtualization and how it can benefit them as data professionals. Participants will gain exposure to all layers of virtualization underneath SQL Server, from storage to hypervisor to the SQL Server instance, no matter if on-premises or in the cloud, and will learn how to construct the entire stack with a strong emphasis on SQL Server performance.

Intended Audience
The intended audience of this course is information workers (both business and IT-centric) involved with architecting a virtualization strategy for SQL Server, or managing business-critical SQL Servers that have already been virtualized.

Course Topics
The following topics will be addressed during the course:
• The Physical Infrastructure Underneath the VM
• Storage and SAN Concepts and Design
• Virtualization and Infrastructure Fundamentals
• The SQL Server Virtual Machine
• Networking, Support, and Licensing
• SQL Server Infrastructure in the Cloud
• High Availability and Disaster Recovery and SQL Server
• Performance Investigation and Tuning of the Entire Stack

A working set of the screens and configuration settings referenced in the session, together with the reference slides and documentation, will be provided to attendees, for strategic reference in working with virtualized SQL Server in their own environments.

Register today! Seats are sure to fill up for these precons!

Mar 292017
 

Watch your memory configuration! You can’t just throw RAM in a physical server and expect it to work right. Depending on your DIMM configuration, you might have accidentally slowed down your memory speed, which will surely slow down your application servers. This speed decrease is virtually undetectable from the OS. Anything that leverages lots of RAM to function, including a database server, can take a substantial performance hit on performance.

An example of this is if you wish to configure 384GB of RAM on a new server. The server has 24 memory slots. You could populate each of the memory slots with 16GB sticks of memory to get to the 384GB total. Or, you could spend a bit more money to buy 32GB sticks of memory and only fill up half of the memory slots. Your outcome is the same amount of RAM. Your price tag on the memory is slightly higher than the relatively cheaper smaller sticks.

In this configuration, your 16GB DIMM configuration runs the memory 22% slower than if you buy the higher density sticks. Check out page 63 of the server build guide for an HPE Proliant DL380 Gen9 server. The fully populated 16GB stick configuration runs the memory at 1866 MHz. If you only fill in the 32GB sticks on half the slots, the memory runs at 2400 MHz.

Database servers, both physical and virtual, use memory as an I/O cache, improving the performance of the database engine by reducing the dependency on slower storage and leveraging the speed of RAM to boost performance. If the memory is slower, your databases will perform worse. Validate your memory speed on your servers, both now and for upcoming hardware purchases. Ensure that your memory configuration yields the fastest possible performance. Your applications will thank you!

Nov 012016
 

I am proud to announce the availability of our new training offerings! We are launching a live web-based version of our one-day SQL Server infrastructure and virtualization master class. We are hosting an all-day training session before the end of the year to help you take advantage of quiet(er) times around the holidays. Select from one of the dates that works best for you and register today!

Tuesday, November 22nd – 10:00am to 6:30pm Eastern

Thursday, December 29th – 10:00am to 6:30pm Eastern

Looking to ramp up on the infrastructure and virtualization underneath your databases? Want to learn how to work effectively with the other silos in your IT organization? This live web-based training will help take you to the next level by understanding the architecture, performance characteristics, and performance tuning opportunities of the storage, physical server hardware, virtualization hypervisor, and VM layers underneath your critical SQL Server databases.

David KleeMicrosoft MVP and VMware vExpert David Klee leads this full-day introduction to getting the most out of the infrastructure underneath your SQL Server environment. The focus of the course is to help those new to the enterprise server infrastructure concepts become familiar with function and purpose of each layer of the architecture around their databases, and how this knowledge can benefit them as data professionals. Participants will gain exposure to all layers of infrastructure and virtualization underneath SQL Server, from storage all the way through to the SQL Server instance, and will learn how to review and engineer the entire stack with a strong emphasis on SQL Server performance.

Intended Audience

The intended audience of this course is information workers (both business and IT-centric) involved with architecting an enterprise IT strategy for SQL Server.

Course Topics

The following topics will be addressed during the course:

  • Datacenter Fundamentals
  • Infrastructure Stack Fundamentals
  • Storage and SAN Concepts and Design
  • Networking and Interconnect Topologies
  • Physical Server Technologies
  • The SQL Server Virtual Machine
  • High Availability and Disaster Recovery
  • Performance Investigation and Tuning of the Entire Stack

Course Objectives

Upon completion of this course, attendees will be able to:

  • Understand key infrastructure concepts
  • Determine the optimal infrastructure configuration for best performance
  • Review their own enterprise infrastructure for performance bottlenecks
  • Construct a SQL Server virtual machine template with best practices for agility and performance
  • Understand how and what stack metrics to benchmark and baseline to ensure proper objective performance measurement
  • Identify and remediate common infrastructure-related SQL Server performance killers
  • Understand proper capacity management techniques for long term management

Prerequisites

Attendees with the following experiences will benefit the most from this course:

  • Basic familiarity with the core components of a server
  • SQL Server installation and configuration
  • Exposure to SQL Server performance metric measurement and identification

A working set of the screens and configuration settings referenced in the session, together with the reference slides and documentation, will be provided to attendees, for strategic reference in working with SQL Server and the infrastructure underneath in their own environments.

Seats are limited in this training offering, so register now!

Tuesday, November 22nd – 10:00am to 6:30pm Eastern

Thursday, December 29th – 10:00am to 6:30pm Eastern

Oct 212016
 

vmworld-2016This past week I had the pleasure of attending our first VMworld Europe conference, held this week at the Fira Gran Via in Barcelona, Spain. What a great experience!

20161017_085815

The key takeaways from the conference are quite interesting. The announcements from last and this week indicate some challenges in portions of their strategy going forward. This photo from the speaker room sums up our thoughts on the state of VMware right now – many of the important puzzle pieces are in place, but there are some significant holes that need to be addressed for a more complete picture.

20161018_130601-copy

Sessions

20161018_103107-copyFirst up, I presented a session entitled “Performance Tuning and Monitoring for Virtualized Database Servers” with Thomas LaRock from Solarwinds. We talked about the need to monitor each layer at and underneath the databases for becoming proactive in active performance troubleshooting and monitoring. We presented to a packed room and fielded a number of great questions afterwards.

A little later in the day, I presented a session entitled “Performance Perspectives” for the vBrownBag TechTalks, where I talked about how VM admins need to understand that measuring performance statistics only from the hypervisor presents only a portion of the actual performance of the system.

[embedyt] http://www.youtube.com/watch?v=KhYxiZPSvGE[/embedyt]

At the end of the day, I sat on a panel session with Patric Chang and Jonathan Flynn from SanDisk, and Jase McCarty from EMC entitled “Running Business Critical Applications and the Software Defined Data Center on Hyper-Converged Infrastructure and VSAN” where we discussed the implications of business-critical applications and their intersection with hyperconverged and all-flash systems.

vmworld-sandisk-session_photo-small

Wednesday

20161019_104942-copyWednesday was also a day filled with presentations and meetings. I started with a fun and action-packed session with Michael Corey entitled “Monster VMs (Database Virtualization) Doing IT Right” where we gave a rapid-fire stream of useful tips and tricks on maintaining maximum performance of virtualized SQL Server and Oracle VMs.

20161019_110304-copy

That afternoon I presented a repeat session of the “Performance Tuning and Monitoring for Virtualized Database Servers” to another near capacity room. Thank you to all of the die-hard attendees that came for the repeat session when you could not get into the previous session.

Next, VMworld has a lounge area that they call “Meet the Experts“. Michael Corey and I chatted with numerous attendees about their unique challenges with virtualizing business-critical apps, especially databases, and hopefully our answers will help them go and solve some of their concerns!

20161019_073942-copy

Announcements

At this conference, VMware announced the next version of their flagship virtualization suite, vSphere 6.5. The whole list of updates and improvements are found here. I feel that this release is a solid evolutionary step towards the future of the on-premises software defined datacenter. I am exceptionally happy about the new REST-based API for managing the environment. We’ve got some ideas that we’re working on where this will come in handy! VMware says general availability of this release is scheduled for later on this year.

However, the overarching buzz at the conference was from VMware’s announcement last week. VMware is partnering with Amazon AWS to provide the means to extend the on-prem VMware deployments to the Amazon public cloud in order to create a seamless hybrid cloud approach. The reaction from the attendees was mixed. I am going to save my thoughts on this announcement for an upcoming blog post. It does open a lot of questions about target platform performance, database licensing implications, and operational management.

Barcelona

I did get to venture outside of the convention center and briefly explore Barcelona on Monday with several other SQL Server presenters. Barcelona is an incredible city. I wish I had a full month to just go exploring!

Thanks for having me speak at the conference VMware! This was my first trip to VMworld Europe, but it will certainly not be the last!

20161017_144823 20161017_151920 20161017_110839

Oct 172016
 

lisa16_button_neat_180I am proud to present that Deji Akomolafe from VMware Corporation, Cody Chapman from our team at Heraflux, and I will be presenting at this year’s USENIX LISA16 conference in Boston, MA, from December 4-9th!

The three of us are presenting a half-day workshop entitled “Designing Your VMware Virtual Infrastructure for Optimal Performance, Resilience and Availability – Straight from the Source” from 1:30p – 5:00p on Monday, December 5th.

akomolafe_deji_100x115  chapman_cody_100x115  klee_david_100x115

Continue reading »

Sep 122016
 

Bottlenecks in your testing tools, infrastructure, or methodology might just hurt your load test results, and the results can skew your metrics.

For example…

A few weeks ago we were running an iperf load test to see what improvements (if any) could be made to the configuration of a Windows Server networking stack. Setting up iperf for maximum throughput testing is easy.

iperf1

iperf2

The base test proved pretty fast, too!

iperf3

This testing was performed on our new 10GbE lab environment without jumbo frames enabled in the guest. We’ve got a Cisco Nexus 5k underneath this environment, so it is nice and fast.

We wanted to see the performance improvements with keeping this stream of data on the backplane of one physical server in our VM lab. We moved both of these VMs to the same host. Performing the same test should improve performance. I’ve seen upwards of a 16x performance improvement on a 1GbE network, so we should see at least a decent improvement on 10GbE, right?

The next test is after we moved them to the same host.

iperf4

It’s virtually (no pun intended) identical.

 

OK. So, let’s try to make Windows a bit more efficient.

autotunedisable

The results show a bit of improvement.

autotunedisablediperf

But… why did it stop there? 

OK…

Looking at the system state showed us an instant bottleneck. One vCPU was maxed out, while the other was stagnant. This imbalance is one key reason why you should rarely trust a single CPU average metric coming from your monitoring tools.

ProcessorUsage

The load test utility had maxed out the compute resource that it had available, due to internal limitations within iperf itself.  It’s a shame that this utility is not multi-threaded, because I think we could have a much greater result of improvement on this system.

Monitor the utilities that you’re using to do load testing, because limitations like this might skew your results!